Highway Capacity Manual 3rd Edition

HIGHWAY
CAPACITY
MANUAL
Special Report 209
Third Edition
TRANSPORTATION RESEARCH BOARD

National Research Council
Washington, D.C. 1998
1998 TRANSPORTATION RESEARCH BOARD Transportation Research Board Special Report 209
EXECUTIVE COMMITTEE
Subscriber Categories
IA planning and administration
OFFICERS IIA highway and facility design
Chairwoman: Sharon D. Banks, General Manager, AC Transit IVA highway operations, capacity, and traffic control
Vice Chairman: Wayne Shackelford, Commissioner, Georgia Depart- VI public transit
ment of Transportation
Executive Director: Robert E. Skinner, Jr., Transportation Research Transportation Research Board publications are available by ordering indi-
Board vidual publications directly from the TRB Business Office, through the Internet
at http://www.nas.edu/trb/index.html, or by annual subscription through organi-
MEMBERS zation or individual affiliation with TRB. Affiliates and library subscribers are
eligible for substantial discounts. For further information, contact the Transpor-
Thomas F. Barry, Jr., Secretary of Transportation, Florida Department of tation Research Board Business Office, National Research Council, 2101 Con-
Transportation stitution Avenue, N.W., Washington, D.C. 20418 (telephone 202-334-3214;
Brian J. L. Berry, Lloyd Viel Berkner Regental Professor, University of Texas fax 202-334-2519; or e-mail kpeterse@nas.edu).
at Dallas
Sarah C. Campbell, President, TransManagement, Inc. NOTICE
E. Dean Carlson, Secretary, Kansas Department of Transportation The project that is the subject of this report was approved by the Governing
Joanne F. Casey, President, Intermodal Association of North America Board of the National Research Council, whose members are drawn from
John W. Fisher, Director, ATLSS Engineering Research Center, and Professor the councils of the National Academy of Sciences, the National Academy of
of Civil and Environmental Engineering, Lehigh University Engineering, and the Institute of Medicine. The members of the committee
Gorman Gilbert, Director, Institute for Transportation Research and Education, responsible for the report were chosen for their special competence and with
North Carolina State University regard for appropriate balance.
Delon Hampton, Chairman and CEO, Delon Hampton & Associates, Chartered This report has been reviewed by a group other than the authors according
Lester A. Hoel, Hamilton Professor, Department of Civil Engineering, Univer- to procedures approved by a Report Review Committee consisting of members
sity of Virginia of the National Academy of Sciences, the National Academy of Engineering,
James L. Lammie, Director, Parsons Brinckerhoff, Inc. and the Institute of Medicine.
Thomas F. Larwin, General Manager, San Diego Metropolitan Transit Develop-
ment Board
Bradley L. Mallory, Secretary of Transportation, Commonwealth of ©1985, 1992, 1994, 1998 by the Transportation Research Board
Pennsylvania All rights reserved. First edition 1950
Jeffrey J. McCaig, President and CEO, Trimac Corporation Third edition 1985
Joseph A. Mickes, Chief Engineer, Missouri Department of Transportation Printed in the United States of America
Marshall W. Moore, Director, North Dakota Department of Transportation
Andrea Riniker, Executive Director, Port of Tacoma First printing, August 1985
John M. Samuels, Vice President—Operations Planning and Budget, Norfolk Second printing, December 1985
Southern Corporation Third printing, June 1987
Les Sterman, Executive Director, East-West Gateway Coordinating Council Fourth printing, June 1993
James W. van Loben Sels, Director, California Department of Transportation Fifth printing, October 1994
(Past Chairman, 1996) Sixth printing, April 1998
Martin Wachs, Director, University of California Transportation Center, and
Professor of Civil Engineering and City and Regional Planning, University
of California Library of Congress Cataloging in Publication Data
David L. Winstead, Secretary, Maryland Department of Transportation National Research Council. Transportation Research Board.
David N. Wormley, Dean of Engineering, Pennsylvania State University (Past Highway capacity manual. 3rd ed. ©1998.
Chairman, 1997) p. cm. — (Special report ; 209)
Mike Acott, President, National Asphalt Pavement Association (ex officio) Includes index.
Joe N. Ballard (Lt. Gen., U.S. Army), Chief of Engineers and Commander, ISBN 0-309-06450-3
U.S. Army Corps of Engineers (ex officio) 1. Highway capacity—Handbooks, manuals, etc. I. Series: Special
Andrew H. Card, Jr., President and CEO, American Automobile Manufacturers report (National Research Council (U.S.) Transportation Research
Association (ex officio) Board); 209.
Kelley S. Coyner, Acting Administrator, Research and Special Programs Ad- HE336.H48H54 1998 ISSN 0360-859X 98-5965
ministration, U.S. Department of Transportation (ex officio) 388.3’ 14—dc21 CIP
Mortimer L. Downey, Deputy Secretary, Office of the Secretary, U.S. Depart-
ment of Transportation (ex officio)
Francis B. Francois, Executive Director, American Association of State High- The Transportation Research Board is a unit of the National Re-
way and Transportation Officials (ex officio) search Council, which serves the National Academy of Sciences and
David Gardiner, Assistant Administrator, Office of Policy, Planning and Evalua- the National Academy of Engineering. The Board’s mission is to
tion, Environmental Protection Agency (ex officio)
promote innovation and progress in transportation by stimulating and
Jane F. Garvey, Administrator, Federal Aviation Administration, U.S. Depart-
ment of Transportation (ex officio) conducting research, facilitating the dissemination of information,
John E. Graykowski, Acting Administrator, Maritime Administration, U.S. De- and encouraging the implementation of research results. The
partment of Transportation (ex officio) Board’s varied activities annually draw on approximately 4,000 engi-
Robert A. Knisely, Deputy Director, Bureau of Transportation Statistics, U.S. neers, scientists, and other transportation researchers and prac-
Department of Transportation (ex officio) titioners from the public and private sectors and academia, all of
Gordon J. Linton, Administrator, Federal Transit Administration, U.S. Depart- whom contribute their expertise in the public interest. The program
ment of Transportation (ex officio) is supported by state transportation departments, federal agencies
Ricardo Martinez, Administrator, National Highway Traffic Safety Administra-
including the component administrations of the U.S. Department of
tion, U.S. Department of Transportation (ex officio)
Walter B. McCormick, President and CEO, American Trucking Associations, Transportation, and other organizations and individuals interested in
Inc. (ex officio) the development of transportation.
William W. Millar, President, American Public Transit Association (ex officio) The National Research Council was organized by the National
Jolene M. Molitoris, Administrator, Federal Railroad Administration, U.S. De- Academy of Sciences in 1916 to associate the broad community of
partment of Transportation (ex officio) science and technology with the Academy’s purpose of furthering
Karen Borlaug Phillips, Senior Vice President, Policy, Legislation, and Eco- knowledge and advising the federal government. Functioning in ac-
nomics, Association of American Railroads (ex officio) cordance with general policies determined by the Academy, the
Valentin J. Riva, President, American Concrete Pavement Association (ex
Council has become the principal operating agency of both the Na-
officio)
George D. Warrington, Acting President and CEO, National Railroad Passen- tional Academy of Sciences and the National Academy of Engi-
ger Corporation (ex officio) neering in providing services to the government, the public, and the
Kenneth R. Wykle, Administrator, Federal Highway Administration, U.S. De- scientific and engineering communities. The Council is administered
partment of Transportation (ex officio) jointly by both the Academies and the Institute of Medicine.
Updated December 1997

Contents
Foreword................................................................................................................................................. v
Contributors and Acknowledgments .............................................................................................. vii
Figures, Photographs, and Tables..................................................................................................... xv
PART I
PRINCIPLES OF CAPACITY
Chapter 1 Introduction, Concepts, and Applications............................................................................. 1-1
Chapter 2 Traffic Characteristics ........................................................................................................... 2-1
PART II
FREEWAYS
Chapter 3 Basic Freeway Sections......................................................................................................... 3-1
Chapter 4 Weaving Areas ...................................................................................................................... 4-1
Chapter 5 Ramps and Ramp Junctions .................................................................................................. 5-1
Chapter 6 Freeway Systems ................................................................................................................... 6-1
PART III
RURAL AND SUBURBAN HIGHWAYS
Chapter 7 Multilane Rural and Suburban Highways ............................................................................ 7-1
Chapter 8 Two-Lane Highways ............................................................................................................. 8-1
PART IV
URBAN STREETS
Chapter 9 Signalized Intersections......................................................................................................... 9-1
Chapter 10 Unsignalized Intersections..................................................................................................... 10-1
Chapter 11 Arterial Streets....................................................................................................................... 11-1
Chapter 12 Transit Capacity..................................................................................................................... 12-1
Chapter 13 Pedestrians ............................................................................................................................. 13-1
Chapter 14 Bicycles.................................................................................................................................. 14-1
APPENDIX A
Glossary ................................................................................................................................. A-1
Symbols ................................................................................................................................. A-5
INDEX
iii Updated December 1997

Foreword
The Highway Capacity Manual (HCM) continues to provide a resource for technical information that is used by
transportation planners, designers, and operators. The materials contained in the HCM represent a collection of
state-of-the-art techniques for estimating capacity and determining level of service for many transportation facilities
and modes. These techniques have been developed and enhanced through funded research projects and through
review of the research results by the Committee on Highway Capacity and Quality of Service.
The contents of this edition of the HCM represent the consensus view of the committee as to the best available
techniques for determining capacity. However, this manual does not establish a legal standard for highway design or
construction. Throughout the manual, sound engineering judgment supplemented by field observations is encouraged.
Throughout previous editions of the manual, many transportation professionals have contributed to the develop-
ment of highway capacity analysis techniques. These efforts were documented in the 1994 update to the manual
and are repeated here to recognize the accomplishments of these professionals.
The first Highway Capacity Manual was published in 1950 as a joint venture between the Highway Research
Board’s Committee on Highway Capacity and the Bureau of Public Roads. O. K. Normann served as committee
chairman and William Walker as secretary. This edition, the first international document on the broad subject of
capacity, provided definitions of key terms, a compilation of maximum observed flows, and the initial fundamentals
of capacity. Analytical procedures were included for uninterrupted-flow facilities, signalized intersections, weaving
sections, and ramps.
The second edition of the manual was published in 1965 by the Highway Research Board and authored by the
Committee on Highway Capacity. It was dedicated to O. K. Normann, who had provided leadership to the committee
from its inception in 1944 until his death in 1964. Carl C. Saal had become committee chairman and Arthur A.
Carter, Jr., continued to serve as secretary. During the final stages of the preparation of the manual, a five-person
task group was assigned by the Bureau of Public Roads to work full time on the project. The 1965 manual was a
significant extension of the 1950 edition and is most noted for its introduction of the level-of-service concept.
The third edition of the manual was published in 1985 by the Transportation Research Board and authored by
the Committee on Highway Capacity and Quality of Service chaired by Carlton C. Robinson, with Charles W. Dale
as secretary. Credit is also due Robert C. Blumenthal and James H. Kell, who served as committee chairmen and
provided leadership between the publication of the 1965 and the 1985 editions. Again, the breadth and depth of
the previous manuals were extended. The 1985 edition is perhaps most noted for the extension into facilities other
than highways, refinements to the level-of-service concept, and the accompanying computer software.
When the 1994 update was published by the Transportation Research Board, it provided new analytical procedures
in response to the increased levels of research and professional interest in this topic. The committee was chaired
by Adolf D. May, with Wayne Kittelson as secretary. This update contained revisions to portions of 8 of the 14
chapters to include current speed-flow relationships, revised capacity values, and new analytical procedures. In
addition, greater emphasis was placed on describing the principles of capacity and on defining the capacity and
level-of-service terms.
v Updated December 1997
vi
This 1997 update of the HCM has been published to make the most current procedures available to the user
community in a timely fashion. It is recognized that the relatively short time between updates of the manual causes
some difficulty in users’ ability to incorporate the new procedures into their practice; however, the committee has
chosen to publish this update to make the results of a significant amount of new research available in a timely
manner.
The current update includes extensive revisions to Chapters 3, 9, 10, and 11. In addition, Chapters 1, 4, 5, 6, and
7 have been modified to make them consistent with other revised chapters.
Chapter 3, Basic Freeway Sections, includes a revised procedure for determining capacity on the basis of density.
Capacity values under ideal flow conditions now vary by free-flow speed. This chapter also provides a ‘‘preview’’
of the proposed format for the HCM 2000 that is currently being developed. Your comments on this format are
requested.
Chapter 9, Signalized Intersections, includes findings from recent research on actuated traffic signals. The delay
equation is modified to account for signal coordination, oversaturation, variable length analysis periods, and the
presence of initial queues at the beginning of an analysis period. The level-of-service measure has been changed
from average stopped delay to total (control) delay. Adjustments have been made to the permitted left-turn movement
model and to the left-turn equivalency table.
Chapter 10, Unsignalized Intersections, has been completely revised to incorporate the results of a nationwide
research project in the United States at two-way and four-way stop-controlled intersections. Modified delay formulas
and new level-of-service thresholds are provided for both two-way and four-way stop-controlled intersections. In
addition, the impact on capacity at a two-way stop-controlled intersection due to the presence of an upstream traffic
signal can be determined. Procedures are provided to account for flared approaches, upstream signals, pedestrian
crossings, and two-stage gap acceptance (where vehicles seek refuge in a median before crossing a second stream
of traffic).
Chapter 11, Arterial Streets, incorporates the changes to the Signalized Intersections chapter that affect Chapter
11. In addition, a new arterial classification is established for high-speed facilities. The delay equation is modified
to take into account the effect of upstream signalized intersections on platoon arrivals.
Despite the extensive improvements incorporated into the 1997 update of the HCM, plans are under way for a
complete revision of the HCM in 2000. The content, format, and delivery system for HCM 2000 will be made more
accessible to users in both paper and multimedia (CD-ROM) formats. Ongoing research in freeway weaving, freeway
systems, two-lane highways, transit capacity, bicycle and pedestrian capacity, interchange ramp terminals, and
enhanced procedures for transportation planning will be included in HCM 2000.
This 1997 update and the upcoming HCM 2000 will represent a major milestone in the ongoing efforts of
researchers and practitioners to provide a practical guide for capacity analysis techniques for all who use them. The
efforts of the funding agencies, research institutions, the academic community, and users from the public and private
sectors are gratefully acknowledged. The Highway Capacity and Quality of Service Committee invites your com-
ments and suggestions regarding this 1997 update as we enhance our ability to design, operate, and plan for improved
transportation facilities.
For the Committee on Highway Capacity and Quality of Service
John D. Zegeer
Chairman

Contributors and Acknowledgments
This report is the result of the coordinated efforts of many individuals, research organizations, and government agencies.
Although responsibility for the content of the Highway Capacity Manual lies with the Committee on Highway Capacity and
Quality of Service, its preparation was accomplished through the efforts of the following groups and individuals:
TRB COMMITTEE ON HIGHWAY CAPACITY AND QUALITY OF SERVICE

Committee Members as of February 1, 1985
Carlton C. Robinson, Chairman, Highway Users Federation for Safety and Mobility
Charles W. Dale, Secretary, Federal Highway Administration
Donald S. Berry, Evanston, Illinois
Robert C. Blumenthal, Blumenthal Associates (Chairman, 1971–1977)
James B. Borden, California Department of Transportation
Fred W. Bowser, Pennsylvania Department of Transportation
V. F. Hurdle, University of Toronto, Canada
James H. Kell, JHK & Associates (Chairman, 1977–1983)
Frank J. Koepke, Northwestern University
Jerry Kraft, New Jersey Turnpike Authority
Walter H. Kraft, Edwards & Kelcey, Inc.
Joel P. Leisch, Jack E. Leisch & Associates
Adolf D. May, Jr., University of California
William R. McShane, Polytechnic Institute of New York
Carroll J. Messer, Texas A&M University System
Guido Radelat, Federal Highway Administration
Hubert M. Shaver, Jr., Virginia Department of Highways and Transportation
Alexander Werner, Alberta Transportation Department, Canada
Robert H. Wortman, University of Arizona
David K. Witheford, Transportation Research Board Staff Representative
Other Committee Members During 1985 Manual Preparation Period

Brian L. Allen, McMaster University
George W. Black, Jr., Gwinnett County, Georgia
Arthur A. Carter, Jr., Federal Highway Administration
Joseph W. Hess, Bethesda, Maryland
Jack A. Hutter, Jack E. Leisch & Associates
Thomas D. Jordan, Skycomp Data Corporation
Paul D. Kiser, City of Salt Lake
Herbert S. Levinson, University of Connecticut
Louis E. Lipp, Colorado Department of Highways
Edward B. Lieberman, KLD Associates, Inc.
Louis J. Pignataro, Polytechnic Institute of New York
Frederick D. Rooney, California Department of Transportation
Stephen E. Rowe, Los Angeles Department of Transportation
vii Updated December 1997

viii
John L. Schlaefli, TRACOR, Inc.

Gerald W. Skiles, Cambria, California
Jeffrey M. Zupan, New Jersey Transit
Committee Members as of December 1, 1994

Adolf D. May, Jr., Chairman, University of California
Wayne K. Kittelson, Secretary, Kittelson & Associates, Inc.
Rahmi Akcelik, Australian Road Research Board Ltd.
James A. Bonneson, University of Nebraska
Werner Brilon, Ruhr University, Germany
Kenneth G. Courage, University of Florida
Rafael E. DeArazoza, Florida Department of Transportation
Richard G. Dowling, Dowling Associates
Daniel B. Fambro, Texas A&M University System
Ronald K. Giguere, Federal Highway Administration
Mariano Gullón Löw, Centro de Estudios de Carreteras, Madrid, Spain
Fred L. Hall, McMaster University, Canada
Douglas W. Harwood, Midwest Research Institute
Michael D. Kyte, University of Idaho
Joel P. Leisch, Glenview, Illinois
Douglas S. McLeod, Florida Department of Transportation
John F. Morrall, University of Calgary, Canada
Barbara K. Ostrom, Berkeley, California
Ronald C. Pfefer, Northwestern University
James L. Powell, DeLeuw, Cather & Co.
William R. Reilly, Catalina Engineering
Roger P. Roess, Polytechnic University
Nagui M. Rouphail, North Carolina State University
Ronald C. Sonntag, Wisconsin Department of Transportation
Alex Sorton, Northwestern University
Dennis W. Strong, Strong Concepts
Stan Teply, University of Alberta, Canada
Pierre Yves Texier, INRETS, France
Rod J. Troutbeck, Queensland University of Technology, Australia
Thomas Urbanik II, Texas A&M University System
John D. Zegeer, Barton-Aschman Associates, Inc.
Richard Cunard, Transportation Research Board Staff Representative
Dan Rosen, NCHRP Staff Representative
Committee Members as of December 31, 1997

John D. Zegeer, Chairman, Kittelson & Associates, Inc.
Richard G. Dowling, Secretary, Dowling Associates, Inc.
James A. Bonneson, Texas A&M University System
Werner Brilon, Ruhr University of Bochum, Germany
Robert W. Bryson, City of Milwaukee
Alan R. Danaher, Kittelson & Associates, Inc.
Lily Elefteriadou, Pennsylvania State University
Albert L. Grover, Albert Grover & Associates, Inc.
Chris Hoban, The World Bank
Wayne K. Kittelson, Kittelson & Associates, Inc.
Adolf D. May, Jr., University of California
Barbara K. Ostrom, EBA Engineering, Inc.
ix
James L. Powell, DeLeuw, Cather & Company

Erik O. Ruehr, Valley Research and Planning Associates
Rikke Rysgaard, Danish Road Directorate
James M. Schoen, Catalina Engineering, Inc.
Richard Cunard, Transportation Research Board Staff Representative
Ray Derr, NCHRP Staff Representative
The work of the following individuals in subcommittees of the Committee on Highway Capacity and Quality of Service contributed
immeasurably to the effectiveness of the committee in accomplishing its goals:
Subcommittee Members as of February 1, 1985

Charles M. Abrams, JHK & Associates
Frank E. Barker, Chicago Transit Authority
Seth S. Barton, New Jersey Department of Transportation
Richard Bowman, Beiswenger Hoch & Associates, Inc.
John P. DiRenzo, Peat, Marwick, Mitchell & Co.
Paul Eng-Wong, Snavely King & Associates
Thomas C. Ferrara, California State University
A. Reed Gibby, California State University
William Haussler, Edwards & Kelcey, Inc.
Joseph W. Hess, Bethesda, Maryland
Paul P. Jovanis, Northwestern University
Joseph M. Kaplan, National Safety Council (Los Angeles Chapter)
Wayne K. Kittelson, CH2M Hill
Herbert S. Levinson, University of Connecticut
C. John MacGowan, National Highway Traffic Safety Administration
Ralph J. Meller, St. Louis, Missouri
David R. Merritt, Federal Highway Administration
Panos G. Michalopoulos, University of Minnesota
Timothy R. Neuman, Jack E. Leisch & Associates
Martin R. Parker, Jr., M.R. Parker & Associates, Inc.
William R. Reilly, JHK & Associates
Roger P. Roess, Polytechnic Institute of New York
Richard Rogers, California Department of Transportation
Frederick D. Rooney, California Department of Transportation
Gilbert T. Satterly, Jr., Purdue University
Frederick S. Scholz, Roger Creighton Associates, Inc.
Steven R. Shapiro, Goodell-Grivas, Inc.
Joseph H. Sinnott, System Design Concepts, Inc.
Frank C. Tecca, Municipality of Anchorage, Alaska
Linda Turnquist, California Department of Transportation
Kenneth H. Voigt, Southeastern Wisconsin Regional Planning Commission
Mark R. Virkler, University of Missouri
Subcommittee Members as of December 1, 1994

Rahmi Akcelik, Australian Road Research Board Ltd.
Donald S. Berry, Evanston, Illinois
James A. Bonneson, University of Nebraska
Ulrich Brannolte, PTV GMBH, Germany
Werner Brilon, Ruhr University, Germany
Joonho Byun, Federal Highway Administration
Michael J. Cassidy, University of California
x

Richard G. Dowling, Dowling Associates
Lily Elefteriadou, Germen Associates
Joseph Fazio, Chicago, Illinois
Kay Fitzpatrick, Texas A&M University System
A. Reed Gibby, California State University
Glenn M. Grigg, City of Cupertino, California
Albert L. Grover, Albert Grover & Associates
Wayne E. Haussler, Edwards & Kelcey, Inc.
VanOlin F. Hurdle, University of Toronto, Canada
Dane Ismart, Federal Highway Administration
Paul P. Jovanis, University of California
R. Ian Kingham, Victoria, Canada
Frank J. Koepke, S/K Transportation Consultants, Inc.
Raymond A. Krammes, Texas A&M University System
B. Kent Lall, Portland State University
Jim C. Lee, Lee Engineering
Joel P. Leisch, Glenview, Illinois
Herbert S. Levinson, New Haven, Connecticut
Feng-Bor Lin, Clarkson University
George F. List, Rensselaer Polytechnic University
Charles W. Manning, Roger Creighton Associates, Inc.
Joseph F. Marek, Clackamas County, Oregon
William R. McGrath, Fort Myers, Florida
William R. McShane, Polytechnic University
David R. Merritt, Federal Highway Administration
Carroll J. Messer, Texas A&M University System
Leonard Newman, Emeryville, California
Michael P. O’Rourke, Eng-Wong Taub and Associates
Barbara K. Ostrom, Berkeley, California
James L. Powell, DeLeuw, Cather & Co.
William A. Prosser, Federal Highway Administration
William R. Reilly, Catalina Engineering
Frederick Rooney, California Department of Transportation
Erik O. Ruehr, JHK & Associates
James M. Schoen, Catalina Engineering
Alex Skabardonis, University of California
Ronald C. Sonntag, Wisconsin Department of Transportation
Marian Tracz, Cracow Technical University, Poland
Thomas Urbanik II, Texas A&M University System
Mark A. Vandehey, Kittelson & Associates, Inc.
R. A. Vincent, Transport Research Laboratory, Great Britain
Mark R. Virkler, University of Missouri-Columbia
Kenneth H. Voigt, HNTB Corporation
xi
Subcommittee Members as of December 31, 1997
Subcommittee on General Concepts and Definitions

Barbara K. Ostrom, Chairwoman, EBA Engineering, Inc.
Thomas Urbanik II, Texas Transportation Institute
Subcommittee on Freeways and Multilane Highways

Adolf D. May, Jr., Chairman, University of California
Michael J. Cassidy, University of California
Michael Church, California Department of Transportation
Joseph Fazio, Illinois Institute of Technology
Abdul-Rahman Hamad, H.W. Lochner, Inc.
Joel P. Leisch, Private Consultant
Ronald C. Pfefer, Northwestern University Traffic Institute
William R. Reilly, Catalina Engineering, Inc.
Bruce W. Robinson, Kittelson & Associates, Inc.
Fred Rooney, California Department of Transportation
James M. Schoen, Catalina Engineering, Inc.
Ronald C. Sonntag, Marquette University
Andrzej P. Tarko, Purdue University
Michelle Thomas, Federal Highway Administration
Jose Ulerio, Polytechnic University
Thomas Urbanik II, Texas Transportation Institute
Subcommittee on Interchange Ramp Terminals

James L. Powell, Chairman, DeLeuw, Cather & Company
James A. Bonneson, Texas A&M University System
Michael Church, California Department of Transportation
F. Thomas Creasey, Wilbur Smith & Associates
Janice Daniel, Georgia Institute of Technology
Michael F. Holling, Transcore
Joel K. Marcuson, Sverdrup Civil, Inc.
Scott J. Parker, Edwards and Kelcey, Inc.
Frederick Rooney, California Department of Transportation
Subcommittee on Signalized Intersections

Dennis W. Strong, Chairman, Strong Concepts
Rahmi Akcelik, ARRB Transport Research, Ltd.
Rahim F. Benekohal, University of Illinois
Donald S. Berry, Private Consultant
Glenn M. Grigg, Private Consultant
David J. P. Hook, Hook Engineering, Inc.
John D. Leonard II, Georgia Institute of Technology
Feng-Bor Lin, Clarkson College
Pawan Maini, University of Colorado at Denver
Carroll J. Messer, Texas Transportation Institute
Elena Prassas, Polytechnic University
xii
Bruce W. Robinson, Kittelson & Associates, Inc.

Subcommittee on Unsignalized Intersections

Michael D. Kyte, Chairman, University of Idaho
Werner Brilon, Ruhr University of Bochum, Germany
Joonho Byun, Federal Highway Administration
Mitzi M. Dobersek, Wisconsin Department of Transportation
Aimee Flannery, Pennsylvania State University
Glenn M. Grigg, Private Consultant
Wayne E. Haussler, Goodkind & O’Dea, Inc.
Dane Ismart, Louis Berger & Associates, Inc.
R. Ian Kingham, Graeme & Murray
George F. List, Rensselaer Polytechnic Institute
Charles Manning, Creighton Manning, Inc.
Joseph F. Marek, Clackamas County Department of Transportation
Michael P. O’Rourke, Eng-Wong-Taub & Associates, Inc.
John Sampson, Jeffares & Green, Inc.
Zong Zhong Tian, Kittelson & Associates, Inc.
Marion Tracz, Cracow Technical University, Poland
Kenneth H. Voigt, HNTB Corporation
Andrew Wolfe, Union College
Subcommittee on Urban and Suburban Arterials

Daniel B. Fambro, Chairman, Texas A&M University System
Janice R. Daniel, Georgia Tech University
Joel K. Marcuson, Sverdrup Corporation
Andrzej P. Tarko, Purdue University
Mark A. Vandehey, Kittelson & Associates, Inc.
Subcommittee on Two-Lane Roads

Douglas W. Harwood, Chairman, Midwest Research Institute
Jan L. Botha, San Jose State University
Greg M. Laragan, Idaho Department of Transportation
David J. Lovell, University of Maryland, College Park
Carroll J. Messer, Texas Transportation Institute
Guido Radelat, Private Consultant
Davey Warren, Federal Highway Administration
Al Werner, Reid Crowther Consultants, Ltd.

xiii
Subcommittee on Transit Systems

Alan R. Danaher, Chairman, Kittelson & Associates, Inc.
Tara Bartee, Florida Department of Transportation
Howard Benn, Private Consultant
William Hoey, Private Consultant
Herbert S. Levinson, Transportation Consultant
Pat McLoughlin, Metropolitan Transit Authority, Los Angeles
David Miller, Parsons Brinckerhoff
Kevin St. Jacques, Wilbur Smith & Associates
Ken Stanley, Oahu Transit
Joe Goodman, Federal Transit Administration
Joel Volinski, University of South Florida
Subcommittee on Planning Applications

Douglas S. McLeod, Chairman, Florida Department of Transportation
Jim Altenstadter, PIMA Association of Governments
F. Thomas Creasey, Wilbur Smith & Associates
Richard G. Dowling, Dowling Associates, Inc.
Kurt Eichin, Florida Department of Transportation
Abdul-Rahman Hamad, H.W. Lochner, Inc.
David Hook, Lee Engineering, Inc.
John Karachepone, Kittelson & Associates, Inc.
William R. McShane, Polytechnic University
Terrel Shaw, Reynolds, Smith & Hills, Inc.
Subcommittee on Pedestrians and Bicycles

Alex Sorton, Chairman, Northwestern University
Patrick Allen, California Department of Transportation
Hein Botma, Delft University of Technology, The Netherlands
W. Jeffrey Davis, The Citadel
Joseph Fazio, Illinois Institute of Technology
Bruce Landis, Sprinkle Consulting Engineering, Inc.
John LaPlante, TY Lin Bascor
Joseph S. Milazzo, North Carolina State University
Virginia Sisiopiku, Michigan State University
Mark R. Virkler, University of Missouri-Columbia
Thomas Walsh, City of Madison, Wisconsin
Subcommittee on User Liaison and Interpretations

Wayne K. Kittelson, Chairman, Kittelson & Associates, Inc.
Robert S. Foyle, North Carolina State University
Shahram Malek, Viggen Corporation, Inc.
Charles E. Wallace, University of Florida
xiv
Subcommittee on Research Coordination

Fred L. Hall, Chairman, McMaster University, Canada
Jim Clark, Federal Highway Administration
Alan R. Danaher, Kittelson & Associates, Inc.
Richard G. Dowling, Dowling Associates, Inc.
Pawan Maini, University of Colorado at Denver
Larry F. Sutherland, Ohio Department of Transportation
Last to be acknowledged among the volunteer contributors to this edition are the unnamed users of draft materials and TRB
Circulars published and distributed during the period of the manual’s development. Their interest and support were a constant
stimulus to both committee and research activities.
Perhaps last in the process, but not least among those who made this document possible, are staff members of the Transportation
Research Board. Naomi Kassabian, Norman Solomon, and David Stearman, Editors, worked with the researchers to produce
the final manuscript. Other design and production supervision was provided by Nancy A. Ackerman, Director of Reports and
Editorial Services. Ray Derr, NCHRP Projects Engineer, and Richard Cunard, Engineer of Traffic and Operations, provided
indispensable staff support to the committee and its subcommittees.

FIGURES Page
2–1 Typical relationship between time mean and space mean speed...................................................... 2-4
2–2 Generalized relationships among speed, density, and rate of flow on uninterrupted flow
facilities ........................................................................................................................................... 2-6
2–3 Conditions at traffic interruption in an approach lane of a signalized intersection......................... 2-7
2–4 Concept of saturation flow rate and lost time ................................................................................... 2-8
2–5 Motor vehicle registrations................................................................................................................. 2-11
2–6 Rural Interstate travel by vehicle type............................................................................................... 2-14
2–7 Annual vehicle miles of travel ........................................................................................................... 2-14
2–8(a) Examples of monthly traffic volume variations showing monthly variations in traffic for a free-
way in Minnesota............................................................................................................................ 2-17
2–8(b) Examples of monthly traffic volume variations showing relative traffic volume trends by route
type on rural roads in Lake County, Illinois ................................................................................. 2-18
2–9 Examples of daily traffic variation by type of route......................................................................... 2-19
2–10 Daily variation in traffic by vehicle type (I-494, 4-lanes, in Minneapolis-St. Paul) ....................... 2-19
2–11 Examples of hourly traffic variations for rural routes in New York State ...................................... 2-20
2–12 Repeatability of hourly traffic variations for four 2-lane arterials in Toronto, Ontario, Canada.... 2-20
2–13 Ranked hourly volumes on Minnesota highways.............................................................................. 2-21
2–14 Ranked hourly volume distribution showing indistinct knee for Kentucky location in 1977......... 2-21
2–15 Relationship between short-term and hourly flows........................................................................... 2-22
2–16 Distribution of power-to-mass ratios of passenger cars .................................................................... 2-24
2–17 On-highway passenger car characteristics ......................................................................................... 2-24
2–18 Nationwide speed trends through 1975 and 1993 ............................................................................. 2-25
2–19 Speed variation by hour of day for I-35W in Minneapolis, weekdays, in relation to volume
variations ......................................................................................................................................... 2-26
2–20 Speed variation by hour of day for I-35W, Minneapolis, Saturdays, in relation to volume
variations ......................................................................................................................................... 2-27
2–21 Observed speed-flow relationship on a San Diego freeway in 6-min sampling intervals (Inter-
state Highway 8, 1987)................................................................................................................... 2-29
2–22 Observed speed-flow relationship on an Ontario freeway in 5-min sampling intervals (Queen
Elizabeth Way near Toronto, 1987)............................................................................................... 2-29
2–23 Observed speed-flow relationship at Caldecott Tunnel in 15-min sampling intervals (California
State Highway 24, 1990)................................................................................................................ 2-29
2–24 Speed-flow relationship for two-lane rural highways ....................................................................... 2-30
2–25 Time headway distribution for Long Island Expressway.................................................................. 2-31
2–26 Comparison of various research results on queue discharge headways ........................................... 2-31
3-1 Example of basic freeway section...................................................................................................... 3-2
3-2 Speed-flow relationships..................................................................................................................... 3-4
3-3 Queue discharge and congested flow................................................................................................. 3-5
3-4 LOS criteria......................................................................................................................................... 3-10
3-5 Worksheet for analysis of basic freeway sections............................................................................. 3-14
I.3-1 Sample solution for composite grade................................................................................................. 3-37
I.3-2 Performance curves for standard trucks (200 lb/hp) ......................................................................... 3-38
4–1 Formation of a weaving section......................................................................................................... 4-2
4–2 Measuring length of a weaving section ............................................................................................. 4-2
4–3 Type A weaving areas ........................................................................................................................ 4-3
4–4 Type B weaving areas ........................................................................................................................ 4-3
4–5 Type C weaving areas ........................................................................................................................ 4-4
4–6 Construction and use of weaving diagrams....................................................................................... 4-10
4–7 Weaving flows in a multiple weave formed by a single merge followed by two diverges............ 4-11
4–8 Weaving flows in a multiple weave formed by two merge points followed by a single diverge .. 4-12
4–9 Weaving area for calculation 1 .......................................................................................................... 4-12
4–10 Weaving area and flows for calculation 2......................................................................................... 4-13
5–1 On- and off-ramp influence areas ...................................................................................................... 5-2
5–2 Critical ramp junction values ............................................................................................................. 5-3
5–3 Models for predicting V12 for on-ramps............................................................................................. 5-5
5–4 Models for predicting V12 for off-ramps ............................................................................................ 5-6
xv Updated December 1997
xvi
FIGURES Page
5–5 Worksheet for the analysis of ramp-freeway terminals .................................................................... 5-10
5–6 Typical two-lane on-ramp .................................................................................................................. 5-11
5–7 Common geometries for two-lane off-ramps..................................................................................... 5-11
5–8 Major merge areas .............................................................................................................................. 5-13
5–9 Major diverge areas ............................................................................................................................ 5-13
5–10 Worksheet for Calculation 1 .............................................................................................................. 5-15
5–11 Freeway section for Calculation 2 ..................................................................................................... 5-16
5–12(a) Worksheet for Calculation 2 (first ramp) .......................................................................................... 5-17
5–12(b) Worksheet for Calculation 2 (second ramp)...................................................................................... 5-18
5–14(a) Worksheet for Calculation 3 (first ramp) .......................................................................................... 5-20
5–14(b) Worksheet for Calculation 3 (second ramp)...................................................................................... 5-21
5–17 Equivalent four-lane segment for Calculation 5................................................................................ 5-24
6–1 Sample design problem ...................................................................................................................... 6-3
6–2 Likely design for sample problem ..................................................................................................... 6-4
6–3 Consideration of multiple weave........................................................................................................ 6-4
6–4 Consideration of multiple weave........................................................................................................ 6-5
6–5 Graphic representation of overall level of service ............................................................................ 6-6
6–6 Effects of breakdown illustrated ........................................................................................................ 6-7
6–7 Illustration of ramp-metering need..................................................................................................... 6-8
6–8 Plot of cumulative ramp demand and output .................................................................................... 6-8
6–9 Potential for hidden bottlenecks......................................................................................................... 6-9
6–10 Phases of a traffic incident................................................................................................................. 6-9
6–11 Range of observed work zone capacities—work crew at site .......................................................... 6-10
6–12 Cumulative distribution of observed work-zone capacities .............................................................. 6-11
6–13 Sample calculation—queue analysis for a work zone....................................................................... 6-13
6–14 Example for analysis of HOV lane impact........................................................................................ 6-15
7–1 Speed-flow relationships on multilane highways .............................................................................. 7-4
7–2 Density-flow relationships on multilane highways............................................................................ 7-5
7–3 Speed-flow curves with LOS criteria................................................................................................. 7-8
7–4 Example of graphic solution using speed-flow curves...................................................................... 7-15
7–5 Worksheet for operational and design analysis ................................................................................. 7-17
7–6 Worksheet for planning analysis ........................................................................................................ 7-20
7–7 Illustration of solution to Calculation 1—general segment .............................................................. 7-22
7–8 Illustration of solution to Calculation 1—grade segment ................................................................. 7-23
7–9 Illustration of solution to Calculation 2—level segment .................................................................. 7-25
7–10 Illustration of solution to Calculation 2—grade segment ................................................................. 7-26
7–11 Illustration of solution to Calculation 3............................................................................................. 7-27
8–1 Speed-flow and percent time delay-flow relationships for two-lane rural highways ...................... 8-4
8–2 Speed reduction curve for a 200-lb/hp truck..................................................................................... 8-13
8–3 Speed reduction curve for a 300-lb/hp truck..................................................................................... 8-13
8–4 Worksheet for operational analysis of general terrain segments ...................................................... 8-15
8–5(a) Worksheet for operational analysis of specific grades on two-lane highways (page 1).................. 8-16
8–5(b) Worksheet for operational analysis of specific grades on two-lane highways (page 2).................. 8-16
8–6 Use of third lane for passing lanes .................................................................................................... 8-19
8–7 Worksheet summarizing solution to calculation 1 ............................................................................ 8-22
8–8 Worksheet summarizing solution to calculation 2 ............................................................................ 8-22
8–9 Worksheet for calculation 4 (pages 1 and 2) .................................................................................... 8-25
9-1 Relationship among actual green, lost-time elements, extension of effective green, and
effective green................................................................................................................................. 9-4
9-2 Protected-plus-permitted signal phasing ............................................................................................ 9-5
9-3 Operational analysis procedure .......................................................................................................... 9-9
9-4 Input data needs for each analysis lane group .................................................................................. 9-10
xvii
FIGURES Page
9-5 Typical lane groups for analysis ........................................................................................................ 9-13
9-6 Permitted left turn............................................................................................................................... 9-19
9-7 Through-car equivalents, ELI, for permitted left turns (1 ) ................................................................ 9-20
9-8(a) Green time adjustments for protected-plus-permitted phasing: standard case and Case 2.............. 9-23
9-8(b) Green time adjustments for protected-plus-permitted phasing: Cases 3 and 4 ................................ 9-24
9-8(c) Green time adjustments for protected-plus-permitted phasing: Case 5 ............................................ 9-25
9-9 Critical lane group determination: leading and lagging green phase plan with exclusive left-turn
lanes................................................................................................................................................. 9-25
9-10 Critical lane group determination: leading and lagging green phase plan with addition of permit-
ted left turn in Phase 2B ................................................................................................................ 9-26
9-11 Critical lane group determination: complex multiphase signal......................................................... 9-27
9-12 Queue accumulation polygons............................................................................................................ 9-31
9-13 Worksheet information flow............................................................................................................... 9-35
9-14 Input Module Worksheet .................................................................................................................... 9-36
9-15 Volume Adjustment Module Worksheet............................................................................................ 9-39
9-16 Saturation Flow Rate Module Worksheet.......................................................................................... 9-40
9-17 Supplemental Worksheet for Permitted Left Turns: Multilane Approach........................................ 9-41
9-18 Supplemental Worksheet for Permitted Left Turns: Single-Lane Approach.................................... 9-42
9-19 Capacity Analysis Module Worksheet............................................................................................... 9-44
9-20 LOS Module Worksheet..................................................................................................................... 9-47
9-21 Supplemental Uniform Delay Worksheet for Left Turns from Exclusive Lanes with Primary
and Secondary Phases..................................................................................................................... 9-49
9-22 Planning Method Input Worksheet .................................................................................................... 9-51
9-23 Planning Method Lane Volume Worksheet....................................................................................... 9-52
9-24 Planning Method Signal Operations Worksheet................................................................................ 9-53
9-25 Planning method worksheet relationships.......................................................................................... 9-54
9-26 Alternative computations using operational analysis ........................................................................ 9-59
9-27 Input Module Worksheet for Calculation 1....................................................................................... 9-61
9-28 Volume Adjustment Module Worksheet for Calculation 1 .............................................................. 9-62
9-29 Saturation Flow Rate Module Worksheet for Calculation 1............................................................. 9-63
9-30(a) Supplemental left-turn worksheet for EB and WB approaches (multilane)..................................... 9-64
9-30(b) Supplemental left-turn worksheet for NB and SB approaches (single lane).................................... 9-65
9-31 Capacity Analysis Module Worksheet for Calculation 1.................................................................. 9-66
9-32 LOS Module Worksheet for Calculation 1........................................................................................ 9-67
9-33 Saturation Flow Adjustment Module Worksheet with no lane utilization factor for
Calculation 1 ................................................................................................................................... 9-68
9-34 LOS Module Worksheet with no lane utilization factor for Calculation 1...................................... 9-68
9-35 LOS Module Worksheet with timing modifications for Calculation 1 ............................................ 9-69
9-39 Lane Volume Worksheet for Calculation 2....................................................................................... 9-72
9-40 Signal Operations Worksheet for Calculation 2 ................................................................................ 9-73
9-41 Supplemental left-turn worksheet for Calculation 2.......................................................................... 9-74
9-44 Supplemental Uniform Delay Worksheet for Calculation 2 ............................................................. 9-76
9-45 Capacity Analysis Module Worksheet for protected-only phasing for Calculation 2 ..................... 9-76
9-46 LOS Module Worksheet for protected-only phasing for Calculation 2 ........................................... 9-77
9-47 Queue accumulation polygons for protected and protected-plus-permitted phasing for
Calculation 2 ................................................................................................................................... 9-77
9-48 Supplemental Worksheet for Permitted Left Turns: permitted-plus-protected (lagging) left-turn
phasing for Calculation 2 ............................................................................................................... 9-78
9-49 Capacity Analysis Module Worksheet for permitted-plus-protected (lagging) left-turn phasing
for Calculation 2 ............................................................................................................................. 9-78
9-50 Supplemental Uniform Delay Worksheet for permitted-plus-protected (lagging) left-turn phas-
ing for Calculation 2....................................................................................................................... 9-79
9-51 LOS Module Worksheet for permitted-plus-protected (lagging) left-turn phasing for
Calculation 2 ................................................................................................................................... 9-79
xviii
FIGURES Page
9-52 Queue accumulation polygons for protected and permitted-plus-protected phasing for Calcula-
tion 2 ............................................................................................................................................... 9-80
9-56 Supplemental Worksheet for Permitted Left Turns for Calculation 3.............................................. 9-83
9-59 Supplemental Uniform Delay Worksheet for Calculation 3 ............................................................. 9-85
9-60 LOS Module Worksheet with revised signal timing for Calculation 3............................................ 9-85
9-61 Planning Method Input Worksheet for Calculation 4 ....................................................................... 9-86
9-64 Lane Volume Worksheet for Calculation 4 with geometric modifications...................................... 9-89
9-65 Signal Operations Worksheet for Calculation 4 with geometric modifications............................... 9-90
9-67 Planning Method Input Worksheet for Calculation 5 ....................................................................... 9-91
9-70 Lane Volume Worksheet with additional EB right-turn lane for Calculation 5 .............................. 9-93
9-71 Signal Operations Worksheet with additional EB right-turn lane for Calculation 5 ....................... 9-93
9-72 Lane Volume Worksheet with NB and SB split-phase operation for Calculation 5 ....................... 9-94
9-73 Signal Operations Worksheet with NB and SB split-phase operation for Calculation 5 ................ 9-94
9-74 Lane Volume Worksheet with added NB and SB left-turn lanes for Calculation 5 ....................... 9-95
9-75 Signal Operations Worksheet with added NB and SB left-turn lanes for Calculation 5 ................ 9-95
I.9-1 Left-turn bay length versus turning volume ...................................................................................... 9-98
II.9-1 Phase plans for pretimed and traffic-actuated control....................................................................... 9-100
II.9-2 Dual-ring concurrent phasing scheme with assigned movements .................................................... 9-100
II.9-3 Worksheet 1: traffic-actuated control input data ............................................................................... 9-104
II.9-4 Queue accumulation polygon illustrating two methods of green time computation........................ 9-107
II.9-5 Convergence of green time computation by elimination of green time deficiency......................... 9-109
II.9-6 Queue accumulation polygon for permitted left turn from exclusive lane ...................................... 9-110
II.9-7 Queue accumulation polygon for permitted left turn from shared lane ........................................... 9-110
II.9-8 Queue accumulation polygon for protected-plus-permitted left-turn phasing with exclusive left-
turn lane .......................................................................................................................................... 9-111
II.9-9 Queue accumulation polygon for permitted-plus-protected left-turn phasing with exclusive left-
turn lane .......................................................................................................................................... 9-111
II.9-10 Worksheet 2: traffic-actuated timing computations........................................................................... 9-112
II.9-11 Traffic-actuated control data for multiphase example....................................................................... 9-114
II.9-12 LOS results for multiphase example.................................................................................................. 9-114
III.9-1 Field intersection control delay worksheet ........................................................................................ 9-118
III.9-2 Sample application of intersection control delay worksheet............................................................. 9-120
III.9-3 Sample application with residual queue at end ................................................................................. 9-121
IV.9-1 Field Saturation Flow Rate Study Worksheet ................................................................................... 9-123
VI.9-1 Case III: supplemental delay with initial oversaturation demand clearing in T. [Supplemental de-
lay per vehicle (d3) in seconds = 1,800Qbt/cT.]............................................................................. 9-139
VI.9-2 Case IV: supplemental delay with initial oversaturation demand decreasing in T. [Supplemental
delay per vehicle (d3) in seconds = 3,600Qb/c − 1,800T[1 − Min(1, X).].................................... 9-139
VI.9-3 Case V: supplemental delay with initial oversaturation demand increasing in T. [Supplemental
delay per vehicle (d3) in seconds = 3,600Qb/c.] ............................................................................ 9-139
VI.9-4 Demand profile for multiple-period analysis (15-min periods) ........................................................ 9-141
VI.9-5 Delay model components for multiple-period analysis..................................................................... 9-143
10-1 Traffic streams at TWSC intersection: (a) four-leg intersection; (b) T-intersection........................ 10-6
10-2 Definition and computation of conflicting volumes.......................................................................... 10-8
10-3 Potential capacity, two-lane roadway................................................................................................. 10-12
10-4 Potential capacity, four-lane roadway................................................................................................ 10-13
10-5 Adjustment to major left-turn, minor-through impedance factor (3) ............................................... 10-14
10-6 Platoon dispersion from upstream signalized intersections............................................................... 10-17
10-7 Upstream signalized intersection........................................................................................................ 10-17
xix
FIGURES Page
10-8 Platoon dispersion model [adapted from Bonneson and Fitts (12)]................................................. 10-18
10-9 Various platoon overlap cases: best case—platoons completely overlap so unplatooned period
is maximum; worst case—platoons alternate so unplatooned period is minimum; average
case—one-half of subordinate platoon is subsumed by dominant platoon .................................. 10-19
10-10 Intersection with two-stage gap acceptance process ......................................................................... 10-20
10-11 Capacity approximation at intersections with flared minor-street approach .................................... 10-21
10-12 Estimation of 95th-percentile queue length ....................................................................................... 10-22
10-13 Average control delay......................................................................................................................... 10-24
10-14 Queue-versus-delay relationship......................................................................................................... 10-26
10-15 TWSC intersection capacity and LOS computational procedures .................................................... 10-28
10-16 Traffic volumes for Sample Calculation A1...................................................................................... 10-47
10-21 Definition of intersection approaches ................................................................................................ 10-59
10-22 Saturation headway conditions for Vehicle 2.................................................................................... 10-59
10-23 Case 1: vehicles on subject approach only........................................................................................ 10-60
10-24 Case 2: vehicles on subject and opposing approaches...................................................................... 10-60
10-25 Case 3: vehicles on subject and conflicting approaches ................................................................... 10-60
10-26 Case 4: vehicles on subject and two other approaches..................................................................... 10-61
10-27 Case 5: vehicles on all approaches .................................................................................................... 10-61
10-28 Two-phase operation analogy............................................................................................................. 10-61
10-29 Four-phase operation analogy ............................................................................................................ 10-62
10-30 Configuration for Formulation 1 ........................................................................................................ 10-62
10-31 Configuration for Formulation 2 ........................................................................................................ 10-63
10-32 Flow for AWSC procedures............................................................................................................... 10-68
10-33 Traffic volumes for Sample Calculation B1...................................................................................... 10-76
10-37 Basic roundabout ................................................................................................................................ 10-82
10-38 Analysis of one roundabout leg ......................................................................................................... 10-83
10-39 Roundabout approach capacity........................................................................................................... 10-84
10-40 Flow stream definitions ...................................................................................................................... 10-85
10-41 Traffic volumes for Sample Calculation C1...................................................................................... 10-87
10-42 Worksheet for Sample Calculation C1 .............................................................................................. 10-88
10-43 Sample Calculation C1 capacity and volume analysis...................................................................... 10-89
11-1 Typical time-space trajectories of vehicles on one-lane arterial segment ........................................ 11-3
11-2 Arterial LOS method .......................................................................................................................... 11-5
11-3 Design categories ................................................................................................................................ 11-7
11-4 Types of segments .............................................................................................................................. 11-9
11-5 Worksheet for summary of arterial intersection delay estimates...................................................... 11-13
11-6 Worksheet for computation of arterial level of service .................................................................... 11-14
11-7 Speed profile by arterial section ........................................................................................................ 11-15
11-8 Arterial LOS calculation process ....................................................................................................... 11-17
11-9 Calculation 2, description: using worksheet for summary of arterial intersection delay estimates 11-18
11-10 Calculation 2, description: using worksheet for summary of arterial level of service .................... 11-18
11-11 Calculation 2, solution: using worksheet for summary of arterial intersection delay estimates ..... 11-21
11-12 Calculation 2, solution: using worksheet for computation of arterial level of service.................... 11-22
11-13 Speed profile for Calculation 2, southbound traffic.......................................................................... 11-23
11-14 Calculation 3, description: using worksheet for summary of arterial intersection delay estimates 11-24
11-17 Speed profile for Calculation 3, northbound traffic .......................................................................... 11-27
11-18 Sample calculation speed as a function of arterial flow rate............................................................ 11-28
11-19 Calculation 5 speed as a function of arterial flow rate on two different segment lengths.............. 11-28
11-21 Speed profile for Calculation 6 .......................................................................................................... 11-31
xx
FIGURES Page
11-24 Speed profile for Calculation 7 .......................................................................................................... 11-34
11-25 Arterial geometry for Calculation 10................................................................................................. 11-34
11-26 Calculation 10, solution: using worksheet for summary of arterial intersection delay estimates ... 11-35
11-27 Calculation 10, solution: using worksheet for computation of arterial level of service.................. 11-36
11-28 Arterial geometry for Calculation 11................................................................................................. 11-37
11-29 Calculation 11, solution: using worksheet for summary of arterial intersection delay estimates ... 11-38
11-30 Calculation 11, solution: using worksheet for computation of arterial level of service.................. 11-39
11-31 Speed profile for Calculation 11, eastbound traffic .......................................................................... 11-40
12–1 Example of freeway person-capacity ................................................................................................. 12-5
12–2 The two-dimensional nature of transit level of service as related to transit capacity ..................... 12-7
12–3 Bus stop capacity related to dwell times and loading positions....................................................... 12-22
12–4 Typical CBD busway line-haul passenger volumes .......................................................................... 12-27
13–1 Relationships between pedestrian speed and density ........................................................................ 13-4
13–2 Relationships between pedestrian flow and space............................................................................. 13-4
13–3 Relationships between pedestrian speed and flow ............................................................................ 13-4
13–4 Relationships between pedestrian speed and space........................................................................... 13-5
13–5 Preemption of walkway width............................................................................................................ 13-5
13–6 Typical free-flow walkway speed distribution .................................................................................. 13-7
13–7 Cross-flow traffic—probability of conflict ........................................................................................ 13-8
13–8 Illustration of walkway levels of service........................................................................................... 13-9
13–9 Minute-by-minute variations in pedestrian flow................................................................................ 13-10
13–10 Relationship between platoon flow and average flow ...................................................................... 13-11
13–11 Levels of service for queuing areas ................................................................................................... 13-12
13–12 Pedestrian movements at a street corner............................................................................................ 13-13
13–13 Worksheet for walkway analysis........................................................................................................ 13-15
13–14 Illustration of solution to walkway problem...................................................................................... 13-16
13–15 Intersection corner geometrics and pedestrian movements............................................................... 13-17
13–16 Intersection corner condition 1—minor street crossing .................................................................... 13-18
13–17 Intersection corner condition 2—major street crossing..................................................................... 13-19
13–18 Worksheet for crosswalk analysis ...................................................................................................... 13-20
13–19 Worksheet for street corner analysis.................................................................................................. 13-21
13–20 Worksheet for street corner analysis of sample calculation.............................................................. 13-23
13–21 Worksheet for crosswalk analysis of sample calculation.................................................................. 13-25
14–1 Illustration of right-turn conflicts with bicycles and pedestrians...................................................... 14-2
PHOTOGRAPHS
Vehicles shying away from both roadside and median barriers ............................................................................ 3-6
Ideal conditions of lane width and lateral clearance .............................................................................................. 3-6
Formation of large gaps in front of slow-moving trucks climbing the grade ....................................................... 3-7
Formation of large gaps in front of trucks or other heavy vehicles on relatively level terrain ........................... 3-8
LOS A ...................................................................................................................................................................... 3-12
LOS B....................................................................................................................................................................... 3-12
LOS C....................................................................................................................................................................... 3-12
LOS D ...................................................................................................................................................................... 3-13
LOS E....................................................................................................................................................................... 3-13
LOS F ....................................................................................................................................................................... 3-13
Divided multilane highway in a rural environment................................................................................................ 7-3
Divided multilane highway in a suburban environment......................................................................................... 7-3
Undivided multilane highway in a rural environment............................................................................................ 7-3
Undivided multilane highway in a suburban environment..................................................................................... 7-3
Bridge pier in center of normally undivided suburban multilane highway........................................................... 7-11
Inadequate shoulder and obstructions on roadway ................................................................................................. 7-11
Ideal divided multilane highway ............................................................................................................................. 7-11
Undivided multilane highway with no obstructions ............................................................................................... 7-11
Typical views of two-lane two-way highways in rural environments................................................................... 8-3
Typical use of paved shoulders—slow-moving vehicle uses shoulder of a two-lane rural highway, permitting
faster vehicles to pass .......................................................................................................................................... 8-19
xxi
PHOTOGRAPHS Page
Design categories: top left, typical high speed design; top right, typical suburban design; bottom left, typical
intermediate design; bottom right, typical urban design .................................................................................... 11-7
TABLES
1–1 Organization of manual ...................................................................................................................... 1-2
1–2 Primary measures of effectiveness for level of service definition.................................................... 1-5
1–3 Adjustment factors used for analyses ................................................................................................ 1-8
1–4 Analysis techniques ............................................................................................................................ 1-9
2–1 Maximum annual average daily traffic reported on selected Interstate routes (1990) .................... 2-12
2–2 Reported maximum one-way hourly volumes on selected freeways................................................ 2-13
2–3 Reported maximum lane volumes on selected freeways .................................................................. 2-14
2–4 Reported maximum one-way volumes for selected multilane highways ......................................... 2-15
2–5 Reported maximum volumes on selected two-lane rural highways ................................................. 2-15
2–6 Reported maximum one-way volumes on selected urban arterials .................................................. 2-16
2–7 Directional distribution characteristics............................................................................................... 2-22
2–8 Observed values of K and D on selected freeways and expressways .............................................. 2-23
2–9 Lane distribution by vehicle type....................................................................................................... 2-23
2–10 National spot speed trends for 55-mph facilities............................................................................... 2-25
2–11 Average speed by day vs. night and lane in mph ............................................................................. 2-27
2–12 Average speed by lane in mph........................................................................................................... 2-28
2–13 Observed saturation flow rates at signalized intersections................................................................ 2-32
2–14 Capacity by facility type .................................................................................................................... 2-34
3-1 LOS criteria for basic freeway sections............................................................................................. 3-11
3-2 Passenger-car equivalents on extended general freeway segments .................................................. 3-16
3-3 Passenger-car equivalents for trucks and buses on specific upgrades.............................................. 3-17
3-4 Passenger-car equivalents for recreational vehicles on specific upgrades........................................ 3-18
3-5 Passenger-car equivalents for trucks and buses on specific downgrades......................................... 3-18
3-6 Adjustment factors for lane width...................................................................................................... 3-21
3-7 Adjustment factors for right-shoulder lateral clearance .................................................................... 3-21
3-8 Adjustment factors for number of lanes ............................................................................................ 3-22
3-9 Adjustment factors for interchange density ....................................................................................... 3-22
4–1 Configuration type vs. minimum number of required lane changes ................................................ 4-4
4–2 Parameters affecting weaving area operation .................................................................................... 4-5
4–3 Constants of prediction for weaving intensity factor, W................................................................... 4-7
4–4 Criteria for unconstrained vs. constrained operation of weaving areas............................................ 4-8
4–5 Limitations on weaving sections ........................................................................................................ 4-8
4–6 Level-of-service criteria for weaving areas ....................................................................................... 4-9
4–7 Results of weaving analysis: Sample Calculation 6.......................................................................... 4-19
5–1 Capacity values for merge and diverge areas.................................................................................... 5-7
5–2 Level-of-service criteria for ramp-freeway junction areas of influence ........................................... 5-7
5–3 Models for prediction of density in ramp influence areas ................................................................ 5-8
5–4 Models for prediction of speed in ramp influence areas .................................................................. 5-8
5–5 Determination of V5 for right-hand ramps on 10-lane freeways....................................................... 5-12
5–6 Approximate capacity of ramp roadways .......................................................................................... 5-14
6–1 Measured average work-zone capacities............................................................................................ 6-10
6–2 Summary of observed capacities for some typical operations.......................................................... 6-11
6–3 Capacity of long-term construction sites with portable concrete barriers ........................................ 6-12
7–1 Level-of-service criteria for multilane highways............................................................................... 7-8
7–2 Adjustment for median type............................................................................................................... 7-10
7–3 Adjustment for lane width.................................................................................................................. 7-10
7–4 Adjustment for lateral clearance ........................................................................................................ 7-10
7–5 Access-point density adjustment ........................................................................................................ 7-10
7–6 Number of access points for general development environments .................................................... 7-11
7–7 Passenger-car equivalents on extended general multilane highway segments ................................. 7-12
7–8 Passenger-car equivalents for trucks and buses on uniform upgrades ............................................. 7-13
7–9 Passenger-car equivalents for recreational vehicles on uniform upgrades ....................................... 7-13
7–10 Passenger-car equivalents for trucks on downgrades ........................................................................ 7-14
7–11 Service flow rates in vehicles per lane for use in planning analysis ............................................... 7-19
8–1 Level-of-service criteria for general two-lane highway segments .................................................... 8-5
8–2 Level-of-service criteria for specific grades ...................................................................................... 8-6
xxii
TABLES Page
8–3 Peak-hour factors for two-lane highways based on random flow .................................................... 8-7
8–4 Adjustment factors for directional distribution on general terrain segments ................................... 8-9
8–5 Adjustment factors for the combined effect of narrow lanes and restricted shoulder width .......... 8-9
8–6 Average passenger-car equivalents for trucks, RVs, and buses on two-lane highways over gen-
eral terrain segments....................................................................................................................... 8-9
8–7 Values of v/c ratio vs. speed, percent grade, and percent no passing zones for specific grades.... 8-10
8–8 Adjustment factor for directional distribution on specific grades .................................................... 8-11
8–9 Passenger-car equivalents for specific grades on two-lane rural highways, E and Eo..................... 8-12
8–10 Maximum AADTs vs. level of service and type of terrain for two-lane rural highways ............... 8-14
8–11 Spacing of passing lanes on two-lane highways ............................................................................... 8-20
8–12 Length of turnouts on two-lane highways ......................................................................................... 8-21
9-1 Level-of-service criteria for signalized intersections......................................................................... 9-7
9-2 Relationship between arrival type and platoon ratio (Rp) ................................................................. 9-11
9-3 Default values for use in operational and planning analyses............................................................ 9-12
9-4 Default lane utilization adjustment factors ........................................................................................ 9-12
9-5 Adjustment factor for average lane width (fw)................................................................................... 9-14
9-6 Adjustment factor for heavy vehicles (fHV)........................................................................................ 9-14
9-7 Adjustment factor for approach grade (fg) ......................................................................................... 9-14
9-8 Adjustment factor for parking (fp)...................................................................................................... 9-15
9-9 Adjustment factor for bus blockage (fbb)............................................................................................ 9-15
9-10 Adjustment factor for area type (fa) ................................................................................................... 9-15
9-11(a) Adjustment factor for right turns (fRT): formulas............................................................................... 9-15
9-11(b) Adjustment factor for right turns: factors .......................................................................................... 9-16
9-12 Adjustment factor for left turns (fLT).................................................................................................. 9-17
9-13 Progression adjustment factor (PF).................................................................................................... 9-29
9-14 Recommended k values for lane groups under actuated and pretimed control................................ 9-29
9-15 Intersection status criteria for signalized intersection planning analysis.......................................... 9-32
9-16 Shared-lane left-turn adjustment computations for planning-level analysis..................................... 9-54
9-17 Phase plan summary for planning analysis........................................................................................ 9-57
9-18 Service flow rate solutions for Calculation 6 .................................................................................... 9-96
I.9-1 Left-turn bay length adjustment factors............................................................................................. 9-98
II.9-1 Comparison of traffic-actuated controller settings for multiphase example..................................... 9-115
III.9-1 Acceleration-deceleration delay correction factor ............................................................................. 9-119
VI.9-1 Selection of delay model variables by case....................................................................................... 9-140
10-1 Critical gaps tc and follow-up times tf for passenger cars at TWSC intersections .......................... 10-11
10-2 Relative pedestrian-vehicle hierarchy ................................................................................................ 10-15
10-3 Pedestrian impedance factors ............................................................................................................. 10-15
10-4 Platoon dispersion factor, a (12, 1) ................................................................................................... 10-18
10-5 Proportion of study period for each flow regime for average case.................................................. 10-19
10-6 Proportion of study period unblocked for each minor movement for average case........................ 10-19
10-7 Level-of-service criteria...................................................................................................................... 10-25
10-8 Example left-turn delay calculation ................................................................................................... 10-26
10-9 Impedance and capacity calculations ................................................................................................. 10-48
10-10 Delay, queue length, and level of service.......................................................................................... 10-49
10-12 Conflicting flows ................................................................................................................................ 10-50
10-13 Lane usage by approach ..................................................................................................................... 10-50
10-14 Critical gap and follow-up time by movement.................................................................................. 10-51
10-17 Upstream signal parameters................................................................................................................ 10-52
10-18 Computation 1: queue clearance time (A5a) ..................................................................................... 10-52
10-19 Computation 2: proportion of TWSC intersection time blocked (A5b)........................................... 10-53
10-20 Computation 3: platoon events and proportion unblocked (A5c)..................................................... 10-53
10-21 Computation 4 (single-stage process) (A5d) ..................................................................................... 10-54
10-22 Computation 5 (single-stage process) (A5e)...................................................................................... 10-54
10-25 Lane usage by approach ..................................................................................................................... 10-54
xxiii
TABLES Page
10-26 Shared-lane capacities......................................................................................................................... 10-55
10-28 Two-stage gap acceptance: Step 3 ..................................................................................................... 10-56
10-29 Two-stage gap acceptance: Step 4 ..................................................................................................... 10-57
10-30 Shared-lane capacities......................................................................................................................... 10-57
10-31 Flared minor-street approach calculations ......................................................................................... 10-58
10-33 Probability of degree-of-conflict case ................................................................................................ 10-64
10-34 Degree-of-conflict cases for two-lane approach intersections .......................................................... 10-65
10-35 Degree-of-conflict cases for three-lane approach intersections ........................................................ 10-65
10-36 Number of vehicles by approach for degree-of-conflict cases, multilane AWSC intersections
(two-lane approach intersections)................................................................................................... 10-65
10-37 Occupied lane combinations for degree-of-conflict cases, multilane AWSC intersections (two-
lane approach intersections) ........................................................................................................... 10-66
10-38 Level-of-service criteria...................................................................................................................... 10-67
10-39 Geometry group .................................................................................................................................. 10-68
10-40 Saturation headway adjustment factors by geometry group ............................................................. 10-68
10-41 Probability of aj .................................................................................................................................. 10-69
10-42 Saturation headway values by case and geometry group.................................................................. 10-75
10-43 Departure headways............................................................................................................................ 10-78
10-44 Capacity, delay, and level of service ................................................................................................. 10-78
10-45 Saturation headways ........................................................................................................................... 10-79
10-46 Departure headways............................................................................................................................ 10-79
10-47 Service times....................................................................................................................................... 10-79
10-49 Saturation headway............................................................................................................................. 10-80
10-50 Degrees of utilization and departure headways................................................................................. 10-80
10-51 Service times....................................................................................................................................... 10-80
10-53 Critical gap and follow-up time ......................................................................................................... 10-84
10-54 Effects of changes in critical gap and move-up time........................................................................ 10-89
11-1 Arterial levels of service .................................................................................................................... 11-4
11-2 Aid in establishing arterial classification........................................................................................... 11-8
11-3 Arterial classification according to functional and design categories .............................................. 11-8
11-4 Segment running time per mile.......................................................................................................... 11-9
11-5 Relationship between arrival type and platoon ratio (Rp) ................................................................. 11-11
11-6 Uniform delay (d1) progression adjustment factor (PF).................................................................... 11-11
11-7 Recommended k-values for lane groups under actuated and pretimed control................................ 11-12
11-8 Recommended I-values for lane groups with upstream signals........................................................ 11-12
11-9 Computations for Sample Calculation 4 ............................................................................................ 11-27
11-10 Computations for Sample Calculation 5 ............................................................................................ 11-29
11-11 Input data for Sample Calculation 10 ................................................................................................ 11-34
11-12 Input data for Sample Calculation 11 ................................................................................................ 11-37
12–1 Peak-hour use of public transport by persons entering or leaving the central business district ..... 12-3
12–2 Important terms in transit capacity .................................................................................................... 12-3
12–3 Factors that influence transit capacity................................................................................................ 12-6
12–4 Characteristics of typical transit vehicles—United States and Canada ............................................ 12-8
12–5 Passenger loading standards and levels of service for bus transit vehicles ..................................... 12-8
12–6 Passenger loading standards and levels of service for urban rail vehicles ...................................... 12-9
12–7 Typical space requirements for seated and standing passengers ...................................................... 12-9
12–8 Passenger car equivalency of urban buses at signalized intersections ............................................. 12-11
12–9 Passenger boarding and alighting times related to service conditions ............................................. 12-12
12–10 Typical bus passenger boarding and alighting service times for selected bus types and door
configurations.................................................................................................................................. 12-13
12–11 Suggested bus flow service volumes for planning purposes ............................................................ 12-13
12–12 Suggested bus passenger service volumes for planning purposes .................................................... 12-14
12–13 Observed peak-hour passenger volumes on U.S. and Canadian rapid transit systems.................... 12-15
12–14 Observed peak-hour passenger volumes on street car and light rail systems in United States
and Canada...................................................................................................................................... 12-16
xxiv
TABLES Page
12–15 Typical rail transit capacities.............................................................................................................. 12-17
12–16 Estimated maximum capacity of bus stops........................................................................................ 12-20
12–17 Levels of service for bus stops .......................................................................................................... 12-21
12–18 Typical service levels, single stop, no passing.................................................................................. 12-21
12–19 Efficiency of multiple linear bus berths ............................................................................................ 12-21
12–20 Estimated capacity of on-line bus stops by number of berths.......................................................... 12-22
12–21 Bus berth passenger capacity equations and illustrative examples................................................... 12-24
12–22 Maximum load point hourly passengers per effective berth at the busiest station—uninterrupted
flow conditions................................................................................................................................ 12-25
12–23 Maximum load point hourly passengers per effective berth at busiest station—interrupted flow
conditions ........................................................................................................................................ 12-26
12–24 Illustrative bus capacity guidelines for CBD busways...................................................................... 12-27
12–25 Busway service volumes at maximum load points ........................................................................... 12-27
12–26 Typical arterial street bus service volumes at maximum load point ................................................ 12-28
12–27 Berth requirements at bus stops ......................................................................................................... 12-29
12–28 Significant examples of bus priority treatments—United States and Canada.................................. 12-31
12–29 Summary of illustrative planning guidelines for bus priority treatments......................................... 12-33
12–30 Summary and applications of transit capacity equations .................................................................. 12-35
12–31 Basic transit capacity variables .......................................................................................................... 12-37
12–32 Summary and applications of transit capacity figures and tables..................................................... 12-38
12–33 Guidelines for application—planning parameters.............................................................................. 12-39
12–34 Person-capacity of a freeway lane for varying bus volumes ............................................................ 12-40
12–35 Anticipated peak-hour buses at transit center.................................................................................... 12-43
12–36 Bus berth requirements, year-1985 .................................................................................................... 12-43
12–37 Bus berth requirements, year-2000 .................................................................................................... 12-44
I.12–1 Reported theoretical bus lane capacities ............................................................................................ 12-49
I.12–2 Observed peak-hour bus volumes on streets and freeways .............................................................. 12-50
I.12–3 Observed bus volumes on urban limited access facilities 1972–1976 conditions ........................... 12-51
I.12–4 Peak-hour bus volumes on urban arterials, 1972–1976 conditions .................................................. 12-52
I.12–5 Observed bus volumes on urban arterials, 1978–1984 ..................................................................... 12-54
I.12–6 Observed passengers at major bus terminals..................................................................................... 12-54
I.12–7 Observed peak bus berth volumes and flow rates at bus terminals ................................................. 12-55
II.12–1 Observed peak-hour passenger volumes on streetcar and LRT lines—Europe ............................... 12-55
II.12–2 Rapid transit car and train capacities................................................................................................. 12-56
II.12–3 Theoretical rail rapid transit equations .............................................................................................. 12-58
III.12–1 Typical CBD service times per passenger......................................................................................... 12-59
III.12–2 Observed rail transit station dwell times, 1980 ................................................................................. 12-59
III.12–3 Bus boarding and alighting times in selected urban areas................................................................ 12-60
III.12–4 Means and variances of observed passenger service time distributions........................................... 12-60
13–1 Observed pedestrian flow rates in urban areas.................................................................................. 13-1
13–2 Fixed obstacle width adjustment factors for walkways .................................................................... 13-6
13–3 Pedestrian level of service on walkways ........................................................................................... 13-8
14–1 Passenger-car equivalent for bicycles ................................................................................................ 14-2
14–2 Reported one-way and two-way high volumes of bicycle facilities................................................. 14-3

chapter 1
INTRODUCTION, CONCEPTS, AND APPLICATIONS
CONTENTS
i. introduction .......................................................................................................................................................................... 1-1

Importance of Capacity ......................................................................................................................................................... 1-1
Purpose of Manual ................................................................................................................................................................ 1-2
Scope of Manual.................................................................................................................................................................... 1-2
Organization of Manual ........................................................................................................................................................ 1-3
ii. concepts .................................................................................................................................................................................. 1-3
Capacity and Levels of Service ............................................................................................................................................ 1-3
Capacity ............................................................................................................................................................................ 1-3
Levels of Service.............................................................................................................................................................. 1-4
Factors Affecting Capacity and Level of Service ................................................................................................................ 1-5
Ideal Conditions ............................................................................................................................................................... 1-5
Roadway Conditions ........................................................................................................................................................ 1-5
Traffic Conditions ............................................................................................................................................................ 1-6
Control Conditions ........................................................................................................................................................... 1-6
Technology ....................................................................................................................................................................... 1-7
Summary ................................................................................................................................................................................ 1-7
iii. applications............................................................................................................................................................................ 1-9
Models of Traffic Flow......................................................................................................................................................... 1-9
Levels of Analysis................................................................................................................................................................. 1-9
Operational Analysis ........................................................................................................................................................ 1-9
Design............................................................................................................................................................................... 1-10
Planning Analysis............................................................................................................................................................. 1-10
Precision................................................................................................................................................................................. 1-10
Field Data .............................................................................................................................................................................. 1-10
Summary ................................................................................................................................................................................ 1-11
I. INTRODUCTION
This publication is the third update of the third edition of chapters and interim documents have been superseded by the
the Highway Capacity Manual (HCM). The first manual was updated chapters identified in Table 1-1.
published in 1950 by the then Bureau of Public Roads as a
guide to the design and operational analysis of highway facilities.
In 1965, the then Highway Research Board, under the guidance
of its Highway Capacity Committee, published the second edition. IMPORTANCE OF CAPACITY
The third edition, published in 1985, reflected more than 2
decades of comprehensive research conducted by a variety of
research agencies under the sponsorship of a number of organiza- The capacity of a transportation facility reflects its ability to
tions, primarily the National Cooperative Highway Research accommodate a moving stream of people or vehicles. It is a mea-
Program and the Federal Highway Administration. Its develop- sure of the supply side of transportation facilities. Level of service
ment was guided by the Transportation Research Board Commit- is a measure of the quality of flow.
tee on Highway Capacity and Quality of Service. As the result Capacity and level-of-service (LOS) estimates are needed for
of continuing research in capacity, the third edition of HCM most traffic engineering and transportation planning decisions and
has been updated again to incorporate numerous findings of actions.
studies conducted since it was published. Previously published Capacity analysis addresses questions such as the following:
1-1 Updated December 1997

1-2 principles of capacity
Table 1-1. Organization of Manual service for individual key elements of transportation systems (i.e.,
chapter description/facility type update for various types of transportation facilities). They have been de-
veloped from a wide range of research studies conducted during
1 Introduction, Concepts, and 1997
Applications
the past 45 years. They reflect North American operating experi-
2 Traffic Characteristics 1994 ence and may not be representative of traffic, transit, and pedestrian
operations in other parts of the world.
Uninterrupted Flow Facilities
3 Basic Freeway Sections 1997
4 Weaving Areas 1997 SCOPE OF MANUAL
5 Ramps and Ramp Junctions 1997
6 Freeway Systems 1997
This manual presents operational, design, and planning capacity
7 Multilane Rural and 1997
Suburban Highways analysis techniques for a broad range of transportation facilities.
8 Two-Lane Highways 1985 It provides procedures for analyzing streets and highways, bus and
rail transit, and pedestrian and bicycle facilities. Note that the
Interrupted Flow Facilities
manual does not address systems of facilities or overall mobility.
9 Signalized Intersections 1997 Nevertheless, the results of these analyses may be used as informa-
10 Unsignalized Intersections 1997 tion for the assessment of broad system issues, such as congestion
11 Arterial Streets 1997 management.
12 Transit Capacity 1985
Facilities are classified into two categories of flow: uninter-
13 Pedestrians 1985
14 Bicycles 1985 rupted and interrupted.
Uninterrupted flow facilities have no fixed elements, such as
NOTE: The Metric Analysis Reference Guide (MARG) is available to assist users
of the Highway Capacity Manual in the conversion of English units. This guide traffic signals, that are external to the traffic stream and may
includes correct metric symbols and provides tables, figures, formulas, and interrupt the traffic flow. Traffic flow conditions result from the
worksheets for metric calculations. interactions among vehicles in the traffic stream and between vehi-
cles and the geometric and environmental characteristics of the
T What is the quality of service provided by an existing facility roadway.
during peak periods, and how much traffic increase can be Interrupted flow facilities have fixed elements that may interrupt
tolerated? the traffic flow. Such elements include traffic signals, stop signs,
T What types of roadway or transit facilities are needed to ac- and other types of controls. These devices cause traffic to stop
commodate a given level of person or vehicle flow? periodically (or slow significantly), irrespective of how much traf-
T What lane configurations are needed for various levels of fic exists.
average daily traffic on freeways or arterial roads? ‘‘Uninterrupted flow’’ and ‘‘interrupted flow’’ are terms that
T What highway or street designs (and hence capacities) are describe the type of facility, not the quality of traffic flow at any
needed to serve a planned development? given time. Thus a freeway experiencing extreme congestion is
T How many buses or railcars are needed to serve peak direction still an uninterrupted flow facility because the causes of congestion
flow at the maximum load point, and can these transit vehicles be are internal to the traffic stream.
passed through the busiest station or other point of constriction? Freeways and their components operate under the purest form
T How wide must the sidewalk be on a street with high pedes- of uninterrupted flow. Not only are there no fixed interruptions to
trian activity, and would the holding space at street corners of a traffic flow, but access is controlled and limited to ramp locations.
signalized intersection be sufficient? Multilane highways and two-lane highways may also operate under
uninterrupted flow in long segments between points of fixed inter-
Four primary traffic engineering activities depend on capacity ruptions. In general, where signal spacing exceeds 2 mi, uninter-
and LOS analyses: rupted flow may exist between the signals. Where signal spacing
1. When new facilities are planned or existing facilities are to is less than 2 mi, the facility is classified as an arterial, and flow
be expanded, their size in terms of width or number of lanes must is considered to be interrupted. On multilane and two-lane high-
be determined. ways, it is often necessary to examine points of fixed interruption
2. When existing facilities are considered for upgrading, either as well as uninterrupted flow segments.
by widening or by traffic operational changes, their operational The analysis of interrupted flow facilities must account for the
characteristics and service levels must be assessed. impact of fixed interruptions. A traffic signal, for example, limits
3. When new developments are planned, capacity and LOS the time available to various movements in an intersection. Capac-
analyses are needed to identify necessary traffic and roadway ity is limited not only by the physical space provided, but also by
changes and to help define cost responsibilities. the time of use that is available to various component movements
4. Studies of operating conditions and levels of service provide in the traffic stream.
base values for determining changes in road-user costs, fuel con- The procedures in this manual do not explicitly address opera-
sumption, air pollutant emissions, and noise. tions of closely spaced signalized intersections. Under such condi-
tions, several unique characteristics must be considered, including
spillback potential from the downstream intersection to the up-
PURPOSE OF MANUAL stream intersection, effects of downstream queues on upstream
saturation flow rate, and unusual platoon dispersion or compression
The parameters and procedures in this manual provide a system- between intersections. An example of such closely spaced opera-
atic and consistent basis for assessing the capacity and quality of tions is signalized ramp terminals at urban interchanges. Queue
introduction, concepts, and applications 1-3
interactions between closely spaced intersections may seriously on the overall operation. In Chapter 11, Arterial Streets, the analy-
distort the procedures in this manual. sis permits an evaluation of the performance of an arterial roadway
Transit, pedestrian, and bicycle flows are generally considered through a series of intersections.
to be interrupted. Uninterrupted flow may exist under certain cir- Each of the procedural chapters is generally organized in four
cumstances, such as in a long busway without stops or a long distinct parts:
pedestrian corridor. However, in most situations, capacity is lim- 1. Introduction: The basic characteristics, concepts, and philos-
ited by stops elsewhere along the facility. ophies of capacity analysis as applied to the subject facility are
described.
2. Methodology: The basic components of the analysis proce-
ORGANIZATION OF MANUAL dure to be applied to the specific facility are presented. Equations
and tabular and graphic information needed to complete the analy-
The third edition of HCM contains 14 chapters. This third update sis are included.
of the third edition involves nine of these chapters. Table 1-1 3. Procedures for application: Step-by-step instructions for
shows how the various chapters are organized according to facility applying capacity analysis computations are presented. Procedures
type and identifies those that have been updated. are specified for operational analysis, design, and planning, al-
In Chapter 1, the role and importance of capacity analysis are though not all chapters contain these three analysis levels. Work-
described, basic concepts are presented, and general guidelines for sheets are provided for most computational procedures and are
application are provided. explained in detail.
In Chapter 2, Traffic Characteristics, basic variables related to 4. Sample calculations: A variety of example applications,
capacity are identified and their values and relationships as ob- showing all computations required for analysis, and detailed dis-
served throughout North America are discussed. cussions of results and interpretations are presented. Sample calcu-
Chapters 3 through 14 are the basic procedural chapters of the lations are provided for the full range of potential applications in
manual. They are organized according to the facility types pre- each chapter.
sented in Table 1-1. Chapters 3 through 8 cover uninterrupted flow Many chapters have separate sections headed by the foregoing
facilities, with Chapters 3 through 6 treating freeways and their titles. In some chapters, sections are combined for clarity of presen-
components and Chapters 7 and 8 dealing with multilane and tation. Where sections are combined, section titles clearly indicate
two-lane highways, respectively. Chapters 9 through 14 focus on where material is located.
interrupted flow facilities and their components, including signal- The organization of the procedural chapters allows frequent us-
ized and unsignalized intersections, arterial streets, and transit, ers to focus on step-by-step instructions without having to read or
pedestrian, and bicycle facilities. scan an entire chapter. All users of this manual, however, should
Two chapters treat the roadways as operating systems, and the read the entire chapter being used at least once to become familiar
analyses focus on the effect of individual segments or components with the concepts, applications, and interpretations of the
on the overall performance and operation of the roadway. In procedures.
Chapter 6, Freeway Systems, the analysis reflects a coordinated As an additional convenience for frequent users, some chapters
evaluation of basic freeway segments, weaving areas, and ramp contain an appendix in which figures and worksheets are repro-
junctions. The results of the analysis show the effects of operations duced (some to a larger scale than that appearing in the text) for
in one component on other freeway segments or components and ease of use.
II. CONCEPTS
CAPACITY AND LEVELS OF SERVICE tions are defined for each type of facility and are related to amounts
of traffic that can be accommodated at each level.
A principal objective of capacity analysis is the estimation of The two principal concepts of this manual—capacity and levels
the maximum number of people or vehicles that can be accommo- of service—are defined in the following sections.
dated by a given facility in reasonable safety within a specified
time period. However, because facilities generally operate poorly
at or near capacity, they are rarely planned to operate in this range. Capacity
Accordingly, capacity analysis also provides a means of estimating
the maximum amount of traffic that can be accommodated by a
facility while prescribed operational qualities are maintained. The capacity of a facility is defined as the maximum hourly
Capacity analysis is, therefore, a set of procedures for estimating rate at which persons or vehicles can reasonably be expected to
the traffic-carrying ability of facilities over a range of defined traverse a point or uniform section of a lane or roadway during a
operational conditions. It provides tools for the analysis of existing given time period under prevailing roadway, traffic, and control
facilities and for the planning and design of improved or future conditions.
facilities. Vehicle capacity represents the maximum number of vehicles
The definition of operational criteria is accomplished by intro- that can pass a given point during a specified period under prevail-
ducing the concept of levels of service. Ranges of operating condi- ing roadway, traffic, and control conditions. This definition as-
sumes no influence of downstream traffic operation, such as perception by motorists and passengers. The descriptions of indi-
backing up of traffic over the analysis point. vidual levels of service characterize these conditions in terms of
Person capacity represents the maximum number of people that such factors as speed and travel time, freedom to maneuver, traffic
can pass a given point during a specified period under prevailing interruptions, and comfort and convenience.
conditions. It is commonly used in evaluating public transit ser- Six levels of service are defined for each type of facility for
vices, high-occupancy-vehicle lanes, or pedestrian facilities. Real- which analysis procedures are available. They are given letter
istic occupancy is a critical factor for transit and other vehicles. designations, from A to F, with LOS A representing the best
Several important points in the definition of capacity warrant operating conditions and LOS F the worst. Each level of service
clarification. represents a range of operating conditions.
1. Capacity is defined for prevailing roadway, traffic, and con- The volume of traffic that can be served under the stop-and-go
trol conditions, which should be reasonably uniform for any sec- conditions of LOS F is generally accepted as being lower than
tion of facility analyzed. Any change in the prevailing conditions that possible at LOS E; consequently, service flow rate E is the
will result in a change in the capacity of the facility. The definition value that corresponds to the maximum flow rate, or capacity, on
assumes that good weather, good pavement conditions, and no the facility. For most design or planning purposes, however, ser-
incidents exist. vice flow rates D or C are usually used because they ensure a
2. Capacity normally refers to a ‘‘point or uniform segment’’ more acceptable quality of service to facility users.
of the facility. Capacity analysis is conducted for segments of a Levels of service for uninterrupted and interrupted flow facilities
facility having uniform traffic, roadway, and control conditions. vary widely in terms of both the user’s perception of service quality
Because capacity depends on these factors, segments with different and the operational variables used to describe them. Chapters 3
prevailing conditions will have different capacities. Capacity of a through 13 of this manual contain detailed descriptions of the
route or system can be inferred from the analysis procedures but levels of service that are defined for each facility type.
is not explicitly discussed. The point or segment with the poorest
operating conditions often determines the overall levels of service Service Flow Rates
for the facility.
3. Capacity refers to a rate of vehicular or person flow during The procedures attempt to establish or predict the maximum
a specified period, which is most often a peak 15-min period. rate of flow that can be accommodated by various facilities at each
Capacity does not refer to the maximum volume that can be accom- level of service, except LOS F, for which flows are unstable. Thus
modated during an hour. This definition recognizes the potential each facility has five service flow rates, one for each level of
for substantial variations in flow during an hour and focuses the service (A through E).
analysis on intervals of maximum flow. The service flow rate is the maximum hourly rate at which
4. Capacity is defined on the basis of ‘‘reasonable expectancy.’’ persons or vehicles can reasonably be expected to traverse a point
That is, a stated capacity for a given facility is a rate of flow that or uniform section of a lane or roadway during a given period
can be repeatedly achieved during peak periods for which sufficient under prevailing roadway, traffic, and control conditions while a
demand exists and that can be achieved on facilities with similar designated level of service is maintained. The service flow rates
characteristics throughout North America. It is not the absolute max- are generally based on a 15-min period. Typically, the hourly flow
imum rate of flow ever observed on such a facility type. Driver char- rate is defined as four times the peak 15-min volume.
acteristics vary from region to region and the absolute maximum rate Note that service flow rates are discrete values, whereas levels
of flow may vary from day to day and from location to location. of service represent a range of conditions. Because the service
Occasionally, measured rates of flow at some locations will exceed flow rates are defined as maximums for each level of service, they
the defined capacity of the facility type. Such rates, however, are effectively define flow boundaries between the various levels of
usually not sustainable or cannot be achieved repeatedly. service.
5. Capacity may be defined in terms of persons per hour or
vehicles per hour, depending on the type of facility involved. The Measures of Effectiveness
concept of person flow is important in making strategic decisions
about transportation modes in heavily traveled corridors and in For each type of facility, levels of service are defined on the
defining the role of transit and high-occupancy-vehicle priority basis of one or more operational parameters that best describe the
treatments. The concepts of person capacity and person flow weigh operating quality for the facility type. Although the concept of
each type of vehicle in the traffic stream by the number of occu- level of service attempts to address a wide range of operating
pants they carry. For example, an arterial street carrying 600 vehi- conditions, limitations on data collection and availability make it
cles per hour with 1.5 persons per vehicle will have a movement impractical to treat the full range of operational parameters for
capability of 900 people per hour; 50 buses per hour, each with every type of facility. The parameters selected to define levels of
an average of 40 persons per bus, would carry 2,000 persons per service for each facility type are called measures of effectiveness
hour for a total person flow of 2,900 persons per hour. As the and represent available measures that best describe the quality of
number of transit vehicles in the traffic stream increases, the num- operation on the subject facility type. Table 1-2 presents the pri-
ber of vehicles that can pass a given point decreases, but the person mary measures of effectiveness used to define levels of service
flow may increase, although possibly at reduced service quality. for each facility type. Each level of service represents a range of
conditions, as defined by a range in the parameter(s) presented in
Levels of Service the table.
Effectiveness and LOS criteria are not defined for bicycles. The
The concept of levels of service uses qualitative measures that treatment of bicycles herein is limited to their impact on other
characterize operational conditions within a traffic stream and their vehicular flow at critical points in the street and highway system.
Table 1-2. Primary Measures of Effectiveness for way, traffic, or control conditions. Vehicle control and technology
LOS Definition represent conditions that change in the long term.
type of facility measure of effectiveness
Freeways Roadway Conditions
Basic freeway segments Density (pc/mi/ln)
Weaving areas Density (pc/mi/ln) Roadway factors include geometric conditions and design ele-
Ramp junctions Flow rates (pcph) ments. In some cases, these factors influence the capacity of a
Multilane highways Density (pc/mi/ln) road, whereas in others, the factors may affect a measure of effec-
Free-flow speed (mph)
tiveness, such as speed, while not affecting the capacity or maxi-
Two-lane highways Time delay (percent)
Signalized intersections Average control delay (sec/veh) mum flow rate that can be carried by the facility.
Unsignalized intersections Average control delay (sec/veh) Roadway factors include the following:
Arterials Average travel speed (mph)
Transit Load factor (pers/seat, veh/hr,
T The type of facility and its development environment.
people/hr) T Lane widths.
Pedestrians Space (sq ft/ped) T Shoulder widths and lateral clearances.
T Design speed.
T Horizontal and vertical alignments.
T Availability of queueing space at intersections.
FACTORS AFFECTING CAPACITY AND LEVEL OF SERVICE
The type of facility is critical. The existence of uninterrupted
flow, the presence of medians, and other major facility type factors
Ideal Conditions significantly affect flow characteristics and capacity. The develop-
ment environment has also been found to affect the performance
Many of the procedures in this manual provide a formula or of two-lane roadways, multilane highways, and signalized
simple tabular or graphic presentation for a set of specified stan- intersections.
dard conditions, which must be adjusted to account for any prevail- Lane and shoulder widths can have a significant impact on
ing conditions not matching those specified. The conditions so traffic flow. Narrow lanes cause vehicles to travel closer to each
defined are often ideal conditions. other laterally than most drivers would prefer. Motorists compen-
In principle, an ideal condition is one for which further improve- sate by slowing down or observing larger longitudinal spacing for
ment will not achieve any increase in capacity. Ideal conditions a given speed, which effectively reduces capacity, service flow
assume good weather, good pavement conditions, users familiar rates, or both.
with the facility, and no incidents impeding traffic flow. Specific Narrow shoulders and lateral obstructions have two important
ideal conditions are identified in each chapter. Examples of ideal impacts. Many drivers will steer away from roadside or median
conditions are given below for uninterrupted flow facilities and objects they perceive to pose a hazard. This action brings them
for intersection approaches. laterally closer to vehicles in adjacent lanes and causes the same
Ideal conditions for uninterrupted flow facilities include the reactions as those exhibited in narrow lanes.
following: Restricted design speeds affect operations and level of service;
T Lane widths of 12 ft. drivers are forced to travel at somewhat reduced speeds and to be
T Clearance of 6 ft between the edge of the travel lanes and the more vigilant in reacting to the harsher horizontal and vertical
nearest obstructions or objects at the roadside and in the median. alignments resulting from a reduced design speed. In extreme
T Design speed of 70 mph for multilane highways, 60 mph for cases, the capacity of multilane facilities has been found to be
two-lane highways. affected by low design speeds.
T Only passenger cars in the traffic stream. The horizontal and vertical alignment of a highway depends
T Level terrain. greatly on the design speed used and the topography through which
the roadway must be constructed. Procedures for uninterrupted
Ideal conditions for intersection approaches include the flow facilities categorize the general terrain of a highway as
following: follows:
T Lane widths of 12 ft. T Level terrain: Any combination of grades and horizontal and
T Level grade. vertical alignment that allows heavy vehicles to maintain approxi-
T No curb parking on the intersection approaches. mately the same speed as passenger cars; this terrain generally
T Only passenger cars in the traffic stream and no local transit includes short grades of no more than 1 to 2 percent.
buses stopping in the travel lanes. T Rolling terrain: Any combination of grades and horizontal or
T All vehicles traveling straight through the intersection. vertical alignment that causes drivers of heavy vehicles to reduce
T Intersection located in a non-central business district area. speeds to substantially below those of passenger cars, but does not
T No pedestrians. require operation at crawl speeds for any significant length of time.
T At signalized intersection approaches, green signal available T Mountainous terrain: Any combination of grades and hori-
at all times. zontal and vertical alignment that causes drivers of heavy vehicles
to operate at crawl speeds for significant distances or at frequent
In most capacity analyses, prevailing conditions are not ideal,
intervals.
and computations of capacity, service flow rate, or level of service
must include predictive adjustments to reflect this absence of ideal Crawl speed is the maximum sustained speed that heavy vehi-
conditions. Prevailing conditions are generally categorized as road- cles can maintain on an extended upgrade of a given percent.
These definitions are general and depend on the particular mix to pick up or discharge passengers on the road or street. Local
of heavy vehicles in the traffic stream. In general, as terrain be- transit buses make such stops within the confines of the roadway.
comes more severe, capacity and service flow rates are reduced.
This impact is significant for two-lane rural highways, where the There is considerable variation in the characteristics and per-
severity of terrain not only affects the operating capabilities of formance capabilities of vehicles within each class of heavy vehi-
individual vehicles in the traffic stream, but also restricts the oppor- cle, just as there is among passenger cars.
tunities to pass slow-moving vehicles in the traffic stream. Trucks cover a particularly wide range of vehicles, however,
In addition to the general impacts of terrain, isolated upgrades from lightly loaded vans and panel trucks to the most heavily
of significant length may have a substantial effect on operations. loaded coal, timber, and gravel haulers. Individual trucks have
Heavy vehicles slow significantly on such upgrades, creating oper- widely varying operational characteristics based on how heavily
ational difficulties in the traffic stream and inefficient use of the they are loaded. Analysis procedures for each type of facility dis-
roadway. cuss the mix of trucks on each in some detail.
Grades also may have a major impact on the operation of inter- None of the procedures segregate the truck population into sub-
section approaches; vehicles must overcome both the grade and categories for separate computational consideration, although some
the inertia of starting from a stopped position at the same time. analysis procedures allow the user to select various typical trucks
on the basis of the prevailing mix.
Recreational vehicles also cover a broad range of vehicle
Traffic Conditions types, including campers, both self-propelled and towed; motor
homes; and passenger cars or small trucks towing a variety of
Traffic conditions that influence capacities and service levels recreational equipment, such as boat, snowmobile, and motorcycle
include vehicle type and directional or lane distribution. The proce- trailers. Although these vehicles may have considerably better
dures assume that drivers are familiar with the facility. Less effi- operating capabilities than trucks, drivers of such vehicles are
cient use of roadway facilities on weekends or in recreation areas is not professionals, which accentuates the impact of such vehicles’
generally attributed mainly to the lack of specific local knowledge. deficiencies.
Intercity buses are relatively uniform in their performance capa-
Vehicle Type bilities. Urban transit buses are generally not as powerful as inter-
city buses. Their most severe impact on traffic, however, results
Whenever vehicles other than passenger cars (which include from the discharge and pickup of passengers on the roadway. Local
small trucks and vans) exist in the traffic stream, the number of transit buses make such stops at the curb, usually at intersections,
vehicles that can be served is affected. Heavy vehicles are defined along multilane suburban highways, arterials, and city streets.
as vehicles having more than four tires touching the pavement. Where there is no curb parking on the roadway, the stopped bus
Heavy vehicles adversely affect traffic in two ways: blocks a travel lane. Where curb parking does exist, the bus dis-
T They are larger than passenger cars and therefore occupy rupts flow in adjacent travel lanes as it enters and leaves the
more roadway space than passenger cars. bus stop.
T They have poorer operating capabilities than passenger cars,
Directional and Lane Distribution
particularly with respect to acceleration, deceleration, and the abil-
ity to maintain speed on upgrades. In addition to the distribution of vehicles types, two other traffic
The second impact is the more critical. Because heavy vehicles characteristics affect capacity, service flow rates, and level of ser-
cannot keep pace with passenger cars in many situations, large vice: directional distribution and lane distribution. Directional dis-
gaps form in the traffic stream that are difficult to fill by passing tribution has a dramatic impact on two-lane rural highway operation.
maneuvers. These gaps create inefficiencies in the use of roadway Optimum conditions occur when the split of traffic is about 50 per-
space that cannot be completely overcome. This effect is particu- cent in each direction. Capacity declines as the directional split be-
larly deleterious on sustained, steep upgrades, where the difference comes more unbalanced. Capacity analysis procedures for multilane
in operating capabilities is most pronounced, and on two-lane high- highways focus on a single direction of flow. Nevertheless, each
ways, where passing must be accomplished by using the opposing direction of the facility is usually designed to accommodate the peak
travel lane. rate of flow in the peak direction. Typically, morning peak traffic
Heavy vehicles also may affect downgrade operations, particu- occurs in one direction and evening peak traffic occurs in the opposite
larly where downgrades are steep enough to require operation of direction. Lane distribution is also a factor on multilane facilities.
such vehicles in a low gear. In such cases, heavy vehicles again Typically, the shoulder lane of a multilane facility carries less traffic
must operate at speeds slower than those of passenger cars and than other lanes. Analysis procedures assume typical lane distribu-
gaps in the traffic stream will form. tions for most types of facilities.
Heavy vehicles are generally grouped in three categories:
T Trucks: Vehicles involved primarily in the transport of goods Control Conditions
or in the delivery of services (other than public transportation).
T Recreational vehicles: Vehicles operated by private motorists For interrupted flow facilities, the control of the time available
and involved in the transport of recreational equipment or facilities. for movement of specific traffic flows is a critical element affecting
T Buses: Vehicles involved in the transportation of groups of capacity, service flow rates, and level of service. The most critical
people on a for-hire, charter, or franchised transit basis. Buses are type of control on such facilities is the traffic signal. Operations
further categorized as intercity or local transit buses. Intercity (or are affected by the type of control in use, signal phasing, allocation
‘‘through’’) buses operate in a traffic stream without making stops of green time, cycle length, and relationship with adjacent control
measures. All of these terms are discussed in detail in Chapter 9, to the analyst to determine the impact of ITS (if any) in individual
Signalized Intersections. For this introduction, it is sufficient to roadway capacity analysis applications.
note that the traffic signal determines the amount of time available In light of current ITS studies, the following comments may
for movement on various lanes of the intersection. provide some guidance in analyzing the impact of ITS in specific
Stop signs also affect capacity, but in a less deterministic way. roadway capacity situations:
Whereas the traffic signal positively assigns designated times when T In the case of freeways and other uninterrupted flow facilities,
each movement is permitted, the stop sign at a two-way stop- ITS strategies may be able to achieve some decrease in headways,
controlled intersection merely assigns the right-of-way perma- which would tend to increase the capacity of these facilities. In
nently to the major street. Motorists traveling on the minor street addition, improvements in level of service may be achieved even
must find gaps in the major traffic flow through which to execute with no increase in headways if vehicle guidance systems can offer
maneuvers. Thus the capacity of such approaches depends on traf- drivers a greater level of comfort than they currently experience
fic conditions on the major street. in driving conditions with closely spaced headways.
All-way stop control forces drivers to stop and alternately enter T For signalized intersections and arterials, any potential de-
the intersection in rotation. Capacity and operational characteristics crease in headways and headway variability would initially be less
may vary widely depending on traffic demands on the various significant than it would be on uninterrupted flow facilities. The
approaches. major benefits of ITS for signal and arterial operations would be
Other types of controls and regulations can significantly affect more efficient allocation of green time.
capacity, service flow rates, and level of service. Restriction of T At unsignalized intersections, capacity improvements related
curb parking can increase the number of lanes available on a street to ITS would tend to occur if assistance were provided to drivers in
or highway. Turn restrictions can eliminate conflicts at intersec- judging gaps in opposing traffic streams or if gaps were controlled.
tions and increase capacity. Lane use controls can positively allo-
Many of the roadway improvements related to ITS are system-
cate available roadway space to component movements and can
level improvements, such as incident response and driver informa-
be used at intersections and to create reversible lanes on critical
tion systems. Although these improvements will provide benefits
arterials. One-way street routings eliminate conflicts between left
to the overall roadway system, they are not expected to have an
turns and opposing traffic. impact on the methods used to calculate capacity for individual
roadway facilities.
Technology SUMMARY
The various adjustments necessary to account for less-than-ideal

The factors described in the preceding discussion generally re-
conditions are summarized in Table 1-3 for each type of facility.
late to immediate conditions that would reduce roadway capacity
The importance of roadway, control, and traffic characteristics is
below ideal conditions. Emerging transportation technologies un-
twofold.
der the broad heading of intelligent transportation systems (ITS) First, the variables discussed are important factors involved in
are being developed to enhance the safety and efficiency of road- the capacity analysis computations described in this manual. Sec-
way systems. Depending on the particular application, ITS strate- ond, these conditions define the parameters that planners and engi-
gies will tend to increase the safety and performance of roadway neers may consider changing to improve capacity and level of
facilities beyond the levels experienced under current roadway and service.
vehicle control systems. For the purposes of this discussion, ITS The engineer has, to varying degrees, control over the geometric
is considered to include any technology that allows real-time infor- and control parameters discussed. Through construction, recon-
mation to be gathered and used by drivers and traffic control struction, or spot improvements, improvements can be made in
system operators to provide better vehicle navigation, roadway lane widths, shoulder widths, the number of lanes, horizontal and
system control, or both. vertical alignment, and other geometric factors. Through regulation
At the time of publication of this chapter, the implementation and signalization, all of the control variables are subject to alter-
of many ITS strategies is only in its beginning stages. Therefore, ation. These are the tools that can be used to address capacity or
little research has been conducted to determine their impact on service deficiencies. One of the most important uses of the proce-
capacity and safety. The procedures in this manual are considered dures in this manual is in the evaluation of alternative improvement
to relate to roadway facilities without ITS enhancements. It is left plans based on such changes.

Table 1-3. Adjustment Factors Used for Analyses

factors
facility roadway traffic control
Freeways—basic sections T Lane width T Peak-hour factor
T Lateral clearance T Heavy vehicles
T Grade T Driver type
T Number of lanes T Free flow speed
T Interchange density
Freeways—weaving T Same as basic sections, T Same as basic section, plus
except interchange density, T Volume ratio
plus T Weaving ratio
T Configuration
T Length
T Number of lanes
Freeways—ramp junctions T Adjacent ramp configuration T Peak-hour factor T Metering rate
T Number of lanes T Heavy vehicles
Freeways—ramp roadways T Lane width T Heavy vehicles
T Number of ramp lanes
Two-lane highways T Design speed T Directional split
T Percent no passing T Peak-hour factor
T Lane width T Heavy vehicles
T Shoulder width
T Grade
Multilane highways T Same as freeways, basic T Same as freeways, basic
sections, except interchange sections
density, plus
T Development environment
Signalized intersections T Lane width T Lane utilization T Phasing
T Area type T Peak-hour factor T Green time
T Grade T Heavy vehicles T Cycle length
T Number of lanes T Right turns T Signal progression
T Type of lanes T Left turns T Upstream filtering/
T Turning radius T Pedestrian activity metering
T Parking
T Bus stops
Unsignalized intersections T Grade T Peak-hour factor T Stop control
T Number of lanes T Heavy vehicles T Upstream signals
T Type of lanes T Turning movements
Urban arterials T Same as signalized T Same as signalized T Same as signalized
intersections, plus arterial intersections, plus free flow intersections
classification speed
Transit T (Level of service within T Peaking T Length of bus stop
vehicle depends on space per T Fare collection
passenger) practices
T Number of lanes
T Station and stop design
Pedestrians—walkways T Effective width T Peaking
Bicycles—bike lanes T Number of lanes T Turning traffic
T Turning traffic T Percent heavy vehicles
T Percent heavy vehicles

III. APPLICATIONS
MODELS OF TRAFFIC FLOW Table 1-4. Analysis Techniques

level of analysis
Models are used to represent the operation of transportation
type of facility operational design planning
facilities and to facilitate analysis. Field data on traffic characteris-
tics and measures of effectiveness may be considered the simplest Freeways
type of model; equations solved by a pencil-and-worksheet method Basic segments X X X
Weaving areas X X —
are another type. Computer models in the form of software may
Ramp junctions X — —
range from the automation of worksheets to simulations that track Multilane highways X X X
the detailed behavior of individual vehicles in space and time. Two-lane highways X — X
Level of service may be estimated by computer models, pro- Signalized intersections X X X
vided that Unsignalized intersections X X X
Urban or suburban arterials X — X
T Input parameters, such as free-flow speed or saturation flow Transit X X X
rate, are determined in a manner consistent with the procedures Pedestrians X — —
described in this manual; and Bicycles X — —
T The measures of effectiveness estimated by the model are NOTE: X = addressed in this manual; — = not addressed in this manual.
consistent with their definitions in this manual. They must either
be calculated using the procedures described here or verified
and calibrated with field data obtained using methods from this
manual. descriptions to estimate the level of service that exists or is ex-
pected to prevail.
It is the analyst’s responsibility to select the appropriate model For existing facilities, this estimation requires detailed informa-
for solving a given problem. When the analysis requires consider- tion on traffic characteristics, including volumes, peak-hour fac-
ation of the interaction between elements, over-capacity condi- tors, directional distributions, and vehicle type distributions. All
tions, or other situations not currently covered in this manual, use geometric conditions for the facility must be known, including
of models that produce outputs convertible to the measures of number and width of lanes, shoulder clearances, design speeds,
effectiveness used in this manual may be appropriate. The analyst grades, and horizontal and vertical alignments.
must understand the underlying assumptions of the selected mod- Where traffic controls exist, such as at signalized intersections,
el(s) to be able to perform any calibration required, to recognize they must be completely specified, including the type of control,
that differences may result from use of alternative methodologies, cycle length, phasing, green time allocation, and other factors. All
and to present the results in the context of the model used. other types of control must also be specified. For planned or future
facilities, the same type of information is required. It would, how-
LEVELS OF ANALYSIS ever, be based on traffic projections and planned facilities instead
of on field-measured data.
Most of the procedural chapters address three different computa- Operational analysis allows for an evaluation of base year and
tional applications: operational analysis, design, and planning anal- future levels of service on an existing facility. This evaluation,
ysis. The operational analysis is used for the detailed determination however, is not its most powerful use. Operational analysis can
of the operating conditions. It is appropriate mainly for short-term be used to evaluate the level of service that would result from
situations in which the basic factors are well known or can be alternative spot and section improvements to an existing facility.
reasonably estimated. Design procedures, where provided, can be The operational impacts of various improvement measures can be
used to determine specific geometric or control parameters to yield estimated and compared and a rational decision made using the
the desired level of service. The planning analysis is more general, results and other relevant information. Alternative designs for new
but is useful for longer-range determination of the type and size facilities can be similarly evaluated using the operational analysis
of a facility. It should be noted that for any given facility, the approach.
operation, design, and planning applications are based on the same Most of the procedures allow not only a determination of level of
principles and basic method. The selection of an appropriate level service, but also an estimation of the value of critical performance
is the responsibility of the analyst or designer. Availability of the parameters. For a freeway segment, density and speed of the traffic
procedures for all three levels of analysis in this manual is shown stream can be estimated, and for a signalized intersection, average
in Table 1-4. control delay can be estimated. Thus an operational analysis not
only yields a determination of the level of service (which covers
a range of conditions), but it also provides values of operational
Operational Analysis parameters.
An alternative use of operational analysis is to determine the
In this application, known or projected traffic flow rates and service flow rates allowable under varying operational (LOS) as-
characteristics are compared with known or projected highway sumptions. Such analyses are extremely useful in evaluating the
sensitivity of service flow rates to various design or LOS All tabulated service flow rates in this manual have been
assumptions. rounded to the nearest 50 vehicles per hour, and analysts may wish
to round all computational results in this manner as well. Traffic
volume and capacity values rounded only to the nearest vehicle
Design provide a spurious sense of precision and should be avoided.
Because each of the factors used in capacity and LOS analyses
The design application is keyed to a specific objective: to deter- is subject to a ‘‘plus or minus’’ accuracy, the results of such
mine the number of lanes required on a particular facility or in a calculations should not be presented in a way that portrays them
given corridor to provide for a specified level of service. The as absolute and precise values. For example, if a delay estimate
design application of capacity analysis procedures treats this aspect at a signalized intersection is calculated at 25.4 sec, presentation
of the overall ‘‘design process’’ and may also be used to assess of the results should be rounded to the nearest second, and in fact,
the impact of such design variables as lane and shoulder width, the true delay might be in the range of 25 sec 6 10 percent. In
lateral clearance, grades, lane use allocations, and other features. practice, the values calculated for measures of effectiveness might
Detailed data on expected traffic volumes and characteristics are be assumed to have a precision range of 6 5 to 10 percent. Where
required, as is the assumption of geometric standards to be used future traffic volumes are used in the analysis, the source of these
in the design: lane widths, lateral clearances, design speeds, and volumes should be identified when the results are reported.
horizontal and vertical alignment. Design of signal timings may
also be accomplished using the procedures presented in Chapter
9, Signalized Intersections, in an iterative process.
FIELD DATA
Design computations are generally limited in scope and may
result in the generation of alternatives that are subsequently sub-
jected to detailed operational analyses. The procedures in this manual have been calibrated to estimate
performance parameters such as speed, density, and delay on the
basis of existing or forecast traffic volumes. Basic volume charac-
Planning Analysis teristics such as vehicle types, peak-hour factor, directional distri-
bution, and hourly variations are normally required to conduct
Planning analysis represents a broad assessment of the levels capacity and LOS analyses. The analysis procedures are mainly
of service. Capacity and LOS analysis in transportation planning keyed to traffic volume characteristics because volume is the most
addresses such questions as the following: readily and often measured traffic stream parameter and is usually
the easiest to predict for future conditions.
T What is the maximum number of people or vehicles that can
The various performance procedures are based on average con-
be accommodated within a specified time period?
ditions throughout North America. They reflect normalized esti-
T What will be the future level of service on an existing or
mates of capacity and level of service, assuming that a given
planned facility?
facility operates like the national average of facilities with the
T What lane configurations or signalization characteristics are
same physical, control, and traffic characteristics. The relationships
needed for various traffic flow levels on an arterial road considered
between volume and performance are subject to variance because
in a land use plan?
of local driving habits and other factors. Thus estimations of opera-
tional criteria will rarely duplicate exactly field-measured values
The planning applications are frequently intended to produce at specific locations. Therefore field measurements of existing
estimates at the earliest stages of planning when the amount, detail, traffic performance are desirable.
and accuracy of information are limited. Planning procedures are It is possible to measure operational variables directly on ex-
often based on forecasts of average annual daily traffic and on isting facilities. LOS determinations may then be made by compar-
assumed traffic, roadway, and control conditions. Typical charac- ing field-measured values against the defined criteria. This
teristics appear in many chapters and may be used as default values. procedure is discussed in each chapter. It must be done with some
It should be recognized that extensive use of default values may care because criteria are often defined for ideal or other specified
lead to errors where prevailing conditions differ substantially from conditions. For example, the densities defining level of service for
those assumed. Therefore, during subsequent project planning and freeways and multilane highways are specified in passenger cars
development stages, generalized planning applications should be per mile per lane. Field-measured values in vehicles per mile per
refined as more information becomes available. The analytical lane would have to be converted to passenger car units before
process during these later stages may reach the design or opera- comparison with the established criteria.
tional level. Where local data are available in sufficient sample quantities
and in an acceptable form, they may be used to fine tune the
procedures presented herein. Several chapters contain specific rec-
PRECISION ommendations on when and how such adjustments should be made.
Procedures specify certain average relationships and values, deter-
The results of capacity computations are no more precise or mined for average U.S., and in some instances Canadian, condi-
accurate than the information of data used as inputs to the analysis. tions. The procedures can often be made more accurate by
Thus where traffic counts are only accurate to within 5 percent, substituting local calibrations for these average values. Examples
or where projections are subject to even larger errors, computations of local calibrations that could be used include flow-density-speed
cannot be expected to be accurate to the nearest vehicle per hour relationships for multilane facilities and saturation flow rates for
or mile per hour. signalized intersections.
When such substitutions are made, care must be taken that local basis of known or projected roadway, traffic, and control condi-
data and calibrations are for the same base conditions as those tions under current vehicle technology. Performance criteria also
described in this manual. A saturation flow rate for a 10-ft lane can be set at desired levels and corresponding traffic, roadway, or
should not be substituted for this manual’s value applied to a 12-ft control conditions estimated.
lane without consideration of the impact on lane width, grade, and The results of the procedures provide important comparative
other adjustment factors, for example. information to the engineer or planner. These results should be
There is no substitute for correctly collected and adequately used with other relevant information to formulate recommenda-
presented field data. A capacity analysis based on inaccurate road- tions on highway, transit, and pedestrian improvements. Although
way, traffic, and control information will produce erroneous re- some local authorities have made use of the techniques included
sults. The results of computations will not be more accurate than in this manual mandatory for traffic analyses, no computation
the input data on which they were based. based on these procedures should be construed as mandating or
requiring the implementation of a particular improvement or design
alternative. The professional judgment of the engineer or planner
is a necessary input to such decisions. This manual is an important
SUMMARY
guide to decision making, but the results of capacity analysis do not
replace the need to consider local legal, societal, environmental,
HCM contains a set of analysis procedures that provide esti- behavioral, and other specific requirements, constraints, and
mates of the performance of a variety of traffic facilities on the conditions.

chapter 2
TRAFFIC CHARACTERISTICS
CONTENTS
INTRODUCTION .................................................................................................................................................................................... 2-2
PART A. BASIC VARIABLES OF TRAFFIC FLOW ......................................................................................................................... 2-2

i. uninterrupted flow............................................................................................................................................................. 2-2
Volume and Rate of Flow..................................................................................................................................................... 2-2
Speed...................................................................................................................................................................................... 2-3
Density ................................................................................................................................................................................... 2-4
Spacing and Headways.......................................................................................................................................................... 2-5
Mathematical Relationships .................................................................................................................................................. 2-5
Relationships Among Basic Variables.................................................................................................................................. 2-5
ii. interrupted flow.................................................................................................................................................................. 2-6
Signalized Intersections......................................................................................................................................................... 2-6
Saturation Flow Rate and Lost Time.................................................................................................................................... 2-7
Unsignalized Intersections..................................................................................................................................................... 2-8
Delay ...................................................................................................................................................................................... 2-9
Speed...................................................................................................................................................................................... 2-9
PART B. OBSERVED VALUES ........................................................................................................................................................... 2-9

i. national roadway traffic trends .................................................................................................................................... 2-10
ii. volumes and flow rates ..................................................................................................................................................... 2-10
Freeways ................................................................................................................................................................................ 2-10
Multilane Highways .............................................................................................................................................................. 2-10
Rural Two-Way, Two-Lane Highways................................................................................................................................. 2-11
Urban Arterials ...................................................................................................................................................................... 2-11
iii. volume characteristics ....................................................................................................................................................... 2-12
Temporal Variations .............................................................................................................................................................. 2-12
Seasonal and Monthly Variations.................................................................................................................................... 2-15
Daily Variations ............................................................................................................................................................... 2-15
Hourly Variations............................................................................................................................................................. 2-16
Peak Hour and Design Hour............................................................................................................................................ 2-16
Subhourly Variations in Flow.......................................................................................................................................... 2-18
Spatial Distributions .............................................................................................................................................................. 2-19
Directional Distribution.................................................................................................................................................... 2-19
Lane Distribution.............................................................................................................................................................. 2-20
Traffic Composition .............................................................................................................................................................. 2-21
Impact of Weather on Maximum Volumes.......................................................................................................................... 2-21
iv. speed characteristics ........................................................................................................................................................... 2-24
National Speed Trends .......................................................................................................................................................... 2-24
Speed Variation by Time of Day.......................................................................................................................................... 2-26
Speed Variation by Lane and Day Versus Night................................................................................................................. 2-27
v. measured relationships for uninterrupted flow ......................................................................................................... 2-28
Speed-Flow Relationships ..................................................................................................................................................... 2-28
Freeways........................................................................................................................................................................... 2-28
Multilane and Two-Lane Rural Highways...................................................................................................................... 2-29
2-1 Updated October 1994

Density-Flow Relationships .................................................................................................................................................. 2-29

Headway Distributions and Random Flow........................................................................................................................... 2-30
vi. interrupted flow facilities................................................................................................................................................ 2-31
Saturation Flow and Lost Time at Signalized Intersections ................................................................................................ 2-31
Gap Acceptance and Saturation Flow at Unsignalized Intersections .................................................................................. 2-33
vii. summary .................................................................................................................................................................................. 2-33
REFERENCES ......................................................................................................................................................................................... 2-34
INTRODUCTION
Chapter 1 identified the objectives and the contents of the man- teristics on roadway operations, and therefore on the planning and
ual and explained the basic concepts of capacity and level of design requirements of transportation facilities, as well as to note
service. The focus of this chapter is on basic characteristics of the variation from national averages that occur because of unique
uninterrupted and interrupted traffic flow. Transit, pedestrian, and local conditions.
bicycle characteristics are discussed in Chapters 12, 13, and 14, The procedures of this manual are based on calibrated ‘‘national
respectively. average’’ traffic characteristics observed over a range of facilities
In Part A of this chapter, variables such as volume and rate of of each type. Observations of these characteristics at specific loca-
flow, speed, density, spacing and headways, saturation flow rates, tions will vary somewhat from national averages because of local
lost times, and delay are introduced. All of these variables, impor- driving habits and unique features of the local driving environment.
tant from the capacity assessment point of view, are discussed In this chapter the characteristics that have been observed are
separately for uninterrupted and interrupted traffic flow. addressed, and they are related to the values used in capacity
In Part B a sampling of national observations of key capacity analysis procedures of subsequent chapters. Information on traffic
and level-of-service variables, including some measured or fitted parameters not explicitly used in analysis procedures, but whose
relationships among them and their variation in time and space, is impact on capacity and level of service is important, is also
presented. It is important to recognize the impact of these charac- presented.
PART A. BASIC VARIABLES OF TRAFFIC FLOW
Three basic variables—volume or rate of flow, speed, and den- Section A-I deals with the variables of uninterrupted traffic flow,
sity—can be used for the description of traffic states on any road- including the conceptual relationships among them. Section A-II
way facility. The procedures of the manual apply volume or traffic explains the concepts of traffic operations on interrupted flow
flow as a variable common to both uninterrupted and interrupted facilities. Principles of transit and pedestrian flow and measures
traffic flow, but speed and density are used mostly for uninter- of effectiveness used to analyze capacity and level of service for
rupted flow. Some variables related to flow rate, such as spacing transit and pedestrian facilities are defined in Chapters 12 and 13,
and headway, are also used for both types of facilities, whereas respectively.
other variables, such as saturation flow or gap, are specific to
interrupted flow.
I. UNINTERRUPTED FLOW
The operational state of any given traffic stream on an uninter- VOLUME AND RATE OF FLOW
rupted traffic flow facility is defined by three primary measures:
T Volume and/or rate of flow, Volume and rate of flow are two measures that quantify the
T Speed, and amount of traffic passing a point on a lane or roadway during
T Density. a designated time interval. These terms are defined as follows:
Spacing and headways are directly related to these primary T Volume—the total number of vehicles that pass over a given
measures. point or section of a lane or roadway during a given time interval;
Updated October 1994
traffic characteristics 2-3
volumes may be expressed in terms of annual, daily, hourly, or flow, when vehicles arrive at a rate of 4,800 vph—even though
subhourly periods. volume is less than capacity over the full hour. This is a serious
T Rate of flow—the equivalent hourly rate at which vehicles situation, because the dynamics of dissipating a breakdown may
pass over a given point or section of a lane or roadway during a extend the effects of congestion up to several hours beyond the
given time interval less than 1 hr, usually 15 min. time of the breakdown. These dynamics are discussed in greater
detail in Chapter 6.
Volume and flow are the variables used to quantify demand, that Peak rates of flow are related to hourly volumes through the
is, the number of vehicle occupants or drivers (usually expressed use of the peak-hour factor. This factor is defined as the ratio of
as the number of vehicles) who desire to use a given facility during total hourly volume to the peak rate of flow within the hour:
a specific time period. Congestion influences demand patterns, and
observed volumes are sometimes more a reflection of capacity Hourly volume
PHF = (2-1)
constraints than of true demand. Peak rate of flow (within the hour)
The distinction between volume and rate of flow is important.
Volume is an actual number of vehicles observed or predicted to If 15-min periods are used, the PHF may be computed as
be passing a point during a time interval. Rate of flow represents PHF = V/(4 × V15) (2-2)
the number of vehicles passing a point during a time interval less
than 1 hr, but expressed as an equivalent hourly rate. A rate of where
flow is found by taking the number of vehicles observed in a
PHF = peak-hour factor,
subhourly period and dividing it by the time (in hours) over which
V = hourly volume (vph), and
they were observed. Thus, a volume of 100 vehicles observed in
V15 = volume during the peak 15 min of the peak hour (veh/
a 15-min period implies a rate of flow of 100 veh/0.25 hr or
15 min).
400 vph.
The following examples further illustrate the difference between Where the peak-hour factor is known, it may be used to convert
hourly volumes and flow rates. a peak-hour volume to a peak rate of flow, as follows:
Example 1: Employee Parking Lot Exit to Highway v = V/PHF (2-3)
Time Period Volume (veh) Rate of Flow (vph) where
5:00–5:15 0 0 v = rate of flow for a peak 15-min period (vph),
5:15–5:30 200 800 V = peak-hour volume (vph), and
5:30–5:45 200 800 PHF = peak-hour factor.
5:45–6:00 0 0
5:00–6:00 400 Equation 2-3 need not be used to estimate peak flow rates where
traffic counts are available. The chosen count interval must allow
Volumes were observed for four consecutive 15-min periods. The the identification of the maximum 15-min flow period. The rate
total volume is 400 vehicles, or the sum of the four counts. The may then be directly computed as 4 times the maximum 15-min
rate of flow, however, is zero in two 15-min intervals and 800 count.
vph in the two other 15-min intervals. Thus, a design based on Many of the procedures use this conversion to allow computa-
the hourly volume (400 vph) would not prove adequate. tions to focus on the peak flow period within the peak hour.
Example 2: Highway
The following traffic counts were made during an hour-long study SPEED
period:
Whereas traffic volumes provide a method of quantifying capac-
Time Period Volume (veh) Rate of Flow (vph) ity values, speed (or its reciprocal—travel time) is an important
5:00–5:15 1,000 4,000 measure of the quality of traffic service provided to the motorist.
5:15–5:30 1,200 4,800 It is used as an important measure of effectiveness defining levels
5:30–5:45 1,100 4,400 of service for many types of facilities, such as rural two-lane
5:45–6:00 1,000 4,000 highways, arterials, freeway weaving sections, and others.
5:00–6:00 4,300 Speed is defined as a rate of motion expressed as distance per
unit time, generally as miles per hour (mph) or kilometers per
Volumes were observed for four consecutive 15-min periods. hour (km/hr). In characterizing the speed of a traffic stream, some
The total volume for the hour is the sum of these counts, or 4,300 representative value must be used, because there is generally a
veh, or 4,300 vph (since they were observed for 1 hr). The rate broad distribution of individual speeds that may be observed in
of flow, however, varies within each 15-min period. the traffic stream. For the purposes of this manual, the speed
During the 15-min period of maximum flow, the rate of flow measure used is average travel speed. This measure is used be-
is 1,200 veh/0.25 hr, or 4,800 vph. Note that 4,800 vehicles do cause it is easily computed from observation of individual vehicles
not pass the point in question during the study hour, but they do within the traffic stream and because it is the most statistically
pass the point at that rate for 15 min. relevant measure in relationships with other variables. Average
Consideration of peak flow rates is critically important in capac- travel speed is computed by taking the length of the highway or
ity analysis. If the capacity of the above segment of highway were street section or segment under consideration and dividing it by
4,500 vph, it would break down during the peak 15-min period of the average travel time of vehicles traversing the segment. Thus,
if travel times t1, t2, t3, . . ., tn are measured for n vehicles traversing
a segment of length L, the average travel speed would be
L nL
S= = (2-4)
n n
o
i=1
ti /n o
i=1
ti
where
S = average travel speed (mph),
L = length of the highway segment (mi),
ti = travel time of the ith vehicle to traverse the section (hr), and
n = number of travel times observed.
Consider the following travel times observed for vehicles trav-

ersing a 1-mi segment of highway: 1.0 min (0.0167 hr), 1.2 min
(0.0200 hr), 1.7 min (0.0283 hr), and 1.1 min (0.0183 hr). The
average travel time is found as (0.0167 + 0.0200 + 0.0283 + Figure 2-1. Typical relationship between time mean and space
0.0183)/4 = 0.0208 hr. The average travel speed is the distance mean speed. (Source: Ref. 1)
(1 mi) divided by this time, or
S = 1.0 mi/0.0208 hr = 48 mph Figure 2-1 shows a typical relationship between time mean and
space mean speeds. Space mean speed is always slower than time
The travel times used in this computation include stopped delays mean speed, with the difference decreasing as the absolute value
due to fixed interruptions or traffic congestion. They are total of speed increases. This relationship is based on statistical analysis
travel times to traverse the defined roadway length. of observed data and is useful, because time mean speeds are often
Several different speed parameters can be applied to a traffic easier to measure in the field than space mean speeds.
stream. These include the following: For capacity analysis, speeds are best measured by observing
travel times over a known length of highway. For uninterrupted
1. Average running speed—This is also called ‘‘space mean
flow facilities operating in the range of stable flow, the length taken
speed’’ in the literature. It is a traffic stream measurement based
may be as short as several hundred feet for ease of observation.
on the observation of vehicle travel times traversing a section of Radar meters or other devices can be used to measure speeds
highway of known length. It is defined as the length of the segment at a point. Such speeds, when averaged, yield a time mean speed.
divided by the average running time of vehicles to traverse the It is possible to compute a space mean speed for a very short
segment. ‘‘Running time’’ includes only time that vehicles spend segment of highway using radar or other observations of individual
in motion. vehicle speeds by calculating the harmonic, rather than the arith-
2. Average travel speed—This is also a traffic stream measure metic, mean of the observations.
based on travel time observations over a known length of highway. When used as a measure of effectiveness, speed criteria must
It is defined as the length of the segment divided by the average recognize driver expectations and roadway function. Thus, a driver
travel time of vehicles traversing the segment, including all stopped expects a higher speed on a freeway than on an urban arterial.
delay times. It is also a ‘‘space mean speed,’’ because the use of Lower speeds will be tolerated on a roadway with more severe
average travel times effectively weights the average according to horizontal and vertical alignment, since drivers will not be comfort-
the length of time a vehicle occupies the defined roadway segment able driving at extremely high speeds. Level of service criteria
or ‘‘space.’’ reflect these and other points.
3. Space mean speed—This is a statistical term frequently
used in the literature to denote an average speed based on the
DENSITY
average travel time of vehicles to traverse a segment of roadway.
It is called a ‘‘space’’ mean speed because the use of average
Density is defined as the number of vehicles occupying a given
travel time essentially weights the average according to the
length of a lane or roadway at a particular instant. For the purposes
length of time each vehicle spends in the defined roadway
of computations in this manual, density is averaged over time and
segment, or ‘‘space.’’
is usually expressed as vehicles per mile (vpm).
4. Time mean speed—This is the arithmetic average of speeds Direct measurement of density in the field is difficult, requiring
of vehicles observed passing a point on a highway and is also a vantage point from which significant lengths of highway can
referred to as the ‘‘average spot speed.’’ Individual speeds of be photographed, videotaped, or observed. It can be computed,
vehicles passing a point are recorded and are arithmetically however, from the average travel speed and rate of flow, which
averaged. are more easily measured.
Most of the procedures using speed as a measure of effectiveness D = v/S (2-5)
in this manual use average travel speed as the defining parameter.
where
For uninterrupted flow facilities not operating at LOS F, the aver-
age travel speed is equal to the average running speed. v = rate of flow (vph),
S = average travel speed (mph), and between pairs of vehicles. The speed would be that of the second
D = density (vpm). vehicles in an individual pair of vehicles. Flow rate is related to
the average headway of the traffic stream:
Thus, a highway segment with a rate of flow of 1,000 vph and
an average travel speed of 50 mph would have a density of Flow rate (vph) = 3,600 (sec/hr)/headway (sec/veh) (2-8)
D = 1,000 vph/50 mph = 20 vpm

RELATIONSHIPS AMONG BASIC VARIABLES
Density is a critical parameter for uninterrupted flow facilities
because it characterizes the quality of traffic operations. It de- Equation 2-5 cites the basic relationship among the three vari-
scribes the proximity of vehicles to one another and reflects the ables describing an uninterrupted traffic stream. Although the
freedom to maneuver within the traffic stream. equation v = S × D algebraically allows for a given rate of flow
Roadway occupancy is frequently used as a surrogate for density to occur at an infinite number of combinations of speed and den-
in control systems because it is easier to measure. Occupancy in sity, there are additional relationships restricting the variety of
space is defined as the proportion of roadway length covered by flow conditions that may exist at any given location.
vehicles, and occupancy in time identifies the proportion of time Figure 2-2 shows a generalized representation of these relation-
a roadway cross section is occupied by vehicles. Under the assump- ships (2), which form the philosophical basis for the capacity
tion of a homogeneous traffic flow, or for a known traffic flow analysis of uninterrupted flow facilities. Although more sophisti-
composition, these two types of occupancy can be taken as equal cated theories of traffic flow exist, a linear speed-density relation-
and used for the derivation of density. ship simplifies the discussion. The flow-density function is placed
directly below the speed-density relationship because of their com-
mon horizontal scales, and the speed-flow function is placed next
SPACING AND HEADWAYS
to the speed-density relationship because of their common vertical
scales. Speed is represented by space mean speed.
Spacing is defined as the distance between successive vehicles The actual form of these functions depends on the prevailing
in a traffic stream, as measured from front bumper to front bumper. traffic and roadway conditions on the roadway segment under
Headway is the time between successive vehicles as they pass a study and on the length of the segment considered in the determina-
point on a lane or roadway, also measured from front bumper to tion of density. Although the diagrams in Figure 2-2 show continu-
front bumper. ous curves, it is unlikely that the full range of the functions will
These characteristics are considered to be ‘‘microscopic,’’ since be found at any particular measurement location. Surveyed data
they relate to individual pairs of vehicles within the traffic stream. usually show discontinuities in which a part of these curves is not
Within any traffic stream, both spacing and headway of individual present. May (2) illustrates and discusses the reasons for these
vehicles are distributed over a range of values, which are generally gaps.
related to the speed of the traffic stream and prevailing conditions. The curves of Figure 2-2 illustrate a number of significant
In the aggregate, these ‘‘microscopic’’ parameters are related to points. Note that a zero rate of flow occurs under two very different
the ‘‘macroscopic’’ flow parameters density and rate of flow. conditions:
Headways are used as part of the Chapter 8 methodology to
estimate percent time delay in a two-lane rural highway traffic 1. When there are no cars on the facility, density is zero,
stream. Defined as the percentage of total time vehicles are delayed and rate of flow is also zero. Speed is purely theoretical for
in an involuntary queue on a two-lane highway, ‘‘percent time this condition and would be whatever the first driver would
delay’’ is estimated as the percentage of vehicle headways less select—presumably a high value. This speed is represented by
than or equal to 5 sec. Sf in the graphs.
2. When density becomes so high that all vehicles stop (speed
is zero), the rate of flow is also zero, because there is no movement
MATHEMATICAL RELATIONSHIPS and vehicles cannot ‘‘pass’’ a point on the roadway. The density
at which all movement stops is called jam density, denoted by Dj
Spacing is a distance measure, in feet. It can be measured di- in the diagrams.
rectly by measuring the distance between common points on suc-
Between these two extreme points, the dynamics of traffic flow
cessive vehicles at a particular instant. This generally requires
produce a maximizing effect. As density increases from zero, rate
complex aerial photographic techniques, so that spacing is usually
of flow also increases, since more vehicles are on the roadway.
derived from other direct measurements. Headway, on the other
While this is happening, speed begins to decline (because of the
hand, can be more easily measured using stopwatch observations
interaction of vehicles). This decline is virtually negligible at low
as vehicles pass a point on the roadway.
and medium densities and rates of flow. As density continues to
The average vehicle spacing in a traffic stream is directly related
increase, these generalized curves suggest that speed decreases
to the density of the traffic stream:
significantly before the capacity is achieved. Capacity is reached
Density = 5,280/spacing (2-6) when the product of density and speed results in the maximum
rate of flow. This condition is shown as optimum speed So (often
The relationship between average spacing and average headway called critical speed), optimum density Do (sometimes referred to
in a traffic stream is dependent on speed: as critical density), and maximum flow vm.
Headway (sec/veh) = spacing (ft/veh)/speed (ft/sec) (2-7) The slope of any ray line drawn from the origin of the speed-
flow curve to any point of the curve represents density, based
This relationship also holds for individual headways and spacings on Equation 2-5. Similarly, a ray line in the density-flow graph
Figure 2-2. Generalized relationships among speed, density, and rate of flow on uninterrupted flow facilities. (Based on May, Ref. 2)
represents speed. As examples, Figure 2-2 shows the average free or breakdown flow. Sudden changes in the state of traffic (i.e., in
flow speed and speed at capacity, as well as optimum and jam speed, density, and rate of flow) may occur. LOS A through E are
densities. The three diagrams shown in the figure are obviously defined on the low-density, high-speed side of the curves, with
redundant, since if any one relationship is known, the other two the maximum flow boundary of LOS E placed at capacity, whereas
are uniquely defined. Whereas the speed-density function is used LOS F, used to describe congested traffic, is represented by the
mostly for theoretical work, the other two are used in the manual high-density, low-speed part of the functions.
to define the levels of service. Examples of recent measurements of the speed-flow relation-
As shown in Figure 2-2, any rate of flow other than capacity ships are shown and discussed in Part B. Although they feature
can occur under two different conditions, one with a high speed overall trends similar to the functions in Figure 2-2, they differ in
and low density and the other with high density and low speed. details, mostly because they are based on flow rates averaged over
The high-density, low-speed side of the curves represents forced longer time intervals.
II. INTERRUPTED FLOW
Interrupted flow is more complex than uninterrupted flow be- tant additional point is the screenline at which the traffic volume
cause of the time dimension involved in the allocation of space to or rate of flow is surveyed. Traditional intersection traffic counts
conflicting traffic streams. Flow on an interrupted flow facility is yield only the number of vehicles that have departed through the
usually dominated by points of fixed operation, such as traffic intersection. The maximum flow is therefore limited to the capacity
signals and stop signs. These control measures have differing im- of the facility. Where demand exceeds capacity and a growing
pacts on overall flow. A detailed discussion of flow at signalized queue is forming, it is advisable to survey traffic demand further
intersections is contained in Chapter 9, and information for stop upstream, before the influence of the congestion.
signs is presented in Chapter 10. Arterial roadway flow is discussed From the capacity computation point of view, speed and density
in Chapter 11. are less important than on uninterrupted flow facilities.
The operational state of traffic at an interrupted traffic flow
facility is defined by the following measures:
T Volume and/or rate of flow; SIGNALIZED INTERSECTIONS
T Saturation flow and/or departure headways;
T Control variables: parameters of stop or signal control;
The most significant source of fixed interruptions on interrupted
T Gaps available in the conflicting traffic streams; and
flow facilities is a traffic signal. At traffic signals, flow in each
T Delay.
movement or set of movements is periodically halted. Thus, move-
The discussion of volume and rate of flow in the first part of this ment on a given set of lanes is only possible for a portion of total
chapter is also applicable to interrupted flow facilities. An impor- time, because the signal prohibits movement during some periods.
Only the time during which the signal is effectively green is avail- After some number of vehicles, N in Figure 2-3, the effect of
able for movement. For example, if one set of lanes at a signalized the start-up reaction and acceleration has dissipated. Successive
intersection receives a 30-sec effective green time out of a 90-sec vehicles now move through past the stop line at a steady speed
total cycle, only 30/90 or 1/3 of total time is available for move- until the last vehicle in the original queue has passed. The headway
ment on the subject lanes. Thus, out of each hour of real time, for these vehicles will be relatively constant.
only 20 min is available for flow on the lanes. Provided the lanes In Figure 2-3, this constant average headway is denoted as h
could accommodate a maximum rate of flow of 1,500 vph if the and is achieved after N vehicles. The headways for the first N
signal displayed green for a full hour, they could accommodate a vehicles are, on the average, greater than h and are expressed as
total rate of flow of only 500 vph, since only one-third of each h + ti, where ti is the incremental headway for the ith vehicle due
hour is available as green. to the start-up reaction and acceleration. As i increases from 1 to
Because signal timings are subject to change, it is convenient N, ti decreases.
to express capacities and service flow rates for signalized intersec- Figure 2-4 shows a conceptual plot of headways measured as
tions in terms of vehicles per hour of green (vphg). In the previous described previously. Although, for practical reasons, the passage
example, the maximum rate of flow would be stated as 1,500 vphg. of the fourth vehicle is used as a starting time for saturation flow
This can be converted to a real-time value by multiplying by the measurements in this manual, N may occur as late as with the sixth
ratio of effective green time to cycle length for the signal. or seventh vehicle (i.e., the start-up and acceleration increment
When the signal turns green, the dynamics of starting a stopped disappears after the sixth or seventh vehicle).
queue of vehicles must be considered. Figure 2-3 shows a queue The value h is defined as the saturation headway and is esti-
of vehicles stopped at a signal. When the signal turns green, the mated as the constant average headway between vehicles occurring
queue begins to move. The headway between vehicles can be after the Nth vehicle in the queue and continuing until the last
observed as they cross the stop line of the intersection. The first vehicle in the initial queue clears the intersection. The saturation
headway would be the elapsed time, in seconds, between the initia- headway is the amount of time consumed by a vehicle that was
tion of the green and the crossing of the rear wheels of the first in the stopped queue as it passes through a signalized intersection
vehicle over the stop line. The second headway would be the
on the green signal, assuming that a continuous queue of vehicles
elapsed time between the crossing of rear wheels of the first and
is available to move through the intersection.
second vehicles over the stop line. Subsequent headways would
The definition of the saturation headway on interrupted flow
be similarly measured.
facilities in the manual is different from that of uninterrupted flow
The driver of the first vehicle in the queue must observe the
headways. For intersections, headway represents the time period
signal change to green and react to the change by releasing the
between the passage of the rear axle of one vehicle and the passage
brake and accelerating through the intersection. The first headway
of the rear axle of the next vehicle over a given cross section on
will be comparatively long as a result of this process. The second
the roadway, whereas the vehicle reference points for uninterrupted
vehicle in the queue follows a similar process, except that the
flow facilities are usually the front bumpers.
reaction and acceleration period can occur while the first vehicle
is beginning to move. The second vehicle will be moving faster
than the first as it crosses the stop line, because it has an additional
vehicle length in which to accelerate. Its headway will still be
SATURATION FLOW RATE AND LOST TIME
comparatively long, but generally less than that of the first vehicle.
The third and fourth vehicles follow a similar procedure, each
achieving a slightly lower headway than the preceding vehicle. Saturation flow rate is defined as the flow rate per lane at which
vehicles can pass through a signalized intersection in such a stable
moving queue. By definition, it is computed as
s = 3,600/h (2-9)
where
s = saturation flow rate (vphgpl),
h = saturation headway (sec), and
3,600 = number of seconds per hour.
The saturation flow rate represents the number of vehicles per

hour per lane that can pass through a signalized intersection if the
green signal were available for the full hour and the flow of vehi-
cles were never halted. This assumes that, in addition to a full
hour of green being available, the average headway of all vehicles
entering the intersection is h sec.
Each time a flow is stopped, it must be started again, and it will
experience start-up reaction and acceleration headways shown in
Figure 2-4 for the first N vehicles. In this figure, the first six
vehicles in the queue experience headways longer than h. The
Figure 2-3. Conditions at traffic interruption in an approach lane increments, ti, are called start-up lost times. The total start-up lost
of a signalized intersection. time for these vehicles is the sum of these increments, or
Figure 2-4. Concept of saturation flow rate and lost time.
N more complex with an increasing number of phases in a signal

l1 =
o
i=1
ti (2-10) cycle.
The amount of lost time affects capacity and delay. The preced-
ing logic suggests that the capacity of the intersection increases
where with increasing cycle length. This is somewhat offset by observa-
l1 = total start-up lost time (sec), and tions that the saturation headway, h, may increase if the length of
ti = lost time for the ith vehicle in queue (sec). continuous green indication becomes very long. Other intersection
features may offset the reduction in capacity due to short cycles,
When a queue of vehicles receives a green signal, it will con- such as turning lanes. Longer cycle lengths increase the number
sume h sec per vehicle plus the start-up lost time, l1, assuming of vehicles in the queues and may cause the left-turn lane to
that there are at least N vehicles in the queue. overflow, thus reducing capacity by blocking through lanes.
Each time a stream of vehicles is stopped, another source of As cycle length is increased, the average stopped-time delay per
lost time is experienced. As one stream of vehicles stops, safety vehicle also tends to increase, assuming that adequate capacity is
requires that there be some clearance time before a conflicting provided. Delay, however, is a complex variable that is affected
stream of traffic is allowed to enter the intersection. During this by many variables, of which cycle length is only one.
period, no vehicles use the intersection. This interval is called Part B of this chapter contains a discussion of the measured
clearance lost time, l2. values of saturation flow, and Chapter 9 presents analytic relation-
In practice, signal cycles provide for this clearance through the ships among saturation headway, saturation flow rate, lost times,
use of ‘‘change intervals,’’ which may include yellow or all red signal timing parameters, and delay.
indications, or both. Drivers generally cannot observe this entire
interval and do use the intersection during some portion of it. The
clearance lost time, l2, is the portion of this change interval that UNSIGNALIZED INTERSECTIONS
is not used by motorists.
The relationship between saturation flow rate and lost times is The driver on a minor street or a driver turning left from the
a critical one. For any given lane or movement, vehicles use the major street of a two-way stop-controlled intersection faces a judg-
intersection at the saturation flow rate for a period of time equaling mental task. A gap must be selected in the priority flow through
the available green time plus the change interval minus the start- which to execute the desired movement. The term gap is com-
up and clearance lost times. Because the lost times are experienced monly used to identify headways in the traffic flow on the roadway
each time a movement is started and stopped, the total amount of with the right-of-way at unsignalized intersections. Gap accep-
time lost over an hour is related to the signal timing. For instance, tance describes the resulting behavior.
if a signal has a 60-sec cycle length, it will start and stop each The capacity of a minor street approach depends on two factors:
movement 60 times per hour, and the total lost time per movement 1. The distribution of available gaps in the major street traffic
will be 60(l1 + l2). If the signal has a 120-sec cycle, each movement stream, and
will be stopped and started 30 times per hour, and the total lost 2. The gap sizes required by minor street drivers to execute
time per movement will be 30(l1 + l2), half as much as the for the their desired movements.
60-sec cycle.
Whereas the preceding discussion suggests that the evaluation The distribution of available gaps in the major street traffic stream
of lost time may be rather simple, its determination becomes much depends on the total volume on the street, its directional distribu-
tion, the number of lanes on the major street, and the degree and average stopped-time delay as the principal measure of effective-
type of platooning in the traffic stream. The gap sizes required by ness in evaluating level of service at signalized intersections and
the minor street drivers depend on the type of maneuver (left, average total delay at unsignalized intersections. In the text, it is
through, right) that must be executed, the number of lanes on the frequently called only delay.
major street, the speed of major street traffic, sight distances, the Stopped-time delay is the time an individual vehicle spends
length of time the minor street vehicle has been waiting, and driver stopped in a queue while waiting to enter a signalized intersection.
characteristics (eyesight, reaction time, age, etc.). The critical gap Average stopped-time delay is the total stopped delay experienced
is the minimum interval between two successive vehicles in the by all vehicles in an approach or lane group of a signalized intersec-
major traffic stream that allows intersection entry to one minor tion during a designated time period divided by the total volume
street vehicle. Note that critical gap has been redefined in this entering the intersection in the approach or lane group during the
update of the manual. When more than one minor street vehicle
same time period, expressed in seconds per vehicle.
uses one major street gap, the time between two subsequent vehi-
At two-way stop-controlled and all-way stop-controlled inter-
cles is called follow-up time. In general, the follow-up time is
sections, total delay is defined as the total elapsed time from when
shorter than the critical gap.
At an all-way stop-controlled intersection, all drivers must come a vehicle joins the queue until the vehicle departs from the stopped
to a complete stop. The decision to proceed is based in part on position at the head of the queue.
the ‘‘rules of the road,’’ which suggests that the driver on the right The use of a similar measure of effectiveness for both signalized
has the right-of-way, and is also a function of the traffic condition and unsignalized intersections provides a means to compare the
on the other approaches. The capacity procedures are based on operation of an intersection under a variety of control conditions.
analyzing each intersection approach independently. The departure Analysis procedures for arterials (Chapter 11) consider both the
headway for the subject approach is defined as the difference travel time between signalized intersections and the delay encoun-
between the successive times of departure of that vehicle and the tered at intersections.
previous departing vehicle on the subject approach. A departure Stopped-time delay is used for signalized intersections because
headway is considered to be a saturation headway if there was it is a reasonably easy parameter to measure and is conceptually
already a vehicle ahead of the given vehicle at the stop line. If simple. Total delay (sometimes called overall delay) involves
traffic is present on one approach only, vehicles depart as rapidly movements at slower speeds on intersection approaches, as vehi-
as individual drivers can safely accelerate into and clear the inter- cles move up in queue position or slow down upstream of an
section. If traffic is present on other approaches, the saturation intersection. Drivers frequently reduce speed when a downstream
headway on the subject approach will increase somewhat, de- signal is red or a queue is present at the downstream intersection
pending on the degree of conflict between the subject approach
approach. Total delay requires the determination of a realistic aver-
vehicles and the vehicles on the other approaches.
age speed for each roadway segment and is implied in the estimates
As at signalized intersections, the vehicle reference points for
of the average travel speed on urban arterial roads.
the determination of saturation headways of the vehicles departing
from the stop line of two- and all-way stop-controlled intersection
approaches are the rear axles of two consecutive vehicles. For the
unobstructed flow of vehicles on the main roadway with the right- SPEED
of-way at two-way stop-controlled intersections, however, the front
bumpers are normally used as reference points, as in other in-
stances of uninterrupted flow. The discussion of speed in the first section of the chapter also
Flow at two- and all-way stop-controlled intersection approaches applies to roadway segments with signalized or unsignalized inter-
and analytic relationships relating critical variables to capacity are sections. For such interrupted flow facilities, segments on which
described in Chapter 10. average travel speed or average running speed is to be determined
should be long enough to include those points of fixed interruption
DELAY of interest. Since travel time lost to flow interruptions is the major
component of the evaluation, speed is generally not relevant. Spot
A critical performance measure on interrupted flow facilities is speed measurements at these facilities are usually used only for
delay. There are several types of delay, but the manual uses only research or enforcement purposes.
PART B. OBSERVED VALUES
The procedures in this manual are based on calibrated ‘‘national addressed in this chapter, and they are related to the values used
average’’ traffic characteristics observed over a range of facilities in the capacity analysis procedures of the subsequent chapters.
of each type. Observations of these characteristics at specific loca- Information on traffic parameters not explicitly used in analysis
tions will vary somewhat from national averages because of the procedures but whose impact on capacity and level of service is
local habits and unique features of the local driving environment. important is also presented in this chapter.
The range of traffic characteristics that have been observed are
I. NATIONAL ROADWAY TRAFFIC TRENDS
The number of motor vehicles in the United States has been the level reported in 1960 as shown in Figure 2-7. Travel grew
steadily increasing, reaching almost 200 million registered vehicles about 54 percent during the 1960s, another 38 percent in the 1970s,
in 1990. The increase during the 10-year period from 1980 repre- and another 41 percent in the 1980s. Travel in urban areas ac-
sented more than 21 percent (Figure 2-5). The number of passenger counted for 1.3 trillion vehicle-miles in 1991, or 60 percent of the
cars increased during that period by more than 22 million, and the total, compared with 44 percent in 1960. The amount of travel in
number of trucks grew by almost 11 million, with most of them urban areas has increased by almost 50 percent in the 1980s, faster
in the light truck category. The number of motorcycles decreased than in rural regions, where the growth was still very significant
from 5.7 million to 4.3 million. at 30 percent.
Automobiles and light trucks and buses on the rural Interstate Traffic congestion, especially in urban and suburban areas, has
system account for about 80 percent of average daily traffic vol- become more severe. As a result, traffic and transportation man-
umes, with heavy trucks and buses representing the remainder agement has gained even more importance. Capacity and level of
(Figure 2-6). Annual travel on the roadways of the United States service analyses are a critical element in the design and evaluation
reached an estimated 2.2 trillion vehicle-miles, or about three times of traffic operations and management.
II. VOLUMES AND FLOW RATES
Capacity is defined in terms of the maximum rate of flow that daily volumes exceeding 20,000 vehicles per lane operate at or
can be accommodated by a given traffic facility under prevailing close to capacity during many hours of an average day. Table 2-
conditions. The determination of capacity involves the observation 2 contains a sample of the maximum reported hourly one-way
of highways of various types operating under high-volume volumes and the average volumes per lane on rural and urban
conditions. freeways in the United States. Most volumes in this table exceed
The direct observation of absolute capacity is difficult to achieve 2,000 vehicles per hour per lane, with several freeways featuring
for several reasons. The recording of a high, or even a maximum, average lane volumes of more than 2,400 vphpl. The highest re-
volume or rate of flow for a given facility does not ensure that a ported lane volumes on selected freeways are given in Table 2-3.
higher flow could not be accommodated at another time. Further, Freeway capacity analysis procedures of this manual use a
capacity is sometimes not a stable operating condition. It has rate of flow of 2,200 pcphpl for freeways with two lanes in
sometimes been estimated by fitting parabolic speed-flow or den- one direction and 2,300 for freeways with three or more lanes
sity-flow curves that included both uncongested and congested in one direction as the basic capacity of such facilities under
conditions. The peak of these curves would define capacity. ideal conditions. These are the average per lane capacities across
Highest reported volume and flow rate observations on various all lanes in a given direction and represent a 200 or 300 pcphpl
types of facilities throughout the United States and Canada are dis- increase over the values used in earlier manuals. Table 2-2
cussed in the following sections. It is noted that these reported obser- contains observations of values higher than this standard, but it
vations may or may not represent the absolute capacities of the should be remembered that these are the maximums reported
subject highways and that they reflect prevailing conditions at the on a given freeway.
locations in question. These observations are a sample of high vol- It should also be noted that an individual lane of a freeway can
carry volumes in excess of 2,200 or 2,300 pcphpl. The highest
umes recorded by state and local highway agencies and do not sug-
reported volumes per lane are given in Table 2-3 for several free-
gest that there are no other facilities experiencing similar, or even
ways with the highest observation close to 2,700 vphpl on a six-
higher, volumes. In some cases, auxiliary lanes may be present,
lane urban freeway. Note that the peak lane volumes may be
resulting in lower actual flows per lane than shown in the tables.
substantially higher than the average volumes per lane.
The data were collected from the literature and from surveys
The recommended values of 2,200 and 2,300 pcphpl should be
conducted by the Committee on Highway Capacity and Quality
considered national averages, around which some variation from
of Service of the Transportation Research Board and by the
region to region and from facility to facility are to be expected.
Federal Highway Administration over a number of years.
MULTILANE HIGHWAYS
FREEWAYS
The observation of multilane rural highways operating under
capacity conditions is difficult, because such operations rarely
The reported average annual daily traffic volumes on selected occur. Table 2-4, however, does contain some data for four-,
Interstate highways are given in Table 2-1. Most of these high- six-, and eight-lane highways in suburban settings operating under
volume freeways are found in the largest metropolitan areas. Daily uninterrupted flow conditions, as well as data for two three-lane
traffic volumes on these heavily used roadways exceed 200,000 bridges. The procedures of this manual assume that the capacity
vehicles per day. A large number of short Interstate sections in of a surface multilane facility is the same as for four-lane freeways
other areas also carry similar volumes. As a rule, all freeways with for uninterrupted flow segments—2,200 pcphpl.
Figure 2-5. Motor vehicle registrations.
RURAL TWO-WAY, TWO-LANE HIGHWAYS URBAN ARTERIALS
High-volume data on two-lane, two-way rural highways in the Since flow on urban arterials is uninterrupted only in the road-
United States and Canada are difficult to obtain. Such highways way segments between intersections, the interpretation of high-
rarely operate at volumes approaching capacity, and thus the obser- volume observations on urban arterials is not as straightforward
vation of capacity operations in the field is extremely complex. as for uninterrupted flow facilities. Signal timing plays a major
A sampling of high-volume observations is given in Table 2-5, role in the capacity of such facilities, limiting the portion of time
but it is emphasized that none of these may be taken to represent that is available for movement along the arterial at critical intersec-
absolute capacity for the facilities shown. In several cases, the tion locations. The volumes reported in Table 2-6 are shown with
volumes noted were accompanied by good operating conditions. the green to cycle time ratios in effect for the subject segments.
European observations on two-lane, two-way rural highways Flow rates in vehicles per hour of green time are estimated by
have been reported at far higher volumes. Volumes of more than taking the reported volumes and dividing by the reported green
2,700 vph have been observed in Denmark, more than 2,800 in over cycle time ratio. These estimates therefore produce a set of
France, more than 3,000 in Japan, and more than 2,450 in Norway. flow observations on a basis comparable with uninterrrupted flow
Some of these volumes have contained significant numbers of facilities. The prevailing conditions on urban arterials may vary
trucks, some as high as 30 percent of the traffic stream (3). greatly, and such factors as curb parking, transit buses, lane widths,
The difficulty in observing capacity operations on two-lane upstream intersections, and similar factors may substantially affect
highways in North America presents problems in suggesting a operations and observed volumes.
standard value for use in computational procedures. The proce- Note that the comparison of maximum flow rates in vehicles
dures for such highways, presented in Chapter 8, are based on a per hour of green per lane varies widely for the various size arteri-
combination of field observations and simulation, which suggested als. These observations did not include such factors as left- and
that a maximum capacity of 2,800 pcph be adopted, total in both right-turn lanes at intersections, which may enhance the capacity
directions under ideal conditions (4). These ideal conditions in- of the intersection approach, nor were other prevailing conditions
clude a 50/50 directional distribution of traffic. Capacity on two- cited. The procedures of Chapter 11 for arterials focus on the issue
lane rural highways varies with directional distribution and reduces of level of service. Capacity of the arterial is generally limited by
as the split moves away from 50/50 to a minimum value of 2,000 the capacity of signalized intersections, with segment characteris-
pcph when the split is 100/0. tics rarely playing a major role in the determination of capacity.

Table 2-1. Maximum Annual Average Daily Traffic Reported on Selected Interstate Routes (1990)
annual average daily
section length average daily traffic per
location (mi) traffic (vpd) lane (vpdpl)
14-Lane Routes
I-405, Los Angeles–Long Beach, California 2.530 328,500 23,464
I-95, New Jersey Turnpike, NE New Jersey 0.610 270,491 19,321
I-95, George Washington Bridge, New York 0.470 270,400 19,314
12-Lane Routes
I-90, Chicago, Illinois 1.030 275,883 22,990
I-5, Seattle-Everett, Washington 1.260 254,172 21,181
I-8, San Diego, California 1.260 253,600 21,133
I-15, San Diego, California 2.880 219,300 18,275
I-280, San Francisco–Oakland, California 1.880 208,900 17,408
I-95, Northeastern New Jersey 1.890 208,768 17,379
10-Lane Routes
I-95, Northeastern New Jersey 1.620 222,229 22,223
I-395, Washington, District of Columbia 0.480 220,455 22,046
I-610, Houston, Texas 1.355 216,390 21,639
H-1, Honolulu, Hawaii 1.690 209,158 20,916
8-Lane Routes
I-285, Atlanta, Georgia 0.210 212,060 26,508
I-635, Dallas–Fort Worth, Texas 4.730 210,497 26,312
I-395, Northern Virginia 1.770 208,590 26,074
6-Lane Routes
I-610, Houston, Texas 0.304 216,390 36,065
SOURCE: Federal Highway Administration
III. VOLUME CHARACTERISTICS
Traffic volumes vary in both space and time. These variations TEMPORAL VARIATIONS
are critical determinants of the way highway facilities are used
and control many of the planning and design requirements for Traffic demand varies by month of the year, by day of the week,
adequately serving traffic demand. by hour of the day, and by subhourly intervals within the hour.
Because traffic volume is not evenly distributed throughout the These variations are important if highways are to effectively serve
day, facilities are often designed to meet peak demands occurring peak demands without breakdown.
for periods as short as 15 min or 1 hr. During other time periods, As discussed in Chapter 6, breakdowns into LOS F operation
highways are often underused. Similarly, traffic does not distribute may occur because of the inability to process demand for periods
equally over available lanes or directions on a given facility. as short as 15 min. The effects of a breakdown may extend far
Whereas the nonuniformity of traffic demand in time and space beyond the time during which demand exceeds capacity and may
produces an inefficient use of available transportation resources, take up to several hours to dissipate. Thus, highways minimally
the spatial and temporal variations are an integral part of the society adequate to handle a peak-hour demand may be subject to break-
and life-style served by those resources. down if flow rates within the peak hour exceed the capacity.
Table 2-2. Reported Maximum One-Way Hourly Volumes on Selected Freeways

total volume avg. vol. per lane
location (vph) (vphpl)
4-Lane Freeways
I-66, Fairfax, Virginia 5301 2650
U.S. 71, Kansas City, Missouri 5256 2628
I-59, Birmingham, Alabama 4802 2401
I-35W, Minneapolis, Minnesota 4690 2345
I-225, Denver, Colorado 4672 2336
I-287, Morris Co., New Jersey 4624 2312
I-295, Washington, D.C. 4480 2240
I-235, Des Moines, Iowa 4458 2229
I-71, Louisville, Kentucky 4446 2223
I-55, Jackson, Mississippi 4436 2218
I-35, Kansas City, Kansas 4398 2199
CA 4, Contra Costa County, California 4342 2171
I-45, Houston, Texas 4240 2120
I-64, Charleston, West Virginia 4152 2077
U.S. 4/NH 16, Newington, New Hampshire 4083 2041
I-564, Norfolk, Virginia 3962 1982
Northern State Parkway, New York 3840 1920
I-93, Windham, New Hampshire 3804 1902
6-Lane Freeways
I-495, Montgomery Co., Maryland 7495 2498
U.S. 6, Denver, Colorado 7378 2459
I-5, Portland, Oregon 7188 2396
CA 17, San Jose, California 6786 2262
Texas 121, Bedford, Texas 6673 2224
I-35E, Dallas, Texas 6611 2203
Garden State Parkway, New Jersey 6608 2203
I-5, Seattle-Everett, Washington 6533 2177
I-15, Salt Lake City, Utah 6357 2119
I-24, Nashville, Tennessee 6280 2093
NJ 3, Secaucus, New Jersey 6251 2083
I-287, Somerset Co., New Jersey 6151 2050
I-290, Hillside, Illinois 6149 2047
I-90, Northwest Tollway, Illinois 6120 2040
I-80, Omaha, Nebraska 6113 2038
Southern State Parkway, New York 5610 1870
8-Lane Freeways
I-635, Dallas, Texas 9090 2272
Garden State Parkway, New Jersey 8911 2228
I-495, Montgomery Co., Maryland 8793 2198
I-25, Denver, Colorado 8702 2175
I-495, Fairfax, Virginia 8610 2152
I-405, Los Angeles, California 8360 2090
I-5, Seattle, Washington 8295 2073
U.S. 50, Sacramento, California 8284 2071
U.S. 59, Houston, Texas 8268 2067
I-80, W. Paterson, New Jersey 6851 1712
I-71, Columbus, Ohio 6682 1670
Tunnels
I-279, Fort Pitt Tunnel, Pittsburgh, Pennsylvania (4-lane) 4278 2139
I-376, Squirrel Hill Tunnel, Pittsburgh, Pennsylvania (4-lane) 3922 1961
I-895, Harbor Tunnel, Baltimore, Maryland (4-lane) 3166 1584
SR 1A, Callahan Tunnel, Boston, Massachusetts (2-lane, half of one-way pair) 3059 1530
I-95, Fort McHenry Tunnel, Baltimore, Maryland (8-lane) 5840 1460
SOURCE: HCQS Survey, Federal Highway Administration, Maryland Transportation Authority, and Callahan Tunnel TSM One Way Toll Project, SG Associates
and Herbert S. Levinson, 1983

Table 2-3. Reported Maximum Lane Volumes on Selected Freeways

avg. vol. per lane vol. in peak lane
location (vphpl) (vphpl)
4-Lane Freeways
I-70, Wheeling, West Virginia — 2552
I-55, Jackson, Mississippi 2218 2542
I-235, Des Moines, Iowa 2229 2466
6-Lane Freeways
8-Lane Freeways
I-70, Columbus, Ohio — 2298
I-71, Columbus, Ohio 1670 2088
SOURCE: HCQS Survey and Federal Highway Administration
Figure 2-6. Rural Interstate travel by vehicle type. (Source: Our Nation’s Highways, Selected Facts and Figures, Federal Highway
Administration, 1992)
Figure 2-7. Annual vehicle miles of travel. (Source: Our Nation’s Highways, Selected Facts and Figures, Federal Highway Administra-
tion, 1992)

Table 2-4. Reported Maximum One-Way Volumes for Selected Multilane Highways
total volume avg. vol. per lane
location (vph) (vphpl)
4-Lane Highways
U.S. 101, Sonoma County, California 4124 2062
Utah 201, Salt Lake City, Utah 3989 1995
SR 17, Bergen County, New Jersey 3776 1888
U.S. 301, Prince Georges County, Maryland 3304 1652
6-Lane Highways
U.S. 46, Passaic County, New Jersey 5596 1865
SR 3, Passaic County, New Jersey 5348 1783
U.S. 1, Essex County, Massachusetts 4776 1592
8-Lane Highways
Almaden Expressway, San Jose, California 5428 1357
SOURCE: HCQS Survey, Federal Highway Administration, and Ref. 33
Table 2-5. Reported Maximum Volumes on Selected Two-Lane Rural Highways

total peak dir.
volume volume off-peak dir.
location (vph) (vph) volume (vph)
Highways
Madera-Olsen Rd., Simi Valley, California 3107 1651 1456
Madera-Olsen Rd., Simi Valley, California 3027 1839 1188
Hwy. 1, Banff, Alberta, Canada 2450 — —
Hwy. 35/115, Kirby, Ontario, Canada 2250 — —
Wasatch Blvd., Salt Lake City, Utah 2198 1504 694
Hwy. 35, Kirby, Ontario, Canada 2050 — —
U.S. 50, Lake Tahoe, California 1796 1386 410
NJ 50, Cape May Co., New Jersey 1714 1445 269
Hwy. 1, Banff-Yoho, Alberta–British Columbia, Canada 1517 — —
Bridges/Tunnels
U.S. 158, Nags Head, North Carolina 3195 — —
Midtown Tunnel, Norfolk/Portsmouth, Virginia 2920 1827 1093
Sagamore Bridge, Hudson, New Hampshire 2701 — —
TH 15, St. Cloud, Minnesota 2242 1146 1096
Underwood Bridge, Hampton, New Hampshire 1960 1041 919
Staley Viaduct, Decatur, Illinois 1919 971 948
SOURCE: HCQS Survey and Federal Highway Administration
Seasonal peaks in traffic demand are also of great importance, 3. Daily traffic patterns vary by month of year most severely
particularly for primarily recreational facilities. Highways serving for recreational routes.
beach resort areas may be virtually unused during much of the
These observations lead to the conclusion that commuter and
year, only to be subject to regular congestion during peak summer
business-oriented travel occurs in more uniform patterns and that
periods.
recreational traffic is subject to the greatest variation among trip-
The sections that follow present observed patterns of time varia-
purpose components of the traffic stream.
tion in traffic demand for various types of facilities in North
The data for Figure 2-8b were collected on the same Interstate
America.
route. One segment is within 1 mi of the central business district
of a large metropolitan area. The other segment is within 50 mi
Seasonal and Monthly Variations of the first but serves a combination of recreational and intercity
business travel. The wide difference in seasonal variation patterns
Seasonal fluctuations in traffic demand reflect the social and for the two segments underscores the effect of trip purpose and
economic activity of the area being served by the highway. Figure may also reflect capacity restrictions on the urban section.
2-8 shows monthly variation patterns observed in Illinois and Min-
nesota. Several significant characteristics are apparent:
Daily Variations
1. Monthly variations are more severe on rural routes than on
urban routes. Volume variations by day of the week are also related to the
2. Monthly variations are more severe on rural routes serving type of highway on which observations are made. Figure 2-9 shows
primarily recreational traffic than on rural routes serving primarily that weekend volumes are lower than weekday volumes for high-
business routes. ways serving predominantly business travel, such as urban free-
Table 2-6. Reported Maximum One-Way Volumes on Selected Urban Arterials

avg. vol- total avg. flow
total ume per flow rate per
volume lane g/C rate lane
location (vph) (vphpl) ratio (vphg) (vphgpl)
4-Lane Arterials
Ill. 83, DuPage Co., Illinois 3819 1910 0.80 4774 2387
So. Virginia St. (US 395), Reno, Nevada 2831 1415 0.62 4566 2282
Tara Blvd., Clayton, Georgia 2137 1068 0.47 4547 2272
Dougall Ave. SB, Windsor, Ontario, Canada 2240 1120 0.60 3733 1867
Antoine, Houston, Texas 2310 1155 0.65 3553 1777
Woodway WB, Houston Texas 2156 1078 0.76 2836 1418
5-Lane Arterials
North Shepard NB, PM, Houston, Texas 2100 1050 0.60 3500 1750
6-Lane Arterials
Col. 2, Denver, Colorado 3435 1145 0.50 6870 2290
US 74/NC 27, Charlotte, North Carolina 4882 1627 0.80 6102 2034
Almaden Expressway, San Jose, California 3960 1320 0.66 6000 2000
Ygnacio Valley Road, Walnut Creek, California 3790 1263 0.65 5831 1943
Southwest Trafficway, Kansas City, Missouri 3492 1164 0.60 5820 1940
U.S. 19, Clearwater, Florida 4305 1435 0.75 5740 1913
Ward Parkway, Kansas City, Missouri 3477 1159 0.61 5700 1900
Seward Highway, Anchorage, Alaska 3177 1059 0.70 4538 1513
8-Lane Arterials
Telegraph Rd., Detroit, Michigan 4400 1100 0.60 7333 1833
FM 1093, Houston, Texas 4500a 1125 0.70 6429 1607
FM 1093, Houston, Texas 4268a 1067 0.70 6097 1524
a
9-ft lanes.
SOURCE: HCQS Survey, Federal Highway Administration, Case Studies in Access Management, Draft Final Report, F. J. Koepke, Jr., and Herbert Levinson,
1992
ways. In comparison, peak traffic occurs on weekends on main The repeatability of hourly variations is of great importance.
rural and recreational access facilities. Furthermore, the magnitude The stability of peak-hour demands affects the feasibility of using
of daily variation is highest for recreational access routes and least such values in design and operational analysis of highways and
for urban commuter routes. other transportation facilities. Figure 2-12 shows data obtained
Figure 2-10 shows the variation in traffic by vehicle type for over a 77-day period in metropolitan Toronto. The shaded area
the shoulder lane of an urban freeway. Truck traffic is the most indicates the range within which one can expect 95 percent of the
severely reduced on weekends. observations to fall. Whereas the variations by hour of the day are
The extent of daily volume variation decreases as volume in- typical for urban areas, the relatively narrow and parallel fluctua-
creases, often reflecting the effect of capacity restrictions on tions among the 77 days indicate the repeatability of the basic
demand. pattern. The observations shown were obtained from detectors
Although the values shown in Figures 2-9 and 2-10 are illustra- measuring one-way traffic only, as evidenced by the single peak
tive of typical patterns that may be observed, they are not meant hour shown for either morning or afternoon.
to substitute for local studies and analyses. The average daily It is again noted that the data of Figures 2-11 and 2-12 are
traffic averaged over a full year is referred to as the annual average typical of observations that can be made. The patterns illustrated,
daily traffic, or AADT, and is often used in forecasting and however, vary in response to local travel habits and environments,
planning. and these examples should not be used as a substitute for locally
obtained data.
Hourly Variations
Peak Hour and Design Hour
Typical hourly variation patterns are shown in Figure 2-11,
where the patterns are related to highway type and day of the Capacity and other traffic analyses focus on the peak hour of
week. The typical morning and evening peak hours are evident traffic volume, because it represents the most critical period for
for urban commuter routes on weekdays. The evening peak is operations and has the highest capacity requirements. The peak-
generally somewhat more intense than the morning peak, as shown hour volume, however, is not a constant value from day to day or
in Figure 2-11. On weekends, urban routes show a peak that is less from season to season.
intense and more ‘‘spread out,’’ occurring early to midafternoon. If the highest hourly volumes for a given location were listed in
Recreational routes also have single daily peaks. Saturday peaks descending order, a large variation in the data would be observed,
on such routes tend to occur in the late morning or early afternoon depending on the type of route and facility under study.
(as travelers go to their recreational destination) and in late after- Rural and recreational routes often show a wide variation in
noon or early evening on Sundays (as they return home). peak-hour volumes. Several extremely high volumes occur on a
1,000th-highest hour. Figure 2-13 includes all hours, not just peak
hours of each day.
It is apparent from these characteristics that traffic engineers
are faced with the need for substantial judgments. Provision of a
recreational facility adequate to handle the highest peak-hour vol-
ume of the year for a given level of service results in gross under-
utilization of capacity during all but a few hours of the year. On
the other hand, providing sufficient capacity for the 30th, 100th,
500th, or other hour would guarantee the occurrence of substantial
congestion and delay during those special event or recreational
peak hours occurring less frequently.
The selection of an appropriate hour for planning, design, and
operational purposes is a compromise between providing an ade-
quate level of service for every (or almost every) hour of the year
and economic efficiency. Customary practice in the United States
is to base rural highway design on an hour between the 30th- and
100th-highest hour of the year. This range generally encompasses
the ‘‘knee’’ of the curve—the area in which the slope of the curve
changes from sharp to flat. For rural highways, the knee has often
been assumed to occur at the 30th-highest hour, which is often
used as the basis for estimates of design hour volume. For urban
roadways, a design hour for the repetitive weekday peak periods
is common. Signal timing analyses frequently use traffic conditions
typical of certain periods of the day or week.
Past studies (7,8) have emphasized the difficulty in locating a
distinct knee on hourly volume curves. Figure 2-14 shows hourly
volumes for all hours of the year at a Kentucky counting station.
The first and third curves illustrate the continuous nature of the
relationship, with no distinct breaks or knee in the decreasing
hourly volume pattern. The second curve shows a rather spread-
out knee, which could easily be located anywhere within the first
100 hr. These curves illustrate the point that arbitrary selection of
Figure 2-8(a). Examples of monthly traffic volume variations a design hour between the 30th- and 100th-highest hour is not a
showing monthly variations in traffic for a freeway in Minnesota. rigid criterion and indicate the need for local data on which to
make informed judgments.
Since the first knee may be followed by another occurring on
different days of the week or month or with different prevailing
few selected weekends or other peak periods, and traffic during
trip purposes, it is advisable to identify the nature of traffic in the
the rest of year is at much lower volumes, even during the peak
hour. This occurs because the traffic stream consists of few daily highest hours.
or frequent users; the major component of traffic is generated by The selection of a design hour must consider the impact of the
seasonal recreational activities and special events. selection on the higher-volume hours that are not accommodated.
Urban routes, on the other hand, show very little variation in The recreational access route curve of Figure 2-14 shows that the
peak-hour traffic. Most users are daily commuters or frequent highest hours of the year have more than twice the volume of the
users, and occasional and special event traffic are at a minimum. 100th hour, whereas the highest hours of an urban radial route are
Furthermore, many urban routes are filled to capacity during each only about 15 percent higher than the volume in the 100th hour.
peak hour, and variation is therefore severely constrained. In many Use of a design criterion set at the 100th hour would create substan-
urban areas, both the a.m. and p.m. peak periods extend for more tial congestion on a recreational access route during the highest-
than 1 hr. volume hours but would have less effect on an urban facility, where
Figure 2-13 shows hourly volume relationships measured on a the variation in peak-hour volumes is less. Another consideration is
variety of highway types in Minnesota. Recreational facilities show the level of service objective. A route designed to operate at LOS
the widest variation in peak-hour traffic, with values ranging from B can absorb larger amounts of additional traffic than a route
30 percent of the AADT occurring in the highest hour of the year designed to operate at LOS D during those extreme hours of the
to about 15.3 percent of AADT occurring in the 200th-highest year with higher volumes than the design hour. As a general guide,
hour of the year and 8.3 percent in the 1,000th-highest hour of the the most repetitive peak volumes may be used for the design, and
year. Main rural facilities also display a wide variation. The highest the level of service during higher-volume periods should be tested
hour is subjected to 17.9 percent of the AADT, decreasing to 10 as to the acceptability of the resulting traffic conditions.
percent in the 100th hour and 6.9 percent in the 1,000th hour. The proportion of AADT occurring in the design hour is often
Urban radial and circumferential facilities show far less variation. referred to as the K-factor. It is expressed as a decimal and varies
The range in percent of AADT covers a narrow band, from approx- on the basis of the hour selected for design or planning application
imately 11.5 percent for the highest hour to 7 to 8 percent for the and the characteristics of the subject route and its development
Figure 2-8(b). Examples of monthly traffic volume variations showing relative traffic volume trends by route type on rural roads in
Lake County, Illinois. (Source: Ref. 51)
environment. Where the K-factor is based on the 30th-highest hour In Figure 2-15 the maximum 5-min rate of flow is 2,232 vph,
of the year, several general characteristics can be noted: whereas the maximum rate of flow for a 15-min period is 1,980
vph. The full hour volume is only 1,622 vph. A design for a peak
1. The K-factor generally decreases as the AADT on a highway
5-min flow rate would result in substantial excess capacity during
increases.
the rest of the peak hour, whereas a design for the peak-hour
2. The reduction rate for high K-factors is faster than that for
volume would result in congestion for a substantial portion of the
lower values.
hour. Note that Figure 2-15 treats discrete 15-min periods for
3. The K-factor decreases as development density increases.
clarity. In practice, the peak 15 min may occur during any 15-min
4. The highest K-factors generally occur on recreational facili-
interval within the hour.
ties, followed by rural, suburban, and urban facilities, in descend-
Consideration of these peaks is important. Congestion due to
ing order.
inadequate capacity occurring for only a few minutes could take
substantial time to dissipate because of the dynamics of breakdown
Subhourly Variations in Flow flow, which are explained in greater detail in Chapter 6. Fifteen-
min flow rates have been selected as the basis for most procedures
Whereas volume forecasts for long-range planning studies are of this manual to incorporate these peak flows.
frequently expressed in units of AADT (vehicles per day), subse- Five-min flow rates have been avoided, since research has
quently reduced to hourly volumes, the analysis of level of service shown them to be statistically unstable. The operational effects of
is based on peak rates of flow occurring within the peak hour. a 5-min flow surge are virtually impossible to predict with any
Most of the procedures in this manual are based on peak 15- certainty.
min rates of flow. Figure 2-15 shows the substantial short-term The relationship between the peak 15-min flow rate and the full
fluctuation in flow rate that can occur during an hour. The data hourly volume is given by the peak-hour factor, defined in Part A
shown are for I-35W in Minneapolis, Minnesota, in 1983. of this chapter.
Figure 2-9. Examples of daily traffic variation by type of route.

Legend: MR curve represents main rural route I-35, Southern
Minnesota, AADT 10,823, 4-lanes, 1980; RA curve represents
recreational access route MN 169, North-Central Lake Region,
AADT 3,863, 2-lanes, 1981; UF curve represents urban freeway,
four freeways in Minneapolis–St. Paul, AADTs 75,000–130,000,
6–8 lanes, 1982. (Source: Minnesota Department of
Transportation) Figure 2-10. Daily variation in traffic by vehicle type. Data for
this figure were collected on I-494, 4-lanes, in Minneapolis–St.
Paul. (Source: Minnesota Department of Transportation)
Whether the design hour was measured, established from the

analysis of peaking patterns, or based on modeled demand, the
peak-hour factor (PHF) is applied to determine design hour flow it at night, may display as much as a 2:1 imbalance in directional
rates. flows. Recreational and rural routes may also be subject to signifi-
Peak-hour factors in urban areas generally range between 0.80 cant directional imbalances, which must be considered in the de-
and 0.98. Lower values signify greater variability of flow within sign process. Table 2-7 gives the directional distribution on various
the subject hour, and higher values signify little flow variation. highway types in Minnesota between 1980 and 1982.
Peak-hour factors over 0.95 are often indicative of high traffic Directional distribution is an important factor in highway
volumes, sometimes with capacity constraints on flow during the capacity analysis. This is particularly true for two-lane rural
peak hour. highways. Capacity and level of service vary substantially on
the basis of directional distribution because of the interactive
nature of directional flows on such facilities. Procedures for
SPATIAL DISTRIBUTIONS
two-lane highways include explicit consideration of directional
distribution.
Whereas traffic volume varies in time, it also varies in space.
Whereas there is no explicit consideration of directional distribu-
The two critical spatial characteristics of interest in capacity analy-
tion in the analysis of multilane facilities, the distribution has a
sis are directional distribution and lane distribution. Volume may
dramatic impact on both design and level of service. As indicated
also vary longitudinally along various segments of a facility, but
in Table 2-7, urban radial routes have been observed to have
this does not explicitly affect capacity analysis computations. Each
up to two-thirds of their peak-hour traffic in a single direction.
facility segment serving different traffic demands must be analyzed
Unfortunately, this peak occurs in one direction during the morning
separately.
and in the other in the evening. Thus, both directions of the facility
must be adequate for the peak directional flow. This characteristic
Directional Distribution has led to the use of reversible lanes on some urban freeways and
arterials.
During any particular hour, traffic volume may be greater in Directional distribution is not a static characteristic. It changes
one direction than in the other. An urban radial route, serving by hours of the day, day of the week, season, and from year to
strong directional demands into the city in the morning and out of year. Development in the vicinity of highway facilities often in-
Figure 2-12. Repeatability of hourly traffic variations for four

2-lane arterials in Toronto, Ontario, Canada. (Source: Ref. 6)
Lane Distribution
Figure 2-11. Examples of hourly traffic variations for rural routes When two or more lanes are available for traffic in a single
in New York State. (Source: Ref. 5) direction, the distribution in lane use varies widely. The lane distri-
bution depends on traffic regulations, traffic composition, speed
and volume, the number of and location of access points, the
origin-destination patterns of drivers, development environment,
and local driver habits.
Because of these factors, there are no ‘‘typical’’ lane distribu-
tions. The procedures of this manual assume an average capacity
duces traffic growth that changes the existing directional
of multilane uninterrupted flow facilities of 2,200 pcphpl. It
distribution.
is recognized that flow in some lanes will be higher and in
The proportion of traffic occurring in the peak direction of travel
others lower. Recent data collected as part of the Highway
during peak hours is often denoted as D. The K-factor, the propor-
Capacity and Quality of Service Committee survey of high-
tion of AADT occurring in the design hour, was discussed pre-
volume facilities indicate no consistency in lane distribution.
viously. These factors are used to estimate the peak-hour traffic
Data indicate that the peak lane on a six-lane freeway, for
volume in the peak direction using the following equation:
example, may be the shoulder, middle, or median lane, de-
pending on local conditions.
DDHV = AADT × K × D (2-11)
Table 2-9 gives lane distribution data for various vehicle types
where
on selected freeways. These are illustrative and are not intended
DDHV = directional design hour volume (vph), to represent ‘‘typical’’ values.
AADT = average annual daily traffic (vpd), The trend indicated in Table 2-9 is reasonably consistent
K = proportion of AADT occurring in the peak direction, throughout North America. Heavier vehicles tend toward the right-
and hand lanes, partially because they may operate at lower speeds
D = proportion of peak hour traffic in peak direction. than other vehicles and partially because of regulations prohibiting
them from using leftmost lanes.
The product of the factors K and D is given for a number of Lane distribution is a critical factor in the analysis of freeway
facilities in Table 2-8. The product gives the proportion of AADT ramp junctions, because the traffic in the shoulder lane forms
occurring in the maximum direction of the peak hour. the merge or diverge volume in conjunction with ramp vehicles.
Figure 2-13. Ranked hourly volumes on Minnesota highways. (Source: Minnesota Department of Transportation)
Procedures for their analysis in Chapter 5 focus on estimating

traffic in the shoulder lane as well as truck presence in the lane.
TRAFFIC COMPOSITION
The fraction of trucks, recreational vehicles, and buses in the

traffic stream is also required to apply the procedures of this man-
ual. Adjustments for these three categories of vehicles, especially
as they relate to grade-climbing capabilities, are given for each of
the procedures in following chapters.
Lighter-weight vehicles dominate the new-car market. Figures
2-16 and 2-17 show trends in passenger car power characteristics
since 1967, with projections to 1995. Whereas the trend is clearly
toward less powerful vehicles (as indicated by the ratio of horse-
power to weight in Figure 2-16), the average 1995 vehicle will
have about 85 percent of the hp/lb of an average 1978 vehicle.
The impact of these changes on capacity and operations is expected
to be minimal.
IMPACT OF WEATHER ON MAXIMUM VOLUMES
There have been relatively few efforts to quantify the effects of

adverse weather on capacity. Some measure of the impact can be
gained from studies conducted on two freeways with automated
data collection systems—the Gulf Freeway (I-45) in Houston (11)
and I-35W in Minneapolis (12). For both freeways, observations
were made on three-lane segments influenced by bottlenecks such
that a history of ‘‘capacity volumes’’ was available. For the Gulf
Figure 2-14. Ranked hourly volume distribution showing indis- Freeway, it was reported that rain significantly reduces capacity
tinct knee for Kentucky location in 1977. (Source: Ref. 7) by 14 to 19 percent compared with clear-weather values.
Figure 2-15. Relationship between short-term and hourly flows. (Source: Minnesota Department of Transportation)
Table 2-7. Directional Distribution Characteristics Results from the I-35W study suggested that even a trace of
precipitation reduced capacity by 8 percent. Each 0.01 in./hr in-
percent traffic in peak directions crease in rainfall resulted in a further decrease of 0.6 percent in
type of facility capacity. When precipitation falls as snow, the impact is even
highest hour
greater: an additional 2.8 percent decrease in capacity for each
of the year urban circ urban radial rural
0.01 in./hr of snow (water equivalent) beyond the initial trace
1st 53 66 57 decrease of 8 percent.
10th 53 66 53
50th 53 65 55 The procedures of this manual do not specifically account for
100th 50 65 52 inclement weather conditions. However, in areas where such con-
SOURCE: Minnesota Department of Transportation, 1980–1982 ditions are prevalent, analysts may wish to modify results to ac-
count for these impacts.
Table 2-8. Observed Values of K and D on Selected Freeways and Expressways

annual volumes in peak
direction
average average
city and 1990 year number daily % 2-way volume
urbanized area count of traffic vehicles aadt per lane
population facility taken lanes (2-way) (1-way) (K × D) (vphpl)
Atlanta, Ga. I-20 E. of CBD at Moreland Ave. 1984 8 99,900 7,794 7.8 1,948
2,157,806 I-20 at Martin Luther King Jr. Drive 1984 8 91,200 (5,198) 5.7 (1975) (1,299)
I-75 S. of CBD at University Ave. 1984 8 146,050 (8,179) 5.6 (1975) (2,045)
I-75 N. of CBD (N. of I-85) 1984 8 82,830 (5,135) 6.2 (1975) (1,284)
I-85 N. of I-75 at Monroe Dr. 1984 8 95,300 6,765 7.1 1,641
Boston, Mass. I-93 N. of I-495 1984 6 76,500 5,200 6.8 1,733

2,775,370 S.E. Expressway at Southampton St. 1982 6 143,300 6,860 4.8 2,286
I-95 E. of Rt. 128 N. of Middlesex 1984 8 125,050 7,282 5.8 1,823
Denver, Colo. I-25 S. of I-70 1984 8 175,000 7,500 4.3 1,875

1,517,977 I-70, Colorado Blvd. to Dahlia 1984 6 114,000 4,650 4.1 1,550
U.S. 6 W. of Federal Blvd. 1985 6 112,000 5,835 5.2 1,945
Detroit, Mich. I-96 Jeffers Freeway at Warren 1980 8 67,600 6,270 9.3 1,568
3,697,529 Lodge at E. Grand Blvd. 1981 6 111,450 4,660 4.2 1,558
Houston, Tex. I-10 E. of Taylor St. 1985 10 151,000 7,600 5.0 1,520
2,901,851 I-10 E. of McCarty 1985 8 110,200 7,530 6.8 1,882
I-610 at Ship Channel 1985 10 103,200 5,540 5.4 1,108
Milwaukee, Wis. N.-S. Freeway at Wisconsin 1984 8 118,080 5,730 4.5 1,342
1,226,293 N.-S. Freeway at Greenfield 1984 8 110,050 6,380 5.8 1,595
E.-W. Freeway at 26th St. 1984 6 121,150 5,700 4.7 1,900
Zoo Freeway at Wisconsin 1984 6 110,730 4,760 4.3 1,581
Airport Freeway at 68th 1984 6 81,020 3,940 4.9 1,313
New York, N.Y. Holland Tunnel 1982 4 73,200 2,700 3.7 1,350
16,044,012
San Francisco, Calif. I-80 Oakland Bay Bridge 1984 10 223,000 8,898 4.0 1,780
3,629,516
Washington, D.C. I-66 Theodore Roosevelt Bridge 1984 6 86,200 (7,413) 8.6 (1975) (2,471)
3,363,061 Anacostia Freeway at Howard Road 1984 6 121,700 (6,085) 5.0 (1975) (2,028)
NOTE: Values in parentheses based on K × D value for the year indicated if different from the year the count was taken.
SOURCE: Characteristics of Urban Transportation Demand—An Update, July 1988, Charles River Associates in association with H. S. Levinson, and Ref. 52
Table 2-9. Lane Distribution by Vehicle Type

percent distribution by lane
b
highway vehicle type lane 1 lane 2 lane 3
Lodge Freeway, Detroit Lighta 29.2 38.4 32.4
SU Trucks 30.8 61.5 7.7
Combinations 88.5 2.9 8.6
All Vehicles 30.9 37.8 31.3
I-95, Connecticut Turnpike Lighta 34.6 40.9 24.5

All Vehicles 37.1 40.4 22.5
I-4, Orlando, Florida All Vehicles 29.9 31.7 38.4

a
Passenger cars, panel trucks, and pickup trucks.
b
Lane 1 = shoulder lane; lanes numbered from shoulder to median.
SOURCE: Ref. 14 and Florida Department of Transportation, 1993

Figure 2-16. Distribution of power-to-mass ratios of passenger cars. (Source: Ref. 9)
Figure 2-17. On-highway passenger car characteristics. (Source: Ref. 10, Figure 2-13)
IV. SPEED CHARACTERISTICS
NATIONAL SPEED TRENDS between average truck and passenger car speed was about 7 to 8
mph. After 1973, this difference was reduced considerably, to
Nationwide speed trends through 1975 are shown in Figure 2- about 2 mph, because of the lower overall speeds being observed.
18a for various vehicle types and in Figure 2-18b for all vehicles Figure 2-18b indicates that speeds have been gradually increas-
on Interstate rural highways through 1991. ing despite the 55-mph speed limit. With the restoration of the 65-
Figure 2-18a, for main rural highways, shows a clear increasing mph speed limit on some of the major roadways, further increases
speed trend from 1942 through the middle of 1972. This reflects in average speeds can be expected.
the better design of both highways and vehicles occurring through- Table 2-10 confirms the increasing speed trends on U.S. high-
out this period. In 1973, in response to a severe fuel shortage, the ways. All of the highways referenced in Table 2-10 had a 55-mph
55-mph national speed limit was introduced, and a sharp decline speed limit in effect.
in speeds was observed. The figure also shows that buses and Aside from the general interest in the speed limit issue, these
passenger cars travel at similar speeds on rural highways, whereas speed trends have an impact on the procedures presented in this
trucks travel at somewhat lower speeds. To 1973, the difference manual. Uninterrupted flow procedures incorporate national aver-
Figure 2-18. Nationwide speed trends through 1975 and 1993. (Source: Ref. 13 and Highway Statistics)
Table 2-10. National Spot Speed Trends for 55-mph Facilities

average speed median speed 85th percentile percent > 55
fiscal year (mph) (mph) speed (mph) mph
Urban Interstate Highways
1985 57.2 57.4 64.0 64.1
1987 58.0 58.0 64.8 67.4
1989 58.9 59.0 66.1 71.3
1991 58.8 58.8 66.1 69.8
Rural Interstate Highways
1985 59.5 59.4 66.1 75.4
1987 59.7 59.7 66.5 73.7
1989 60.1 60.3 67.2 76.8
1991 59.9 59.4 67.2 75.5
Rural Arterials
1985 54.9 55.2 61.7 50.5
1987 55.9 56.1 62.8 54.3
1989 56.2 56.4 63.1 56.0
1991 56.4 56.3 63.1 56.5
Urban Principal Arterials
1985 53.5 53.6 60.5 42.1
1987 54.0 54.1 60.7 44.7
1989 54.6 55.1 61.3 47.7
1991 54.0 53.9 60.8 42.2
NOTE: All highways have 55-mph speed limit.
SOURCE: Highway Statistics, Federal Highway Administration, 1992

age speed-flow and speed-density trends. The exact shape of these I-35W in Minneapolis. Figure 2-19 shows a weekday variation
curves and the calibration of speeds (especially at the free-flow pattern, whereas Figure 2-20 shows a similar distribution for a
end of the relationships) reflect current trends. Curves used in this Saturday.
manual allow for average speeds up to 70 mph, 5 or 15 mph over In these exhibits note that speed remains relatively constant
the usual speed limit, in response to the observed increase in driver- despite significant changes in volume. In Figure 2-19, speed shows
selected speeds under free-flow conditions. a marked response to volume increases only when the volume
exceeds approximately 1,600 vphpl. This trend is illustrated later
and is an important characteristic in all of the procedures of this
SPEED VARIATION BY TIME OF DAY manual. If speed does not vary with rate of flow over a broad
range of flows, it becomes difficult to use speed as the sole measure
Figures 2-19 and 2-20 show variations of speed with time of of effectiveness defining level of service. This important character-
day, along with hourly volume variations, over a 24-hr period for istic is the major reason that such measures as density and percent
Figure 2-19. Speed variation by hour of day for I-35W in Minneapolis, weekdays, in relation to volume variations. (Source: Minnesota
Department of Transportation)
Figure 2-20. Speed variation by hour of day for I-35W, Minneapolis, Saturdays, in relation to volume variations. (Source: Minnesota
Department of Transportation)
time delay have been introduced as primary measures of effective- SPEED VARIATION BY LANE AND DAY VERSUS NIGHT
ness for uninterrupted flow facilities, with speed playing a second- Table 2-11 gives a comparison of speeds by day versus night
ary role. conditions on the Connecticut Turnpike near Bridgeport. The table
The speeds in Figures 2-19 and 2-20 are also virtually the shows that day/night variations are slight, on the order of 1 mph.
same, despite significantly lower volumes on weekends. This Variations by lane are considerably greater, a factor indicated in
is a reflection of driver populations and trip purpose effects. Table 2-12 for a number of other facilities.
Saturday drivers may be less familiar with the facility, or, if Level of service speed criteria in the manual refer to average
familiar, they do not drive with the same sense of urgency values across all lanes of the facility or all lanes in one direction
devoted to the daily commute to work. Procedures of this of the facility. The data indicate that drivers in general are using
manual also take this into account by introducing adjustments the lanes of multilane facilities as intended—slower drivers to the
for driver population types in several chapters. right and faster drivers in the middle and median lanes.

Table 2-11. Average Speed by Day vs. Night and Lane in mph
lane 1a lane 2 lane 3
vehicle type day night day night day night
Passenger cars 49.5 48.8 57.7 57.4 65.1 61.6
Trucks 47.5 46.4 54.3 54.6 59.4 58.1
Percent trucks in lane (15.0) (17.3) (7.5) (13.0) (0.7) (5.4)
a
SOURCE: Ref. 14
Table 2-12. Average Speeds by Lane in mph

avg. volume
per lane
location lane 1a lane 2 lane 3 lane 4 (vph)
N.J. Turnpike 46 55 60 — 1120
Conn. Turnpike 49 57 64 — 692
L.I. Expwy., N.Y. 52 56 57 — 1460
I-8, San Diego 49 51 58 62 1503
44 48 53 55 2386
SR 94, San Diego 50 53 57 56 1282
47 49 52 49 2168
I-4, Orlando, Florida 56 61 61 — —
a
SOURCE: Refs. 14 and 15, California Department of Transportation, 1984, and Florida Department of Transportation, 1993
V. MEASURED RELATIONSHIPS FOR UNINTERRUPTED FLOW
Part A of this chapter introduced the generalized basic form of on an empirical rather than on an analytical approach to defining
the relationships among speed, flow rate, and density for uninter- the shape of the curves and the values of capacity.
rupted flow facilities. Rarely is it possible to observe these charac-
teristics, especially at flow rates approaching capacity, under ideal
conditions. Practically all data collected for the calibration of such SPEED-FLOW RELATIONSHIPS
relationships are subject to the influences of changing environmen-
tal conditions, nonhomogeneity of vehicles in the traffic stream,
and (particularly for urban facilities) lack of complete isolation Freeways
from ramps and interchanges.
The shape and calibration of such relationships are important, Research attention has largely focused on the speed-flow rela-
because they provide the basis for the selection of measures of tionship. These two variables are the traffic stream characteristics
effectiveness and the definition of level of service ranges for unin- most often measured and have been traditionally used in the assess-
terrupted flow facilities. Such relationships also serve to estimate ment of traffic operations.
the capacity of uninterrupted flow facilities and the operating con- The general shape of the speed-flow data tend to be similar
ditions under which it occurs. Estimation of capacity requires clear regardless of the location within North America. Figure 2-21 shows
identification of the maximum flow point on a speed-flow or speed- data from San Diego, California (21). Figure 2-22 shows data from
density curve, a process fraught with difficulty because of the near Toronto, Ontario, Canada (22). The lack of data points in the
stochastic nature of the observations near capacity. Even under low-flow, high-speed region of the curve reflects the fact that these
ideal conditions, observations of capacity flows will not be con- surveys took place during busy morning peak periods. In both
stant but will form a distribution of values. It is not possible to cases the data were collected in 30-sec intervals, but the data were
take a single measurement and know with certainty where it fits averaged over 6-min intervals in San Diego and 5-min intervals
within that distribution. in Toronto. Both figures show high speeds up to what may be
In recognition of such difficulties, many researchers have devel- capacity flows and a cluster of points at a speed half of that at
oped analytical models describing these relationships. These mod- capacity. Figure 2-21 (San Diego) also includes more points on
els have sometimes been used to identify the capacity of a highway what may be considered the lower branch of the theoretical curve.
by extrapolation of data for both uncongested and congested condi- Other studies support that general pattern with some variations:
tions. The results of many research projects show the difficulties Figure 2-23 from a Northern California study (23) shows only the
of extrapolation from these models in the vicinity of capacity flows uncongested top portion of the curve with a gradual but minor
(16–20). Work within the last decade has therefore focused more decrease of speed from 60 to 50 mph with increasing flow. The
Figure 2-23. Observed speed-flow relationship at Caldecott Tun-

nel in 15-min sampling intervals (California State Highway 24,
1990). (Source: Ref. 23)
Figure 2-21. Observed speed-flow relationship on a San Diego
freeway in 6-min sampling intervals (Interstate Highway 8, 1987). have different characteristics. The relationship between speed and
(Source: Ref. 21) flow found on a two-lane rural highway in Alberta, Canada (35),
is shown in Figure 2-24. The curve shows a virtually constant
speed for two-way flows up to 2,400 pcph, and the entire speed
range is only 59 to 50 mph for the full range of flows. Speeds of
50 mph are not unusual at capacity of two-lane highways. Most
of the new speed-flow data for multilane flow also suggest that
capacity occurs at a critical speed in the vicinity of 50 mph (33).
It should be remembered, however, that capacity of a two-lane
highway occurs at a total flow of between 2,000 and 2,800 pcph
depending on directional distribution, whereas for multilane high-
ways the flow at capacity is 2,200 pcphpl. Speeds on multilane
highways for similar per lane flows (1,000 to 1,400 pcphpl) are
well over 50 mph. The capacity of two-lane highways is more
influenced by interactions between directional flows than by road-
way space availability. As a result, other measures have been
proposed as primary level of service criteria for multilane and two-
lane roadways. They are discussed in Chapters 7 and 8.
Figure 2-22. Observed speed-flow relationship on an Ontario
freeway in 5-min sampling intervals (Queen Elizabeth Way near DENSITY-FLOW RELATIONSHIPS
Toronto, 1987). Different data symbols represent different survey
days. (Source: Ref. 22) A number of early analytical efforts paid considerable attention
to density-flow relationships. The main problem with this approach
to the capacity problem is in the direct measurement of density.
maximum flow has been reached at slightly over 2,200 vph. Other
In most instances, its values were calculated from observed speeds
recent studies also lend support to the general form of the relation-
and flows using Equation 2-4. Some researchers have fit continu-
ship shown in these figures (24–28). Some studies suggested that
ous curves through density-flow data, yielding a single maximum
there is in fact a drop in the maximum observed flows with the
flow rate. Others have projected discontinuous curves through data,
onset of congestion (29–31), whereas others failed to identify this
with one curve treating stable flow points and another unstable or
effect (32). Note that these curves differ from the functions used
forced flow points. In these cases two maxima are achieved, one
in the 1985 Highway Capacity Manual.
for each curve. All such models indicate that the maximum flow
In measuring the speed-flow relationship, it is important to use
rate for the stable flow curve is considerably higher than that for
appropriate time intervals, since they strongly influence the form of
the unstable flow curve, perhaps as much as 200 vph higher. This
the curve, especially around the capacity flow and in the congested
is an interesting feature that projects a discontinuity in flow near
region. Five-min intervals are recommended as the shortest time
capacity, the point at which flow breaks down. It also explains the
base for practical purposes.
difficulty in recovering from a breakdown, since the maximum
flow that can be achieved from an unstable flow condition is less
Multilane and Two-Lane Rural Highways than that for stable flow. A paper by Easa and May (20) contains
several sample calibrations and illustrations of density-flow data.
Whereas the bulk of the data for uninterrupted flow come from It has been shown, however, that determination of density from
freeways, these conditions also occur on rural multilane and two- speed and flow leads to a biased estimate of the relationships
lane roads. In many respects, traffic flow on multilane roadways (36,37). The emphasis on the density-flow relationship has there-
is similar to that on freeways (33,34). Two-lane roads, however, fore decreased recently. Occupancy in time, which is readily avail-
Figure 2-24. Speed-flow relationship for two-lane rural highways. (Source: Ref. 4)
able from many freeway control systems, is sometimes used. The Figure 2-25 shows relatively few headways less than 1.0 sec.
general shape of the occupancy-flow relationship is similar to the A vehicle traveling at 60 mph (88 ft/sec) would have a spacing of
density-flow curve. 88 ft with a 1.0-sec headway, and only 44 ft with a 1⁄2-sec headway.
This effectively reduces the space between vehicles (rear bumper
to front bumper) to only 25 to 30 ft and would be extremely
HEADWAY DISTRIBUTIONS AND RANDOM FLOW difficult to maintain and would allow little margin for driver error.
Drivers react to this intervehicle spacing, which they perceive
directly, rather than to the traditional front bumper–to–front
At any given lane flow rate, the mean or average headway is
bumper measures used by traffic engineers. The latter include the
the reciprocal of flow rate. Thus, at a flow of 1,200 vphpl, the
length of the vehicle, which became smaller for passenger cars in
average headway is 3,600/1,200 or 3 sec. Vehicles do not, how- the vehicle mix of the 1980s. If drivers maintain essentially the
ever, travel at constant headways. Vehicles tend to travel in groups, same intervehicle spacing and car lengths continue to get shorter,
or platoons, with varying headways between successive vehicles. some increases in capacity could conceivably result.
An example of the distribution of headways observed on the Long If traffic flow were truly random, small headways (less than 1.0
Island Expressway is shown in Figure 2-25. Lane 3 has the most sec) would occur frequently. Several mathematical models have
uniform headway distribution, as evidenced by the range of values been developed that recognize the absence of small headways in
and the high frequency of the modal value—the peak of the distri- most traffic streams as described by Gerlough and Huber (15).
bution curve. The distribution of Lane 2 is similar to that of Lane These models have been useful in developing simulation models
3, with slightly greater scatter (range from 1⁄2 to 9.0 sec). Lane 1 of traffic flow, thereby extending research on flow characteristics
shows a much different pattern: it is far more dispersed, with beyond conditions that can be observed and monitored in the field.
headways ranging from 1⁄2 to 12.0 sec, and the frequency of the Traffic flow in urban areas is rarely purely random. Traffic signals
modal value is only about one-third of that for the other lanes. and other controls regulate flows, and the trip generation character-
This reflects the lower flow usually occurring in Lane 1 (shoulder istics of adjacent land generally produce trips in a nonrandom
lane) and the driver desires of Lane 1 users. fashion.
Figure 2-25. Time headway distribution for Long Island Expressway. (Source: Ref. 40)
VI. INTERRUPTED FLOW FACILITIES
SATURATION FLOW AND LOST TIME AT SIGNALIZED wheels as measurement benchmark. Some other national practices
INTERSECTIONS apply different definitions or measurement techniques of saturation
flow (41–44). For that reason, the values quoted in international
The basic concepts of saturation headway and saturation flow literature are not quite comparable (38). The Canadian survey
rate and of start-up and change interval lost times were introduced technique (42,43), however, allows the estimation of saturation
flow rates for situations with queues as short as four to five vehi-
in Part A. The empirical studies referenced in this section span
cles. Saturation flow rates cited in various sources may also be
more than 40 years, from Greenshields in 1946 to the latest re-
influenced by the choice of vehicle positions included and by the
search before the printing of this update.
definition of lost time (45).
Table 2-13 summarizes the results of representative past and
Although most studies of intersection discharge headways have
recent studies. The table indicates that saturation headways have
focused on the observation of the first 10 to 12 vehicles, there is
been becoming shorter in the last decade and, consequently, satura-
tion flow rates have been increasing. This trend has been observed
by both practicing professionals and researchers (38). In the table,
saturation headway ranges from a low of 1.8 sec to a high of 2.4
sec—corresponding to a range of saturation flow rate of 2,000 to
1,500 vphgpl.
Figure 2-26 shows vehicle headway by position in the queue
resulting from several past studies. It shows that, in most studies,
the saturation headway does not become established until the sixth
or seventh vehicle in the queue, indicating that the first five or six
vehicles experience some start-up lost time. In discussing the re-
sults of Figure 2-26, Berry and Gandhi (40) noted that the variation
in discharge headways of the first several vehicles depended on
the choice of a screenline for measuring headways rather than any
real difference in the observed headways. Stop lines or curb lines
have been used in combination with the front bumper, front or rear
axles, or rear bumper. Caution is therefore advisable in comparing
values of discharge headways from different studies. The update Figure 2-26. Comparison of various research results on queue
of Chapter 9 uses only the stop line as the screenline and rear discharge headways. (Source: Ref. 39)
Table 2-13. Observed Saturation Flow Rates at Signalized Intersections

saturation flow
measurement
starting with saturation
date of queue position start-up lost saturation flow headway
source study city or state sample size number time (sec) rate (pcphgpl) (sec)
Gerlougha 1967 Los Angeles, Santa Monica, 6 Int. 5 2.05 1470 2.45
California
Carstensb 1971 Ames, Iowa 4 Int. 4 0.75 1572 2.29
Kingc 1976 Nationwide 5 — 1682 2.14
Agentd 1983 Lexington, Kentucky 4 1.40 1651 2.18
Leee 1986 Lawrence, Kansas 5 3.04 1827 1.97
Molinaf 1986 Austin, Dallas, Houston 8 hr 6 — 2000 1.8
3 Int.
Zegeerg 1986 Chicago, Houston, Los 7 Int. 5 1.31 1875 1.92
Angeles
Fambroh 1987 Houston (peak) 30 hr 5 — 1896 1.9
2 Int.
Fambroh 1987 Houston (off peak) 30 hr 5 — 1832 1.97
2 Int.
Fambroh 1987 Los Angeles (peak) 34 hr 5 — 1936 1.86
2 Int.
Fambroh 1987 Los Angeles (off peak) 34 hr 5 — 1785 2.02
2 Int.
Prevedourosi 1988 Chicago 6.25 hr 4 — 2000 1.8
10 Int.
Roessj 1988 California, New York, 5 Int. 5 — 1791 2.01
Texas
(single lane)
Roessj 1988 California, Illinois, New 7 Int. 5 — 1937 1.86
York, Texas
(multilane)
Shortk 1989 College Station 30 hr 4 1.31 1905 1.89
2 Int.
Gastonl 1991 Dallas 25 hr 5 — 1910 1.88
4 Int.
Zegeerm 1992 Florida 16 Int. — — 1840 1.96
a
Gerlough, D.L., and Wagner, F.A., ‘‘Improved Criteria for Traffic Signals at Individual Intersections.’’ NCHRP Report 32 (1967).
b
Carstens, R.L., ‘‘Some Parameters at Signalized Intersections.’’ Traffic Engineering (Aug. 1971).
c
King, G., and Wilkinson, M., ‘‘Relationship of Signal Design to Discharge Headway, Approach Capacity, and Delay.’’ Transportation Research Record 615,
Transportation Research Board, Washington, D.C. (1973).
d
Agent, K., and Crabtree, J., Analysis of Lost Times at Signalized Intersections. Report, Kentucky Transportation Research Program, University of Kentucky,
Lexington, Ky. (1983).
e
Lee, J., and Chen, L.R., ‘‘Engineering Headway at Signalized Intersections in Small Metropolitan Area.’’ Transportation Research Record 1091, Transportation
Research Board, Washington, D.C. (1986).
f
Molina, C.J., Messer, C.J., and Fambro, D.B., Passenger Car Equivalencies for Large Trucks at Signalized Intersections. TTI Research Report 397-2, Texas
Transportation Institute, Texas A&M University, College Station, Tex. (1987).
g
Zegeer, J.D., ‘‘Field Validation of Intersection Capacity Factors.’’ Transportation Research Record 1091, Transportation Research Board, Washington, D.C.
(1986).
h
Fambro, D.B., Chang, E.P.C., and Messer, C.J., ‘‘Effects of the Quality of Traffic Signal Progression on Delay.’’ NCHRP Report 339 (1991).
i
Prevedouros, P.D., and Jovanis, P.P., ‘‘Validation of Saturation Flows and Progression Factors for Traffic Actuated Signals.’’ Transportation Research Record
1194, Transportation Research Board, Washington, D.C. (1988).
j
Roess, R.P., Papayannoupoulis, J.M., Ulerio, J.M., and Levinson, H.S., Levels of Service in Shared-Permissive Left-Turn Lane Groups at Signalized Intersec-
tions. Report DTFH 61-87-C-00012, Transportation Training and Research Center, Polytechnic University, Brooklyn, N.Y. (1989).
k
Short, J.T., ‘‘Effects of Dips and Bumps on Saturation Flow Rates at Signalized Intersections.’’ Thesis, Texas A&M University, College Station, Tex. (1989).
l
Gaston, G.D., ‘‘An Operational Analysis of Protected Lead-Lag Left Turn Phasing.’’ Thesis, Texas A&M University, College Station, Tex. (1991).
m
Zegeer, J., Data Collected for Florida Department of Transportation (1992).

some indication that the saturation headway may increase some- tations. A data collection technique to measure saturation flow is
what when green time becomes quite long. This effect implies that described in an appendix of Chapter 9. Signalized intersection
green phases longer than 40 or 50 sec may not be proportionally procedures of this manual rely heavily on saturation headway and
as efficient as those in the normal range (43). lost time calibrations as a means of describing the use of available
Zegeer (46) has shown the significance of prevailing conditions green time.
of lane width, parking, transit interference, pedestrian interference,
turning movements, flow composition, signal progression, and GAP ACCEPTANCE AND SATURATION FLOW AT
other factors, all of which influence saturation flow values. UNSIGNALIZED INTERSECTIONS
For ideal conditions, including 12-ft lanes, all through vehicles,
all passenger cars, no parking, no transit interference, and low Two variables are used to estimate capacity flow rate at two-
pedestrian volumes, the procedures of Chapter 9 recommend a way stop-controlled intersections: the critical gap and the follow-
up time (50).
saturation flow rate of 1,900 pcphgpl, corresponding to a saturation
Typical values of the critical gap for an urban two-way stop-
flow headway of 1.9 sec. This represents an increase of 100 pcphpl
controlled intersection on a four-lane road with a 30-mph speed
compared with the 1985 manual.
limit range from 5.0 sec for the left-turning traffic from the major
Start-up lost times were also measured during the studies men-
street to 6.5 sec for the left-turning traffic from the minor street.
tioned in Table 2-13 and other research projects (47,48) for a
Follow-up gaps range from 2.1 to 3.4 sec for the same maneuvers.
variety of conditions, including city size (population), location In effect, the follow-up gap is a saturation flow headway, since it
within the city, signal timing, speed limit, and other factors. Typi- is the gap between consecutive minor street vehicles using a long
cal values observed range from 1.0 to about 2.0 sec. Corresponding major street gap.
values of change-interval lost time range from 1.2 to 2.8 sec. The Saturation flow at a stop line of an all-way stop-controlled inter-
length of the change interval (yellow + all red) has a significant section depends mostly on the presence of vehicles at other inter-
effect on the value observed. The latest research suggests that the section approaches. When no traffic was present on other
lost time associated with a phase may be up to 3.0 sec shorter intersection approaches, the saturation flow rate on a single lane
than its change interval (49). approach was measured at 1,100 vph (50), whereas for an intersec-
The variation in the data in Table 2-13 and the importance of tion with four evenly loaded approaches and ideal conditions,
prevailing conditions suggest that local data collection to determine 2,000 vph was achieved. These saturation flows correspond to the
saturation flow rate and lost time can lead to more accurate compu- departure headways of 3.3 and 7.2 sec, respectively.
VII. SUMMARY
The range and use of important highway traffic characteristics 2-14 presents a summary. In terms of passenger car units, capacity
in capacity analysis have been addressed in this chapter. It is under ideal conditions is characterized by 2,200 pcphpl for uninter-
emphasized that these characteristics are not uniform or constant rupted flow along four-lane freeways and multilane highways and
throughout North America and that variations due to local driving 2,300 pcphpl on freeways with six or more lanes. On two-lane
habits and environments are to be expected. Direct measurement rural highways, capacity ranges from 2,000 to 2,800 pcph total for
of such characteristics may be used to fine-tune or improve the both directions of flow, depending on the directional split of vol-
results of the analysis procedures of this manual, which are based ume. At signalized intersections, 1,900 passenger car units (pcu)
on observed national averages. The values and relationships pre- can depart from the stop line for each hour of green time, on a
sented in this chapter provide a backdrop to capacity and service per lane basis. The capacity of minor approaches of unsignalized
levels discussed in the following chapters. intersections is influenced by the type of control and by the com-
The expression ‘‘capacity’’ depends on the units being observed peting traffic flows. It varies from 500 to 1,100 vph.
(vehicles, passenger car units, pedestrians), the time period, and Because specific local situations are seldom ideal, downward
the area of the facility being considered (lane, width in feet, area). adjustments are normally made to account for actual operating
Each facility type has specific units for expressing capacity. Table conditions. The following chapters detail these procedures.

Table 2-14. Capacity by Facility Type

units of capacity (ideal
facility units time perioda area flow conditions)
Freeway
Basic section, four Passenger cars Hour Lane pcphplb 2,200
lanes
Basic section, six or Passenger cars Hour Lane pcphpl 2,300
more lanes
Weaving area Passenger cars Hour Lane pcphpl 1,900
Ramp junction Passenger cars Hour Merge or diverge pcph 2,000
area
One-lane ramp Passenger cars Hour Ramp roadway pcph 1,700
Multilane highway Passenger cars Hour Lane pcphpl 2,200
Two-lane highway Passenger cars Hour Both lanes pcph 2,800c
Signalized intersection Passenger cars Hour of Lane pcphgpl 1,900d
green
Unsignalized intersection
Two-way stop con- Passenger cars Hour Lane or pcph 1,060e
trolled movement
All-way stop con- Vehicles Hour Entering lane vph 500–1,100f
trolled
Urban arterialg
Exclusive transit bus Buses Hour Lane bphpl 90–120
lane on urban arterial
with stops
Pedestrian walkway Pedestrians Minute Foot of effective p/min/ft 25
width
Bikeway Bicycles Hour Lane bike/hr 2,150h
a
Time periods of 1 hr are usually based on a peak 15-min volume expanded to an ‘‘hourly rate of flow.’’
b
Passenger cars per hour per lane.
c
For 50-50 volume split by direction.
d
Saturation flow rate, in passenger cars per hour of green per lane.
e
Potential capacity with no conflicting volume.
f
Depending on volume distribution from conflicting approaches.
g
Capacity usually measured and controlled by most restrictive signalized intersection.
h
Middle of reported range.
REFERENCES
1. Drake, J., Schofer, J., and May, A., ‘‘A Statistical Analysis Transportation Research Record 596, Transportation Re-
of Speed-Density Hypotheses.’’ Highway Research Record search Board, Washington, D.C. (1976).
154, Transportation Research Board, Washington, D.C. 7. Crabtree, J., and Deacon, J., ‘‘Highway Sizing.’’ Transpor-
(1967). tation Research Record 869, Transportation Research Board,
2. May, A.D., Traffic Flow Fundamentals. Prentice-Hall, Engle- Washington, D.C. (1982).
wood Cliffs, N.J. (1990) pp. 284–315.
8. Werner, A., and Willis, T., ‘‘Cost-Effective Level of Service
3. Two-Lane Rural Roads: Design and Traffic Flow. Organiza-
and Design Criteria.’’ Transportation Research Record 699,
tion for Economic Cooperation and Development, Paris,
France (1972). Transportation Research Board, Washington, D.C. (1979).
4. Messer, C., Two-Lane, Two-Way Rural Highway Capacity. 9. Glauz, W., Harwood, D., and St. John, A., ‘‘Projected
Final Report, NCHRP Project 3-28A, Texas Transportation Vehicle Characteristics Through 1995.’’ Transportation Re-
Institute, College Station, Tex. (1983). search Record 772, Transportation Research Board, Washing-
5. Transportation and Traffic Engineering Handbook. 2nd Edi- ton, D.C. (1980).
tion, Prentice-Hall, Englewood Cliffs, N.J. (1982). 10. Highway Statistics—Summary to 1975. Federal Highway Ad-
6. McShane, W., and Crowley, K., ‘‘Regularity of Some ministration, Washington, D.C. (1975).
Detector-Observed Arterial Traffic Volume Characteristics.’’ 11. Jones, E., Goolsby, M., and Brewer, K., ‘‘The Environmen-
tal Influence of Rain on Freeway Capacity.’’ Highway Re- Transportation Research Record 1194, Transportation Re-
search Record 321, Transportation Research Board, search Board, Washington, D.C. (1988).
Washington, D.C. (1970). 29. Agyemang-Duah, K., and Hall, F.L., ‘‘Freeway Capacity
12. Ries, G., Impact of Weather on Freeway Capacity. Minnesota Drop and the Definition of Capacity.’’ Transportation Re-
Department of Transportation, Minneapolis, Minn. (1981). search Record 1320, Washington, D.C. (1991).
13. Hurdle, V., and Datta, P., ‘‘Speeds and Flows on an Urban 30. Hall, F.L., and Agyemangh-Duah, K., ‘‘Some Isues Re-
Freeway: Some Measurements and a Hypothesis.’’ Transpor- garding the Numerical Value of Capacity.’’ In Highway Ca-
tation Research Record 905, Transportation Research Board, pacity and Level of Service, Branolte, U. (ed.), Balkema,
Washington, D.C. (1983). Rotterdam, Netherlands (1991).
14. Huber, M., and Tracy, J., ‘‘Operating Characteristics of 31. Banks, J.H., ‘‘The Two-Capacity Phenomenon: Some Theo-
Freeways.’’ NCHRP Report 60 (1968). retical Issues.’’ Transportation Research Record 1320, Trans-
15. Gerlough, D., and Huber, M., Traffic Flow Theory. Special portation Research Board, Washington, D.C. (1991).
Report 165, Transportation Research Board, Washington, D.C. 32. Urbanik, T., Hinshaw, W., and Barnes, K., ‘‘Evaluation of
(1975). High-Volume Urban Texas Freeway.’’ Transportation Re-
16. Drake, J., Schofer, J., and May, A., ‘‘A Statistical Analysis search Record 1320, Transportation Research Board, Wash-
of Speed-Density Hypotheses.’’ Vehicular Traffic Science, ington, D.C. (1991).
American Elsevier, New York, N.Y. (1967). 33. Reilly, W.R., Harwood, D.W., Schoen, J.M., Kuehl, R.O.,
17. Drake, J., Schofer, J., and May, A., ‘‘A Statistical Analysis Bauer, K., and St. John, A.D., Capacity and Level of Service
of Speed-Density Hypotheses.’’ Highway Research Record Procedures for Multilane Rural and Suburban Highways. Fi-
154, Transportation Research Board, Washington, D.C. nal Report, NCHRP Project 3-33, JHK & Associates and Mid-
(1967). west Research Institute (May 1989).
18. Ceder, A., ‘‘Investigation of Two-Regime Traffic Flow Mod- 34. Roess, R., McShane, W., and Pignataro, L., ‘‘Freeway
els at the Micro- and Macroscopic Levels.’’ Thesis, University Level of Service: A Revised Approach.’’ Transportation Re-
of California at Berkeley, Berkeley, Calif. (1975). search Record 699, Transportation Research Board, Washing-
19. Ceder, A., and May, A.D., ‘‘Further Evaluation of Single- ton, D.C. (1980).
and Two-Regime Traffic Models.’’ Transportation Research 35. Morrall, J., Two-Lane, Two-Way Highway Capacity—Cana-
Record 567, Transportation Research Board, Washington, dian Data and Input to NCHRP 3-28A. Report, Texas Trans-
D.C. (1976). portation Institute, College Station, Tex. (1983).
20. Easa, S.M., and May, A.D., ‘‘Generalized Procedures for 36. Duncan, N.C., ‘‘A Note on Speed/Flow/Concentration Rela-
Estimating Single- and Two-Regime Traffic Flow Models.’’ tions.’’ Traffic Engineering and Control, London, England
Transportation Research Record 772, Transportation Re- (1976) p. 34.
search Board, Washington, D.C. (1980). 37. Duncan, N.C., ‘‘A Further Look at Speed/Flow/Concentra-
21. Banks, J.H., ‘‘Freeway Speed-Flow-Concentration Relation- tion.’’ Traffic Engineering and Control, London, England
ship: More Evidence and Interpretations.’’ Transportation Re- (1979) p. 482.
search Record 1225, Transportation Research Board, 38. Teply, S., and Jones, A.M., ‘‘Saturation Flow: Do We Speak
Washington, D.C. (1989). the Same Language?’’ Transportation Research Record 1320,
22. Hall, F.L., and Hall, L.M., ‘‘Capacity and Speed-Flow Transportation Research Board, Washington, D.C. (1991).
Analysis of the QEW in Ontario.’’ Transportation Research 39. King, G., and Wilkinson, M., ‘‘Relationship of Signal De-
Record 1287, Transportation Research Board, Washington, sign to Discharge Headway, Approach Capacity, and Delay.’’
D.C. (1990). Transportation Research Record 615, Transportation Re-
23. Chin, H.C., and May, A.D., ‘‘Examination of the Speed- search Board, Washington, D.C. (1976).
Flow Relationship at the Caldecott Tunnel.’’ Transportation 40. Berry, D., and Gandhi, P., ‘‘Headway Approach to Intersec-
Research Record 1320, Transportation Research Board, Wash- tion Capacity.’’ Highway Research Record 453, Transporta-
ington, D.C. (1991). tion Research Board, Washington, D.C. (1973).
24. Wemple, E.A., Morris, A.M., and May, A.D., ‘‘Freeway 41. Akcelik, R., Traffic Signals: Capacity and Timing Analysis.
Capacity and Level of Service Concepts.’’ In Highway Capac- Research Report ARR 123, Australian Road Research Board,
ity and Level of Service, Branolte, U. (ed.), Balkema, Rotter- Victoria, Australia (1981).
dam, Netherlands (1991). 42. Teply, S. (ed.), Canadian Capacity Guide for Signalized In-
25. Allen, B.L., Hall, F.L., and Gunther, M.A., ‘‘Another tersections (1st edition). Institute of Transportation Engineers,
Look at Identifying Speed-Flow Relationships on Freeways.’’ District 7, Canada, and University of Alberta, Edmonton, Al-
Transportation Research Record 1005, Transportation Re- berta, Canada (1984).
search Board, Washington, D.C. (1985). 43. Teply, S., Allingham, D., Richardson, D., and Stephen-
26. Gunther, M.A., and Hall, F.L., ‘‘Transitions in the Speed son, B., Factual Draft, Second Edition of the Canadian Ca-
Flow Relationship.’’ Transportation Research Record 1091, pacity Guide for Signalized Intersections. Institute of
Transportation Research Board, Washington, D.C. (1986). Transportation Engineers, District 7, Canada (1993).
27. Hurdle, V.F., and Datta, P.K., ‘‘Speeds and Flows on an 44. Richtlinien fuer Lichtsignalanlagen, Forschungsgesellschaft
Urban Freeway: Some Measurements and a Hypothesis.’’ fuer Strassen- und Verkehrswesen, Köln, Germany (1992).
Transportation Research Record 905, Transportation Re- 45. Bonneson, J.A., ‘‘Study of Headway and Lost Time at
search Board, Washington, D.C. (1983). Single-Point Urban Interchanges.’’ Transportation Research
28. Persaud, B.N., and Hurdle, V.F., ‘‘Some New Data that Record 1365, Transportation Research Board, Washington,
Challenge Some Old Ideas About Speed-Flow Relationships.’’ D.C. (1992).
46. Zegeer, J.D., ‘‘Field Validation of Intersection Capacity Fac- 50. Transportation Research Circular 373: Interim Materials on
tors.’’ Transportation Research Record 1091, Transportation Unsignalized Intersection Capacity. Transportation Research
Research Board, Washington, D.C. (1986). Board, Washington, D.C. (1991).
47. Agent, K., and Crabtree, J., Analysis of Saturation Flow 51. Mutanyi, T., ‘‘A Method of Estimating Traffic Behavior on
at Signalized Intersections. Report, Kentucky Transportation All Routes in a Metropolitan County.’’ Highway Research
Research Program, University of Kentucky, Lexington, Ky. Record 41, Transportation Research Board, Washington, D.C.
(May 1982). (1963).
48. Agent, K., and Crabtree, J., Analysis of Lost Times at Sig- 52. Levinson, H., Characteristics of Urban Transportation De-
nalized Intersections. Report, Kentucky Transportation Re- mand—A Handbook for Transportation Planners. Urban Mass
search Program, University of Kentucky, Lexington, Ky. Transportation Administration, Washington, D.C. (1978).
(Feb. 1983).
49. Bonneson, J., ‘‘Change Interval Timing and Lost Time for
Single Point Urban Interchanges.’’ Journal of Transportation
Engineering, Vol. 118, No. 5 (1992).

Highway Capacity Manual 1997
Chapter 3
BASIC FREEWAY SECTIONS
CONTENTS
I. INTRODUCTION .................................................................................................. 3-1

Definitions ............................................................................................... 3-1
Freeway Facilities ............................................................................ 3-1
Freeway Capacity Terms................................................................. 3-1
Base Conditions for Freeway Capacity ........................................... 3-2
Flow Characteristics................................................................................ 3-2
Ideal Conditions for Freeway Flow .................................................. 3-3
Speed-Flow Relationship ................................................................. 3-3
Free Flow.................................................................................. 3-3
Queue Discharge and Congested Flow ................................... 3-4
Factors Affecting Free-Flow Speed ................................................. 3-4
Lane Width and Lateral Clearance .......................................... 3-4
Number of Lanes ...................................................................... 3-5
Interchange Density.................................................................. 3-6
Other Factors............................................................................ 3-7
Vehicle Equivalents ......................................................................... 3-7
Driver Population ............................................................................. 3-8
II. METHODOLOGY ................................................................................................. 3-8
Performance Measures........................................................................... 3-8
Levels of Service .................................................................................... 3-8
Basic Relationships................................................................................. 3-14
Determination of Flow Rate ............................................................. 3-14
Peak-Hour Factor ..................................................................... 3-15
Heavy-Vehicle Adjustment Factor ............................................ 3-15
Extended General Freeway Segments ............................. 3-15
Specific Grades ................................................................. 3-16
Equivalents for Extended General Freeway Segments ........... 3-16
Level Terrain ..................................................................... 3-16
Rolling Terrain ................................................................... 3-16
Mountainous Terrain ......................................................... 3-16
Equivalents for Specific Upgrades ........................................... 3-16
Equivalents for Specific Downgrades....................................... 3-18
Equivalents for Composite Grades .......................................... 3-19
Computation of Heavy-Vehicle Factor ..................................... 3-19
Driver Population Adjustment ................................................... 3-19
Determination of Free-Flow Speed ................................................. 3-19
Field Measurement ................................................................... 3-20
Estimation Guidelines ............................................................... 3-20
Ideal Free-Flow Speed ...................................................... 3-20
Lane Width ........................................................................ 3-21
Lateral Clearance .............................................................. 3-21
Number of Lanes............................................................... 3-21
Interchange Density .......................................................... 3-22
Determination of Level of Service ................................................... 3-22
III. APPLICATIONS ................................................................................................... 3-23
Segmenting the Freeway ........................................................................ 3-23
Computational Steps............................................................................... 3-24
Planning Analysis.................................................................................... 3-25
Updated December 1997 Page 3-i Basic Freeway Sections

HOV Facilities ......................................................................................... 3-25
Tools for Analysis ................................................................................... 3-25
IV. EXAMPLE PROBLEMS ......................................................................................... 3-26
Example Problem 1 ................................................................................ 3-26
Example Problem 2 ................................................................................ 3-28
Example Problem 3 ................................................................................ 3-30
Example Problem 4 ................................................................................ 3-32
Example Problem 5 ................................................................................ 3-34
V. REFERENCES .................................................................................................... 3-36
APPENDIX I. Precise Procedure for Determining Passenger-Car Equivalents
of Trucks on Composite Upgrades ................................................................ 3-37
APPENDIX II. Worksheet for Analysis of Basic Freeway Sections................... 3-40
FIGURES AND TABLES
Figure 3-1. Example of Basic Freeway Section........................................................... 3-2

Figure 3-2. Speed-Flow Relationships ......................................................................... 3-4
Figure 3-3. Queue Discharge and Congested Flow .................................................... 3-5
Figure 3-4. LOS Criteria ............................................................................................... 3-10
Figure 3-5. Worksheet for Analysis of Basic Freeway Sections.................................. 3-14
Figure I.3-1. Sample Solution for Composite Grade.................................................... 3-37
Figure I.3-2. Performance Curves for Standard Trucks (200 lb/hp) ............................ 3-38
Table 3-1. LOS Criteria for Basic Freeway Sections ................................................... 3-11
Table 3-2. Passenger-Car Equivalents on Extended General
Freeway Segments....................................................................................................... 3-16
Table 3-3. Passenger-Car Equivalents for Trucks and Buses
on Specific Upgrades ................................................................................................... 3-17
Table 3-4. Passenger-Car Equivalents for Recreational Vehicles on Specific
Upgrades ...................................................................................................................... 3-18
Table 3-5. Passenger-Car Equivalents for Trucks and Buses on Specific
Downgrades.................................................................................................................. 3-18
Table 3-6. Adjustment Factors for Lane Width ............................................................ 3-21
Table 3-7. Adjustment Factors for Right-Shoulder Lateral Clearance......................... 3-21
Table 3-8. Adjustment Factors for Number of Lanes .................................................. 3-22
Table 3-9. Adjustment Factors for Interchange Density .............................................. 3-22
Basic Freeway Sections Page 3-ii Updated December 1997

I. INTRODUCTION
The procedures in this chapter are intended to analyze the capacity, level of
service, lane requirements, and effects of traffic and design features of basic freeway
sections.
DEFINITIONS
Freeway Facilities
A freeway may be defined as a divided highway with full control of access and
two or more lanes for the exclusive use of traffic in each direction.
Freeways provide uninterrupted flow. There are no signalized or stop-controlled Freeways provide
at-grade intersections, and direct access to and from adjacent property is not permitted. uninterrupted flow.
Access to and from the freeway is limited to ramp locations. Opposing directions of
flow are continuously separated by a raised barrier, an at-grade median, or a raised
traffic island.
Operating conditions on a freeway primarily result from interactions among
vehicles and drivers in the traffic stream and between vehicles and their drivers and the
geometric characteristics of the freeway. Operations can also be affected by
environmental conditions, such as weather or lighting conditions, by pavement
conditions, and by the occurrence of traffic incidents.
A tollway or toll road is similar to a freeway, except that tolls are collected at Toll road is similar to a
designated points along the facility. Although the collection of tolls does involve freeway.
interruptions to traffic, these facilities may generally be treated as freeways. However,
special attention should be given to the unique characteristics, constraints, and delays
caused by toll collection facilities.
A freeway consists of three component parts:
T Basic freeway sections: Segments of the freeway that are outside of the Basic sections are outside
influence area of ramps or weaving areas. the influence of ramps or
T Weaving areas: Segments of the freeway where two or more vehicle flows weaving.
must cross each other’s path along a length of the freeway. They are usually formed
when merge areas are followed by diverge areas. They are also formed when an on-
ramp is followed by an off-ramp and the two are connected by an auxiliary lane (for
analysis of weaving areas, see Chapter 4).
T Ramp junctions: Points at which on- and off-ramps join the freeway. The
junction formed at this point is an area of turbulence because of concentrations of merging
or diverging vehicles (for analysis of ramps and ramp junctions, see Chapter 5).
Figure 3-1 illustrates a basic freeway section. The integration of these three
component parts into a freeway facility is covered in Chapter 6.
Freeway Capacity Terms

T Freeway capacity: the maximum sustained 15-min rate of flow, expressed in
passenger cars per hour per lane (pcphpl), that can be accommodated by a uniform
freeway segment under prevailing traffic and roadway conditions in a specified
direction.
T Traffic characteristics: any characteristic of the traffic stream that may affect
capacity, free-flow speed, or operations, including the percentage composition of the
traffic stream by vehicle type and the familiarity of drivers with the roadway.
T Roadway characteristics: the geometric characteristics of the freeway segment
under study, including the number and width of lanes, right-shoulder lateral clearance,
interchange spacing, vertical alignment, and lane configurations.
T Free-flow speed: the mean speed of passenger cars under low to moderate
flow rates that can be accommodated on a uniform freeway section under prevailing
roadway and traffic conditions.
It should be noted that capacity analysis is based on freeway segments with
uniform traffic and roadway conditions. If any of these prevailing conditions change
Updated December 1997 Page 3-1 Basic Freeway Sections

See Chapter 4, Weaving FIGURE 3-1. EXAMPLE OF BASIC FREEWAY SECTION
Areas.
See Chapter 5, Ramps and
Ramp Junctions.
See Chapter 6, Freeway
Systems.
significantly, the capacity of the segment and its operating conditions change as well.
Therefore, each uniform segment should be analyzed separately.
Base Conditions for Freeway Capacity
Conditions under which the full capacity of a basic freeway section is achieved are
good weather, good visibility, and no incidents or accidents. When one or more of
these conditions fail to exist, the speed, level of service, and capacity of the freeway
section all tend to be reduced.
FLOW CHARACTERISTICS
Traffic flow within basic freeway sections can be highly varied depending on the
conditions at upstream and downstream bottleneck locations that constrict the flow into
and out of the freeway section. Bottlenecks can be created by ramp merge and
weaving areas, lane drops, maintenance and construction activities, accidents, and
objects in the road. An incident does not have to block a travel lane to create a
bottleneck. Disabled vehicles in the median or shoulder can influence traffic flow
within freeway lanes.
Recent freeway research has resulted in a better understanding of the
characteristics of freeway flow relative to the influence of upstream and downstream
bottlenecks. Traffic flow within a basic freeway segment can generally be categorized
into three flow types: free flow, queue discharge flow, and congested flow. Each flow
type can be defined within general speed-flow-density ranges and represents different
conditions on the freeway.
T Free flow represents traffic flow that is unaffected by upstream or downstream
conditions. This flow regime is generally defined within a speed range of 55 to
75 mph at low to moderate flow rates and a range of 45 to 65 mph at high flow rates.
T Queue discharge flow represents traffic flow that has just passed through a
bottleneck and is accelerating back to the free-flow speed of the freeway. Queue
discharge flow is characterized by relatively stable flow as long as the effects of
another bottleneck downstream are not present. This flow type is generally defined
within a narrow range of flows, 2,000 to 2,300 pcphpl, with speeds typically ranging
from 35 mph up to the free-flow speed of the freeway section. Lower speeds are
typically observed just downstream of the bottleneck. Depending on horizontal and
vertical alignment, queue discharge flow usually accelerates back to the free-flow
Basic Freeway Sections Page 3-2 Updated December 1997
speed of the facility within 1⁄2 to 1 mi downstream from the bottleneck. Recent studies
suggest that the queue discharge flow rate from the bottleneck is lower than the
maximum flows observed before a breakdown. A general rule of thumb for this drop
in flow rate is approximately 5 percent.
T Congested flow represents traffic flow that is influenced by the effects of a
downstream bottleneck. Traffic flow in the congested regime can vary over a broad
range of flows and speeds depending on the severity of the bottleneck. Queues may
extend many thousands of feet upstream from the bottleneck. Freeway queues differ
from queues at intersections in that they are not static or ‘‘standing.’’ On freeways
vehicles move slowly through a queue, with periods of stopping and movement.
Ideal Conditions for Freeway Flow
The specific speed-flow-density relationship depends on prevailing traffic and
roadway conditions for the basic freeway section in question. The basic characteristics
described here were established for the following ideal conditions:
T Minimum lane widths of 12 ft;
T Minimum right-shoulder lateral clearance between the edge of the travel lane
and the nearest obstacle or object influencing traffic behavior of 6 ft (minimum
median lateral clearance is 2 ft);
T Traffic stream consisting of passenger cars only;
T Ten or more lanes (in urban areas only);
T Interchanges spaced every 2 mi or more;
T Level terrain, with grades no greater than 2 percent; and
T Driver population dominated by regular and familiar users of the facility.
These ideal conditions represent the highest type of basic freeway section, one with a
free-flow speed of 70 mph or greater. It should be noted that these conditions are
considered ideal only from the point of view of free-flow speed, capacity, and level of
service, and that the term ‘‘ideal’’ has no connotation with respect to safety or other
factors.
Speed-Flow Relationship
Free Flow
Figure 3-2 describes the speed-flow relationships for free flow on basic freeway
sections.
All recent freeway studies indicate that speed on freeways is insensitive to flow if
the flow is low to moderate. This is reflected in Figure 3-2, which shows speed to be
constant for flows up to 1,300 pcphpl for a 70-mph free-flow speed. For freeways with
a lower free-flow speed, the region over which speed is insensitive to flow extends to
even higher flow rates. Thus, free-flow speed is easily measured in the field as the
average speed of passenger cars when flow rates are less than 1,300 pcphpl. Field
determination of free-flow speed is easily accomplished by performing travel time or
spot speed studies during periods of low flows.
Note that although Figure 3-2 shows only curves for free-flow speeds of 75, 70,
65, 60, and 55 mph, curves representing any free-flow speed between 75 and 55 mph
can be obtained by interpolation. Also, the speed-flow curve representing a 75-mph
free-flow speed, which corresponds with the recent increase in the posted speed limit
on many rural freeway sections throughout the United States, shown by a dashed line,
is not based on empirical field research but was created by extrapolation from the
70-mph free-flow speed curve. Capacity at free-flow speeds greater than or equal to
70 mph is considered to be 2,400 pcphpl.
Research leading to these speed-flow curves found that a number of factors affect
free-flow speed, including number of lanes, lane width, lateral clearance, and
interchange density or spacing. Other factors believed to influence free-flow speed, but
for which little is known quantitatively, include horizontal and vertical alignment,
speed limit, level of enforcement, lighting conditions, and weather.
Under ideal traffic and geometric conditions, freeways will operate with capacities
as high as 2,400 pcphpl. This capacity is typically achieved on freeways with free-flow
speeds of 70 mph or greater. As the free-flow speed decreases, there is a slight
decrease in capacity. For example, the capacity of a basic freeway section with a free-
flow speed of 55 mph is expected to be approximately 2,250 pcphpl.
The average speed of passenger cars at flow rates that represent capacity is
expected to range from 53 mph (free-flow speeds of 70 mph or greater) to 50 mph for
a section with a 55-mph free-flow speed. Note that the higher the free-flow speed, the
greater the drop in speed as flow rates move toward capacity. Thus, for a 70-mph free-
flow speed, there is a 17-mph drop from low-volume conditions to capacity conditions.
The drop is only 5 mph for a freeway with a 55-mph free-flow speed.
As indicated in Figure 3-2, the point at which an increase in flow rate begins to
affect the average passenger car speed varies from 1,300 to 1,750 pcphpl. Speed will
begin to be reduced at 1,300 pcphpl for free-flow speeds of 70 mph or greater. For
lower-speed facilities, the free-flow speed begins to diminish at higher flow rates.
Curves are based on FIGURE 3-2. SPEED-FLOW RELATIONSHIPS

research conducted in 1992–
1995 under NCHRP Project
3-45 (1).
Queue Discharge and Congested Flow

Unlike free flow, queue discharge and congested flow have not been extensively
studied, and these traffic flow types can be highly variable. However, freeway research
performed since 1990 has provided valuable insight into possible speed-flow
relationships that describe these two flow regimes. Figure 3-3 presents one suggested
relationship and is intended for informational purposes only.
This relationship is not included in the level of service (LOS) procedures in this
chapter, which address free-flow conditions only. Users of this manual are cautioned
that although the alternative relationship in Figure 3-3 may provide a general predictive
model for speed under queue discharge and congested flows, it should be considered
conceptual at best. Further research is needed to better define flow in these two
regimes.
Factors Affecting Free-Flow Speed
Factors affecting free-flow Recent research has found that the free-flow speed on a freeway depends on the
speed. traffic and roadway conditions present on a given facility. These conditions are
described in the following sections.
Lane Width and Lateral Clearance
Lateral clearance is When lane widths are less than 12 ft, drivers are forced to travel closer to one
measured from edge of travel another laterally than they would normally desire. The effect of restricted lateral
lane to curb, guardrail, or clearance is similar. When objects are located too close to the edge of the median and
other physical obstruction. roadside lanes, drivers in these lanes will shy away from them, positioning themselves
FIGURE 3-3. QUEUE DISCHARGE AND CONGESTED FLOW
further from the lane edge. This restricted lateral clearance has the same effect as
narrow lanes: it forces drivers closer together laterally. Drivers have been found to
compensate by reducing their speed. The closeness of objects has been found to have a
greater effect on drivers in the right shoulder lane than on those in the median lane.
Drivers in the median lane appear to be unaffected by lateral clearance when
minimum clearance is 2 ft, whereas drivers in the right shoulder lane are affected when
lateral clearance is less than 6 ft. Illustration 3-1 shows the effects of lane width and
lateral clearance on lateral placement of vehicles. Illustration 3-2 shows a freeway
section considered ideal with respect to lane width and lateral clearance.
Number of Lanes
The number of lanes on a freeway section influences free-flow speed. As the
number of lanes increases, so does the opportunity for drivers to position themselves to
avoid slower-moving traffic. In typical freeway driving, traffic tends to be distributed
across lanes according to speed. Traffic in the median lane or lanes typically moves
faster than in the lane adjacent to the right shoulder. Thus, a four-lane freeway (two
lanes in each direction) provides less opportunity for drivers to move around slower
traffic than does a freeway with 6, 8, or 10 lanes. The effect of decreased
maneuverability is to reduce the average speed of vehicles in the traffic stream.

Illustration 3-1.
Vehicles shy away from both

roadside and median barriers,
driving as close to the lane
marking as possible. The
existence of narrow lanes
compounds the problem,
making it difficult for two
vehicles to travel alongside
each other.
Illustration 3-2.
Ideal conditions of lane width

and lateral clearance: the
concrete median barrier does
not cause vehicles to shift
their lane position and
therefore would not be
considered an obstruction.
Interchange Density
Research findings from Merging and weaving associated with interchanges affect the speed of traffic.
NCHRP Project 3-45 (1). Freeway sections with closely spaced interchanges, such as those in heavily developed
urban areas, operate at lower free-flow speeds than sections on suburban or rural
freeways where interchanges are less frequent. Recent research that formed the basis
for the analysis procedures presented in this chapter found that speeds generally
decrease with increasing frequency of interchanges. The ideal average interchange
spacing over a reasonably long freeway section (5 to 6 mi) is 2 mi or more. The
minimum average interchange spacing over a sustained length of freeway that is
considered possible, though highly unlikely, is 1⁄2 mi.
Other Factors
Design speed, a factor found in previous LOS guidelines, is not included in this Horizontal and vertical
chapter. The design speed of the primary physical elements of a freeway can affect geometry may influence free-
travel speed. In particular, horizontal and vertical alignments of a highway may flow speed.
contribute to the free-flow speed of a given freeway section. Although factors
describing the effects of these physical features on free-flow speed are not included in
this chapter, if a freeway has somewhat extreme horizontal or vertical conditions, the
user is encouraged to determine free-flow speed from field observation and field study
rather than to rely on the estimation of free-flow speed given in this chapter.
Vehicle Equivalents
The concept of vehicle equivalents is based on observations of freeway conditions
in which the presence of heavy vehicles, including trucks, buses, and recreational
vehicles (RVs), creates less-than-ideal conditions. These nonideal conditions include
longer and more frequent gaps of excessive lengths both in front of and behind heavy
vehicles. Also, the speed of vehicles in adjacent lanes and their spacing may be
affected by these generally slower-moving large vehicles. Finally, physical space taken
up by a large vehicle is typically two to three times greater in terms of length than that
of a typical passenger car. To allow the method for estimating freeway capacity to be
based on a consistent measure of flow, each heavy vehicle is converted into the
equivalent number of passenger cars. The conversion results in a single value for flow
rate in terms of passenger cars per hour per lane. The conversion factor used depends
on the proportion of heavy vehicles present in the traffic stream as well as the length
and severity of the upgrade or downgrade.
Illustrations 3-3 and 3-4 show the effects of trucks and other heavy vehicles on
freeway traffic.
Illustration 3-3.
Large gaps form in front of

slow-moving trucks climbing
the grade.

Illustration 3-4.
Even on relatively level

terrain, large gaps in front of
trucks or other heavy vehicles
are unavoidable.
Driver Population
Among the ideal conditions defined for freeway flow is a driver population
consisting primarily of commuters. It has been noted in several studies across the
nation that non-commuter-oriented driver populations do not display the same
characteristics as commuter driver populations. For recreational traffic streams,
capacities have been observed to be as much as 20 percent lower than for commuter
traffic traveling on the same section. An effect on free-flow speed has not been
reported. This effect on capacity, however, is highly variable and should be locally
calibrated.
II. METHODOLOGY
PERFORMANCE MEASURES
A basic freeway section can be characterized by three performance measures:
density in terms of passenger cars per mile per lane, speed in terms of mean passenger
car speed, and volume-to-capacity ratio. Each of these measures is an indication of
how well or how poorly traffic flow is being accommodated by the freeway.
The assigned primary performance measure used to provide an estimate of level of
service is density. The three measures of speed, density, and flow or volume are
interrelated. When two of these measures are known, the third can be solved for.
LEVELS OF SERVICE
Although speed is a major indicator of service quality to drivers, freedom to
maneuver within the traffic stream and proximity to other vehicles are equally
noticeable concerns. These other concerns are related to the density of the traffic
stream. Furthermore, unlike speed, density increases as flow increases up to capacity,
resulting in a measure of effectiveness that is sensitive to a broad range of flows.
Density is used to define For these reasons, density is the parameter used to define levels of service for
level of service. basic freeway sections. The ranges of density used to define levels of service are as
follows:

Level of Service Density Range (pc/mi/ln)

A 0–10.0
B 10.1–16.0
C 16.1–24.0
D 24.1–32.0
E 32.1–45.0
F > 45.0
For any given level of service, the maximum allowable density is somewhat lower
than that for the corresponding level of service on multilane highways. This reflects
the higher service quality drivers expect when using freeways as compared with
surface multilane facilities. This does not imply that under similar conditions an at-
grade multilane highway will perform better than a freeway with the same number of
lanes. For any given density, a freeway will carry higher flow rates at higher speeds
than will a comparable multilane highway.
Although the specification of maximum densities for LOS A through D is based
on the collective professional judgment of the members of the Committee on Highway
Capacity and Quality of Service, the upper value shown for LOS E (45 pc/mi/ln) is
not. That value is the density at which capacity occurs for different free-flow speeds; it
is the maximum density at which sustained flows at capacity are expected to occur.
LOS criteria for basic freeway sections are provided in Table 3-1 for free-flow
speeds of 75, 70, 65, 60, and 55 mph. To be within a given level of service, the
density criterion must be met. In effect, under ideal conditions, these are the speeds
and flow rates expected to occur at the designated densities. Local variations in driving
behavior, however, may cause some variance from these expectations.
It should be noted that the LOS F operations observed within a queue are the
result of a breakdown or bottleneck at a downstream point. LOS F is also used to
describe conditions at the upstream point of the breakdown or bottleneck as well as the
operations within the queue that forms behind it.
Failure, breakdown, congestion, and LOS F occur when queues begin to form on
the freeway. Density tends to increase sharply within the queue and may be expected
to be considerably higher than the maximum value of 45 pc/mi/ln for LOS E.
Figure 3-4 shows the relationship among speed, flow, and density for basic
freeway sections. It also shows the definition of the various levels of service using
density boundary values.
Operational characteristics for the six levels of service are shown in Illustrations
3-5 through 3-10. The levels of service were defined to represent reasonable ranges in
the three critical flow variables: speed, density, and flow rate.
LOS A describes free-flow operations. Free-flow speeds prevail. Vehicles are
almost completely unimpeded in their ability to maneuver within the traffic stream.
Even at the maximum density for LOS A, the average spacing between vehicles is
about 530 ft, or 26 car lengths, which affords the motorist a high level of physical and
psychological comfort. The effects of incidents or point breakdowns are easily
absorbed at this level.
LOS B represents reasonably free flow, and free-flow speeds are maintained. The
lowest average spacing between vehicles is about 330 ft, or 17 car lengths. The ability
to maneuver within the traffic stream is only slightly restricted, and the general level
of physical and psychological comfort provided to drivers is still high. The effects of
minor incidents and point breakdowns are still easily absorbed.
LOS C provides for flow with speeds at or near the free-flow speed of the
freeway. Freedom to maneuver within the traffic stream is noticeably restricted at LOS
C, and lane changes require more care and vigilance on the part of the driver.
Minimum average spacings are in the range of 220 ft, or 11 car lengths. Minor
incidents may still be absorbed, but the local deterioration in service will be
substantial. Queues may be expected to form behind any significant blockage.
LOS D is the level at which speeds begin to decline slightly with increasing flows.
In this range, density begins to increase somewhat more quickly with increasing flow.
Freedom to maneuver within the traffic stream is more noticeably limited, and the
FIGURE 3-4. LOS CRITERIA
driver experiences reduced physical and psychological comfort levels. Even minor
incidents can be expected to create queueing, because the traffic stream has little space
to absorb disruptions. Minimum average vehicle spacings are about 165 ft, or eight car
lengths.
At its highest density value, LOS E describes operation at capacity. Operations at
this level are volatile, there being virtually no usable gaps in the traffic stream.
Vehicles are spaced at approximately six car lengths, leaving little room to maneuver
within the traffic stream at speeds that are still over 49 mph. Any disruption to the
traffic stream, such as vehicles entering from a ramp or a vehicle changing lanes, can
establish a disruption wave that propagates throughout the upstream traffic flow. At
capacity, the traffic stream has no ability to dissipate even the most minor disruptions,
and any incident can be expected to produce a serious breakdown with extensive
queueing. Maneuverability within the traffic stream is extremely limited, and the level
of physical and psychological comfort afforded the driver is poor.
LOS F describes breakdowns in vehicular flow. Such conditions generally exist
within queues forming behind breakdown points. Such breakdowns occur for a number
of reasons:
Conditions that create LOS F. T Traffic incidents cause a temporary reduction in the capacity of a short
segment, so that the number of vehicles arriving at the point is greater than the number
of vehicles that can move through it.
T Points of recurring congestion exist, such as merge or weaving areas and lane
drops where the number of vehicles arriving is greater than the number of vehicles
discharged.
T In forecasting situations, any location where the projected peak-hour (or other)
flow rate exceeds the estimated capacity of the location presents a problem.
Note that in all cases, breakdown occurs when the ratio of demand to actual
capacity or the ratio of forecast demand to estimated capacity exceeds 1.00.
Operations immediately downstream of such a point, however, are generally at or near
TABLE 3-1. LOS CRITERIA FOR BASIC FREEWAY SECTIONS
Maximum Minimum Maximum
Level of Density Speed Service Flow Maximum v/c
Service (pc/mi/ln) (mph) Rate (pcphpl) Ratio
Free-Flow Speed = 75 mph
A 10.0 75.0 750 0.31
B 16.0 75.0 1,200 0.50
C 24.0 71.0 1,704 0.71
D 32.0 65.0 2,080 0.87
E 45.0 53.0 2,400 1.00
F >45.0 <53.0 <2,400 <1.00
A 10.0 70.0 700 0.29
B 16.0 70.0 1,120 0.47
C 24.0 68.0 1,632 0.68
D 32.0 64.0 2,048 0.85
E 45.0 53.0 2,400 1.00
F var var var var
A 10.0 65.0 650 0.28
B 16.0 65.0 1,040 0.44
C 24.0 64.5 1,548 0.66
D 32.0 62.0 1,984 0.84
E 45.0 52.0 2,350 1.00
F var var var var
A 10.0 60.0 600 0.26
B 16.0 60.0 960 0.42
C 24.0 60.0 1,440 0.63
D 32.0 58.0 1,856 0.81
E 45.0 51.0 2,300 1.00
F var var var var
A 10.0 55.0 550 0.24
B 16.0 55.0 880 0.39
C 24.0 55.0 1,320 0.59
D 32.0 54.5 1,744 0.78
E 45.0 50.0 2,250 1.00
F var var var var
capacity, and downstream operations improve (assuming that there are no additional
downstream bottlenecks) as discharging vehicles move away from the bottleneck.
It should be noted that LOS F operations within a queue are the result of a
breakdown or bottleneck at a downstream point. LOS F is also used to describe both
conditions at the point of the breakdown or bottleneck and the operations within the
queue that forms upstream.
Whenever LOS F conditions exist, there is the potential for these conditions to
extend upstream for significant distances. A prerequisite for valid analyses using these
procedures is the assumption that the section under consideration is free from
downstream effects that promulgate upstream. In such cases, upstream operations will
reflect the effect of the downstream bottleneck and will not be as indicated by the
procedures of this chapter.
Illustration 3-5. LOS A
Illustration 3-6. LOS B
Illustration 3-7. LOS C

Illustration 3-8. LOS D
Illustration 3-9. LOS E
Illustration 3-10. LOS F

FIGURE 3-5. WORKSHEET FOR ANALYSIS OF BASIC FREEWAY SECTIONS
80
12
130
70
145
160 10 Analysis
Average Passenger-Car Speed

Type Input Output

60
175
50 80
A B C D E I vP, FFS LOS
(km/hr)
(mph)
40
60 II vP, LOS, FFS S
30
40
III FFS, LOS vP
20
IV vP, LOS N
20
10
0 0
0 200 400 600 800 100 1200 1400 1600 1800 2000 2200 2400
Flow Rate (pcphpl)
General Information
Analyst Date Performed
Agency or Company Analysis Type
I II III IV
Site Information Flow Rate (vP)

Highway/Dir. Travel ____________ ET __________________ Tables 3-2, 3-3, 3-5
From/To ____________ ER __________________ Tables 3-2, 3-4
Jurisdiction ____________ fHV __________________ 1
Analysis Time Period ____________ 1 + PT(ET - 1) + PR(ER - 1)
fp __________________ (1.0 - 0.85)
Analysis Year ____________ vP pcphpl V
__________________
Traffic and Roadway Conditions (PHF x N x fHV x fp)
Volume, V ____________ vph Free-Flow Speed (FFS)
Speed, S ____________ mph FFSi ___________mph
Lane Width, LW ____________ ft fLW ___________mph Table 3-6
Number of Lanes, N ____________
fLC ___________mph Table 3-7
Rt-Shoulder Lat. Clear., LC ____________ ft
fN ___________mph Table 3-81
Peak-Hour Factor, PHF ____________
fID ___________mph Table 3-9
Interchange Density, ID ____________
FFS ___________mph (est.) FFSi - fLW - fN - fLC - fID
% Trucks and Buses, PT _______________
% RVs, PR or
_______________
General Terrain FFS ___________mph (measured)
Level Rolling Mountainous
Level of Service (LOS)
Specific Grade
Length ____________ mi Density, D _______________ pc/mi/ln vP/S
Up/Down ____________ % LOS _______________ Table 3-1
Driver Type 1
For rural freeway sections, f = 0
N
Commuter/Wk Day Recreational/Wk End
Using the basic speed-flow curves (see Figure 3-4), the relationships between
levels of service, flow, and speed can be analyzed. Figure 3-5 provides a worksheet
that can be used for analysis of basic freeway sections.
BASIC RELATIONSHIPS
The determination of level of service for a basic freeway section generally
involves three components:
T Flow rate,
T Free-flow speed, and
T Level of service.
Determination of each of these components is described in the following sections.
Determination of Flow Rate

The hourly flow rate must reflect the effects of heavy vehicles, the temporal
variation of traffic flow during an hour, and the characteristics of the driver population.
These effects are reflected by adjusting hourly volume counts or estimates, typically
reported in vehicles per hour (vph), to arrive at an equivalent passenger-car flow rate
in passenger cars per hour (pcph). The equivalent passenger-car flow rate is calculated
using the heavy-vehicle and peak-hour adjustment factors and is reported on a per lane
basis, or in passenger cars per hour per lane. Equation 3-1 is used to calculate the
equivalent passenger-car flow rate. Convert vph to pcph using
heavy-vehicle and peak-hour
V factors.
vp = (3-1)
PHF × N × fHV × fp
where
vp = 15-min passenger-car equivalent flow rate (pcphpl),
V = hourly volume (vph),
N = number of lanes,
fHV = heavy-vehicle adjustment factor, and
fp = driver population factor.
Peak-Hour Factor
The peak-hour factor represents the temporal variation in traffic flow during an
hour. Observations of traffic flow consistently indicate that the flow rates found in the
peak 15-min period within an hour are not sustained during the entire hour.
Application of the peak-hour factor in Equation 3-1 accounts for this phenomenon.
On freeways, typical peak-hour factors range from 0.80 to 0.95. Lower peak-hour
factors are characteristic of rural freeways or off-peak conditions. Higher factors are
typical of urban and suburban peak-hour conditions. Users are encouraged to apply
available data to develop peak-hour factors suitable to local conditions. If local data
are unavailable, 0.85 and 0.90 may be used for rural and urban-suburban peak-hour
factors, respectively.
Heavy-Vehicle Adjustment Factor

Freeway traffic volumes that include a mix of vehicle types must be adjusted to an
equivalent flow rate expressed in passenger cars per hour per lane. This adjustment is
made using the factor fHV.
Adjustments for the presence of heavy vehicles in the traffic stream apply for Heavy-Vehicle Adjustment
three vehicle types: trucks, buses, and RVs. There is no evidence to indicate any Factor (fHV).
differences in performance characteristics between the truck and bus populations on
freeways, so trucks and buses are treated identically.
The factor fHV is found using a two-step process. First, the passenger-car
equivalent for each truck or bus and RV is found for the traffic and roadway
conditions under study. These equivalency values, ET and ER, for trucks or buses and
RVs, respectively, represent the number of passenger cars that would use the same
amount of freeway capacity as one truck or bus or RV under prevailing roadway and
traffic conditions. Second, using the values of ET and ER and the percentage of each
type of vehicle in the traffic stream (PT and PR), the adjustment factor fHV can be
computed.
The effect of heavy vehicles on traffic flow depends on grade conditions as well Truck and RV equivalency
as on traffic composition. Passenger-car equivalents can be selected for one of three values ET and ER.
conditions: extended general freeway segments, specific upgrades, and specific
downgrades.
Extended General Freeway Segments. It is often possible to consider an
extended length of freeway containing a number of upgrades, downgrades, and level
segments as a single uniform segment. This is possible when no one grade is long
enough or steep enough to have a significant effect on the overall operation of the
general segment. As a guideline, extended general segment analysis can be used where
no one grade of 3 percent or greater is longer than 1⁄4 mi or where no one grade of less
than 3 percent is longer than 1⁄2 mi.
Extended segment used Specific Grades. Any grade less than 3 percent that is longer than 1⁄2 mi or any
when no one grade (3 percent grade of 3 percent or more that is longer than 1⁄4 mi must be analyzed as a separate
or greater) is longer than 1⁄4 segment because of its significant effect on traffic flow.
mi or when no one grade
(less than 3 percent) is longer Equivalents for Extended General Freeway Segments
than 1⁄2 mi.
Whenever extended general segment analysis is used, the terrain of the freeway
must be classified as level, rolling, or mountainous.
Level Terrain. Any combination of grades and horizontal or vertical alignment
that permits heavy vehicles to maintain the same speed as passenger cars is classified
as level terrain. This classification generally includes short grades of no more than
2 percent.
Rolling Terrain. Any combination of grades and horizontal or vertical
alignment that causes heavy vehicles to reduce their speeds substantially below those
of passenger cars but that does not cause heavy vehicles to operate at crawl speeds for
any significant length of time or at frequent intervals is classified as rolling terrain.
Mountainous Terrain. Any combination of grades and horizontal or vertical
alignment that causes heavy vehicles to operate at crawl speeds for significant
distances or at frequent intervals is classified as mountainous terrain.
‘‘Crawl speed’’ is the maximum sustained speed that trucks can maintain on an
extended upgrade of a given percent. If any grade is long enough, trucks will be forced
to decelerate to crawl speed, which they will then be able to maintain for extended
distances. Appendix I contains truck performance curves illustrating crawl speed and
length of grade.
Appendix I shows truck The exact categorization of terrain depends on the characteristics of the terrain
performance curves. itself and the prevailing mix of heavy vehicles. Grades that cause large trucks to
operate at crawl speed, for example, may not have the same effect on RVs.
Table 3-2 shows passenger car equivalents for extended general freeway segments.
It should be noted that it is extremely difficult to use extended general segment
analysis for ‘‘mountainous terrain’’ as defined here without violating the guidelines for
using the general terrain methodology (i.e., having no grade greater than 3 percent
longer than 1⁄4 mi). To a lesser extent, the same statement may be made with respect to
rolling terrain. The equivalence values shown in Table 3-2 are most useful in the
planning stage of analysis, when specific alignments are not known but approximate
capacity computations are still needed.
TABLE 3-2. PASSENGER-CAR EQUIVALENTS ON EXTENDED GENERAL FREEWAY SEGMENTS

Type of Terrain
Category Level Rolling Mountainous
ET for trucks and buses 1.5 3.0 6.0
ER for recreational vehicles 1.2 2.0 4.0
Equivalents for Specific Upgrades

Any freeway grade of more than 1⁄2 mi for grades less than 3 percent or 1⁄4 mi for
grades of 3 percent or more should be considered as a separate segment. For such
segments, analysis procedures must consider the upgrade conditions and whether the
grade is a single, isolated grade of constant percentage or part of a series of grades
forming a composite segment. The performance of heavy vehicles on significant grades
varies considerably among the categories of vehicles and among the individual vehicles
of a particular category.
Weight-to-horsepower ratios. Several studies have indicated that freeway truck populations have an average
weight-to-horsepower ratio ranging between 125 and 150 lb/hp. The procedures in this
chapter adopt passenger-car equivalents calibrated for a mix of trucks and buses in this
range in the traffic stream. RVs vary considerably in both type and characteristics.
These vehicles range from cars with trailers of various types to self-contained mobile
campers. In addition to the variability of the vehicles, their drivers are not
professionals, and their degree of skill in handling such vehicles covers a broad range.
Typical weight-to-horsepower ratios of RVs range from 30 to 60 lb/hp.
Tables 3-3 and 3-4 give values of ET and ER for specific upgrade sections requiring
separate analysis. These factors vary with the percent of grade, length of the grade, and
percent of trucks and buses in the traffic stream. The maximum values of ET and ER
occur when there are only a few such vehicles in the traffic stream. The equivalents
decrease as the number of heavy vehicles increases, because these vehicles tend to
form platoons and have operating characteristics that are more uniform as a group than
those of passenger cars.
The length of grade is generally taken from a profile of the highway in question Establishing length of grade.
and typically includes the straight portion of the grade plus some portion of the vertical
curves at the beginning and end of the grade. It is recommended that one-fourth of
the length of the vertical curves at the beginning and end of the grade be included in the
length of the grade. Where two consecutive upgrades are present, one-half of the length
of the vertical curve between them is assigned to the length of each upgrade.
TABLE 3-3. PASSENGER-CAR EQUIVALENTS FOR TRUCKS AND BUSES ON SPECIFIC UPGRADES
Passenger-Car Equivalent, ET
Grade Length Percent Trucks and Buses

(%) (mi) 2 4 5 6 8 10 15 20 25
<2 All 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5
1
2 0– ⁄4 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5
1
⁄4–1⁄2 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5
1 3
⁄2– ⁄4 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5
3
⁄4–1 2.5 2.0 2.0 2.0 1.5 1.5 1.5 1.5 1.5
1–11⁄2 4.0 3.0 3.0 3.0 2.5 2.5 2.0 2.0 2.0
>11⁄2 4.5 3.5 3.0 3.0 2.5 2.5 2.0 2.0 2.0
3 0–1⁄4 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5
1
⁄4–1⁄2 3.0 2.5 2.5 2.0 2.0 2.0 2.0 1.5 1.5
1 3
⁄2– ⁄4 6.0 4.0 4.0 3.5 3.5 3.0 2.5 2.5 2.0
3
⁄4–1 7.5 5.5 5.0 4.5 4.0 4.0 3.5 3.0 3.0
1–11⁄2 8.0 6.0 5.5 5.0 4.5 4.0 4.0 3.5 3.0
>11⁄2 8.5 6.0 5.5 5.0 4.5 4.5 4.0 3.5 3.0
4 0–1⁄4 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5
1
⁄4–1⁄2 5.5 4.0 4.0 3.5 3.0 3.0 3.0 2.5 2.5
1 3
⁄2– ⁄4 9.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5
3
⁄4–1 10.5 8.0 7.0 6.5 6.0 5.5 5.0 4.5 4.0
>1 11.0 8.0 7.5 7.0 6.0 6.0 5.0 5.0 4.5
5 0–1⁄4 2.0 2.0 1.5 1.5 1.5 1.5 1.5 1.5 1.5
1
⁄4–1⁄3 6.0 4.5 4.0 4.0 3.5 3.0 3.0 2.5 2.0
1 1
⁄3– ⁄2 9.0 7.0 6.0 6.0 5.5 5.0 4.5 4.0 3.5
1 3
⁄2– ⁄4 12.5 9.0 8.5 8.0 7.0 7.0 6.0 6.0 5.0
3
⁄4–1 13.0 9.5 9.0 8.0 7.5 7.0 6.5 6.0 5.5
>1 13.0 9.5 9.0 8.0 7.5 7.0 6.5 6.0 5.5
6 0–1⁄4 4.5 3.5 3.0 3.0 3.0 2.5 2.5 2.0 2.0
1
⁄4–1⁄3 9.0 6.5 6.0 6.0 5.0 5.0 4.0 3.5 3.0
1 1
⁄3– ⁄2 12.5 9.5 8.5 8.0 7.0 6.5 6.0 6.0 5.5
1 3
⁄2– ⁄4 15.0 11.0 10.0 9.5 9.0 8.0 8.0 7.5 6.5
3
⁄4–1 15.0 11.0 10.0 9.5 9.0 8.5 8.0 7.5 6.5
>1 15.0 11.0 10.0 9.5 9.0 8.5 8.0 7.5 6.5
NOTE: If the length of grade falls on a boundary, apply the longer category; interpolation may be used to find equiva-
lents for intermediate percent grades.

Critical point for analysis. In the analysis of upgrades, the critical point is usually at the end of the grade,
where heavy vehicles presumably have the maximum effect on operations. However, if
a ramp junction is located mid-grade, the point of the merge or diverge would also be
a critical point for analysis. In the case of composite grades, the point at which heavy
vehicles are traveling slowest is the critical point for analysis. If a 5 percent upgrade is
followed by a 2 percent upgrade, it is reasonable to assume that the end of the 5
percent portion would be critical, since heavy vehicles would be expected to accelerate
on the 2 percent portion of the grade.
Equivalents for Specific Downgrades

There is little specific information on the effect of heavy vehicles on traffic flow
on downgrades. In general, if the downgrade is not severe enough to cause trucks to
shift into low gear, it may be treated as a level terrain segment, and passenger-car
equivalents are selected accordingly. When more severe downgrades occur, trucks must
often use low gear to avoid gaining too much speed and running out of control. In
such cases, their effect on traffic flow is greater than it would be on level terrain.
Table 3-5 gives values of ET. For RVs, downgrades may be treated as level terrain.
TABLE 3-4. PASSENGER-CAR EQUIVALENTS FOR RECREATIONAL VEHICLES ON SPECIFIC UPGRADES

Passenger-Car Equivalent, ER
Grade Length Percent RVs

(%) (mi) 2 4 5 6 8 10 15 20 25
<2 All 1.2 1.2 1.2 1.2 1.2 1.2 1.2 1.2 1.2
1
3 0– ⁄2 1.2 1.2 1.2 1.2 1.2 1.2 1.2 1.2 1.2
>1⁄2 2.0 1.5 1.5 1.5 1.5 1.5 1.2 1.2 1.2
4 0–1⁄4 1.2 1.2 1.2 1.2 1.2 1.2 1.2 1.2 1.2
1
⁄4–1⁄2 2.5 2.5 2.0 2.0 2.0 2.0 1.5 1.5 1.5
>1⁄2 3.0 2.5 2.5 2.0 2.0 2.0 2.0 1.5 1.5
5 0–1⁄4 2.5 2.0 2.0 2.0 1.5 1.5 1.5 1.5 1.5
1
⁄4–1⁄2 4.0 3.0 3.0 3.0 2.5 2.5 2.0 2.0 2.0
>1⁄2 4.5 3.5 3.0 3.0 3.0 2.5 2.5 2.0 2.0
6 0–1⁄4 4.0 3.0 2.5 2.5 2.5 2.0 2.0 2.0 1.5
1
⁄4–1⁄2 6.0 4.0 4.0 3.5 3.0 3.0 2.5 2.5 2.0
>1⁄2 6.0 4.5 4.0 4.0 3.5 3.0 3.0 2.5 2.0
NOTE: If the length of grade falls on a boundary, apply the longer category; interpolation may be used to find equiva-
lents for intermediate percent grades.
TABLE 3-5. PASSENGER-CAR EQUIVALENTS FOR TRUCKS AND BUSES ON

SPECIFIC DOWNGRADES
Passenger-Car Equivalent, ET
Grade Length Percent Trucks/Buses

(%) (mi) 5 10 15 20
<4 All 1.5 1.5 1.5 1.5
4 ≤4 1.5 1.5 1.5 1.5
4 >4 2.0 2.0 2.0 1.5
5 ≤4 1.5 1.5 1.5 1.5
5 >4 5.5 4.0 4.0 3.0
≥6 ≤4 1.5 1.5 1.5 1.5
>6 >4 7.5 6.0 5.5 4.5

Equivalents for Composite Grades
The vertical alignment of most freeways results in a continuous series of grades. It
is often necessary to determine the effect of a series of significant grades in
succession. Consider the following example: a 2 percent grade of 1⁄2 mi is followed
immediately by a 4 percent grade of 1⁄2 mi. The analysis problem of interest is the
maximum effect of heavy vehicles, which would occur at the end (top) of the
4 percent segment.
The most straightforward technique is to compute the average grade to the point in Average grade technique.
question. The average grade is defined as the total rise in feet from the beginning of
the composite grade divided by the length of the grade in feet.
Passenger-car equivalents for this composite grade would be found for a 3 percent Appendix I gives detailed
grade, 1 mi long. The average-grade technique is an acceptable approach for grades in example of composite grade.
which all subsections are less than 4 percent or the total length of the composite grade
is less than 4,000 ft. For more severe composite grades, a detailed technique that uses
vehicle performance curves and equivalent speeds to determine the effective simple
grade for analysis is presented in Appendix I.
For the example cited above, the total rise is (2,640 × 0.02) + (2,640 × 0.04) =
165.4 ft. The average grade is 165.4/5,280 = 0.03 or 3 percent.
Computation of Heavy-Vehicle Factor

Once the values of ET and ER have been found, the determination of the Calculate heavy-vehicle factor
adjustment factor, fHV, is straightforward: to reflect proportion of trucks,
buses, and RVs, and grade
1 or terrain.
fHV = (3-2)
1 + PT (ET − 1) + PR (ER − 1)
where
ET, ER = passenger-car equivalents for trucks or buses and RVs in the traffic stream,
PT, PR = proportion of trucks or buses and RVs in the traffic stream, and
fHV = heavy-vehicle adjustment factor.
In many cases, trucks will be the only heavy vehicle type present in the traffic
stream to a significant degree. Where the percentage of RVs is small compared with
the percentage of trucks, it is sometimes convenient to consider all heavy vehicles to
be trucks. Thus, a traffic stream consisting of 10 percent trucks and 2 percent RVs
might be analyzed as having 12 percent trucks. It is generally acceptable to do this
where the percentage of trucks and buses in the traffic stream is at least 5 times the
percentage of RVs present.
Driver Population Adjustment

The traffic stream characteristics that are the basis for the procedure described here
are representative of regular drivers who are substantially commuters or are familiar with
the facility. It is generally accepted that traffic streams with different characteristics (i.e.,
recreational drivers) use freeways less efficiently. Whereas data are sparse and reported
results vary substantially, significantly lower capacities have been reported on weekends,
particularly in recreational areas. It may generally be assumed that the reduction in
capacity extends to service volumes for other levels of service as well.
The adjustment factor fp is used to reflect this effect. The values for fp range from
1.0 to 0.85. In general, the analyst should select 1.0, which reflects commuter traffic
(i.e., familiar users), unless there is sufficient evidence or it is the analyst’s judgment
that a lesser value reflecting more recreational traffic characteristics should be applied.
When greater accuracy is needed, comparative field studies of commuter and
recreational traffic flow and speeds are recommended.
Determination of Free-Flow Speed

Free-flow speed is the mean speed of passenger cars measured under low to Measure or estimate free-flow
moderate flows (up to 1,300 pcphpl). For a specific section of freeway, speeds are speed.
virtually constant in this range of flow rates. Two methods can be used to determine
the free-flow speed of the basic freeway section being studied: field measurement and
estimation with guidelines provided in this chapter. The field measurement procedure
is provided for users who prefer to gather these data directly. However, field
measurement is not necessary for application of this procedure.
Field Measurement
The free-flow speed of a basic freeway section can be determined directly from a
speed study conducted in the field. If field-measured data are used, no subsequent
adjustments are made to the free-flow speed. The speed study should be conducted at a
representative location within the freeway section being evaluated; for example, a
segment on an upgrade or downgrade should not be selected within a section that is
generally considered level. Any speed measurement technique that has been found
acceptable for other types of traffic engineering speed studies may be used.
The speed study should be conducted when flows are low (up to 1,300 pcphpl).
Weekday off-peak hours are generally good times to observe low to moderate flow
rates. The speed study should measure the speeds of all passenger cars or use a
systematic sample (e.g., every 10th passenger car). The speed study should also
measure at least 100 passenger-car speeds across all lanes. Further guidance on the
conduct of speed studies is found in standard traffic engineering publications, such as
the Manual of Traffic Engineering Studies published by the Institute of Transportation
Engineers.
The average of all passenger-car speeds measured in the field under low-volume
conditions can be used directly as the free-flow speed of the freeway section. This
speed reflects the net effects of all conditions at the study site that influence speed,
including those considered in this procedure (lane width, lateral clearance, number of
lanes, and interchange density) as well as others such as speed limit and vertical and
horizontal alignment.
Highway agencies with ongoing speed-monitoring programs or with existing speed
data on file may prefer to use those data rather than conduct a new speed study or use
an indirect method to estimate speed. Such data can be used directly if collected in
accordance with the procedures presented above. Data that include both passenger cars
and heavy vehicles can probably be used for level terrain or moderate downgrades but
should not be used for rolling or mountainous terrain.
Estimation Guidelines
Estimate free-flow speed if If field measurement of free-flow speed is not possible, the free-flow speed can be
field measurement is not estimated indirectly on the basis of the physical characteristics of the freeway section
possible. being studied. These physical characteristics include lane width, right-shoulder lateral
clearance, number of lanes, and interchange density. Equation 3-3 is used to estimate
the free-flow speed of a basic freeway section:
FFS = FFSi − fLW − fLC − fN − fID (3-3)
where
FFS = estimated free-flow speed (mph);

FFSi = estimated ideal free-flow speed, 70 or 75 mph;
fLW = adjustment for lane width from Table 3-6 (mph);
fLC = adjustment for right-shoulder lateral clearance from Table 3-7 (mph);
fN = adjustment for number of lanes from Table 3-8 (mph); and
fID = adjustment for interchange density from Table 3-9 (mph).
Ideal Free-Flow Speed. Estimation of a free-flow speed for an existing or

future freeway section being studied is accomplished by adjusting ideal free-flow speed
downward to reflect the influence of four factors: lane width, lateral clearance, number
of lanes, and interchange density. Thus, the analyst is required to select an appropriate
ideal free-flow speed (FFSi) as a starting point.
FFSi represents the mean speed for passenger cars under low to moderate flow
rates and the ideal conditions defined previously. The research on which these
procedures are based found that FFSi ranged between 70 and 75 mph. In using this
procedure to estimate FFS, it is recommended that an FFSi of 70 or 75 mph be
selected. As a rule of thumb, 75 mph can generally be considered to represent FFSi on
a rural freeway, whereas on urban and suburban freeways, FFSi is best represented by
70 mph. The analyst should use careful judgment if it is believed that a higher or
lower FFSi is more representative of freeway traffic flow in a given area. Again,
measurement of actual free-flow speeds on the freeway section being studied or an
area freeway with similar features is highly encouraged.
Lane Width. The ideal condition for lane width is 12 ft or greater. When the
average lane width across all lanes within a freeway section is less than 12 ft, the ideal
free-flow speed (e.g., 70 mph) is reduced. Adjustment factors to reflect the effect of
narrower average lane width are provided in Table 3-6. These factors were developed
on the basis of studies conducted on multilane highways. Adjustment factors are
provided only for average lane widths of 11 and 10 ft. It should be noted that freeway
sections with average lane widths below 11 ft are generally considered rare.
Lateral Clearance. Ideal lateral clearance is 6 ft or greater on the right side and Adjustment for lateral
2 ft or greater on the median or left side. When the right-shoulder lateral clearance is clearance reflects right
less than 6 ft, the ideal free-flow speed is reduced. Adjustment factors to reflect the shoulder width only.
effect of narrower right-shoulder lateral clearance are provided in Table 3-7. No
adjustment factors are available to reflect the effect of median lateral clearance less
than 2 ft; however, lateral clearance on either the right or left sides less than 2 ft is
considered rare. Considerable judgment must be used in determining whether objects
or barriers along the right side of a freeway section present a true obstruction. Such
obstructions may be continuous, such as a retaining walls, concrete barriers, or
guardrails, or may be periodically occurring objects, such as light supports or bridge
abutments. In some cases, drivers may become accustomed to certain types of
obstructions, in which case their effect on traffic flow may become negligible.
Number of Lanes. Freeway sections with five or more lanes (in one direction) Adjustment for number of
are considered ideal with respect to number of lanes. When fewer lanes are present, the lanes is not applicable to
ideal free-flow speed is reduced. Table 3-8 provides adjustment factors to reflect the rural freeway segments.
effect of number of lanes on ideal free-flow speed. When number of lanes is
determined, only mainline lanes, both basic and auxiliary, should be considered. High-
occupancy-vehicle (HOV) lanes should not be included.
The adjustment factors in Table 3-8 are based exclusively on data collected on
urban and suburban freeway sections and do not reflect conditions on rural freeways,
TABLE 3-6. ADJUSTMENT FACTORS FOR LANE WIDTH

Lane Reduction in Free-Flow
Width (ft) Speed fLW (mph)
≥12 0.0
11 2.0
10 6.5
TABLE 3-7. ADJUSTMENT FACTORS FOR RIGHT-SHOULDER LATERAL CLEARANCE

Reduction in
Free-Flow Speed, fLC (mph)
Right Shoulder Lateral Lanes in One Direction

Clearance (ft) 2 3 4
≥6 0.0 0.0 0.0
5 0.6 0.4 0.2
4 1.2 0.8 0.4
3 1.8 1.2 0.6
2 2.4 1.6 0.8
1 3.0 2.0 1.0
0 3.6 2.4 1.2

TABLE 3-8. ADJUSTMENT FACTORS FOR NUMBER OF LANES
Number of Lanes Reduction in Free-Flow
(One Direction) Speed fN (mph)
≥5 0.0a
4 1.5
3 3.0
2 4.5
a
For rural freeway sections, fN = 0.0.
TABLE 3-9. ADJUSTMENT FACTORS FOR INTERCHANGE DENSITY

Interchanges Reduction in Free-Flow
per Mile Speed fID (mph)
≤0.50 0.0
0.75 1.3
1.00 2.5
1.25 3.7
1.50 5.0
1.75 6.3
2.00 7.5
which typically carry two lanes in each direction. Therefore, in using Equation 3-3 to
estimate the free-flow speed of a rural freeway section, the value of the adjustment for
number of lanes, fN, should be 0.
A 6-mi section is used to Interchange Density. The ideal interchange density is 0.5 interchange per mile,
determine interchange or 2-mi interchange spacing. Ideal free-flow speed is reduced when interchange density
density. is greater. Adjustment factors to reflect the effect of interchange density are provided
in Table 3-9. Interchange density is determined over a 6-mi section of freeway (3 mi
upstream and 3 mi downstream) in which the freeway section being studied is located.
An interchange is defined as having at least one on-ramp. Therefore, interchanges that
have only off-ramps would not be considered in determining interchange density.
Interchanges considered should include typical interchanges with arterials or highways
and major freeway-to-freeway interchanges.
Determination of Level of Service

The level of service on a basic freeway section can be determined directly from
Figure 3-4 on the basis of the free-flow speed and the flow rate. The procedure is as
follows:
Step 1. Define and segment the freeway section as appropriate.
Step 2. On the basis of the measured or estimated free-flow speed on the freeway
segment, construct an appropriate speed-flow curve of the same shape as
the typical curves shown in Figure 3-2. The curve should intercept the
y-axis at the free-flow speed.
Step 3. On the basis of the flow rate, vp, read up to the free-flow speed curve
identified in Step 2 and determine the average passenger-car speed and
level of service corresponding to that point.
Step 4. Determine the density of flow as
D = vp /S (3-4)
where
D = density (pc/mi/ln),
vp = flow rate (pcphpl), and
S = average passenger-car speed (mph).
The level of service can also be determined using the density ranges provided in
Table 3-1.
III. APPLICATIONS
The methodology presented in this chapter may be used for operational analysis,
design, or planning of basic freeway sections. The analyst should refer to Chapter 4,
Weaving Areas, and Chapter 5, Ramps and Ramp Junctions, for additional information
regarding definition of the freeway section being analyzed to confirm that the
procedures provided in this chapter are indeed appropriate.
Operational analysis involves the consideration of a known present or projected
freeway for which all geometric variables and traffic variables can be specified. The
analysis results in the determination of a level of service, as well as a probable
operating speed and density.
In design, a forecast demand volume is used to determine the number of lanes
needed to provide the desired level of service.
The objectives of a planning analysis are to determine the number of lanes needed
to deliver a desired level of service. The difference between planning and design
analysis is the level of detail available as input. In planning, annual average daily
traffic (AADT) is generally known with very little detail concerning composition
and/or hourly distribution. A planning analysis consists of making an estimate of
directional design-hour volume (DDHV) and then applying typical (i.e., default) values
for inputs to a design-type analysis.
These procedures do not present a separate computational application for planning,
since computations for both operational analysis and design are relatively simple and
straightforward.
Applications can be set into four general analysis types, each having a target Analysis types.
output with the remaining parameters being known and being used as inputs.
Type I analysis resolves the primary question, ‘‘Within what level of service does
the freeway operate?’’ Thus, hourly flow rate (vp) and free-flow speed (FFS) are
needed as inputs. This analysis type is used for operational studies.
Type II analysis produces an estimate of speed (S) as the output. For the input, vp,
level of service, and FFS must be used. Typically, this type of analysis is applied
when travel time is the parameter of interest, for example, in feasibility or economic
studies. Environmental studies (air and noise) would also depend on estimates of mean
speed.
For Type III analysis, an estimate of vp (pcphpl) is computed. As known inputs,
both level of service and FFS are required. Typical applications include comparing the
estimate of vp with the year-by-year forecast volumes for the freeway section. The
timing of future improvements to maintain a specified level of service can be identified
in this application.
Type IV analysis results in an estimate of the number of lanes, N, required for a
given set of conditions. As known inputs, vp, FFS, and level of service are required.
This application has traditionally been called ‘‘design.’’ Also, when the analyst is using
AADT data as a starting point, this same analysis can be termed ‘‘planning analysis.’’
SEGMENTING THE FREEWAY

Any capacity or LOS analysis requires that the freeway section to be analyzed
have uniform traffic and roadway conditions. A number of locations on any freeway
form natural boundaries for identifying uniform segments. Any on-ramp or off-ramp is
such a boundary, since the volume of freeway traffic changes at each. The beginning
and end of specific simple or composite grades also act as boundaries between uniform
freeway segments. Any point at which the traffic or roadway conditions change should
be used as a boundary between uniform segments, each of which should be analyzed
separately.
In addition to the natural boundaries created by on-ramps and off-ramps, changes
in the following characteristics generally dictate that the freeway section under analysis
be segmented:
T Number of lanes,
T Right-shoulder lateral clearance that would result in a lower estimated free-
flow speed,
T Grade of 2 percent or more or a constant upgrade over 4,000 ft long, and
T Speed limit.
COMPUTATIONAL STEPS
The worksheet for computations is shown in Figure 3-5. For all analysis types, the
analyst enters data in the General Information and Site Information portions of the
worksheet.
Type I Analysis. Find LOS. For Type I analysis, all information in the Traffic and Roadway Conditions section
Enter speed-flow graph with of the worksheet is entered except speed, S (which typically will be a secondary
vp to find LOS. output). Flow is then computed with the aid of the tables of passenger-car equivalents.
Free-flow speed (FFS) is estimated by applying adjustments for four factors to an ideal
FFS determined by the analyst. Finally, level of service is derived (using vp) from the
speed-flow graph at the top of the worksheet by intersecting the specific curve that has
been selected or constructed for the freeway section being analyzed.
This point of intersection identifies the level of service and also (on the vertical
axis of the graph) the estimated speed, S. If the analyst requires a value for density
(D), it is calculated as vp /S.
Type II Analysis. Find speed For Type II analysis, all information in the Traffic and Roadway Conditions
(S). Enter speed-flow graph section is filled in by the analyst with the exception of speed, S. The flow rate
with vp to find S. calculations are performed by using the heavy-vehicle equivalents and the free-flow
speed is estimated by using the adjustment factors in Tables 3-6 through 3-9.
Alternatively, free-flow speed may have been directly measured in the field; then the
direct measurement can be entered on the worksheet. With the free-flow speed
established, the specific speed-flow curve for the section being analyzed is constructed
or selected. The point of intersection between the flow rate, vp, and the appropriate
speed-flow curve gives the speed, S. As secondary outputs, level of service can be read
directly from the graph and density can be computed using flow rate and speed.
Type III Analysis. Find vp. The objective of Type III analysis is to estimate the flow rate in passenger cars
Enter graph at intersection per hour per lane given a set of traffic, roadway, and free-flow speed conditions.
points of LOS lines and Typically a desired level of service is stated and entered on the worksheet. Then the
facility-specific speed-flow free-flow speed of the section is established by using either the ideal speed (70 or
curve. Read vp on horizontal 75 mph) and the four adjustment factors or a free-flow speed directly measured in the
axis.
field. Once this facility-specific speed-flow curve has been established, the analyst can
determine what flow rate is achievable with the given level of service. This flow rate
would be the maximum flow rate achievable or allowable for the given level of
service. Also directly available from the graph is the estimated average passenger-car
speed. Finally, if a value for density is required, it can be directly calculated by using
the estimated flow rate and the average speed.
Type IV Analysis. Find Type IV analysis is used to establish the required number of lanes in a design
number of lanes (N). Enter application. A planning estimate of the required number of lanes can also be made.
graph at intersecting points of The key to Type IV analysis is the establishment of an hourly volume on which the
specific speed-flow curve and design is to be based. All information except number of lanes can be entered in the
LOS threshold. Read vp on Traffic and Roadway Conditions portion of the worksheet. A free-flow speed, either
horizontal axis. Use equation
computed or measured directly, is entered on the worksheet, and the appropriate curve
for vp and solve for N.
representative of the free-flow speed is established on the graph. The required or
desired level of service is also entered. Then the analyst takes the intersection of the
LOS threshold and the appropriate speed-flow curve and projects downward to the
horizontal axis of the graph to establish the maximum flow rate per lane. Using this
maximum allowable flow rate in the equation for flow rate on the worksheet, the
analyst can solve for number of lanes, N. Since, in this case, N will be a fractional or
decimal number, the analyst has the ability to use judgment on whether to round
upward or round slightly downward to estimate the number of lanes required in one
direction. Note that again density is easily calculated using vp and S.
The number of lanes on a specific freeway section depends not only on the desired
level of operation, but also on the continuity of lanes with adjacent sections and along
the freeway system. Frequent adding or dropping of lanes along a freeway is not
practical, yet can be considered within critical freeway sections. On specific grades,
the need for a larger number of lanes on an upgrade than on a downgrade carrying the
same traffic volume is a clear indication that a climbing lane is required.
PLANNING ANALYSIS
Planning-level analysis can be done using the analysis worksheet. Planning
information on future AADT is required and can be converted to an estimated DDHV
using the known or forecast values of proportion of AADT occurring during the peak
hour, K, and directional distribution factor, D:
DDHV = AADT × K × D (3-5)
Most highway traffic agencies have data on K and D, which tend to be regional
for a particular class of highway. Normal values for K range from about 0.08 in dense
urban areas to as high as 0.15 to 0.20 in rural areas. D varies from about 52 percent to
as high as 80 percent in some rural situations. Traffic during a peak hour is rarely
distributed evenly in both directions, even on urban circumferential routes.
Planning applications typically are used to estimate either the level of service or
the number of lanes required to carry a specified amount of traffic. The analyst
typically has few, if any, of the input values required for Types I, II, III, and IV
analyses. The following default values are suggested for planning analysis:
Volume, V (from DDHV)
ET = 1.5
ER = 1.2
% Trucks and buses = 5 percent
% RVs = 2 percent
fp = 1.0
fHV = 0.939
fLW = 0.0
fLC = 0.0
fN = 3.0 mph
fID = 0.0
FFSi = 70 mph
FFS = 67 mph
HOV FACILITIES
These procedures apply only to the analysis of traffic conditions on the freeway
mainline and are not generally intended for use in analyzing capacity and level of
service of HOV lanes. However, when the HOV facility has two or more lanes in each
direction during all or part of the day and access to the HOV facility from adjacent
general use freeway lanes is limited (e.g., 1.0-mi or greater access point spacing), these
procedures may be used to analyze the sections of the HOV facility between access
points. These procedures should not be used to analyze the operation of HOV facilities
composed of one lane in each direction. Further guidelines for the analysis of HOV
facilities are provided in Chapter 6, Freeway Systems.
TOOLS FOR ANALYSIS
The worksheet shown in Figure 3-5 and provided in Appendix II can be used to
perform analysis Types I, II, III, and IV. Tables that provide adjustment factors for
free-flow speed estimation and those providing passenger-car equivalents for heavy
vehicles are used to provide entries to the worksheet.

IV. EXAMPLE PROBLEMS
EXAMPLE PROBLEM 1
The Freeway Existing four-lane freeway, urban area, very restricted geometry,
rolling terrain, 65-mph speed limit.
The Question What is the level of service during the peak hour?
The Facts
√ Four lanes (two in each direction)
√ 11-ft lanes
√ 2-ft lateral clearance on left and right sides
√ 2,000-vph peak-hour volume (one direction)
√ Rolling terrain
√ 5 percent trucks
√ PHF = 0.92
√ Interchange density = 1.0/mi
√ Mostly commuter traffic
Comments
√ Assume no buses and no RVs since none indicated.
√ Assume ideal FFS of 70 mph considering freeway type and geometry.
Outline of Solution All input parameters are known; thus no default values are
required. Demand will be computed in terms of pcphpl, a free-flow speed will be
estimated, and the level of service determined from the speed-flow graph. If desired, an
estimate of passenger-car speed is directly available from the graph, and a value for
density can be calculated using speed and flow rate.
Steps
1. Convert volume (vph) to flow rate V
vp =
(pcphpl) (PHF)(N)(fHV)(fp)
fp = 1.0 (commuter traffic)
2,000
vp =
0.92 × 2 × fHV × 1.0
2. Find fHV (no buses or RVs) 1
fHV =
(use Table 3-2 for ET) 1 + PT (ET − 1)
1
fHV =
1 + 0.05(3 − 1)
fHV = 0.909
3. Find vp 2,000
vp =
(0.92)(2)(0.909)
vp = 1,196 pcphpl
4. Compute free-flow speed (using FFS = FFSi−fLW−fLC−fN−fID

Tables 3-6, 3-7, 3-8, and 3-9) FFS = 70 − 2.0 − 2.4 − 4.5 − 2.5
FFS = 58.6 mph (or round to 59 mph)
5. Determine level of service (using

Table 3-1)
The Results Level of service = C
Summary Other outputs are speed (S = 59 mph) and density (20.3 pc/mi/ln)
calculated as vp /S, or 1,196/59.

Worksheet for Example Problem 1
80
120
1300
70
1450
1600 100

60
1750
50 80
Analysis
A B C D E Type Input Output
(km/hr)
(mph)
40
60 I vP, FFS LOS
30
II vP, LOS, FFS S
40
20
III FFS, LOS vp
20
10
IV vP, LOS N
0 0
0 200 400 600 800 1000 1200 1400 1600 1800 2000 2200 2400
Flow Rate (pcphpl)
General Information
S. Stevens Date Performed 3/18/97
Analyst
Binghamton Co. Analysis Type x
Agency or Company
I II III IV
Site Information Flow Rate (vP )

Highway/Dir. Travel SR 210
___________ ET 3.0
____________ Tables 3-2, 3-3, 3-5
From/To Adams St./Lincoln Ave.
___________ ER ____________ Tables 3-2, 3-4
Jurisdiction Binghamton Co.
___________ fHV ____________
0.909 1
1 + PT(ET - 1) + PR(ER - 1)
Analysis Time Period PM Peak Hour
___________
Analysis Year 1997
___________ fp 1.0
____________ (1.0 - 0.85)
vP V
Traffic and Roadway Conditions 1196
____________ pcphpl (PHF x N x f x fp)
HV
Volume, V 2,000
___________ vph
Free-Flow Speed (FFS)
Speed, S ___________ mph 70
FFSi _________mph
Lane Width, LW 11
___________ ft 2.0
fLW _________mph Table 3-6
Number of Lanes, N 2
___________ 2.4
fLC _________mph Table 3-7
2
Rt-Shoulder Lat. Clear., LC ___________ ft fN 4.5
_________mph Table 3-81
Peak-Hour Factor, PHF 0.92
___________ 2.5
fID _________mph Table 3-9
Interchange Density, ID 1.0
___________ 59
FFS _________mph (est.) FFSi - fLW - fN - fLC - fID
% Trucks and Buses, PT 5
___________ or
% RVs, PR 0
___________ FFS _________mph (measured)
General Terrain
Level x Rolling Mountainous Level of Service (LOS)
Specific Grade 20.3
Density, D ____________ pc/mi/ln vP/S
Length ___________ mi LOS C
__________ Table 3-1
1For rural freeway sections, f = 0
N
Up/Down ___________ %
Driver Type
x Commuter/Wk Day Recreational/Wk End

EXAMPLE PROBLEM 2
The Freeway New suburban freeway being designed, level terrain.
The Question How many lanes are needed to provide LOS D during peak
periods?
The Facts
√ 4,000-vph peak-hour volume (one direction)
√ Level terrain
√ 3 percent RVs
√ PHF = 0.85
√ Interchange density = 1.50/mi
√ Lane width and lateral clearance to be designed to high standards
Comments
√ Assume commuter traffic, since the freeway is suburban; thus, fp = 1.0.
√ Assume ideal FFS of 70 mph considering the freeway type (i.e., suburban).
Outline of Solution No default values are required. The flow rate per lane, vp, to
be accommodated at LOS D for four-, six-, and eight-lane freeways will be determined
and compared with the demand flow rate.
Steps
1. Convert volume (vph) to flow rate V
vp =
(pcphpl) (PHF)(N)(fHV)(fp)
2. Find fHV 1
fHV =
(use Table 3-2 for ET and ER) 1 + PT (ET − 1) + PR(ER − 1)
1
fHV =
1 + (0.15)(1.5 − 1) + 0.03(1.2 − 1)
fHV = 0.925
3. For four-lane freeway 4,000

vp =
(0.85)(2)(0.925)(1.0)
vp = 2,544 pcphpl
4. For six-lane freeway 4,000

vp =
(0.85)(3)(0.925)(1.0)
vp = 1,696 pcphpl
5. For eight-lane freeway 4,000

vp =
(0.85)(4)(0.925)
vp = 1,272 pcphpl
6. Eliminate the four-lane freeway

since demand per lane exceeds
capacity (2,400 pcphpl for an ideal
section)
7. Compute free-flow speed for six-lane FFS = FFSi − fLW − fLC − fN − fID
and eight-lane freeways (using FFS = 70−0.0−0.0−3.0−5.0 (three lanes)
Tables 3-6, 3-7, 3-8, and 3-9) FFS = 62 mph (three lanes)
FFS = 70−0.0−0.0−1.5−5.0 (four lanes)
fHV = 63.5 mph (four lanes)
Table continues

8. Use Table 3-1 to determine that six-

lane freeway operates at LOS D,
eight-lane freeway at LOS C
The Results A six-lane freeway will meet operational objective of LOS D during
peak periods.
Summary Other outputs are speed = 61 mph and density = 27.8 pc/mi/ln (from
1,696/61). Also, an eight-lane design would operate at LOS C; speed = 63.5 mph,
density = 20.0 pc/mi/ln (from 1,272/63.5).
80
120
1300
70
1450
1600 100 Average Passenger-Car Speed
Analysis
Type Input Output
60
1750
50 80
A B C D E I vP , FFS LOS
(km/hr)
(mph)
40
60 II vP , LOS, FFS S
30
40
III FFS, LOS vP
20
N=3
N=4
IV vP, LOS N
20
10
0 0
0 200 400 600 800 1000 1200 1400 1600 1800 2000 2200 2400
Flow Rate (pcphpl)
General Information
Analyst MJM Date Performed 2/7/96
Agency or Company TRANSCON Analysis Type x
I II III IV
Site Information Flow Rate (v P)

Highway/Dir. Travel WBWB SR 101
____________ ET 1.5
__________________ Tables 3-2, 3-3, 3-5
From/To I-10/I-17
____________ ER 1.2
__________________ Tables 3-2, 3-4
Jurisdiction ____________ fHV 0.925
__________________ 1
1 + PT(ET - 1) + PR(ER - 1)
Analysis Time Period Peak Period
____________ fp 1.0
__________________ (1.0 - 0.85)
N= 2 3 4
Analysis Year 2015
____________ vP 2544 1696 1272
__________________ pcphpl V
Volume, V 4,000
____________ vph Free-Flow Speed (FFS)
Speed, S ____________ mph 70
FFSi ___________ mph
Lane Width, LW 12
____________ ft 0
fLW ___________ mph Table 3-6
Number of Lanes, N 2, 3, 4
____________ 0
fLC ___________ mph Table 3-7
10 N= 3 4
Rt-Shoulder Lat. Clear., LC ____________ ft fN 3.0 1.5 mph Table 3-81
___________
____________ fID 5.0
___________ mph Table 3-9
____________ N= 3 4
FFS 62 63.5 mph (est.) FFSi - fLW - fN - fLC - fID
% Trucks and Buses, PT 1.5
_______________ or
% RVs, PR 3
_______________ FFS ___________mph (measured)
General Terrain
x Level Rolling Mountainous
N= 3 4
Specific Grade Density, D _______________
27.8 20.0 pc/mi/ln vP/S
N= 3 4
Length ____________ mi LOS _______________
D C Table 3-1
Up/Down ____________ %
1
N
Driver Type

EXAMPLE PROBLEM 3
The Freeway Existing six-lane freeway, growing urban area, level terrain.
The Question What is the current LOS during peak periods? What LOS will occur
in 3 years? To avoid the condition of demand exceeding capacity, when should a fourth
lane be added in each direction?
The Facts
√ Six lanes
√ 5,000-vph volume (in one direction) (existing)
√ Level terrain
√ PHF = 0.95
√ 5,600-vph volume (in one direction) (in 3 years)
√ Beyond 3 years, traffic grows at 4 percent per year
√ FFS = 65 mph (measured in field)
Comments
√ Since no information is given on possible changes over time, assume that percent
trucks and PHF remain constant.
√ This problem deals with a variety of demand levels and can most easily be solved
by computing the maximum volume that can be accommodated for each level of
service.
√ Free-flow speed need not be estimated since it has been field measured.
√ Assume no buses and no RVs.
√ Assume familiar driver population given the freeway type and area type.
Outline of Solution The maximum volume (vph) that can occur for each LOS will
be computed. The demand volumes will then be compared and an LOS estimated.
Steps
1. Convert the maximum service flow V
vp =
(pcphpl) for each LOS to vph (PHF)(N)(fHV)(fp)
or
vp(PHF) (N) (fHV) (fp) = V
2. Find fHV 1
fHV =
1 + PT (ET − 1)
1
fHV =
1 + 0.10(1.5 − 1)
fHV = 0.952
3. Find vp (maximum for each LOS) from LOS A vp = 650 pcphpl

Table 3-1 LOS B vp = 1,040
(fp = 1.0) LOS C vp = 1,548
LOS D vp = 1,984
LOS E vp = 2,350
4. Compute V (vph) LOS A V = 1,763 vph

LOS B V = 2,822
LOS C V = 4,200
LOS D V = 5,383
LOS E V = 6,376
5. Compare 5,000 vph and 5,600 vph

with above; determine LOS
Table continues

6. When traffic exceeds 6,376 vph, a 5,600 (1.04n) = 6,376

fourth lane in each direction will be n = 3.3 years
needed. A compounding equation is
used
The Results LOS D (existing)

LOS E (in 3 years)
6.3 years (fourth lane needed)
80
120
1300
70
1450
1600 100 Analysis

Type Input Output
60
1750
50 80
(km/hr)
(mph)
40
30
40
III FFS, LOS vP
20
IV vP, LOS N
20
10
0 0
0 200 400 600 800 1000 1200 1400 1600 1800 2000 2200 2400
Flow Rate (pcphpl)
General Information
Susan Collins 5/5/97
Arizona DOT x
I II III IV

Highway/Dir. Travel I-17 NB
____________ ET 1.5
__________________ Tables 3-2, 3-3, 3-5
From/To Dunlap Ave/Northern Ave
____________ ER __________________ Tables 3-2, 3-4
Jurisdiction Arizona DOT
____________ fHV 0.952
__________________ 1
1 + PT(ET - 1) + PR(ER - 1)
Analysis Time Period PM Peak Hour
____________ fp 1.0
__________________ (1.0 - 0.85)
Analysis Year 1997/2000
____________ vP LOS A 650 LOS D 1984 pcphpl V
LOS B 1040 LOS E 2350 (PHF x N x fHV x fp)
Traffic and Roadway Conditions LOS C 1548
1997-5000 Free-Flow Speed (FFS)
Volume, V ____________
2000-5600 vph
Number of Lanes, N ____________ fLC ___________mph Table 3-7
Rt-Shoulder Lat. Clear., LC 3
____________ ft fN ___________mph Table 3-81
Peak-Hour Factor, PHF ____________ fID ___________mph Table 3-9
____________ FFS ___________mph (est.) FFSi - fLW - fN - fLC - fID
_______________ or
% RVs, PR _______________ FFS 65
___________mph (measured)
General Terrain
x Level Rolling Mountainous Level of Service (LOS)
Specific Grade Density, D _______________ pc/mi/ln vP/S
1997 - D
Length ____________ mi LOS _______________
2000 - E Table 3-1
Up/Down ____________ % 1
N
Driver Type

EXAMPLE PROBLEM 4
The Freeway Existing four-lane freeway, rural, ideal lane widths and clearances.
The Question What is the LOS for both upgrade and downgrade during peak
periods?
The Facts
√ Four lanes
√ 2,300 vph (in one direction)
√ Composite grade, 3,000 ft at 3 percent and 2,500 ft at 5 percent
√ PHF = 0.90
√ FFS = 75 mph (measured in field)
Comments
√ Assume no buses and no RVs.
√ Free-flow speed need not be estimated since it has been field measured.
√ The precise procedure for composite grades is used because there is a section
steeper than 4 percent and the total length is greater than 4,000 ft.
√ Since most drivers on a rural freeway section are generally unfamiliar with the
area, a driver population factor of 0.95 is assumed.
Outline of Solution The truck performance curves in Appendix I are used to

develop an equivalent grade (i.e., a constant grade that has the same effect on heavy
vehicles as does the composite grade). The equivalency tables for both upgrades and
downgrades are used to define fHV. The flow rates, vp, are computed, and Table 3-1 is
used to estimate LOS.
Steps
1. Determine equivalent constant grade See detailed description in Appendix I.
Answer = 4.8 percent (1.04 mi)
2. Find vp V
vp =
(PHF)(N)(fHV)(fp)
3. Find fHV (upgrade) (Table 3-3, use in- 1
fHV =
terpolation) 1 + PT (ET − 1)
1
fHV = = 0.562
1 + 0.15(6.2 − 1)
4. Find fHV (downgrade) 1
fHV = = 0.930
(Table 3-5) 1 + 0.15(1.5 − 1)
5. Compute vp (upgrade) 2,300

vp =
(0.90)(2)(0.562)(0.95)
vp = 2,392 pcphpl
6. Compute vp (downgrade) 2,300

vp =
(0.90)(2)(0.930)(0.95)
vp = 1,447 pcphpl
7. Enter Table 3-1 or worksheet with vp,

determine LOS
The Results Using the speed-flow curve of 75 mph, upgrade operates at LOS E
(speed 53.5 mph, density 43.5 pc/mi/ln). Downgrade operates at LOS C (speed 69.5
mph, density 20.8 pc/mi/ln).

Using Appendix I, enter Figure I.3-2 at 3,000 ft. Speed at top of 3 percent grade is
42.5 mph. Intersection of horizontal at 42.5 mph and 5 percent curve implies trucks have
been on 5 percent grade for 1,200 ft. A vertical is drawn at 3,700 ft to the 5 percent
deceleration curve and a horizontal shows a final truck speed of 27.5 mph.
80
120
1300
70
1450
1600 100 Analysis

Type Input Output
60
1750
50 80
(km/hr)
(mph)
40
30
III FFS, LOS vP
Down
Up
40
20
IV vP, LOS N
20
10
0 0
0 200 400 600 800 1000 1200 1400 1600 1800 2000 2200 2400
Flow Rate (pcphpl)
General Information
J. Thompson Date Performed 11/24/96
Analyst
DJP Associates x
I II III IV

Highway/Dir. Travel I-405/SB
____________ ET 6.2
__________________ Tables 3-2, 3-3, 3-5
From/To Genesee/Carmel
____________ Cyn. ER __________________ Tables 3-2, 3-4
Jurisdiction CALTRANS
____________ fHV 0.930
__________________ 1
1 + PT(ET - 1) + PR(ER - 1)
Analysis Time Period AM/PM Peak Hour
____________ fp 0.95
__________________ (1.0 - 0.85)
1996 Up Down
Analysis Year ____________ vP 2392 1447
__________________ pcphpl V
Volume, V 2,300
Rt-Shoulder Lat. Clear., LC ____________ ft fN ___________mph Table 3-81
Peak-Hour Factor, PHF ____________ fID ___________mph Table 3-9
Interchange Density, ID ____________ FFS ___________mph (est.) FFSi - fLW - fN - fLC - fID
% Trucks and Buses, PT _______________ or
% RVs, PR _______________ FFS 75
___________mph (measured)
General Terrain
Level Rolling Mountainous Level of Service (LOS)
Up Down
Specific Grade Density, D _______________
44.7 20.8 pc/mi/ln vP/S
3,000/ 2,500 ft Up Down
Length ____________ LOS _______________
E C Table 3-1
Up/Down 3/5
____________ %
1
N
Driver Type
Commuter/Wk Day x Recreational/Wk End

EXAMPLE PROBLEM 5
The Freeway New urban facility being planned, forecast opening day AADT of
75,000 vpd, rolling terrain.
The Question For opening day volumes, how many lanes will be needed to
provide LOS D during peak periods?
The Facts
√ 75,000 vpd
√ K = 0.09
√ D = 55/45
√ Rolling terrain
Comments
√ Several input variables (FFS, PHF, and percent trucks) are not given. Reasonable
default values (for a moderately dense urban area) are selected as FFS = 65 mph,
PHF = 0.90, and 10 percent trucks.
√ Assume commuter traffic, given the urban characteristics of the area.
Outline of Solution The flow rate, vp, is computed for four-, six-, and eight-lane
alternatives. Table 3-1 is used to find LOS for each vp, using the 65-mph FFS curve.
Steps
1. Convert AADT to design hour volume DDHV = AADT × K × D
DDHV = 75,000 × 0.09 × 0.55
DDHV = 3,713 vph
2. Find fHV (using Table 3-2) 1

fHV =
1 + PT (ET − 1)
1
=
1 + 0.10(3 − 1)
= 0.833
3. Find vp V
vp =
(PHF)(N)(fHV)(fp)
3,713
=
(0.90)(2)(0.833)(1.0)
= 2,476 pcphpl (two lanes)
= 1,651 pcphpl (three lanes)
= 1,238 pcphpl (four lanes)
The Results Three lanes in each direction will provide for LOS D.
Summary With two lanes in each direction, flows exceed capacity. The four-lane
(each direction) option would provide LOS C.

80
120
1300
70
1450
1600 100 Analysis

Type Input Output
60
1750
50 80
(km/hr)
(mph)
40
30
40
III FFS, LOS vP
20
IV vP, LOS N
20
10
0 0
0 200 400 600 800 1000 1200 1400 1600 1800 2000 2200 2400
Flow Rate (pcphpl)
General Information
N. Terry 7/97
TX DOT x
I II III IV

Highway/Dir. Travel SR 805
____________ ET 3.0
__________________ Tables 3-2, 3-3, 3-5
From/To Caldwell Blvd/29th St
____________ ER __________________ Tables 3-2, 3-4
Jurisdiction TX DOT
____________ fHV 0.833
__________________ 1
1 + PT(ET - 1) + PR(ER - 1)
Analysis Time Period Peak Hour
____________ fp 1.0
__________________ (1.0 - 0.85)
2002 N= 2 3 4
Analysis Year ____________ vP 2476 1651 1238
__________________ pcphpl V
(PHF x N x fHV x fp)
Traffic and Roadway Conditions
Volume, V 3,713
Rt-Shoulder Lat. Clear., LC ____________ ft fN ___________mph Table 3-81
____________ fID ___________mph Table 3-9
Interchange Density, ID ____________ FFS ___________mph (est.) FFSi - fLW - fN - fLC - fID
_______________ or
% RVs, PR _______________ FFS 65
___________ mph (measured)
General Terrain
Level x Rolling Mountainous Level of Service (LOS)
Specific Grade Density, D _______________ pc/mi/ln vP/S
N=2 3 4
Length ____________ mi LOS _______________
F D C Table 3-1
Up/Down ____________ % 1
N
Driver Type

V. REFERENCES
The methodology in this chapter is based primarily on the results of a National
Cooperative Highway Research Program (NCHRP) study completed in 1995 (1). Some
adjustment factors, including those for lane width and heavy vehicles, were developed
as part of an NCHRP study of traffic flow on multilane highways (2).
1. Schoen, J., May, A., Reilly, W., and Urbanik, T., Speed-Flow Relationships
for Basic Freeway Sections. Final Report, NCHRP Project 3-45, JHK &
Associates, Tucson, Ariz. (May 1995).
2. Reilly, W., Harwood, D., Schoen, J., et al., Capacity and Level of Service
Procedures for Multilane Rural and Suburban Highways. Final Report,
NCHRP Project 3-33, JHK & Associates, Tucson, Ariz. (1988).
3. ‘‘Basic Freeway Sections (Chapter 3).’’ Special Report 209: Highway
Capacity Manual (third edition), Transportation Research Board, Washington,
D.C. (1994).
4. Hall, F.L., Hurdle, V.F., and Banks, J.H., ‘‘Synthesis of Recent Work on the
Nature of Speed-Flow and Flow-Occupancy (or Density) Relationships on
Freeways.’’ Transportation Research Record 1365, Transportation Research
Board, Washington, D.C. (1992).
5. Urbanik, T., Hinshaw, W., and Barnes, K., ‘‘Evaluation of High-Volume
Urban Texas Freeways.’’ Transportation Research Record 1320,
6. Banks, J.H., ‘‘Flow Processes at a Freeway Bottleneck.’’ Transportation
Research Record 1287, Transportation Research Board, Washington, D.C.
(1990).
7. Hall, F.L., and Hall, L.M. ‘‘Capacity and Speed-Flow Analysis of the Queen
Elizabeth Way in Ontario.’’ Transportation Research Record 1287,
8. Hall, F.L., and Agyemang-Duah, K., ‘‘Freeway Capacity Drop and the
Definition of Capacity.’’ Transportation Research Record 1320,
9. Chin, H.C., and May, A.D., ‘‘Examination of the Speed-Flow Relationship at
the Caldecott Tunnel.’’ Transportation Research Record 1320, Transportation
10. Banks, J., Evaluation of the Two-Capacity Phenomenon as a Basis for Ramp
Metering. Final Report, San Diego State University, San Diego, Calif. (1991).

APPENDIX I
PRECISE PROCEDURE FOR DETERMINING PASSENGER-CAR
EQUIVALENTS OF TRUCKS ON COMPOSITE UPGRADES
In a capacity analysis, an overall average grade can be substituted for a series of
grades if no single portion of the grade is steeper than 4 percent or the total length of
the grade is less than 4,000 ft. For grades outside these limits, that is, those having
both a total length greater than 4,000 ft and portions steeper than 4 percent, the
following technique is recommended. It estimates the continuous grade that would
result in the same final truck speed as the actual series of composite grades. The
solution for this equivalent grade uses performance curves for trucks on grades, which
are included in this appendix.
As noted elsewhere in the chapter, the truck acceleration and deceleration curves
presented and used in this appendix are for a vehicle with an average weight-to-
horsepower ratio of 200 lb/hp. This is a somewhat heavier vehicle than the usual mix
of trucks found on a typical freeway, which averages between 125 and 150 lb/hp. A
conservative approach is taken to reflect the fact that heavier vehicles will have a
greater influence on operations than those that are lighter. The 200-lb/hp vehicle is
used only to determine the equivalent of a composite grade and the passenger-car
equivalent values for trucks are based on the 125- to 150-lb/hp range.
The technique for determining composite grade equivalents is best illustrated by an
example. Consider a composite grade with 5,000 ft of 2 percent grade followed by
5,000 ft of 6 percent grade. If the average grade technique (which is not valid in this
case) were applied, the following results would be obtained:
Total rise = (5,000 × 0.02) + (5,000 × 0.06) = 400 ft
Average grade = 400 4 10,000 = 0.04 or 4 percent
The more precise method recommended in this appendix determines a percent
grade of 10,000 ft that results in the same final speed of trucks as the actual sequence
of grades. The solution for this point is illustrated in Figure I.3-1. A blank set of truck
performance curves is included as Figure I.3-2.
To find the truck speed at the end of the first 5,000 ft of 2 percent upgrade, a
vertical line is drawn from the 5,000-ft point on the horizontal axis to its intersection
with the 2 percent deceleration curve. This is Point 1 in Figure I.3-1. The truck speed
is found by drawing a horizontal line from this point to the vertical axis, which
intersects at Point 2, 47 mph. This is also the speed at which trucks enter the second
5,000 ft of 6 percent grade.
FIGURE I.3-1. SAMPLE SOLUTION FOR COMPOSITE GRADE.

LB/HP).
FIGURE I.3-2. PERFORMANCE CURVES FOR STANDARD TRUCKS (200

The intersection of the horizontal line between Points 1 and 2 and the 6 percent
deceleration curve is found and labeled Point 3. A vertical line is constructed from this
point to the horizontal axis, locating Point 4. This point indicates that at 47 mph,
trucks enter the 6 percent grade as if they had already been on it for 750 ft, starting
from level terrain.
Because trucks will now travel another 5,000 ft on the 6 percent grade, this
distance is added to the 750 ft determined by Point 4 to find Point 5 at 5,750 ft. A
vertical is constructed from this point to the intersection with the 6 percent deceleration
curve (Point 6) to find the final truck speed after the second 5,000 ft of 6 percent
upgrade. A horizontal line drawn from this point intersects the vertical axis at Point 7,
23 mph, the final truck speed.
To find a single constant grade of 10,000 ft that will result in a final truck speed
of 23 mph, a horizontal line is drawn from 23 mph to its intersection with a vertical
line drawn from 10,000 ft, Point 8. The equivalent grade is found to be 6 percent, not
the 4 percent indicated by the average grade technique.
The value of ET would now be selected for a 6 percent grade of 10,000 ft. In
general terms the following steps describe the solution for equivalent grade:
1. Enter the truck performance curves of Figure I.3-2 with initial grade and length
of grade. Find the truck speed at the end of the first grade, which is also the
speed at which trucks enter the second segment.
2. Find the length along the second grade that results in the same speed as that
found in Step 1. Use this point as the starting point along the second grade.
3. Starting with the length found in Step 2, add the length of the second grade,
and find the speed at the end of the second grade.
4. If there are additional grade segments, repeat Steps 1 through 3 for each
subsequent grade until a final speed is found.
5. Enter the truck performance curves with a final truck speed and the total length
of the composite grade to find the equivalent uniform grade percent, which may
be used in finding ET.
Note that this analysis can be applied to any number of successive grades. A given
series of grades may even include downgrade portions, or segments of level terrain.
Such points should not be used as points of demarcation between analysis segments
unless the truck speed can be shown to have returned to 55 mph under free-flow
conditions. For any given set of consecutive grades, it is important to identify the point
at which truck speeds are lowest. The deleterious effect of trucks is most severe at this
point. Thus, the appropriate point at which to evaluate a composite grade may not be
at its end, but at the point of minimum truck speed. For example, if a 4 percent
upgrade of 1 mi were followed by 1⁄2 mi of 2 percent upgrade, the point of minimum
truck speed will be the end of the 4 percent grade, not the subsequent 2 percent grade.
Note also that the procedure uses discrete grade segments and ignores the vertical
curves that join them. This simplifies computations and results in sufficient accuracy
for capacity analysis purposes.

APPENDIX II
WORKSHEET FOR ANALYSIS OF BASIC FREEWAY SECTIONS
80
12
130
70
145
160 10 Analysis

Type Input Output
60
175
50 80
(km/hr)
(mph)
40
30
40
III FFS, LOS vP
20
IV vP, LOS N
20
10
0 0
0 200 400 600 800 100 1200 1400 1600 1800 2000 2200 2400
Flow Rate (pcphpl)
General Information
I II III IV
Site Information Flow Rate (vP)

Highway/Dir. Travel ____________ ET __________________ Tables 3-2, 3-3, 3-5
From/To ____________ ER __________________ Tables 3-2, 3-4
Jurisdiction ____________ fHV __________________ 1
Analysis Time Period ____________ 1 + PT(ET - 1) + PR(ER - 1)
fp __________________ (1.0 - 0.85)
Analysis Year ____________ vP pcphpl V
__________________
Volume, V ____________ vph Free-Flow Speed (FFS)
Number of Lanes, N ____________
fLC ___________mph Table 3-7
Rt-Shoulder Lat. Clear., LC ____________ ft
fN ___________mph Table 3-81
Peak-Hour Factor, PHF ____________
fID ___________mph Table 3-9
Interchange Density, ID ____________
FFS ___________mph (est.) FFSi - fLW - fN - fLC - fID
% Trucks and Buses, PT _______________
% RVs, PR or
_______________
General Terrain FFS ___________mph (measured)
Level Rolling Mountainous
Specific Grade
Length ____________ mi Density, D _______________ pc/mi/ln vP/S
Up/Down ____________ % LOS _______________ Table 3-1
Driver Type 1
N
Commuter/Wk Day Recreational/Wk End

chapter 4
WEAVING AREAS
CONTENTS
i. introduction .......................................................................................................................................................................... 4-2

Weaving Length .................................................................................................................................................................... 4-2
Configuration ......................................................................................................................................................................... 4-2
Type A Weaving Areas ................................................................................................................................................... 4-2
Type B Weaving Areas.................................................................................................................................................... 4-4
Type C Weaving Areas.................................................................................................................................................... 4-4
Determining Configuration Type..................................................................................................................................... 4-4
Weaving Width and Type of Operation ............................................................................................................................... 4-4
Weaving Area Parameters..................................................................................................................................................... 4-5
ii. methodology.......................................................................................................................................................................... 4-6
Prediction of Weaving and Nonweaving Speeds ................................................................................................................. 4-6
Determination of Type of Operation .................................................................................................................................... 4-7
Limits on Weaving Area Operations .................................................................................................................................... 4-8
LOS Criteria .......................................................................................................................................................................... 4-9
iii. procedures for application ................................................................................................................................................ 4-9
Simple Weaving Areas.......................................................................................................................................................... 4-9
Step 1—Establish Roadway and Traffic Conditions ...................................................................................................... 4-9
Step 2—Convert All Traffic Volumes to Peak Flow Rates Under Ideal Conditions ................................................... 4-10
Step 3—Construct Weaving Diagram ............................................................................................................................. 4-10
Step 4—Compute Unconstrained Weaving and Nonweaving Speeds ........................................................................... 4-10
Step 5—Check for Constrained Operation...................................................................................................................... 4-10
Step 6—Compute Average (Space Mean) Speed and Density of All Vehicles in Weaving Area .............................. 4-10
Step 7—Check Weaving Area Limitations..................................................................................................................... 4-10
Step 8—Determine Level of Service .............................................................................................................................. 4-11
Multiple Weaving Areas ....................................................................................................................................................... 4-11
iv. sample calculations ............................................................................................................................................................ 4-12
Calculation 1—Analysis of Major Weaving Area ............................................................................................................... 4-12
Calculation 2—Analysis of Ramp-Weave Section .............................................................................................................. 4-13
Calculation 3—Constrained Operation ................................................................................................................................. 4-14
Calculation 4—Design Application ...................................................................................................................................... 4-16
Calculation 5—Multiple Weaving Area ............................................................................................................................... 4-17
Calculation 6—Sensitivity Analysis with Design Application ............................................................................................ 4-19
v. references .............................................................................................................................................................................. 4-20

4-2 freeways
I. INTRODUCTION
Weaving is defined as the crossing of two or more traffic streams

traveling in the same general direction along a significant length
of highway without the aid of traffic control devices. Weaving
areas are formed when a merge area is closely followed by a
diverge area, or when an on-ramp is closely followed by an off-
ramp and the two are joined by an auxiliary lane.
Weaving areas require intense lane-changing maneuvers as driv-
Figure 4-2. Measuring length of a weaving section.
ers must access lanes appropriate to their desired exit point. Thus,
traffic in a weaving area is subject to turbulence in excess of
that normally present on basic highway sections. This turbulence Application of these procedures to arterial weaving areas is not
presents special operational problems and design requirements that recommended.
are addressed by the procedures in this chapter.
Figure 4-1 shows the formation of a weaving area. If entry and
exit roadways are referred to as ‘‘legs,’’ vehicles traveling from WEAVING LENGTH
Leg A to Leg D must cross the path of vehicles traveling from
Leg B to Leg C. Flows A-D and B-C are, therefore, referred to The requirement that drivers execute lane changes to complete
as weaving flows. Flows A-C and B-D may also exist in the section, many weaving movements introduces a new geometric parameter
but these need not cross the path of other flows, and are referred for consideration—weaving length. The length of the weaving sec-
to as nonweaving flows. Figure 4-1 shows a simple weaving area, tion constrains the time and space in which the driver must make
formed by a single merge point followed by a single diverge point. all required lane changes. Thus, as the length of a weaving area
Multiple weaving areas, formed by one merge followed by two decreases (all other factors being constant), the intensity of lane-
diverges or two merges followed by a single diverge, are discussed changing, and the resulting level of turbulence, increases. The
later in this chapter. measurement of weaving area length is shown in Figure 4-2.
Weaving areas may exist on any type of facility: freeways, Length is measured from the merge gore area at a point where the
multilane highways, two-lane highways (in interchange areas), or right edge of the freeway shoulder lane and the left edge of the
arterials. The methodology presented in this chapter is based on merging lane(s) are 2 ft apart to a point at the diverge gore area
research focusing on freeway facilities. where the two edges are 12 ft apart.
This methodology was developed using research conducted in Procedures in this chapter generally apply to weaving sections
four widely separated studies: by the Bureau of Public Roads in of up to 2,500 ft in length. Weaving may exist in longer sections,
the early 1960s (published in an appendix to the 1965 HCM, but but merging and diverging movements are often separated, with
not used therein); by the Polytechnic University in the early 1970s lane-changing tending to concentrate near merge and diverge gore
(1–3); by Leisch in the 1970s (4), and by JHK & Associates in areas. For longer sections, merge and diverge areas may be sepa-
the early 1980s (5). rately analyzed using the procedures of Chapter 5. Weaving turbu-
The procedures herein have been updated to reflect recent re- lence may exist throughout a long section to some degree, but
search on basic freeway sections (6) and ramp junctions (7) as operations are approximately the same as those for a basic freeway
they affect weaving areas. No new data base, however, has been section.
established, and the updating is based on logical extensions of the
freeway and ramp research. Some of these updated procedures
CONFIGURATION
were developed for a text by Roess et al. (8) and are used herein
with the permission of the publisher.
Because lane-changing is the critical operational feature of
NCHRP Project 3-55(5) is expected to result in an improved
weaving areas, another critical geometric characteristic can drasti-
and expanded methodology by the year 2000.
cally affect performance: configuration. Configuration refers to
Although this chapter is primarily intended for use in analyzing
the relative placement and number of entry lanes and exit lanes
freeway weaving areas, suggestions are made for its application
for the section, and it can have a major impact on how much lane-
to weaving on uninterrupted flow segments of multilane highways.
changing must take place in the section.
The procedures in this chapter deal with three primary types of
weaving configuration. These are referred to as Type A, Type B,
and Type C sections, and are shown in Figures 4-3, 4-4, and 4-5,
respectively. The types are defined in terms of the minimum num-
ber of lane changes that must be made by weaving vehicles as
they travel through the section.
Type A Weaving Areas
Type A weaving areas require that each weaving vehicle make

one lane change in order to execute the desired movement. Figure
Figure 4-1. Formation of a weaving section. 4-3 shows two examples of Type A weaving areas. In Figure 4-3(a),
weaving areas 4-3
Figure 4-3. Type A weaving areas: (a) ramp-weave/one-sided

weave, and (b) major weave.
Figure 4-5. Type C weaving areas: (a) major weave without lane
an on-ramp is followed by an off-ramp, with a continuous auxiliary balance or merging, and (b) two-sided weave.
lane between the ramps. All on-ramp vehicles must make a lane
change out of the auxiliary lane into the shoulder lane of the
freeway, and all off-ramp vehicles must make a lane change from
the shoulder lane of the freeway to the auxiliary lane. Lane changes
because all weaving movements take place on one side of the
to and from the outer lanes of the freeway may also take place
roadway. It should be noted that on-ramps followed by off-ramps
within the section, but these are not mandated or required by the
that are not joined by a continuous auxiliary lane are not considered
weaving movement.
to be weaving areas. They are treated as separate merge and diverge
Sections formed by on-ramp/off-ramp sequences joined by con-
areas and analyzed using the procedures of Chapter 5.
tinuous auxiliary lanes are often referred to as ramp-weave sec-
Figure 4-3(b) illustrates a major weaving section. Major weav-
tions. They may also be referred to as one-sided weaving sections,
ing sections are characterized by three or more entry and exit
roadways having multiple lanes. In Figure 4-3(b), two two-lane
sections join to form a four-lane roadway, only to separate into
two two-lane sections again at the diverge point. Note that all
weaving vehicles must make at least one lane change, regardless
of the direction in which they are weaving.
Figure 4-3(a) and 4-3(b) are similar in that each has a crown
line, that is, a lane line that connects the nose of the entrance gore
area to the nose of the exit gore area. The lane change that each
weaving vehicle must make is across this crown line.
The two sections illustrated differ primarily in the impact of
ramp geometrics on speed. For many ramp-weave sections, the
design speed of ramps is significantly lower than that of the free-
way. Thus, on- or off-ramp vehicles must accelerate or decelerate
as they traverse the weaving section. For major weaving sections,
the design of multilane entry and exit legs is more compatible
with the design of the freeway mainline, and the impact of accelera-
tion and deceleration in the section is minimal. It should be noted,
however, that this difference is not reflected in the procedures in
this chapter because of the relative scarcity of major weave sites
with crown lines and the lack of data concerning operations in
such sites.
Because weaving vehicles in a Type A weaving area must cross
the crown line, weaving vehicles are usually confined to occupying
the two lanes adjacent to the crown line while in the weaving
Figure 4-4. Type B weaving areas: (a) major weave with lane section. Normally, some nonweaving vehicles will also remain in
balance at exit gore, (b) major weave with merging at entrance lanes adjacent to the crown line. Lanes adjacent to the crown
gore, and (c) major weave with merging at entrance gore and lane line are, therefore, generally shared by weaving and nonweaving
balance at exit gore. vehicles. One of the most significant effects of configuration on
4-4 freeways
operations is to limit the maximum number of lanes that weaving Table 4-1. Configuration Type Versus Minimum
vehicles may occupy while traversing the section. Number of Required Lane Changes
Type B Weaving Areas }a
All weaving areas classified as Type B may also be referred to

}b
as major weaving sections, because all involve multilane entry minimum number minimum number of req’d lane
legs or exit legs or both. Two critical characteristics distinguish of req’d lane changes for
Type B weaving areas from all others: changes for weaving mvt. b
weaving mvt. a 0 1 ≥2
1. One weaving movement may be accomplished without mak-
ing any lane changes. 0 Type B Type B Type C
1 Type B type A —
2. The other weaving movement requires at most one lane ≥2 Type C — —
change.
Figure 4-4(a) and (b) show two such weaving areas. In both
illustrations, movement B-C can be made without executing any
lane changes, whereas movement A-D requires only one lane lane balance at the exit gore and no crown line exists. Although
change. In Figure 4-4(a), this is accomplished by providing a such a section is relatively efficient for weaving movements in the
diverging lane at the exit gore. From this lane, a vehicle may direction of the through lane, it cannot efficiently handle large
proceed on either exit leg without making a lane change. This type weaving volumes in the other direction.
of design is also referred to as lane balanced, that is, the number Figure 4-5(b) shows a two-sided weaving area. It is formed
of lanes leaving the diverge is one greater than the number of when a right-hand on-ramp is followed by a left-hand off-ramp or
lanes approaching it. In Figure 4-4(b), a lane from Leg A is merged vice versa. In such cases, the through volume on the freeway is
with a lane from Leg B at the entrance gore area. functionally a weaving movement. Ramp-to-ramp vehicles must
Type B weaving areas are extremely efficient in carrying large cross all lanes of the freeway to execute their desired maneuver.
weaving volumes, primarily because of the provision of a through Freeway lanes are, in effect, through weaving lanes. Ramp-to-
lane for one of the weaving movements. Weaving maneuvers can ramp drivers must execute three lane changes in Figure 4-5(b).
be accomplished with a single lane change from the lane or lanes Although it is technically a Type C configuration, there is little
adjacent to this through lane. Thus, weaving vehicles can occupy information concerning the operation of such sections, and the
a substantial number of lanes in the weaving section and are not methodology of this chapter is only a rough approximation of their
as restricted in this regard as in Type A sections. characteristics. They should generally be avoided in cases where
Figure 4-4(c) shows an unusual configuration in which both a there is any significant ramp-to-ramp volume.
merge of two lanes at the entrance gore and lane balance at the
exit gore are provided. In this case, both weaving movements can
Determining Configuration Type
be made without a lane change. Again, weaving movements can
be made with a single lane change from the two lanes adjacent to
Figures 4-3, 4-4, and 4-5 show the three basic types of weaving
the through lane. Such configurations are usually found on collec-
area configuration. Weaving configuration is determined on the
tor-distributor roadways. Although some weaving movements are
basis of the number of required lane changes that must be per-
accomplished as a merge followed by a diverge, lane changes to
formed by the two weaving flows in the section. This determination
and from lanes adjacent to the through lane yield real weaving
ignores lane changes that are not necessary to the completion of
activity, and these sections are analyzed as weaving areas.
a particular weaving movement. Table 4-1 identifies the configura-
tion type on the basis of lane-changing characteristics.
Type C Weaving Areas
WEAVING WIDTH AND TYPE OF OPERATION
Type C weaving areas are similar to Type B sections in that
one or more through lanes are provided for one of the weaving The third geometric characteristic with a significant impact on
movements. The distinguishing feature between Type B and Type weaving area operations is the width of the weaving area, measured
C sections is the number of lane changes required for the other as the number of lanes in the section. It is, however, not only the
weaving movement. A Type C weaving area is characterized as total number of lanes that affects weaving area operations, but
follows: the proportional use of those lanes by weaving and nonweaving
1. One weaving movement may be accomplished without mak- vehicles.
ing a lane change. The nature of weaving movements creates traffic stream turbu-
2. The other weaving movement requires two or more lane lence and results in the consumption of more of the available
changes. roadway space by a weaving vehicle than by a nonweaving vehicle.
The exact nature of relative space use depends on the relative
Figure 4-5 shows two Type C weaving areas. In Figure 4-5(a), weaving and nonweaving volumes using the weaving area and the
movement B-C does not require lane-changing, whereas movement number of lane changes that weaving vehicles must make. The
A-D requires two lane changes. This type of section is formed latter is, as discussed, dependent on the configuration of the weav-
when there is neither a merging of lanes at the entrance gore nor ing section. Thus, the proportional use of space is dependent not
weaving areas 4-5
only on relative volumes, but also on the configuration of the Where configuration does not restrain weaving vehicles from
weaving area. occupying a balanced proportion of available lanes, the operation
Configuration has a further impact on proportional use of avail- is classified as unconstrained. Average running speeds of weaving
able lanes. The configuration can limit the ability of weaving and nonweaving vehicles generally differ by less than 5 mph,
vehicles to use outer lanes in the section. This limitation is most except in short Type A sections, where acceleration and decelera-
severe in Type A sections, in which all weaving vehicles must tion of ramp vehicles limit their average speed regardless of the
cross a crown line, and is least severe in Type B sections. use of available lanes.
In general, vehicles in a weaving area will make use of available A major component of the procedure presented in this chapter
lanes in such a way that all component flows achieve approxi- is the determination of whether operations in a given section are
mately the same average running speed, with weaving flows constrained or unconstrained. This is discussed in the Methodology
somewhat slower than nonweaving flows. Occasionally, the con- section.
figuration limits the ability of weaving vehicles to occupy the
proportion of available lanes required to achieve this equivalent
or balanced operation. In such cases, weaving vehicles occupy a
WEAVING AREA PARAMETERS
smaller proportion of the available lanes than desired, and non-
weaving vehicles occupy a larger proportion of lanes than for
balanced operation. When this occurs, the operation of the weaving The introductory portions of this chapter have discussed a num-
area is classified as constrained by the configuration. The result ber of parameters that may affect the operation of weaving areas.
of constrained operation is that nonweaving vehicles will operate For convenience, Table 4-2 presents these measures and defines
at significantly higher speeds than weaving vehicles. the symbols that will be used to depict them.
Table 4-2. Parameters Affecting Weaving Area Operation

symbol definition
L................................................................ Length of weaving area, in ft.

LH .............................................................. Length of weaving area, in hundreds of ft.
N ............................................................... Total number of lanes in the weaving area.
Nw .............................................................. Number of lanes used by weaving vehicles in the weaving area.
Nnw............................................................. Number of lanes used by nonweaving vehicles in the weaving area.
v ................................................................ Total flow rate in the weaving area, in passenger car equivalents, in pcph.
vw............................................................... Total weaving flow rate in the weaving area, in passenger car equivalents, in pcph.
vw1 ............................................................. Weaving flow rate for the larger of the two weaving flows, in passenger car equivalents, in pcph.
vw2 ............................................................. Weaving flow rate for the smaller of the two weaving flows, in passenger car equivalents, in pcph.
vnw ............................................................. Total nonweaving flow rate in the weaving area, in passenger car equivalents, in pcph.
VR ............................................................. Volume ratio vw /v.
R ............................................................... Weaving ratio vw2/vw.
Sw .............................................................. Average (space mean) speed of weaving vehicles in the weaving area, in mph.
Snw ............................................................. Average (space mean) speed of nonweaving vehicles in the weaving area, in mph.

4-6 freeways
II. METHODOLOGY
The methodology presented in this chapter has four distinct a (1 + VR)b (v/N)c
W= (4-3)
components: Ld
1. Equations predicting the average running speed of nonweav- where
ing and weaving vehicles in a weaving area based on known road-
way and traffic conditions. Equations are specified for each VR = volume ratio, vw /v;
configuration type, and for unconstrained and constrained v = total flow rate in weaving area (equivalent pcph);
operations. vw = weaving flow rate in weaving area (equivalent pcph);
2. Equations describing the proportional use of available lanes N = number of lanes in weaving area; and
by weaving and nonweaving vehicles, used to determine whether L = length of weaving area (ft).
operations are constrained or unconstrained.
Constants a, b, c, and d are given in Table 4-3. They vary on the
3. Definitions of limiting values of key parameters for each type
basis of three factors:
of weaving configuration, beyond which equations do not apply.
4. Definition of LOS criteria based on average running speeds 1. Whether a weaving speed, Sw, or a nonweaving speed, Snw,
of weaving and nonweaving vehicles. is being predicted;
2. Configuration type (A, B, or C) of the weaving area; and
These components are discussed in the sections that follow.
3. Whether operations are constrained or unconstrained.
In the case of Item 3, initial computations always assume uncon-

PREDICTION OF WEAVING AND NONWEAVING SPEEDS strained operations. The assumption is subsequently tested for
validity.
The heart of the weaving area analysis procedure is the predic- Equations 4-2 and 4-3 yield sensitivities that are consistent with
tion of speeds and density of vehicles within the weaving area. observed operations of weaving areas. Specifically,
Because weaving and nonweaving vehicles may travel at speeds
that are similar or at speeds that are markedly different, average 1. As the length of the weaving section increases, speeds also
(space mean) speeds are predicted separately for weaving and increase as the intensity of lane-changing declines.
nonweaving vehicles. An average speed and density for all vehicles 2. As the proportion of weaving vehicles in the total flow, VR,
is then estimated, and level of service is based on the estimated increases, speeds decrease, reflecting the increased turbulence
density. caused by higher proportions of weaving vehicles in the traffic
The algorithm for prediction of weaving and nonweaving speeds stream.
may be stated in general terms as 3. As the total average flow per lane, v/N, increases, speeds
decrease, reflecting more intense demand.
Smax − Smin 4. Constrained operations will have lower weaving speeds and
Si = Smin + (4-1)
1+W higher nonweaving speeds than similar unconstrained operations
because weaving vehicles are constrained to less space than they
where would need for unconstrained operation, whereas nonweaving ve-
Si = speed of weaving (i = w) or nonweaving (i = nw) vehicles hicles have more.
(mph), 5. Type B sections are the most efficient for handling large
Smin = minimum speed expected in section (mph), weaving flow rates. For high flow rates, weaving speeds are higher
Smax = maximum speed expected in section (mph), and than for similar Type A and C sections.
W = weaving intensity factor. 6. The sensitivity of weaving speed to increasing VR is greatest
for Type A configurations and least for Type B configurations,
For the purposes of these procedures, the minimum speed, Smin, is illustrating the greater efficiency of B sections in handling large
taken to be 15 mph. The maximum speed, Smax, is taken to be the proportions of weaving vehicles in the traffic stream. It also sug-
average of the free-flow speeds of freeway segments entering and gests that Type A sections are effective where the proportion of
leaving the section plus 5 mph. The addition of 5 mph to the free- weaving vehicles in the traffic stream is relatively low.
flow speed adjusts for the tendency of the algorithm to underpredict 7. The sensitivity of weaving speed to increasing length is great-
high speeds. Setting the maximum and minimum speeds in this est for Type A sections, because vehicles are often accelerating
way constrains the prediction range of the algorithm to reasonable or decelerating through the section in this configuration. The sensi-
values. With the assumed maximum and minimum speeds defined, tivity of weaving speed is less for Type B and C configurations,
the algorithm becomes where at least one weaving movement is made without a lane
change.
SFF − 10
Si = 15 + (4-2) It is important to note that Type A configurations are quite
1+W
different from Type B and Type C configurations. Because all
where SFF is the average free-flow speed of the freeway segments weaving vehicles must cross a crown line in Type A sections,
entering and leaving the weaving area. weaving and nonweaving flows tend to become segregated in such
The weaving intensity factor, W, is a measure of weaving activ- segments; weaving vehicles become concentrated in lanes adjacent
ity and its intensity. It is computed as to the crown line, and nonweaving vehicles gravitate to outer lanes.
weaving areas 4-7
Table 4-3. Constants of Prediction for Weaving Intensity Factor, W

general form:
a (1 + VR)b (v/N)c
W=
Ld
constants for constants for
weaving nonweaving
type of speed, Sw speed, Snw
configuration a b c d a b c d
Type A
Unconstrained 0.226 2.2 1.00 0.90 0.020 4.0 1.30 1.00
Constrained 0.280 2.2 1.00 0.90 0.020 4.0 0.88 0.60
Type B
Unconstrained 0.100 1.2 0.77 0.50 0.020 2.0 1.42 0.95
Constrained 0.160 1.2 0.77 0.50 0.015 2.0 1.30 0.90
Type C
Unconstrained 0.100 1.8 0.80 0.50 0.015 1.8 1.10 0.50
Constrained 0.100 2.0 0.85 0.50 0.013 1.6 1.00 0.50
NOTE: All variables are as defined in Table 4-2.
In Type B and C configurations, there is substantial mixing of in the data bases reported by Pignataro et al. (1), Roess et al. (2),
weaving and nonweaving vehicles across a number of lanes. and Reilly et al. (5).
This difference makes Type A sections behave somewhat differ- Type A sections are the most restrictive in terms of the maxi-
ently from Type B or C sections. Speeds tend to be higher in Type mum number of lanes that can be used by weaving vehicles. As
A sections than in Type B or C sections with the same flows, noted previously, weaving vehicles must, in general, confine them-
length, and number of lanes. This does not, however, suggest selves to the two lanes adjacent to the crown line in order to
that Type A sections always operate better than Type B or C execute their desired maneuvers. However, nonweaving vehicles
configurations with similar lengths, widths, and flows. Restrictions will also remain in these lanes, and full use of them by weaving
on the amount of weaving flow that Type A sections can accommo- vehicles is not a reasonable expectation. For Type A sections,
date are more severe than those for other configurations. These weaving vehicles generally use at most 1.4 lanes, regardless of the
restrictions are discussed in a subsequent section (see Table 4-5). total number of lanes available.
Type B sections do not greatly restrict weaving vehicles in their
DETERMINATION OF TYPE OF OPERATION
use of available lanes. Weaving vehicles may occupy up to 3.5
lanes in a Type B section. This is based on the full use of through
The determination of whether a particular section is operating weaving lanes and lanes immediately adjacent to the through lane,
in a constrained or unconstrained state is based on the comparison as well as partial use of outer lanes. Such configurations are most
of two variables: efficient when weaving flows compose substantial portions of the
traffic stream. Because weaving vehicles may filter through most
Nw = number of lanes that must be used by weaving vehicles of the lanes in the segment, nonweaving vehicles tend to share
in order to achieve balanced or unconstrained opera- lanes and are generally unable to segregate themselves from weav-
tion; and ing flows.
Nw (max) = maximum number of lanes that may be used by weav- Type C sections are similar to Type B sections in the provision
ing vehicles for a given configuration. of a through weaving lane. The multiple lane-changing required of
Fractional values for lane requirements of weaving vehicles may one weaving movement, however, restricts the ability of weaving
occur because lanes are shared with nonweaving vehicles. vehicles to use outer lanes of the sections. Thus, in Type C sections,
Cases for which Nw ≤ Nw (max) will be unconstrained, because weaving vehicles can use no more than 3.0 lanes. One exception
there are no impediments to weaving vehicles using the required to this rule is a two-sided weaving area [see Fig. 4-5(b)]. For two-
number of lanes. Where Nw > Nw (max), the configuration con- sided configurations, all freeway lanes are through weaving lanes,
strains weaving vehicles to a smaller number of lanes than that and weaving vehicles may therefore use all lanes without
required for balanced operation. Such cases are constrained and restriction.
will result in average nonweaving vehicle speeds significantly The proportional use of available lanes by weaving vehicles is
higher than average weaving vehicle speeds. again quite different for Type A sections as compared with Type
Table 4-4 contains equations for the computation of Nw and values B and C sections. In Type A sections, more lanes are required by
for Nw (max), both of which vary with the type of configuration. weaving vehicles for balanced operation as length increases. This
The equations for Nw are based on weaving and nonweaving is primarily due to the substantial segregation of weaving and
speeds for unconstrained operation. Computed values are com- nonweaving flows in such sections, and the higher speeds of weav-
pared with the maximum values shown in the third column of ing vehicles that result. As length increases, weaving speeds be-
Table 4-4 to determine whether operations are constrained or un- come quite high, and more space is required by weaving vehicles
constrained. Values of Nw (max) in Table 4-4 reflect observations to maintain these speeds. This characteristic produces, however,
4-8 freeways
Table 4-4. Criteria for Unconstrained Versus Constrained Operation of Weaving Areas
type of no. of lanes req’d for unconstrained max. no. of weaving
configuration operation, Nw lanes, Nw (max)
Type A 2.19 N VR0.571 LH0.234/Sw0.438 1.4
Type B N [0.085 + 0.703VR + (234.8/L) − 0.018(Snw − Sw)] 3.5
Type C N [0.761 − 0.011LH − 0.005(Snw − Sw) + 0.047VR] 3.0 a
NOTE: All variables are as defined in Table 4-2. When Nw ≤ Nw (max), operation is unconstrained. When Nw > Nw (max), operation is constrained.
a
For two-sided weaving areas, all freeway lanes may be used as weaving lanes.
an interesting result. For any given set of flows and number of diverging maneuvers tend to become segregated, and that the pro-
lanes, it is more likely for a Type A section to operate in the cedures of Chapter 5 may be applied. Speeds in longer sections
constrained mode as length is increased. tend to approach those that would be achieved in a basic freeway
Type B and Type C sections show an opposite trend. Increasing section, even where some weaving turbulence exists.
length has a much smaller impact on weaving speed than for The weaving capacity of a Type A section is limited to a flow
Type A sections, primarily because of the mixing of weaving and rate of 2,000 pcph because all weaving vehicles must cross a single
nonweaving flows. As length increases, the proportion of lanes crown line, restricting the number of vehicles that can cross from
required by weaving vehicles for balanced operation decreases, one side of the roadway to the other. Type B and C sections can
and it is less likely that constrained operation will occur. accommodate up to 3,500 pcph and 3,000 pcph, respectively. This
is primarily due to the existence of a through lane for weaving
vehicles and the flexibility in lane use provided by such configura-
LIMITS ON WEAVING AREA OPERATIONS tions. It is critical to note that weaving flow rates higher than these
values cannot normally be sustained within a weaving area within
Table 4-5 gives a number of limitations on the application of the length limitations of Table 4-5.
this methodology that may not be obvious from either the speed As length increases beyond the maximum limits shown, weav-
or lane use equations described previously. These include weaving ing capacity is difficult to define. When the length increases to
capacity, maximum flow rate per lane, and maximum volume and a point where weaving lane-changing is no greater than the
weaving ratios at which the various configuration types generally normal lane-changing that would occur on a basic freeway
operate, as well as length limits beyond which merge and diverge segment, weaving flow rates are constrained only by the total
areas may operate independently. capacity of the freeway. The length required to achieve this,
The interpretation of each of these limitations varies. In the case however, is not precisely defined. Analysts and designers should
of limitations on weaving flow rate, vw, and total flow rate per view with caution any weaving flow rates in excess of the Table
lane, v/N, acceptable operations are unlikely beyond these values. 4-5 values. Changes in the basic design of the freeway system,
They are therefore maximum values that may be accommodated including provision of grade separations, may be considered to
in a weaving section, within the maximum lengths shown in Table accommodate higher flows.
4-5. Limits on volume ratio, VR, and weaving ratio, R, represent A maximum limitation on v/N should also be observed in weav-
values beyond which weaving operations are rarely observed. ing areas within the length limits of Table 4-5. This limitation is
Higher values may occur, but these fall outside the prediction based on the per-lane capacity of the basic freeway section entering
range of the methodology, and results should be taken as approxi- the freeway, and is expressed in Table 4-5 as c-100 for Type A
mate. Length limitations, L, represent the range of the calibration and B configurations, and c-200 for Type C configurations, which
data base. As noted previously, weaving may occur in longer sec- are more restrictive, where c is the capacity per lane under ideal
tions. In such cases it is generally considered that merging and conditions for a basic freeway or multilane highway segment.
Table 4-5. Limitations on Weaving Sections

weaving maximum
capacity maximum maximum weaving
vw (max),a v/N,b VRc maximum length
configuration pcph pcphpl N VR Rd L,e ft
Type A 2,000 c-100 2 1.00 0.50 2,000
3 0.45
4 0.35
5 0.22
Type B 3,500 c-100 0.80 0.50 2,500
Type C 3,000 c-200 0.50 0.40 2,500
a
Section likely to fail at higher weaving flows.
b
Section likely to fail at higher average per-lane flows.
c
Section will likely operate at lower speeds than predicted if VR limit is exceeded.
d
Section will likely operate at lower speeds than predicted if R limit is exceeded.
e
When length exceeds these limits, merge and diverge are treated as isolated junctions and analyzed accordingly.

weaving areas 4-9
The capacity per lane is found from the procedures in Chapter Table 4-6. LOS Criteria for Weaving Areas
3 and is related to the average free-flow speed of the freeway maximum density (pc/mi/ln)
sections entering and leaving the weaving area.
level of freeway multilane and c-d
Limitations on volume ratio, VR, reflect the character of each service weaving area weaving areas
configuration type. Type A sections are intended to handle small
weaving flows comprising a minority of the traffic stream. Because A 10 12
B 20 24
weaving vehicles do not normally use more than 1.4 lanes in such C 28 32
sections, the limiting VR depends on the total number of lanes D 35 36
available and decreases as N increases. Freeway weaving areas E ≤43 ≤40
with Type A configurations generally should not be used where F >43 >40
weaving traffic includes a proportion of total flow larger than that
shown in Table 4-5.
Type C configurations are more generous in handling larger where S is the average (space mean) speed of all vehicles in the
proportions of weaving traffic, but are still not efficient where weaving section in miles per hour, and all other variables are as
weaving flows dominate total flow. Only Type B configurations previously defined. The density is then found:
effectively handle situations in which VR > 0.50 and N > 2.
The weaving ratio, R, is the ratio of the smaller weaving flow v/N
D= (4-5)
to the total weaving flow. Its maximum value is 0.50, which occurs S
when the two weaving flows are equal. Neither Type A nor Type
B configurations have any practical limitation on R, because both where D is the density in passenger cars per mile per lane.
can accommodate equal weaving flows without operational prob- Table 4-6, contains LOS criteria based on density in the weaving
lems. Type C configurations, however, are most efficient where area. Note that criteria are shown for freeways as well as for
weaving flows are unequal because one weaving movement re- multilane highways and collector-distributor (C-D) roadways. The
quires no lane-changing, whereas the other requires two or more procedures in this chapter can be applied to weaving sections on
lane changes. Such sections generally do not operate efficiently multilane highways by using an appropriate free-flow speed in the
when the weaving ratio exceeds 0.40, with the larger flow in the prediction of nonweaving and weaving vehicle speeds. Chapter 7
direction requiring no lane changes. contains procedures for the estimation of free-flow speed on a
multilane highway if field measurements are not available. Multi-
lane criteria may be cautiously applied to C-D roadway weaving
LOS CRITERIA areas, but it is recommended that free-flow speed be measured or
roughly estimated from design speed or speed limit information.
Level of service in weaving areas is related to the average den- In general, these criteria allow for slightly higher densities at
sity of all vehicles in the section. Average density in the weaving any given LOS threshold than on a comparable basic freeway
area is computed by finding the average (space mean) speed of or multilane highway section. This follows the philosophy that
all vehicles in the weaving section and then estimating density as drivers expect higher densities in weaving areas relative to those
total flow divided by average (space mean) speed: on basic freeway or multilane highway segments. The LOS E-
vw + vnw F boundary does not apply this approach. Rather, it is thought
S= (4-4) that breakdown will occur at slightly lower densities than on
vw vnw basic sections because of the additional turbulence resulting from
+
Sw Snw weaving movements.
III. PROCEDURES FOR APPLICATION
SIMPLE WEAVING AREAS a straight-line form. The relative placement of entry and exit
points (A, B, C, D) in the diagram matches the actual site to
ensure proper placement of weaving and nonweaving flows
Procedural steps for the analysis of simple weaving areas are
relative to each other. Flows on the weaving diagram should
given below. Computations are performed in the operational anal-
represent flow rates for the peak 15 min under ideal conditions,
ysis mode; that is, a known or projected situation is analyzed for
expressed in passenger cars per hour. It is also convenient to
the probable level of service. All roadway and traffic conditions
use the weaving diagram as a guide in computing the parameters
must be specified, including weaving length, type of configuration,
used during an analysis.
number of lanes, lane widths, terrain or grade, weaving and non-
The level of service in an existing or projected weaving area is
weaving flow rates by movement, peak-hour factor, and traffic
evaluated using the following computational steps.
composition.
Weaving analysis is made easier through the use of a weaving
Step 1—Establish Roadway and Traffic Conditions
diagram, which is a schematic drawing showing weaving and
nonweaving flows in a weaving area. Figure 4-6 shows such a All existing or projected roadway and traffic conditions must
diagram. Note that the weaving diagram depicts actual flows in be specified. Roadway conditions include the length, number of
4-10 freeways
V
v= (4-6)
PHF × fHV × fw × fp
where
v = flow rate for peak 15 min under ideal conditions (pcph);
V = hourly volume under prevailing condition (veh/hr);
PHF = peak-hour factor;
fHV = heavy-vehicle adjustment factor, determined using the
procedures in Chapter 3 or 7; and
fp = driver population adjustment factor, determined using the
procedures in Chapter 3.
Step 3—Construct Weaving Diagram
A weaving diagram of the type illustrated in Figure 4-6 is now

constructed, with all flows indicated as peak flow rates under ideal
conditions in passenger cars per hour. Critical analysis variables
are identified and computed as shown in Figure 4-6.
Step 4—Compute Unconstrained Weaving and Nonweaving

Speeds
Using weaving intensity factors for the appropriate configuration

from Table 4-3, compute the average (space mean) speed for weav-
ing and nonweaving vehicles. Unconstrained operation is assumed
for this step.
Figure 4-6. Construction and use of weaving diagrams.
Step 5—Check for Constrained Operation
Using the speeds computed in Step 4, estimate the number of

lanes, and type of configuration for the weaving area under study. lanes needed by weaving vehicles to achieve unconstrained opera-
Table 4-1, should be consulted in assigning the type of configuration using the equations in Table 4-4. Compare the computed value
tion. Other roadway features of importance are lane widths and of Nw with the tabulated value of Nw (max) to determine whether
the general terrain or grade conditions for the section. operation is constrained or unconstrained.
Traffic conditions include the distribution of vehicle types in the If Nw ≤ Nw (max), the operation is unconstrained, and the
traffic stream, as well as the peak-hour factor, or peak-hour factors speeds computed in Step 4 are accurate. If Nw > Nw (max), the
where the component flows have differing peaking characteristics. operation is constrained. Values of Sw and Snw must be recomputed
Because the weaving area should be analyzed on the basis of using Equation 4-3 and the constrained weaving intensity factor
peak flow rates for a 15-min interval within the hour of interest, for the appropriate configuration given in Table 4-3.
hourly volumes must be adjusted by dividing by the peak-hour
factor. Such a conversion, however, ignores the fact that the four Step 6—Compute Average (Space Mean) Speed and
component flows in a weaving area may not all peak during the Density of All Vehicles in Weaving Area
same interval. Where possible, weaving flows should be observed
and recorded for 15-min intervals so that critical periods may be Use Equation 4-4 to compute the average (space mean) speed
identified for analysis. Where hourly volumes are available or of all vehicles in the weaving section. The result may be used
projected, it will be assumed that all component flows peak simul- in Equation 4-5 to compute the density in weaving section. Then
taneously—a conservative procedure. The predicted speeds of
weaving and nonweaving vehicles will be lower than those actually vw + vnw
S=
occurring in such cases. It should also be noted that the component vw vnw
movements in a weaving area may not have the same peak-hour +
Sw Snw
factor. Where possible, each flow and its peaking characteristics
should be considered separately. v/N
D=
S
Step 2—Convert All Traffic Volumes to Peak Flow Rates
Under Ideal Conditions
Step 7—Check Weaving Area Limitations
Because all of the speed and lane-use algorithms presented ear- Table 4-5 should be consulted to ensure that none of the limita-
lier are based on peak flow rates under ideal conditions, expressed tions specified are exceeded. Where one or more of these limits
in passenger cars per hour, all component flows must be converted are exceeded, consult the Methodology section of this chapter for
to this basis: the appropriate interpretation.
weaving areas 4-11
Care should be taken in applying the limiting values given weaving movements. Figures 4-7 and 4-8 show the two types
in Table 4-5. Where the weaving capacity is exceeded, it is of multiple weaving areas and where weaving movements are
likely that breakdowns will occur and that LOS F will prevail, most likely to take place. Weaving diagrams may be developed
at least for weaving vehicles. Where limitations on VR or R for each subsegment of the weaving area, each of which can
are exceeded, breakdowns need not occur, but speeds would be be analyzed as a simple weaving area using the procedures
lower than those anticipated by the equations of Table 4-3. specified earlier.
Maximum lengths reflect the limits of the predictive equations. Figure 4-7 depicts a single merge area followed by two diverge
Lengths beyond the values shown may be analyzed as separate areas. The weaving of Movement 5 with Movements 3 and 4 must
merge and diverge areas using the procedures in Chapter 5. It take place in the first segment, because vehicles in Movement 5
would not be expected that speeds within the section would be leave at the first diverge point. The weaving of Movement 2
significantly lower than those for a basic freeway section serving with Movement 3 may take place anywhere in either segment
the same volume. of the section. However, to avoid the turbulence of weaving
that must take place in the first segment, these latter weaving
movements will tend to concentrate in the second segment of
Step 8—Determine Level of Service
the section.
The estimated value of density, D, in the weaving area is com- Figure 4-8 depicts two merge areas followed by a single diverge
pared with the criteria in Table 4-6 to determine the prevailing area. In this case, the weaving of Movements 3 and 4 with Move-
level of service. ment 5 must take place in the second segment of the section,
because Movement 5 enters at the second merge. Although the
weaving of Movements 2 and 3 could take place anywhere in the
MULTIPLE WEAVING AREAS
section, it will tend to be concentrated in the first segment, because
Multiple weaving areas are formed when one merge point is drivers seek to avoid the turbulence of other weaving movements
followed closely by two diverge points or where two merge points in the second segment.
are closely followed by a single diverge point. In such cases, Thus, the analysis of multiple weaving areas involves the con-
several sets of weaving movements take place over the same seg- struction of appropriate weaving diagrams for each subsegment of
ments of freeway, and lane-changing turbulence may be higher the area using Figures 4-7 and 4-8. Once these diagrams are estab-
than that found in simple weaving areas. lished, each subsegment may be analyzed as a simple weaving
Drivers will carefully select where to execute their required area, according to the procedures in this chapter. Limits established
lane changes in a manner that minimizes interference with other in Table 4-5 would apply to the individual subsegments.
Figure 4-7. Weaving flows in a multiple weave formed by a single merge followed by two diverges.
4-12 freeways
Figure 4-8. Weaving flows in a multiple weave formed by two merge points followed by a single diverge.
IV. SAMPLE CALCULATIONS
The following sample calculations illustrate the application and Solution

interpretation of the methodology presented in this chapter.
The calculation is conducted according to the steps outlined in
the Procedures for Application section of this chapter.
CALCULATION 1—ANALYSIS OF MAJOR WEAVING AREA

Description
The existing geometrics and traffic volumes are stated in the
The weaving area illustrated in Figure 4-9 serves the following description. Note that the section is a Type B configuration (see
traffic volumes: A–C = 1,815 veh/hr; A–D = 692 veh/hr; B–C = Table 4-1). Weaving Movements B and C may be made without
1,037 veh/hr; B–D = 1,297 veh/hr. Traffic volumes include 10 a lane change, whereas Movements A–D can be made with a single
percent trucks, and the PHF is 0.91. The section is located in lane change.
generally level terrain, and lane widths are 12 ft. There are no
lateral obstructions. The driver population is composed primarily Step 2—Convert All Traffic Volumes to Peak Flow Rates
of commuters. The observed free-flow speed of the freeway, SFF, Under Ideal Conditions
is 65 mph. At what level of service will the section operate?
All volumes must be converted to peak flow rates under ideal
conditions expressed in passenger cars per hour:
V
v=
PHF × fHV × fp
where
PHF = 0.91 (given);
ET = 1.5 (Table 3-3);
fHV = 0.95 (computed as 1/[1 + 0.10(1.5 − 1)]; and
fp = 1.00.
SFF = 65 mph
Then
A–C = 1,815/(0.91 × 0.95 × 1.00) = 2,100 pcph
Figure 4-9. Weaving area for Calculation 1. A–D = 692/(0.91 × 0.95 × 1.00) = 800 pcph
weaving areas 4-13
B–C = 1,037/(0.91 × 0.95 × 1.00) = 1,200 pcph As 2.4 lanes is less than Nw (max), which is 3.5 lanes for a Type
B–D = 1,297/(0.91 × 0.95 × 1.00) = 1,500 pcph B section, the operation is unconstrained, and the predicted speeds
of Step 4 stand without modification.
Step 3—Construct Weaving Diagram
Step 6—Compute Average (Space Mean) Speed and
A weaving diagram for the calculation is now constructed using Density of All Vehicles in Weaving Area
the converted flow rates of Step 2:
Level of service is found by first computing the average (space
mean) speed of all vehicles and then the average density of all
vehicles in the weaving section:
vw + vnw
S=
vw vnw
+
Sw Snw
2,000 + 3,600
S= = 42.2 mph
2,000 3,600
+
Critical ratios may also be computed for use in analysis: 42.7 42.0
v/N
vw = 1,200 + 800 = 2,000 pcph D=
S
v = 2,000 + 2,100 + 1,500 = 5,600 pcph
R = 800/2,000 = 0.400 5,600/4
D= = 33.2 pc/hr/ln
VR = 2,000/5,600 = 0.357 42.2
From the density criteria in Table 4-6, this is LOS D.
Step 4—Compute Unconstrained Weaving and
Nonweaving Speeds Step 7—Check Weaving Area Limitations
The unconstrained speeds of weaving and nonweaving vehicles
All limiting values of Table 4-5 are met. The results are therefore
may be estimated using Equation 4-3 and the weaving intensity
expected to be valid as indicated.
factors computed from Table 4-3:
a (1 + VR)b (v/N)c
W= CALCULATION 2—ANALYSIS OF RAMP-WEAVE SECTION
Ld
0.10 (1 + 0.357)1.2 (5,600/4)0.77 Description
Ww = = 0.9855
1,5000.5
0.02 (1 + 0.357)2.0 (5,600/4)1.42 The weaving section shown in Figure 4-10 serves the traffic
Wnw = = 1.0385 flows indicated. Lane widths are 12 ft and the section is located
1,5000.95
in level terrain. There are no lateral obstructions. For convenience,
Then, on the basis of a free-flow speed of 65 mph for the freeway, all traffic flows are given in terms of peak flow rates for ideal
weaving and nonweaving speeds may be computed as follows: conditions, expressed in passenger cars per hour. The free-flow
speed of the freeway is observed to be 70 mph through field
SFF − 10
Si = 15 + evaluation. At what level of service will the section operate?
1+W
65 − 10
Sw = 15 + = 42.7 mph
1 + 0.9855
65 − 10
Snw = 15 + = 42.0 mph
1 + 1.0385
Using the unconstrained estimates of weaving and nonweaving

vehicle speeds, the assumption of unconstrained operation is
SFF = 70 mph
checked using the criteria in Table 4-4. From this table, the number
of lanes required by weaving vehicles for unconstrained operation
is computed as follows:
Nw = N [0.085 + 0.703VR + (234.8/L) − 0.018(Snw − Sw)]
Nw = 4[0.085 + 0.703(0.357) + (234.8/1,500)
− 0.018(42.0 – 42.7)] = 2.4 lanes Figure 4-10. Weaving area and flows for Calculation 2.
4-14 freeways
Solution weaving and nonweaving speeds occurs even though there is no

constraint present and reflects the short length of the section and
Step 1—Establish Roadway and Traffic Conditions the fact that most weaving vehicles will be accelerating or deceler-
All prevailing traffic and roadway conditions are specified in ating within the confines of the section.
the calculation description and in Figure 4-10. Note that this is a Step 6—Compute Average (Space Mean) Speed and
Type A configuration because both weaving movements are re- Density of All Vehicles in Weaving Area
quired to make one lane change.
The level of service is found by computing the average (space
Step 2—Convert All Traffic Volumes to Peak Flow Rates mean) speed and then estimating the average density in the section:
vw + vnw
No conversions of stated traffic demands are required, because S=
vw Vnw
they are given in terms of peak flow rates under ideal conditions, +
expressed in passenger cars per hour. Sw Snw
900 + 4,100
Step 3—Construct Weaving Diagram S= = 55.5 mph
900 4,100
+
The weaving diagram is shown in Figure 4-10. Critical ratios 48.1 57.5
may be computed as follows:
v/N
vw = 600 + 300 = 900 pcph D=
S
v = 900 + 4,000 + 100 = 5,000 pcph
VR = 900/5,000 = 0.18 5,000/4
D= = 22.5 pc/hr/ln
R = 300/900 = 0.33 55.5
Step 4—Compute Unconstrained Weaving and From the density criteria of Table 4-6, this is LOS C.
Nonweaving Speeds
Weaving intensity factors are computed from Table 4-3 for as-
sumed unconstrained conditions on a Type A weaving section: None of the limitations indicated in Table 4-5 have been vio-
lated, and the results seem to be appropriate.
a (1 + VR)b (v/N)c
W=
Ld
CALCULATION 3—CONSTRAINED OPERATION
0.226 (1 + 0.18)2.2 (5,000/4)1.0
Ww = = 0.8110
1,0000.9 Description
4.0 1.3
0.02 (1 + 0.18) (5,000/4)
Wnw = = 0.4117 The ramp-weave section shown in Figure 4-11 serves the follow-
1,0001.0
ing demand volumes. A–C = 975 veh/hr; A–D = 650 veh/hr; B–
Then, on the basis of a free-flow speed, SFF, of 70 mph, the C = 520 veh/hr; B–D = 0 veh/hr. Traffic includes 15 percent trucks,
weaving and nonweaving vehicle speeds can be estimated: is composed of daily commuters, and the PHF is 0.85. Twelve-
foot lanes are provided with no lateral obstructions, and the section
SFF − 10 is located in generally rolling terrain. Through field evaluation,
Si = 15 +
1+W the free-flow speed is observed to be 65 mph. What is the expected
level of service for the section?
70 − 10
Sw = 15 + = 48.1 mph
1 + 0.8110
Solution
70 − 10
Snw = 15 + = 57.5 mph
1 + 0.4117 Step 1—Establish Roadway and Traffic Conditions
Step 5—Check for Constrained Operation All roadway and traffic conditions are specified in the calcula-
tion description and Figure 4-11. Note that this is a Type A config-
The assumption of unconstrained operation is checked using the uration because both Movements A–D and B–C require one lane
equations and criteria of Table 4-4: change.
2.19N VR0.571L 0.234
H
Nw =
S 0.438
w
2.19(4)(0.180.571)(100.234)
Nw = = 1.1 lanes
48.10.438
Because this is less than the maximum value of 1.4 lanes, the
SFF = 65 mph
section may be assumed to be unconstrained. The speeds computed
in Step 4 stand without modification. The large difference between Figure 4-11. Weaving area for Calculation 3.
weaving areas 4-15
Step 2—Convert All Traffic Volumes to Peak Flow Rates The estimated speed of weaving and nonweaving vehicles (as-
Under Ideal Conditions suming unconstrained operation) may now be computed as
follows:
The given demand volumes must be converted to peak flow
rates under ideal conditions expressed in passenger cars per hour: SFF − 10
Si = 15 +
1+W
V
v= 65 − 10
PHF × fHV × fp Sw = 15 + = 40.0 mph
1 + 1.292
where
65 − 10
Snw = 15 + = 42.1 mph
PHF = 0.85 (given); 1 + 1.028
ET = 3 (Table 3-3);
fHV = 0.77 = 1/[1 + 0.15(3 − 1)]; and Step 5—Check for Constrained Operation
fp = 1.00.
The assumption of unconstrained operation is now checked us-
Then ing the equations and limits of Table 4-4:
2.19N VR 0.571L 0.234
H
A–C = 975/(0.85 × 0.77 × 1.00) = 1,490 pcph Nw =
A–D = 650/(0.85 × 0.77 × 1.00) = 993 pcph S 0.438
w
B–C = 520/(0.85 × 0.77 × 1.00) = 794 pcph

2.19(3)(0.550.571)(100.234)
B–D = 0 pcph Nw = = 1.6 lanes
40.00.438
As this is greater than Nw(max) of 1.4 lanes for a Type A weaving

Step 3—Construct Weaving Diagram section, the section operates in a constrained mode. The weaving
intensity factors and speeds must therefore be recomputed for the
A weaving diagram for the calculation is now constructed using
constrained case:
the converted flow rates of Step 2:
0.28 (1 + 0.55)2.2 (3,277/3)1.0
Ww = = 1.600
1,0000.9
0.02 (1 + 0.55)4.0 (3,277/3)0.88
Wnw = = 0.863
1,0000.6
and
65 − 10
Sw = 15 + = 36.2 mph
1 + 1.600
65 − 10
Snw = 15 + = 44.5 mph
1 + 0.863
Critical ratios may be computed as follows:
Step 6—Compute Average (Space Mean) Speed and
vw = 993 + 794 = 1,787 pcph Density of All Vehicles in Weaving Area
v = 1,787 + 1,490 = 3,277 pcph
The level of service is found by computing the average (space
VR = 1,787/3,277 = 0.55
mean) speed in the weaving area and then the density. The resultant
R = 794/1,787 = 0.44
density is compared with the criteria of Table 4-6:
vw + vnw
Step 4—Compute Unconstrained Weaving and S=
Nonweaving Speeds vw vnw
+
Sw Snw
From Table 4-3, the weaving intensity factors, W, for assumed
unconstrained flow for a Type A weaving section are 1,787 + 1,490
S= = 39.6 mph
1,787 1,490
+
a (1 + VR)b (v/N)c 36.2 44.5
W=
Ld v/N
D=
0.226(1 + 0.55)2.2 (3,277/3)1.0 S
Ww = = 1.292
1,0000.9 3,277/3
D= = 27.6 pc/hr/ln
4.0 1.3 39.6
0.02 (1 + 0.55) (3,277/3)
Wnw = = 1.028
1,0001.0 From Table 4-6, this is LOS C, though barely.
4-16 freeways
All required roadway and traffic conditions are specified in the

description.
Step 2—Convert All Traffic Volumes to Peak Flow Rates

No conversions are required because all demands are stated as

Figure 4-12. Weaving area for Calculation 4. peak flow rates under ideal conditions, in passenger cars per hour.
Note: Free-flow speed for a design case is generally estimated
from the design speed or speed limit. Step 3—Construct Weaving Diagram
A weaving diagram and critical ratios are shown in Figure 4-12.

In consulting the limiting values of Table 4-5, it is seen that the Step 4—Compute Unconstrained Weaving and
volume ratio (VR) of 0.55 exceeds the maximum value of 0.45 in Nonweaving Speeds
the table. This means that it is likely that the operation is worse
than that indicated by this analysis, although a breakdown is not The unconstrained weaving and nonweaving speeds are ex-
necessarily going to occur. A possible solution is to add one lane pressed for the Type C configuration by using equations from
to the exit ramp, creating a Type 8 configuration that can better Table 4-3:
serve high volume ratios.
a (1 + VR)b (v/N)c
W=
Ld
CALCULATION 4—DESIGN APPLICATION
0.100 (1 + 0.385)1.8 (6,500/5)0.8
Ww = = 1.438
Description 1,5000.50
0.015 (1 + 0.385)1.8 (6,500/5)1.1
A weaving area is being considered at a major junction between Wnw = = 1.853
1,5000.5
two urban freeways. The configuration of entry and exit roadways
is expected to be as shown in Figure 4-12, which also shows the and
expected demand flow rates, expressed as peak flow rates under
ideal conditions in passenger cars per hour. Design constraints SFF − 10
limit the section length to a maximum of 1,500 ft. LOS C design Si = 15 +
1+W
is desired for the section.
70 − 10
Sw = 15 + = 39.6 mph
1 + 1.438
Solution
70 − 10
Snw = 15 + = 36.1 mph
Design of weaving areas is best achieved by trial-and-error anal- 1 + 1.853
ysis of likely design scenarios. Because the length of the section
is limited to 1,500 ft, trial designs will start with this assumed
length. Given the anticipated design of entry and exit roadways,
the most obvious design would be a five-lane section as shown
Using these estimates, the type of operation is checked using
below:
equations and values given in Table 4-4:
Nw = N [0.761 − 0.011LH − 0.005(Snw − Sw) + 0.047VR]

Nw = 5 [0.761 − 0.011(15) − 0.005(36.1 − 39.6) + 0.047(0.385)]
Nw = 3.2 lanes > 3.00 lanes
Because the number of lanes required by weaving vehicles for

unconstrained operation is greater than the maximum number of
lanes that can be achieved in a Type C configuration, the operation
will be constrained, and the speeds must be recomputed.
This design results from simply connecting each of the five
entry lanes with the five exit lanes. Note that the resulting configu- 0.100 (1 + 0.385)2.0 (6,500/5)0.85
ration is Type C, because Movement B–C may be made without Ww = = 2.200
1,5000.50
lane changing, whereas Movement A–D requires a minimum of
two lane changes. The resulting section is now analyzed for the 0.013 (1 + 0.385)1.6 (6,500/5)1.0
Wnw = = 0.735
anticipated level of service. 1,5000.5
weaving areas 4-17
and 0.100 (1 + 0.385)1.2 (6,500/5)0.77

Ww = = 0.954
1,5000.50
70 − 10
Sw = 15 + = 33.8 mph
1 + 2.200 0.02 (1 + 0.385)2.0 (6,500/5)1.42
Wnw = = 0.974
1,5000.95
70 − 10
Snw = 15 + = 49.6 mph
1 + 0.735 and
70 − 10
Step 6—Compute Average (Space Mean) Speed and Sw = 15 + = 45.7 mph
1 + 0.954
Density of All Vehicles in Weaving Area
70 − 10
The level of service is computed by finding the average (space Snw = 15 + = 45.4 mph
1 + 0.974
mean) speed and density in the weaving area. The resulting density
is compared with the criteria in Table 4-6: Step 5: The type of operation is now checked using equations
and values from Table 4-4:
vw + vnw
S= Nw = N [0.085 + 0.703VR + (234.8/L) − 0.018(Snw − Sw)]
vw vnw
+
Sw Snw Nw = 5 [0.085 + 0.703(0.385) + (234.8/1,500)
2,500 + 4,000 − 0.018 (45.4 – 45.7)]

S= = 42.2 mph
2,500 4,000 Nw = 2.6 lanes < 3.50 lanes
+
33.8 49.6
The operation is therefore unconstrained.
v/N Step 6: None of the limiting values of Table 4-5 are exceeded
D= by the trial design.
S
Step 7: The level of service is found by computing the average
6,500/5 (space mean) speed and density in the weaving area. The density
D= = 30.8 pc/hr/ln
42.2 is compared with the criteria of Table 4-6:
From Table 4-6, this is LOS D. This does not meet the design 2,500 + 4,000
S= = 45.5 mph
objective of LOS C. Further, the dramatic difference in speed 2,500 4,000
+
of weaving and nonweaving vehicles suggests that this design is 45.7 45.4
inappropriate for the demand shown. In particular, it is inappropri-
ate to have 1,000 vehicles in the minor weaving direction making 6,500/5
D= = 28.6 pc/hr/ln
two lane changes in 1,500 ft. 45.5
Since the length of the section is limited to 1,500 ft (problem From Table 4-6, this is still LOS D, just missing the maximum of
statement), and it is difficult to see how lanes could be added 28 pc/ml/ln for LOS C, the target level of service. Given the
within the section, a change of configuration seems to be in order. limitation on length, however, there are no practical alternatives
A Type B configuration can be created by adding one lane to exit that would improve operations. The design level of service is
leg D. This also solves the principal problem in the section, because almost achieved.
the 1,000 veh/hr in the minor weaving direction must now only This calculation illustrates, however, the significant improve-
make a single lane change. The resulting section is shown below: ment in operation that can be achieved by providing an appropriate
configuration. In both trial designs, the demand on and length and
width of the weaving section were exactly the same. The provision
of a Type B configuration, however, resulted in better balance
between weaving and nonweaving vehicles and improved density.
CALCULATION 5—MULTIPLE WEAVING AREA
Description
Figure 4-13 shows a multiple weaving area. Peak flow rates for
the sections are
This revised trial design may now be analyzed using the proce- A–X = 900 pcph
dures of this chapter.
B–X = 400 pcph
Step 1: All roadway and traffic conditions have been stated. A–Y = 1,000 pcph
Step 2: All flows are expressed in peak flow rates under ideal
conditions, in passenger cars per hour. B–Y = 200 pcph
Step 3: Figure 4-12 includes a weaving diagram.
C–X = 300 pcph
Step 4: Speed equations are now selected from Table 4-3 for
unconstrained operation on a Type B configuration: C–Y = 100 pcph
4-18 freeways
All geometric conditions are ideal, and the terrain is generally and
level. At what level of service would the section operate?
SFF − 10
SFF = 60 mph (field measurement) Snw = 15 +
1 + 0.02(1 + VR)2.0(v/N)1.42/L0.95
60 − 10
Snw = 15 +
1 + 0.02(1 + 0.56)2.0(2,500/3)1.42/1,0000.95
Snw = 40.4 mph
The number of lanes required by weaving vehicles for uncon-
strained operation is computed using the appropriate equation from
Table 4-4 and compared with the maximum value of 3.5 lanes,
Figure 4-13. Weaving area for Calculation 5.
also obtained from Table 4-4, for Type B configurations:
Solution Nw = N [0.085 + 0.703 VR + (234.8/L) − 0.018(Snw − Sw)]

Nw = 3 [0.085 + 0.703(0.56) + (234.8/1,000)
A multiple weaving section is analyzed as two separate simple
weaving areas. The initial step in the analysis is to construct weav- − 0.018(40.4 – 40.5)]
ing diagrams for the two subsegments of the multiple weaving
area. Because all demands are stated in peak flow rates under ideal Nw = 2.2 lanes < 3.5 lanes
conditions and no conversion computations are required, this is The section is therefore unconstrained.
done immediately. The weaving area under study is of the type None of the limitations of Table 4-5 are exceeded. The harmonic
illustrated in Figure 4-8, that is, two merge areas followed closely average speed and density are computed to find level of service:
by a diverge area. Weaving diagrams are constructed in accordance
with Figure 4-8, as follows: 1,400 + 1,100
S= = 40.5 mph
1,400 1,100
+
40.5 40.4
2,500/3
D= = 20.6 pc/hr/ln
40.5
From Table 4-6, this is LOS C, just missing the LOS B boundary.
Segment 2
Using the same equations as for Segment 1, because both are

Type B configurations,
60 − 10
Sw = 15 +
1 + 0.10(1 + 0.517)1.2(2,900/3)0.77/1,5000.5
Sw = 42.1 mph
Note that both segments of the weaving area are Type B configu-
rations. In Segment 1, Movement A–Y may be made with no lane 60 − 10
Snw = 15 +
changes, whereas Movement B–X requires one lane change. In 1 + 0.02(1 + 0.517)2.0(2,900/3)1.42/1,5000.95
Segment 2, Movements A–Y and B–Y may be made with no lane
Snw = 43.3 mph
changes, but Movement C–X requires a single lane change.
Computations for speed are now done for each segment. Note The number of lanes required by weaving vehicles is
that the first three steps of the procedure have been completed in
the establishment of weaving diagrams for the two segments. Nw = 3 [0.085 + 0.703(0.517) + (234.8/1,500)
− 0.018(43.3 – 42.1)]
Segment 1
Nw = 1.8 lanes < 3.5 lanes
Unconstrained speed predictions are obtained from Table 4-3
for a Type B configuration. To save space, the equation for the Operation is unconstrained.
weaving intensity factor, W, is inserted directly into the speed None of the limitations of Table 4-5 are exceeded. The harmonic
prediction equation: average speed and density are computed to find level of service:
SFF − 10 1,500 + 1,400
Sw = 15 + S= = 42.7 mph
1 + 0.10(1 + VR)1.2(v/N)0.77/L0.5 1,500 1,400
+
60 − 10 42.1 43.3
Sw = 15 +
1 + 0.10(1 + 0.56)1.2(2,500/3)0.77/1,0000.5 2,900/3
D= = 22.6 pc/hr/ln
Sw = 40.5 mph 42.7
weaving areas 4-19
Table 4-7. Results of Weaving Analysis: Sample Calculation 6

type a configurations type b configurations type c configurations
no. of length S, D, pc/ cons. S, D, pc/ cons. S, D, pc/ cons.
lanes (ft) mph mi/ln los ? mph mi/ln los ? mph mi/ln los ?
3 500 36.7 38.2 E Yes 32.0 43.8 F No 29.3 47.8 F No
750 40.9 34.3 D Yes 36.1 38.8 E No 31.6 44.3 F No
1,000 44.0 31.8 D Yes 39.2 35.7 E No 33.4 42.0 E No
1,500 48.5 28.8 D Yes 43.9 31.9 D No 36.0 38.9 E No
2,000 51.7 27.1 C Yes 47.2 29.7 D No 38.0 36.8 E No
2,500 — — — — 49.7 28.2 D No 39.6 35.3 E No
4 500 40.3 26.0 C Yes 36.3 28.9 D No 32.6 32.2 D No
750 44.8 23.4 C Yes 40.9 25.7 C No 35.2 29.9 D No
1,000 48.0 21.9 C Yes 44.2 23.8 C No 37.2 28.3 D No
1,500 52.4 20.0 B Yes 48.8 21.5 C No 40.0 26.2 C No
2,000 55.4 19.0 B Yes 52.0 20.2 C No 42.2 24.9 C No
2,500 — — — — 54.3 19.4 B No 43.8 24.0 C No
5 500 43.4 19.4 B Yes 40.1 21.0 C No 35.4 23.7 C No
750 47.9 17.6 B Yes 44.8 18.8 B No 38.2 21.9 C No
1,000 51.0 16.5 B Yes 48.1 17.5 B No 40.3 20.9 C No
1,500 55.3 15.2 B Yes 52.5 16.0 B No 47.0 17.8 C Yes
2,000 58.0 14.5 B Yes 55.4 15.2 B No 49.1 17.1 C Yes
2,500 — — — — 57.5 14.6 B No 50.6 16.6 C Yes
From Table 4-6, this is LOS C. The analysis shows that the section conditions covering three, four, or five lanes; lengths from 500 to
will operate at speeds in the 40- to 43-mph range, producing densi- 2,500 ft; and all three types of configuration. Although this is a
ties that are in the upper portion of LOS C. time-consuming process, it is easily set up on a programmable
calculator, microcomputer spreadsheet, or any type of computer.
The results of such computations are shown in Table 4-7.
CALCULATION 6—SENSITIVITY ANALYSIS WITH DESIGN
A number of points should be made concerning these results
APPLICATION
and their impact on a final design decision:
1. Before all the potential solutions that yield LOS C are exam-
Description
ined, the configuration of entry and exit legs should be considered.
To provide for LOS C on each of the entry and exit legs, using
A major interchange is to be built to join two major freeways
the criteria for basic freeway sections, each leg would have the
in a suburban area. The issue of handling some of the interchanging
following number of lanes, based on a service flow rate of 1,644
movements in a weaving section is to be investigated. The flows
pcphpl:
in question are shown below, and are given in terms of flow rates
in passenger cars per hour under ideal conditions. Leg Volume (pcph) No. of Lanes for LOS C
A 2,200 2
B 2,000 2
C 2,500 2
D 1,700 2
With this number of entry and exit legs, all five-lane solutions are
impractical and are therefore eliminated from serious consider-
ation. Five lanes is excessive, given that four lanes is adequate to
handle all input and output flows.
2. If the lanes from the above legs are simply connected, a four-
lane Type A configuration results. All of the four-lane Type A
configurations produce LOS C or better. On the other hand, they
Because the interchange joins two future facilities, there is sub-
are also all constrained, indicating a serious imbalance between
stantial flexibility in both the length and width that may be consid-
weaving and nonweaving flows. Further, for four- and five-lane
ered for the section. LOS C operation is desired.
configurations, the VR of 0.405 for this problem exceeds the maxi-
mum indicated in Table 4-5. Thus, the results may be misleading,
Solution and poorer operating conditions would likely result.
3. There is no easy way to create a Type C configuration
Since the length, width, and configuration to be used are open given the number of lanes on entry and exit legs. A Type B
to question, as is the issue of whether or not to use a weaving configuration could be created by adding a lane to Leg C. If
section, many trial computations must be made. Speeds can be this were done, a four-lane section with a length between 750
computed for weaving and nonweaving vehicles for a range of and 2,500 ft could be provided, because all would result in
4-20 freeways
LOS C or better. No length provides LOS C in a three-lane geometry. The additional lane on Leg C would most likely be
design. The final design result would be in this range, with the dropped at some downstream point, because it is not needed to
exact length being determined by economics and specifics of provide for LOS C on that leg.
V. REFERENCES
1. Pignataro, L., et al. NCHRP Report 159: Weaving Areas— 5. Reilly, W., et al. Weaving Analysis Procedures for the New
Design and Analysis. Transportation Research Board, Washing- Highway Capacity Manual. Technical Report, JHK & Associ-
ton, D.C. (1975). ates, Tucson, Ariz. (1983).
2. Roess, R.P., et al. Freeway Capacity Analysis Procedures. 6. Schoen, J., et al. Speed-Flow Relationships for Basic Free-
Final Report, Project No. DOT-FH-11-9336, Polytechnic Uni- way Sections. Final Report, NCHRP Project 3-45, Catalina En-
versity, Brooklyn, N.Y. (1978). gineering, Tucson, Ariz. (1995).
3. Roess, R.P. ‘‘Development of Weaving Procedures for the 1985 7. Roess, R., and Ulerio, J. Capacity of Ramp-Freeway Junc-
Highway Capacity Manual.’’ Transportation Research Record tions. Final Report, NCHRP Project 3-37, Polytechnic Univer-
1112, Transportation Research Board, Washington, D.C. sity, Brooklyn, N.Y. (1993).
(1988). 8. Roess, R.P., McShane, W.R., and Prassas, E.S. Traffic Engi-
4. Leisch, J. Completion of Procedures for Analysis and Design neering, 2nd ed. Prentice-Hall, Simon & Schuster, Salt Lake
of Traffic Weaving Areas. Final Report, Vols. 1 and 2, Federal City, Utah (Jan. 1998).
Highway Administration, Washington, D.C. (1983).

chapter 5
RAMPS AND RAMP JUNCTIONS
CONTENTS
i. introduction .......................................................................................................................................................................... 5-1

Ramp Components ................................................................................................................................................................ 5-2
Operational Characteristics.................................................................................................................................................... 5-2
Length of Acceleration and Deceleration Lanes .................................................................................................................. 5-3
ii. methodology.......................................................................................................................................................................... 5-3
Prediction of Flow Entering Lanes 1 and 2 (V12) ................................................................................................................ 5-3
General Model Structure ................................................................................................................................................... 5-4
Specific Models ................................................................................................................................................................. 5-4
Capacity Values..................................................................................................................................................................... 5-4
Merge Areas ...................................................................................................................................................................... 5-4
Diverge Areas .................................................................................................................................................................... 5-7
Level-of-Service Criteria....................................................................................................................................................... 5-7
Prediction of Density......................................................................................................................................................... 5-8
Prediction of Speed ........................................................................................................................................................... 5-8
Single-Lane On- and Off-Ramps .......................................................................................................................................... 5-9
Special Applications .............................................................................................................................................................. 5-9
Two-Lane On-Ramps ........................................................................................................................................................ 5-9
Two-Lane Off-Ramps........................................................................................................................................................ 5-11
Lane Additions and Lane Drops....................................................................................................................................... 5-11
Effects of Ramp Control ................................................................................................................................................... 5-12
Ramps on 10-Lane Freeway Sections (5 Lanes in Each Direction) ............................................................................... 5-12
Left-Hand Ramps .............................................................................................................................................................. 5-12
Effects of Ramp Geometry ............................................................................................................................................... 5-12
Major Merge Sites............................................................................................................................................................. 5-12
Major Diverge Sites .......................................................................................................................................................... 5-13
Capacity of Ramp Roadways................................................................................................................................................ 5-14
Calculation 1: Isolated On-Ramp.......................................................................................................................................... 5-14
Calculation 2: Consecutive Off-Ramps on Six-Lane Freeway............................................................................................ 5-16
Calculation 3: On-Ramp–Off-Ramp Pair on Eight-Lane Freeway...................................................................................... 5-19
Calculation 4: Two-Lane On-Ramp...................................................................................................................................... 5-23
Calculation 5: Off-Ramp on 10-Lane Freeway.................................................................................................................... 5-23
Calculation 6: Left-Side On-Ramp ....................................................................................................................................... 5-24
v. references .............................................................................................................................................................................. 5-27
I. INTRODUCTION
A ramp may be described as a length of roadway providing an mately analyze ramp junctions on nonfreeway facilities, such as
exclusive connection between two highway facilities. Analysis of expressways, multilane highways, and two-lane highways, pro-
ramp-freeway junctions is the focus of this chapter, and some vided that the junctions involve merging or diverging movements
material on ramp roadways is provided. Ramp-freeway junction not controlled by traffic signals or stop or yield signs. For ramp-
analysis procedures presented herein may be applied to approxi- street junctions controlled by such devices, the procedures of

5-2 freeways
Chapter 9, Signalized Intersections, or Chapter 10, Unsignalized upstream freeway demand is a composite of upstream trip genera-
Intersections, should be applied. tion patterns from a variety of sources.
In the merge area, individual on-ramp vehicles attempt to find
gaps in the traffic stream of the adjacent freeway lane. Since most
RAMP COMPONENTS ramps are on the right side of the freeway, the freeway lane in
which on-ramp vehicles seek gaps is the shoulder lane, designated
A ramp may consist of up to three geometric elements of herein as Lane 1. In this chapter, lanes are numbered 1 to N from
interest: the shoulder to the median.
The action of individual merging vehicles entering the Lane 1
1. The ramp-freeway junction,
traffic stream creates turbulence in the traffic stream in the vicinity
2. The ramp roadway, and
of the ramp. Approaching freeway vehicles move toward the left
3. The ramp-street junction.
to avoid this turbulence. Recent studies (2) have shown that the
A ramp-freeway junction is generally designed to permit high- operational effect of merging vehicles is heaviest in freeway Lanes
speed merging or diverging to take place with a minimum of 1 and 2 and the acceleration lane for a distance extending from
disruption to the adjacent freeway traffic stream. The geometric the physical merge point to 1,500 ft downstream. Figure 5-1 shows
characteristics of ramp-freeway junctions vary. Elements such as the ‘‘influence area’’ for on-ramp junctions. Models presented in
the length and type (taper, parallel) of acceleration or deceleration this chapter focus on operational characteristics within this defined
lane, free-flow speed of the ramp in the immediate vicinity of the influence area.
junction, and sight distances may all influence ramp operations. Interactions are dynamic. Approaching freeway vehicles will
The procedures in this chapter are primarily applicable to high- move left as long as there is capacity to do so. Whereas the intensity
type designs. Nevertheless, some of the models used account ex- of ramp flow generally influences the behavior of freeway vehicles,
plicitly for the effect of acceleration or deceleration lane length general freeway congestion can also limit ramp flow, causing di-
and the free-flow speed of the ramp and can therefore be applied version to other interchanges or routes.
to a range of geometric designs, including some that might be At off-ramps the basic maneuver is a diverge, that is, a single
considered substandard. Geometric design standards for ramps and traffic stream separating into two separate streams. Exiting vehi-
ramp junctions are given by AASHTO (1). cles must occupy the lane adjacent to the off-ramp, Lane 1 for a
Geometric characteristics of ramp roadways also vary from loca- right-hand off-ramp. Thus, as the off-ramp is approached, exiting
tion to location. Ramps may vary in terms of number of lanes vehicles move right. This movement brings about a redistribution
(usually one or two), design speed, grade, and horizontal curvature. of other freeway vehicles, which move left to avoid the turbulence
The design of a ramp roadway is seldom a source of operational of the immediate diverge area. Again, recent studies (2) show that
difficulty unless a traffic incident causes disruption along its the area of most intense turbulence is the deceleration lane plus
length. Ramp-street terminal problems can cause queueing along Lanes 1 and 2 over 1,500 ft extending upstream from the physical
the length of a ramp, but this queueing is generally not related to diverge point (Figure 5-1).
the design of the ramp roadway. Procedures in this chapter treat both ramp and freeway flow
Freeway-to-freeway ramps have two ramp-freeway terminals rates as inputs to an operational analysis of the merge or diverge
and do not have a ramp-street terminal. Many ramps, however, influence area. Thus, design and planning applications become
connect limited-access facilities to local arterials and collectors. trial-and-error computations using the operational analysis tech-
For such ramps, the ramp-street terminal is often a critical element niques as specified. This procedure is logical, because the ramp
in the overall design. Ramp-street junctions can permit uncon- is a point location on an overall facility for which flows are either
trolled merging and diverging movements or take the form of an known or specified.
at-grade intersection. The procedures in this chapter assume that the behavior of merg-
Procedures in this chapter allow for the identification of likely ing or diverging vehicles is unaffected by downstream or upstream
breakdowns at ramp-freeway terminals [Level-of-Service (LOS) constrictions or disruptions. Downstream problems, for example,
F] and for the analysis of operations at ramp-freeway junctions can easily propagate upstream through a merge or diverge area. In
and on ramp roadways at LOS A through E. For analysis of ramp- such cases operations reflect the characteristics of the downstream
street junctions involving an at-grade intersection, consult Chapter
9, Signalized Intersections, or Chapter 10, Unsignalized
Intersections.
Sections addressing special applications, including metered
ramps, ramps on five-lane (one-direction) freeway sections, two-
lane ramps, major merge areas, and major diverge areas, are con-
tained in this chapter.
OPERATIONAL CHARACTERISTICS
A ramp-freeway junction is an area of competing traffic de-

mands for space. Upstream freeway traffic competes for space
with entering on-ramp vehicles in merge areas. On-ramp demand
is usually generated locally, although arterials and collectors may
bring some drivers to the ramp from more distant origins. The Figure 5-1. On- and off-ramp influence areas.
ramps and ramp junctions 5-3
breakdown and would not be expected to conform to the models deceleration (LD) lane. The length of such lanes is measured
presented herein. from the point at which the left edge of the ramp lanes and
the right edge of the freeway lanes converge to the end of the
taper segment connecting the ramp to the freeway. The point
LENGTH OF ACCELERATION AND DECELERATION LANES
of convergence is typically defined by painted markings or
A critical geometric parameter influencing operations at a physical barriers, or both.
merge or diverge area is the length of the acceleration (LA) or
II. METHODOLOGY
As shown in Figure 5-1, the basic approach to the modeling of must be converted to equivalent passenger cars per hour under ideal
merge and diverge areas focuses on an influence area of 1,500 ft conditions for the peak 15 min of the hour of interest. Equation 5-
including the acceleration or deceleration lane and Lanes 1 and 2 1 is used to convert any hourly flow rate in vehicles per hour to
of the freeway. the desired format:
The methodology has three major steps:
Vveh/hr
1. The flow entering Lanes 1 and 2 immediately upstream of Vpcph = (5-1)
PHF fHV fp
the merge influence area or the beginning of the deceleration lane
at an off-ramp is determined. This flow is designated V12. It must where
be known, because it is a major determinant of operating character-
Vpcph = maximum 15-min flow rate in passenger cars per hour
istics within the ramp influence area.
(pcph) under ideal conditions,
2. Critical capacity values are determined, and demand flows
Vveh/hr = hourly volume in vehicles per hour (veh/hr) under prevail-
are compared with these values. The comparison determines
ing conditions,
whether the merge or diverge area is likely to break down. Capacity
is evaluated at two points: (a) the maximum total flow departing
fHV = adjustment factor for heavy vehicles, and
from the merge or diverge area (VFO for on-ramps; VFO + VR for
fp = adjustment factor for driver population.
off-ramps) and (b) the maximum total flow that can reasonably
enter the merge or diverge influence area (VR12 for on-ramps; V12 The following sections detail the three steps of the ramp-freeway
for off-ramps). If demand exceeds either of these two capacity junction operational model.
values, breakdown is likely.
3. The density within the merge or diverge influence area (DR)
and the level of service based on this value are determined. For PREDICTION OF FLOW ENTERING LANES 1 AND 2 (V12)
some situations, the average speed of vehicles within the merge
or diverge area (SR) may also be predicted. Studies have shown that the principal influences on lane distri-
bution of freeway vehicles immediately upstream of merge or
Figure 5-2 shows these key variables and their relationship to
diverge areas are the following:
each other. All aspects of the model and LOS criteria are expressed
in equivalent maximum flow rates in passenger cars per hour for T VF, total freeway flow approaching the merge or diverge area
the peak 15 min of the hour of interest. Therefore, before any of (pcph);
these procedures are applied, all relevant freeway and ramp flows T VR, total ramp flow (pcph);
T LA or LD, total length of the acceleration or deceleration lane
(ft); and
T SFR, free-flow speed of ramp at point of merge or diverge
(mph).
Of these, total freeway flow is easily the dominant factor. Models
are structured to account for this phenomenon without distorting
other relationships. Total ramp flow plays a major role in lane
distribution immediately upstream of off-ramps, because all of the
ramp traffic must be in Lane 1 to access the ramp. For on-ramps,
this parameter has surprisingly little influence on flow entering
Lanes 1 and 2.
The length of the acceleration or deceleration lane also influ-
ences lane distribution. In merge areas, longer acceleration lanes
contribute to lower turbulence levels and lower densities in the
merge influence area. Thus, approaching freeway vehicles are less
likely to move left to avoid the turbulence, and V12 tends to in-
Figure 5-2. Critical ramp junction values. crease. The influence of deceleration lane length is less pronounced
5-4 freeways
in diverge areas. Higher ramp free-flow speeds tend to push drivers must traverse the ramp influence area, because there are no lanes
further left to avoid high-speed merging or diverging. that avoid it.
Lane distribution at a given ramp may also be influenced by The form of each equation in Figures 5-3 and 5-4 is indicative
flows on adjacent upstream and downstream ramps. When nearby of causal interactions among operational and geometric variables
ramps inject vehicles into or remove them from Lane 1, the lane in merge and diverge areas. Equation 2 (Figure 5-3), the general
distribution of total vehicles may be seriously altered. Several equation for six-lane freeways covering single-lane on-ramps, is
variables are critical: quite simple. It suggests that the only variable affecting the propor-
tion of flow remaining in Lanes 1 and 2 immediately upstream of
T VU, total flow on an upstream adjacent ramp (pcph),
the merge is the length of the acceleration lane. By reducing merge
T VD, total flow on a downstream adjacent ramp (pcph),
turbulence, a longer acceleration lane allows more freeway vehi-
T DU, distance to the adjacent upstream ramp (ft), and
cles to remain in Lanes 1 and 2.
T DD, distance to the adjacent downstream ramp (ft).
Equations 3 and 4 (Figure 5-3) deal with merging on six-lane
Whether upstream or downstream adjacent ramps have a signifi- freeways but take into account the effect of upstream adjacent off-
cant influence on lane distribution depends on the size of the ramps and downstream adjacent off-ramps on the subject ramp.
freeway, the specific combination of upstream or downstream These equations should be used only when all variables fall within
ramp (or both), and the distances and flows involved. the limits shown in Figure 5-3. When input variables fall outside
these limits, the general equation for six-lane freeways, Equation
2, should be applied. The general equation is also applied where
General Model Structure upstream or downstream adjacent on-ramps exist; there is no recent
evidence that they affect behavior at the on-ramp in question.
The model form for prediction of V12 immediately upstream of Equation 5 (Figure 5-3) is used for all single-lane right-hand on-
single-lane, right-hand on-ramps is ramps on eight-lane freeways. No separate equations are applied
to account for upstream and downstream adjacent ramp effects.
V12 = VF × PFM (5-2)
Equation 5 indicates that higher ramp flows have a negative impact
where PFM is the proportion of freeway vehicles remaining in Lanes on V12, whereas the proportion of traffic remaining in Lanes 1 and
1 and 2 immediately upstream of an on-ramp and V12 and VF are 2 increases with increasing length of acceleration lane (as in the
as previously defined. case of six-lane freeways) and decreasing free-flow speed of the
This form allows the model to retain the importance of total ramp. The latter suggests that ramp vehicles entering the freeway
freeway flow in determining flow in Lanes 1 and 2, and PFM at higher speeds cause more approaching freeway vehicles to move
expresses the behavioral choices of drivers selecting lanes. In es- out of Lanes 1 and 2.
sence, the model focuses on predicting the proportion of vehicles Equation 7 (Figure 5-4) is the general diverge equation for six-
in Lanes 1 and 2 and applies this to the freeway flow, which is lane freeways. Equation 8 applies to six-lane freeway off-ramps
known or designated. where an upstream adjacent on-ramp is present, whereas Equation
The model for single-lane, right-hand off-ramps must take a 9 is used where a downstream adjacent off-ramp is present. These
different form. V12 for off-ramps is defined immediately upstream equations should be used only when all variables fall within the
of the beginning of the deceleration lane. Thus, V12 must include ranges indicated in Figure 5-4. When they do not, the general
VR, the off-ramp flow. The real issue is the proportion of through Equation 7 should be used. Equation 7 is also used for six-lane
vehicles remaining in Lanes 1 and 2 at this point. A model express- freeway off-ramps where upstream adjacent off-ramps or down-
ing this logic is stream adjacent on-ramps exist. They do not have any significant
influence on off-ramp behavior.
V12 = VR + (VF − VR)PFD (5-3) Equation 10 is used for all single-lane right-hand off-ramps on
This model focuses on predicting the choice to be made by ap- eight-lane freeways. It suggests that the proportion of nonexiting
proaching freeway drivers not exiting at the ramp (i.e., drivers traffic remaining in Lanes 1 and 2 is a constant. Thus, V12 is
with a choice to make). influenced only by VF and VR, which are part of the general
model used.
Specific Models
CAPACITY VALUES
The methodology is based on the results of a National Coopera-

tive Highway Research Program study (2) in which equations for Merge Areas
PFM and PFD were calibrated for different possible configurations,
including width of freeway and upstream and downstream ramp The capacity of merge areas is controlled by either of the follow-
configurations. The data base for the study included 58 sites from ing two criteria: (a) the total flow leaving the merge area on the
seven regions of the United States, each studied for 2 to 4 hr. downstream freeway (VFO) and (b) the maximum flow entering the
Figures 5-3 and 5-4 provide an index to predictive models for merge influence area (VR12).
V12. Figure 5-3 shows the models used in conjunction with single- The total flow leaving the merge area is subject to the constraints
lane right-hand on-ramps and provides a matrix for determining of the downstream freeway section. There is no evidence that
which model applies for a given configuration. Figure 5-4 provides the turbulence of the merge area causes the downstream freeway
similar information for single-lane right-hand off-ramps. capacity to be less than that of a basic freeway segment. Thus, for
Prediction of V12 for four-lane freeways is trivial, since Lanes stable flow operations to exist, the sum of the merging flows
1 and 2 compose the entire freeway in a given direction. Drivers cannot exceed the capacity of the downstream freeway segment.
Figure 5-3. Models for predicting V12 for on-ramps.

5-6 freeways
Figure 5-4. Models for predicting V12 for off-ramps.

Table 5-1. Capacity Values for Merge and Diverge Areas

max flow max flow
entering entering
freeway maximum upstream (VF) or downstream (VFO) freeway flow merge diverge
free-flow (pcph) by no. of lanes in one direction influence influence
speed area (VR12) area (V12)
(mph) 2 3 4 >4 (pcph) (pcph)
70 4,800 7,200 9,600 2,400/ln 4,600 4,400
65 4,700 7,050 9,400 2,350/ln 4,600 4,400
60 4,600 6,900 9,200 2,300/ln 4,600 4,400
55 4,500 6,750 9,000 2,250/ln 4,600 4,400
NOTE: For capacity of ramp roadways, see Table 5-6.
It is possible, however, to experience congestion in the merge ramp-street junction should also be checked using the procedures
influence area even if the capacity of the downstream freeway for signalized intersections (Chapter 9) or those for unsignalized
segment is adequate. Studies (2) have shown that there is a practical intersections (Chapter 10) to ensure that queues will not form and
maximum flow that may enter the merge influence area and still spread upstream on the ramp, affecting traffic operations on the
maintain stable operations. In a ramp merge junction, both the diverge area.
flow in Lanes 1 and 2 and the flow in the on-ramp enter the merge
influence area. Thus,
LEVEL-OF-SERVICE CRITERIA
VR12 = VR + V12
LOS A through E for ramp-freeway terminals are based on the
Table 5-1 shows capacity values for the downstream freeway density in the influence area of the ramp and the expectation that no
flow (VFO) and the merge influence area (VR12). If the demand breakdown will occur. LOS F signifies that a breakdown condition
expected at either point exceeds the capacity values shown, failure, exists or is expected to exist. LOS F occurs whenever demand
or LOS F, is expected to exist. When this is the case, the analysis exceeds the limits indicated in Table 5-1. When none of these
ends, and solutions are sought to alleviate the problem. Where limits is exceeded, no breakdown is expected, and the level of
stable operations are expected (i.e., demand does not exceed capac- service is based on density, as indicated in Table 5-2. Table 5-2
ity at either point), the next step of the analysis—estimation of also gives average speed of vehicles in the ramp influence area as a
density in the merge influence area—is implemented to find the secondary LOS parameter. This is particularly useful in comparing
level of service. these criteria with field data, since density is rarely measured
directly.
Diverge Areas The density values shown for LOS A through E assume stable,
nonbreakdown operations. Studies (2) have shown that there is an
Three capacity values should be checked in a diverge area: (a) the overlap in the density range such that some breakdown operations
total flow that may leave the diverge area, (b) the maximum flow may actually have lower densities than those achieved under stable
that may enter Lanes 1 and 2 immediately before the deceleration operation. This is due to the wavelike movement of vehicles in a
lane, and (c) the capacity of each of the exiting legs of the freeway. queue and the rather short length of the defined ramp influence
The total flow that can leave the diverge area is generally limited area. The model first calls for determination of whether LOS F
by the capacity of the freeway lanes approaching the diverge junc- exists using the maximum flow levels of Table 5-1. Then density
tion. In all appropriate diverge designs, the number of lanes leaving is estimated and the level of service assigned if flow is stable.
the diverge area is either equal to or one greater than the number Except for LOS A, each of the density boundaries is higher than
entering. This departing flow is designated VFO. that of a similar basic freeway section (Chapter 3). This is because
The flow entering Lanes 1 and 2 just upstream of the decelera- (a) drivers expect increased turbulence and greater proximity of
tion lane is simply the flow in Lanes 1 and 2 (V12). This flow other vehicles in a merge or diverge area and (b) drivers are gener-
includes the off-ramp flow. Table 5-1 gives capacity values for ally traveling at somewhat lower speeds at any given per-lane flow
the first two capacity checks. rate in the merge or diverge area than on open freeway.
The third limit is most important because it is the primary reason
for failure of diverge areas. Failure at a diverge is most often Table 5-2. Level-of-Service Criteria for
related to the capacity of one of the exit legs, usually the ramp. Ramp-Freeway Junction Areas of Influence
The capacity of each exit leg must be checked against the expected
maximum density minimum speed
demand. For a downstream freeway leg (at a major diverge area
level of (primary measure) (secondary measure)
there may be two), capacity values may be drawn from Table 5-1 for service (pc/mi/ln) (mph)
the appropriate number of freeway lanes. The capacity of ramp
A 10 58
roadways is discussed later in the chapter. B 20 56
The failure of any of these capacity checks, that is, an expected C 28 52
demand that exceeds the capacities given, indicates that the merge D 35 46
area will fail. In such cases, breakdown and formation of queues E >35 42
a a
are expected to occur. Where an off-ramp terminates at an at-grade F
intersection (either signalized or unsignalized), the capacity of the a
Demand flows exceed limits of Table 5-1.

5-8 freeways
LOS A represents unrestricted operations. Density is low enough Table 5-4. Models for Prediction of Speed in Ramp
to permit merging and diverging maneuvers without disruption to Influence Areas
through vehicles. There is virtually no noticeable turbulence in the item equation or value
ramp influence area, and speeds remain close to the expected basic
Single-Lane On-Ramps, Stable Flow
freeway section level.
At LOS B, merging and diverging maneuvers become noticeable Model SR = SFF − (SFF − 42) MS
1 2
to through drivers, and minimal levels of turbulence exist. Merging LASFR
MS = 0.321 + 0.0039 e(VR12 /1,000) − 0.002
drivers must adjust their speeds to smoothly fill available gaps, as R2
0.60 1,000
do diverging drivers making lane changes within the ramp influ- SE (mph) 2.20
ence area. Speeds of vehicles in the influence area begin to decline Data periods (no.) 132
slightly. Single-Lane Off-Ramps, Stable Flow
At LOS C, average speed within the ramp influence area begins
to decline as the level of merging or diverging turbulence becomes Model SR = SFF − (SFF − 42) DS
DS = 0.883 + 0.00009 VR − 0.013 SFR
noticeable. Both freeway and on-ramp vehicles begin to adjust R2 0.44
their speeds to accommodate smooth merging maneuvers. In di- SE (mph) 2.46
verge areas, vehicles begin to slow to allow lane-changing as off- Data periods (no.) 73
ramp vehicles approach the diverge. Driving conditions are still
relatively comfortable at this level.
At LOS D, turbulence levels become intrusive, and virtually all
vehicles slow to accommodate merging or diverging maneuvers.
Some ramp queues may form at heavily used on-ramps, but free- provide additional lane length over which to disperse the total flow
way operation remains stable. in the influence area.
LOS E represents conditions approaching and reaching capacity The density models of Table 5-3 apply only to cases where no
operation. Speeds reduce to the low 40s (mph), and the turbulence breakdown is occurring or is expected to occur on the basis of
of merging and diverging maneuvers becomes intrusive to all driv- demand flows. Thus, all densities predicted by these models are,
ers in the influence area. Flow levels approach capacity limits, and by definition, in the range of LOS A through E. No models are
small changes in demand or disruptions within the traffic stream available for directly predicting the density of a ramp influence
can cause both ramp and freeway queues to begin forming. area operating under LOS F.
LOS F represents breakdown, or unstable, operation. At this Values of VR, LA, and LD are known inputs. Values of V12 are
level, approaching demand flows exceed the discharge capacity of predicted using the models of Figures 5-3 and 5-4, as previously
the downstream freeway (and ramp, in the case of diverge areas). discussed.
Queues are visibly formed on the freeway and on-ramps and con-
tinue to grow as long as approaching demand exceeds the discharge
capacity of the section. Freeway queues are not the same as inter- Prediction of Speed
section or other stopped queues. Consult Chapter 3 for a more
complete description. Where desired, models are also available for the prediction of
average travel speed (space mean speed) within the ramp influence
area. This may be useful information, but it should not be used as
Prediction of Density
a primary measure of level of service unless density is unavailable.
It is not necessary to estimate the speed of vehicles traversing the
Table 5-3 gives models for prediction of density in merge or
ramp influence area to use this methodology.
diverge influence areas. Independent variables include flows enter-
As in the case of density models, reliable speed predictors are
ing the influence area and length of acceleration or deceleration
not available for unstable flow conditions.
lane. Such lanes have an important effect on density, because they
Table 5-4 gives models for the prediction of average speed of
vehicles within the ramp influence area defined in this chapter.
Speed models are obviously approximate. The R2 values do
Table 5-3. Models for Prediction of Density in Ramp not indicate strong correlations, but the standard errors (SEs) are
Influence Areas reasonable enough for rough estimates of speed. Predicted speeds
item equation or value from these equations should never be used to establish level of
Single-Lane On-Ramp Merge Area service, because the SEs are larger than some of the LOS speed
ranges.
Model DR = 5.475 + 0.00734VR + 0.0078V12 The equations are all based on the concept of maximum and
− 0.00627LA minimum speeds under stable and unstable operation. For stable
R2 0.88
Std. error (pc/mi/ln) 2.68 flow, the maximum speed is the free-flow speed of the freeway
Data periods (no.) 167 (SFF). Because 42 mph has been found to be the dividing line
between stable and unstable flow, it becomes the minimum speed
Single-Lane Off-Ramp Diverge Areas in stable flow models. M and D are merging and diverging intensity
Model DR = 4.252 + 0.0086V12 − 0.009LD factors used to scale the drop from maximum to minimum speed.
R2 0.93 No unstable flow model is presented, but the practical range of
Std. error (pc/mi/ln) 1.75 speeds under LOS F is from a minimum of 10 to 12 mph to a
Data periods (no.) 86 maximum of 42 mph.
SINGLE-LANE ON- AND OFF-RAMPS calibration ranges indicated in Figures 5-3 and 5-4. For all other
cases, the general equations for the size of freeway under consider-
Models for the analysis of single-lane on- and off-ramp termi- ation are used, regardless of whether all variables are within the
nals on freeways were presented and discussed in the previous calibration ranges of these equations. Some caution should be
section. This section provides simple step-by-step procedures for exercised when the general equations are used outside their calibra-
their application. Figure 5-5 shows a worksheet on which the tion ranges. The accuracy of predictions outside these ranges can-
results of such an analysis may be summarized. not be statistically assessed. Nevertheless, given no alternative
As noted, all ramp computations are done in the operational models, they must be used. The user should, however, check the
analysis mode—that is, the geometry and all demand volumes are reasonableness of the results.
specified. The operational analysis determines the likely density Step 4: Find checkpoint flow rates. Once the value of V12 is
in the ramp influence area and therefore the expected level of estimated, it can be combined with known values of VF and VR to
service for the operation specified. Design alternatives are analyzed find the checkpoint flow rates needed to compare with the capacity
through trial-and-error application of this process. Various designs values of Table 5-1.
may be proposed and operational analysis performed to determine For merge areas,
the expected level of service that would result.
VFO = VF + VR
Step 1: Specify geometry and demand volumes. To conduct an
operational analysis, both the geometry and demand volumes must VR12 = VR + V12
be fully specified. A sketch of the geometry of the ramp under
For diverge areas, four checkpoints are needed. The limit on
analysis is entered into the upper portion of the worksheet of Figure
total flow is the capacity of the approaching freeway (VF). Other
5-5. It should show all lanes and their configuration, lane widths,
checkpoints include the expected demand at the diverge influence
the ramp volume (VR) in vehicles per hour, and the upstream ap-
area (V12 ) and the capacity of each exit leg of the diverge (i.e.,
proaching freeway volume (VF) in vehicles per hour. Where up-
VFO, VR).
stream adjacent or downstream adjacent ramp information is
Checkpoint flow rates are compared with the capacity values of
available, it is entered in the areas to the left and right of the
Table 5-1. If existing or expected flows exceed these capacities,
sketch on the worksheet.
LOS F is indicated, and a Y is noted in the appropriate cell. If
Step 2: Convert all demand volumes to flow rates (in passenger
existing or expected flows do not exceed these capacities, stable
cars per hour) under ideal conditions. All demand volumes speci-
flow in the range LOS A through E is expected, and an N is
fied in mixed vehicles per hour for the full hour under consider-
entered in the ‘‘LOS F?’’ cell.
ation must be converted to flow rates (for the peak 15 min of the
Step 5: Determine level of service. If Step 4 has already resulted
hour) in passenger cars per hour under equivalent ideal conditions.
in a determination of LOS F, this step is eliminated.
This is done according to Equation 5-1:
If Step 4 has determined that the level of service is in the range
Vveh/hr A through E, the expected density in the ramp influence area is
Vpcph = computed using the equations of Table 5-3, which are shown on
PHF fHV fp
the worksheet. These equations are valid only when the level of
The following volumes must be converted in this way: VF, VR, service is in the A through E range. Input flow rates must be in
VU, and VD. The peak-hour factor, PHF, is specified on the basis passenger cars per hour under ideal conditions. An LOS determina-
of local demand characteristics. The two adjustment factors are tion is made by comparing the resultant density with the criteria
found using the methods specified in Chapter 3, Basic Freeway in Table 5-2. For additional information, the average speed in
Sections. the ramp influence area may be roughly approximated using the
Step 3: Estimate V12. The flow rate of freeway vehicles re- equations in Table 5-4.
maining in Lanes 1 and 2 immediately upstream of the merge
point or beginning of the deceleration lane is critical. The appro-
priate model is selected from Figure 5-3 (merge areas) or 5-4 SPECIAL APPLICATIONS
(diverge areas) and applied. All input flow rates used in these
equations must be converted to passenger cars per hour under ideal The procedures outlined in the previous section apply to stan-
conditions (i.e., they are taken from the ‘‘conversion’’ section of dard one-lane, right-hand on- and off-ramp freeway terminals.
the worksheet). The results of this computation are entered into the There are a number of special situations requiring modifications of
third section of the worksheet as shown. The appropriate equation the basic procedure. Each is discussed in the following subsections.
number (from Figure 5-3 or 5-4) is also shown so that it may be
checked later.
Where a configuration may be covered by more than one equa- Two-Lane On-Ramps
tion, such as when both an upstream and a downstream adjacent
ramp fall within the appropriate ranges for application, both com- Figure 5-6 shows a typical two-lane freeway on-ramp. It is
putations should be made. The higher resulting value of V12 should characterized by two separate acceleration lanes, each successively
be used. forcing merging maneuvers to the left. Whereas the general intent
Equations dealing with the effects of upstream or downstream of such ramps is to allow higher ramp flows to merge more
adjacent ramps are used only when all variables fall within the smoothly into the traffic stream, studies (2) have not clearly dem-
5-10 freeways
Figure 5-5. Worksheet for the analysis of ramp-freeway terminals.

onstrated whether two-lane on-ramps can effectively serve higher

on-ramp flow rates than similar one-lane ramps.
Two-lane on-ramps entail two modifications of the basic
methodology: (a) the flow remaining in Lanes 1 and 2 immedi-
ately upstream of the on-ramp is generally somewhat higher
than that for one-lane on-ramps in similar situations, and (b)
densities in the merge area are lower than in similar one-lane
on-ramp situations. The latter modification is primarily due to
the existence of two acceleration lanes and the generally longer
distance over which the two acceleration lanes extend. The
effectiveness of two-lane on-ramps, then, is that higher ramp
flows are handled more smoothly and at better levels of service
than if the same flows were carried on a one-lane ramp with
a conventional merge design.
In computing V12 for two-lane on-ramps, the standard expression
is used:
Figure 5-7. Common geometries for two-lane off-ramps.
V12 = VF (PFM)
However, the formula for PFM given in Figure 5-3 is replaced by In computing V12, the general equation for diverge areas is used:
the following:
V12 = VR + (VF − VR)PFD
T For four-lane freeways, PFM = 1.0000.
However, PFD is not found from the equations in Figure 5-4. It is
T For six-lane freeways, PFM = 0.5550.
determined as follows:
T For eight-lane freeways, PFM = 0.2093.
T For four-lane freeways, PFD = 1.000.
In computing the expected density in the ramp influence area,
T For six-lane freeways, PFD = 0.450.
the standard equation of Table 5-3 is applied, except that the length
T For eight-lane freeways, PFD = 0.260.
of the acceleration lane, LA, is replaced by the effective length of
the acceleration lane, LAeff, as follows; In estimating the density in the ramp influence area, the standard
equation of Table 5-3 is still applied. Where the geometry is of
LAeff = 2LA1 + LA2 (5-4)
the type shown in Figure 5-7(a), the length of the deceleration
where LA1 and LA2 are as defined in Figure 5-6. lane, LD, is replaced by the effective length of the deceleration
The capacity values governing maximum flow rates for VFO and lane, LDeff, as follows:
VR12 are not affected by the use of a two-lane on-ramp. The capacity
LDeff = 2LD1 + LD2 (5-5)
of the downstream freeway section continues to control the total
output capacity of the merge, and the number of vehicles that may Where the geometry is of the type shown in Figure 5-7(b), the
enter the influence area on Lanes 1 and 2 of the freeway is not standard density equation is applied without modification.
enhanced by the existence of a two-lane on-ramp. The capacity As in the case of two-lane on-ramps, the basic capacity
values of Table 5-1 apply unchanged. limitations for two-lane off-ramps are not different from those
of one-lane off-ramps. The control on total output (VFO + VR)
is the capacity of the upstream basic freeway section, since this
Two-Lane Off-Ramps
capacity limits the total flow that can be delivered to the diverge.
No evidence suggests that the maximum value of V12 is affected
Two-lane off-ramps have two general types of geometry, as
by whether the off-ramp has one or two lanes. Thus, the capacity
shown in Figure 5-7. In the first, two deceleration lanes are succes-
values of Table 5-1 are applied without modification.
sively introduced. In the second, only a single deceleration lane
Whereas the total flow that can be discharged through a two-
is used, with drivers in Lane 1 of the freeway permitted to directly
lane off-ramp section is not different from that of a one-lane off-
access the second lane of the ramp without a deceleration lane.
ramp, the distribution of the discharge flow between freeway and
As in the case of two-lane on-ramps, the existence of a two-
ramp is most certainly affected. A two-lane off-ramp can handle
lane off-ramp influences the flow rate in Lanes 1 and 2, and the
significantly greater ramp flows than a one-lane off-ramp. Assum-
resulting density in the influence area is reduced if the geometry
ing that there is no more stringent limitation at the other ramp
shown in Figure 5-7(a) is used.
terminus, a two-lane off-ramp can accommodate flows of up to
4,000 pcph. One-lane off-ramps have a significantly lower capac-
ity, as is detailed in a later section. One-lane off-ramps most often
fail because of insufficient ramp capacity, not because of any factor
related to the diverge area itself.
Lane Additions and Lane Drops
Sometimes on-ramps are associated with lane additions and off-

Figure 5-6. Typical two-lane on-ramp. ramps with lane drops. Where a single-lane ramp results in a lane
5-12 freeways
addition or deletion, the capacity of the ramp is governed by its Left-Hand Ramps
geometry, as indicated in Table 5-6 in a later section of this chapter.
Where a two-lane ramp results in a lane addition or deletion, Although not normally recommended, left-hand ramps do exist
the section should usually be treated as a major merge or diverge on some freeways and occur frequently on collector-distributor
according to procedures described later. roadways. When this happens, the ramp influence area covers the
same length as that for right-hand ramps but now encompasses
the two left lanes plus an acceleration or deceleration lane. Whereas
Effects of Ramp Control for right-hand ramps a critical computation is the estimation of
V12, for left-hand ramps the two left lanes are of interest. For a
For the purposes of this chapter, procedures are not modified four-lane freeway, this remains V12 and there is no difficulty. For
in any way to account for the local effect of ramp control, except a six-lane freeway, the entering flow of interest is V23, and for an
for the limitation the ramp meter may have on VR. eight-lane freeway it is V34. Although no direct method is available
for the analysis of left-hand ramps, some rational modifications can
be applied to right-hand ramp methodologies to produce reasonable
Ramps on 10-Lane Freeway Sections (5 Lanes in Each results.
Direction) The following procedure is suggested: compute V12 using stan-
dard procedures for right-hand ramps. Then
Although they are not common, sections of 10-lane freeway do
exist in parts of the United States, and procedures must be devel- T For left-hand ramps on four-lane freeways, V12 = V12.
oped for handling the right-hand on- and off-ramps that may be T For left-hand on-ramps on six-lane freeways, V23 = 1.12V12.
placed on such sections. The general approach is a simple one: T For left-hand off-ramps on six-lane freeways, V23 = 1.05V12.
the flow rate in Lane 5 of the freeway (V5) is estimated and de- T For left-hand on-ramps on eight-lane freeways, V34 = 1.20V12.
ducted from the total approaching freeway flow. This becomes the T For left-hand off-ramps on eight-lane freeways, V34 = 1.10V12.
effective approaching freeway flow (VFeff) for an equivalent eight-
The remaining computations for density or speed (or both) may
lane freeway section. The analysis proceeds using the standard
continue; V12 is replaced with V23 or V34 as appropriate. All capacity
procedures for eight-lane freeways.
values remain unchanged. These procedures have been adapted
from Leisch (3) using judgment.
VFeff = VF − V5 (5-6)
where V5 is estimated using the criteria of Table 5-5. Effects of Ramp Geometry
Values for V5 in advance of on-ramps are taken from a recent
study (2). Values estimated in advance of off-ramps are taken from The procedures in this chapter explicitly consider the effect of
a 1974 report (3). The values may appear somewhat incongruous the length of the acceleration or deceleration lane and the free-
in that Lane 5 flows are predicted to be heavier in advance of on- flow speed of the ramp roadway on the performance of ramp-
ramps than off-ramps. Whereas this is partially due to the time terminal influence areas. The latter is a surrogate variable that is
difference between the two studies, off-ramp values reflect the fact affected by many related factors, including design speed of various
that none of the off-ramp flow will be in Lane 5, which will lower segments of the ramp roadway, relative grades, sight distance, and
the expected proportion of total approaching freeway flow in Lane others. No models are available that explicitly consider each of
5. Off-ramp values are, however, somewhat more conservative, these factors as an operational variable.
particularly at freeway flows under 4,000 pcph, where no Lane 5 Drew (4) demonstrated, using gap acceptance models, that the
flow is anticipated. gap acceptance capacity of an on-ramp vehicle would be reduced
by as much as 90 percent when a 2-degree angle of convergence
and a 1,200-ft acceleration lane were worsened to 10 degrees and
400 ft, respectively. The user is cautioned that Drew’s use of ‘‘gap
acceptance capacity’’ is in no way related to the definition of
capacity used in these procedures.
Table 5-5. Determination of V5 for Right-Hand Ramps
on 10-Lane Freeways More recent studies (2) show that improved geometric details
do not influence capacity at all, but rather help create better and
total freeway flow, flow in lane 5, smoother merging and diverging operations.
VF (pcph) V5 (pcph)
Approaching Right-Hand On-Ramps
>8,500 2,500 Major Merge Sites
7,500–8,499 0.285VF
6,500–7,499 0.270VF A major merge is one in which two primary roadways, each
5,500–6,499 0.240VF with multiple lanes, merge to form a single freeway segment.
<5,500 0.220VF
The merging roadways may originate in a freeway interchange
Approaching Right-Hand Off-Ramps or from an arterial or rural highway. Major merges are different
>7,000 0.200VF from one- and two-lane on-ramps in that each of the merging
5,500–7,000 0.150VF roadways is generally at or near freeway design standards and
4,000–5,499 0.100VF there is no clear ‘‘ramp’’ or acceleration lane involved in the
<4,000 0 merge.
Figure 5-9. Major diverge areas.
Figure 5-8. Major merge areas. the number of lanes leaving the diverge area is one greater than
the number of lanes on the freeway segment approaching the di-
verge. In Figure 5-9(b), the number of lanes leaving the diverge
Such major merge areas come in a variety of geometries, all of is the same as the number of lanes approaching it.
which fall into two general categories, as shown in Figure 5-8. In Major diverge areas differ from one- and two-lane off-ramps in
merges of the type shown in Figure 5-8(a), the number of lanes that both diverging roadways have multiple lanes and are built to
departing from the merge area is one less than the total number reasonably high freeway design standards. The diverging roadways
of lanes approaching it. This is accomplished by having the right may be part of a freeway-to-freeway interchange or may eventually
lane of the left merging leg and the left lane of the right merging connect with arterial or other surface roadway systems.
leg combine to form a single lane. In geometries of the type shown Once again, the major form of analysis is a comparison of
in Figure 5-8(b), the number of lanes departing the merge is the entering and departing demand flows (converted to passenger cars
same as the total number of lanes approaching it. per hour under ideal conditions) with the capacities of the ap-
There are no good models of performance for major merge proaching freeway lanes and the departing legs. At major diverges,
areas. The analysis of merge areas, therefore, is limited to a operational problems are most often created by insufficient capac-
check of capacities on the approaching legs and the departing ity on one or more of the departing legs.
freeway lanes. The capacity of each entering leg and the departing Whereas there is no performance model for major merge areas,
freeway lanes is computed using the values in Table 5-1. The a relatively simple model has been developed (2) to predict density
capacity of each entering leg is compared with the peak demand across all freeway lanes in a 1,500-ft length immediately in ad-
flow on each (converted to passenger cars per hour under ideal vance of a major diverge:
conditions), whereas the capacity of the departing freeway lanes
VF
is compared with the sum of the two peak entering demands D = 0.0175 (5-7)
(also converted to passenger cars per hour under ideal conditions). N
Problems in major merge areas generally result from insufficient
where D is the average density across all freeway lanes in a
capacity of the downstream freeway section.
1,500-ft range immediately upstream of the diverge (pc/mi/ln) and
N is the number of freeway lanes immediately upstream of the
Major Diverge Sites diverge (R2 = 0.56, SE = 1.74 pc/mi/ln, 21 data periods).
This model was calibrated with a very small data base and
Like major merge areas, major diverge areas occur in one of should be used with caution. It can be used to give a general
two types of geometry. Shown in Figure 5-9, they are basically estimate of density in the major diverge area and to establish level
the reverse of the two major merge geometries. In Figure 5-9(a), of service using the criteria in Table 5-2.
5-14 freeways
CAPACITY OF RAMP ROADWAYS Table 5-6. Approximate Capacity of Ramp Roadways
free-flow capacity (pcph)

Because most operational problems occur at ramp terminals speed of ramp, single-lane two-lane
(either the ramp-freeway terminal or the ramp-street terminal), SFR (mph) ramps ramps
there is little information regarding the operational characteristics >50 2,200 4,400
of ramp roadways themselves. Some basic design standards exist 41-50 2,100 4,100
in AASHTO policies (1), but they are not related to specific opera- 31-40 2,000 3,800
tional characteristics. In the 1970s, Leisch (3) adapted this material 21-30 1,900 3,500
to provide a broader set of criteria that were, again, unrelated to <21 1,800 3,200
specific operational characteristics. Thus, information presented in
higher level of service but will not increase the capacity of the
this section is for general guidance only.
merge, which is controlled by the capacity of the downstream
Ramp roadways differ from the freeway mainline in the follow-
freeway section. For higher on-ramp flows, a two-lane on-ramp
ing ways:
must be used in conjunction with a lane addition and a major
1. Ramps are roadways of limited length and width (often just merge configuration.
one lane). Two-lane off-ramps can accommodate higher ramp flows
2. The free-flow speed of the ramp is frequently lower than that through the diverge area than single-lane off-ramps, although high
of the roadways it connects, particularly the freeway. observations are in the 4,000-pcph range. Such high off-ramp
3. On single-lane ramps, where passing is not possible, the flows, however, often leave the continuing freeway section with
adverse effect of trucks and other slow-moving vehicles is more relatively low per-lane flow rates. A major diverge configuration
pronounced than on a multilane roadway. can be considered and may more effectively balance the per-lane
4. Acceleration and deceleration often take place on the ramp flows on each departing leg.
itself. Even where a single-lane merge or diverge configuration is used,
5. At ramp-street junctions, queueing may develop on the ramp, there are several reasons to consider widening the ramp to two
particularly if the ramp-street junction is signalized. lanes outside the terminal areas, including the following:
1. When the ramp is longer than 1,000 ft, a second lane allows
Table 5-6 gives approximate criteria for the capacity of ramp
drivers to pass stalled or slow-moving vehicles. This can also be
roadways. These capacities are based on recent studies (2) and
accomplished with a single-lane ramp and a paved shoulder of 8
previously noted work conducted in the 1970s (3).
ft or more.
Table 5-6 gives the approximate capacity of the ramp roadway
2. When queues are expected to form at signalized and other
itself, not the ramp-freeway terminal. There is no evidence, for
ramp-street terminals, an additional ramp lane provides additional
example, that a two-lane on-ramp freeway terminal can accommo-
storage capacity.
date any more vehicles than a one-lane ramp terminal without the
3. When the ramp has a steep grade or other minimal geomet-
addition of a lane (in which case the configuration becomes a
rics, a second ramp lane again allows drivers to pass slow-moving
major merge area).
vehicles.
Thus, it is unlikely that two-lane on-ramps can accommodate
more than 2,200 pcph through the merge area itself. The two-lane In such cases, the two-lane ramp is tapered to a single lane in
configuration will achieve a merge with less turbulence and a advance of the ramp-freeway terminal.
CALCULATION 1: ISOLATED ON-RAMP Solution
A sketch of this section is shown in Figure 5-10, the worksheet

Problem for this calculation.
The first computation must be the conversion of all demand
An on-ramp on a four-lane freeway with standard 11.8-ft (3.6-m) volumes to flow rates in passenger cars per hour under ideal condi-
lane widths and adequate clearances serves a demand of 550 vph tions. For each demand flow, the PHF is given as well as informa-
(5 percent trucks). The freeway mainline approaching the ramp tion that allows the determination of fHV and fp. These factors are
carries 2,500 vph (10 percent trucks). The terrain is level, PHF is selected according to the procedures in Chapter 3. The driver
0.90, and the ramp has an acceleration lane with a total length of population factor, fp, is 1.00, because no special characteristics are
750 ft. Free-flow speeds are 60 mph for the freeway and 45 mph noted. For level terrain, the passenger car equivalent for trucks is
for the ramp. Drivers are primarily regular users of the facility. 1.5 per truck, yielding an fHV of 1/[1 + 0.10 (1.5 − 1)] = 0.952 for
At what level of service would this ramp be expected to operate? freeway volume and 1/[1 + 0.05 (1.5 − 1)] = 0.976 for ramp
Figure 5-10. Worksheet for Calculation 1.

5-16 freeways
volume. The PHF for both volumes is given as 0.90. Thus, the
adjusted demand flow rates are
2,500
VF = = 2,918 pcph
0.90 (0.952) (1.00)
550
VR = = 626 pcph
0.90 (0.976) (1.00)
The remainder of this calculation uses these converted demand
flow rates as inputs. All flow rate results are in terms of
passenger cars per hour during the peak 15 min of the hour of
interest. Figure 5-11. Freeway section for Calculation 2.
Figure 5-3 indicates that V12 should be computed using
PFM = 1.00 (for a four-lane freeway), and V12 = VF = 2,918 pcph.
Two capacity values must now be checked. The total down-
stream freeway flow rate leaving the merge area is 2,918 + sonable conclusions concerning the operation of the entire section
626 = 3,544 pcph. This flow is less than the capacity for a from this information.
four-lane freeway with a 60-mph free-flow speed (4,600 pcph, The first computation involves converting the freeway volume
Table 5-1), so no problem is anticipated. The total flow entering and the ramp volumes to equivalent passenger cars per hour under
the ramp influence area is also 3,544 pcph, which is also less ideal conditions for the peak 15-min period within the hour. Ad-
than the capacity of 4,600 pcph for such flows (Table 5-1). justment factors are drawn as appropriate from Chapter 3. The
The operation is therefore expected to be stable; that is, no driver population factor, fp, is 1.00, because there is no indication
queues will form under the conditions of this calculation. of a nonstandard driver population. The passenger car equivalent,
The expected density in the merge influence area is now com- ET, for rolling terrain is 3.00, yielding fHV = 1/[1 + 0.05 (3 − 1)] =
puted using the on-ramp equation in Table 5-3: 0.909 for all movements. Then
DR = 5.475 + 0.00734VR + 0.0078V12 − 0.00627LA 4,500
VF = = 5,211 pcph
DR = 5.457 + 0.00734 (626) + 0.0078 (2,918) − 0.00627 (750) 0.95 (0.909) (1.00)
DR = 28 pc/mi/ln 300
VR1 = = 347 pcph
0.95 (0.909) (1.00)
From Table 5-2, this density exactly equals the upper limit for
LOS C. For supplemental information, the expected average speed 500
VR2 = = 579 pcph
of vehicles can be estimated using the on-ramp, stable-flow equa- 0.95 (0.909) (1.00)
tion in Table 5-4:
By implication, the freeway flow rate immediately upstream of
SR = SFF − (SFF − 42) Ms the second ramp is 5,211 − 347 = 4,864 pcph. Figure 5-12a and
5-12b shows worksheets for each of the two ramps.
Ms = 0.321 + 0.0039e(3,544/1,000) − 0.002 (225 × 70/1,000) = 0.39
The first ramp is an off-ramp with an adjacent downstream off-
SR = 60 − [(60 − 42) × (0.39)] = 53 mph ramp. Figure 5-4 suggests that Equation 7 or 9 be used for this
case. In checking the range of applicability for Equation 9, the
The results of this analysis indicated that the on-ramp displayed ramp flow of 347 pcph falls well outside the range of calibration,
would be expected to operate at an acceptable level of service, 502 to 696 pcph. The methodology suggests that in such cases,
LOS C. No unusual turbulence or queueing would be expected the general equation for a six-lane freeway be used. Thus, Equation
to occur. 7 would be appropriate:
V12 = VR + (VF − VR ) PFD
CALCULATION 2: CONSECUTIVE OFF-RAMPS ON SIX-LANE
FREEWAY PFD = 0.760 − 0.000025 (5,211) − 0.000046 (347) = 0.614
V12 = 347 + (5,211 − 347) (0.614) = 3,333 pcph
Problem
Capacity values for the first ramp are now checked. The total
Figure 5-11 illustrates the section under study in this calculation. flow leaving the diverge area is 5,211 pcph, which is lower than
Two consecutive off-ramps are spaced at 750 ft on a section of the capacity for a six-lane freeway with a 60-mph free-flow speed
freeway in generally rolling terrain. All other pertinent information (6,900 pcph, Table 5-1). The demand flow V12 is 3,333 pcph,
is shown in Figure 5-11. What is the expected level of service which is lower than the capacity for vehicles entering the diverge
through this section? influence area, 4,400 pcph (Table 5-1). The off-ramp itself will
carry 347 pcph. The capacity of a single-lane off-ramp with a free-
flow speed of 35 mph is 2,000 pcph (Table 5-6). Thus, none of
Solution the capacity values are exceeded by present demands, and the
operation is expected to be stable.
The solution to this problem involves separate analysis of the The expected density in the ramp influence area is computed
operation of each ramp-freeway terminal and the drawing of rea- using the following single-lane off-ramp equation:
Figure 5-12a. Worksheet for Calculation 2 (first ramp).

5-18 freeways
Figure 5-12b. Worksheet for Calculation 2 (second ramp).

DR = 5.252 + 0.0086 (3,333) − 0.009 (500) = 28 pc/mi/ln Solution
From Table 5-2, this is at the upper limit of LOS C. The off-ramp The unique feature of this problem is that the 1,500-ft influence
stable-flow equation from Table 5-4 can be used to get a general areas of the two ramps will substantially overlap. In such a case,
estimate of the average speed in the ramp influence area. This operations would be dominated by the analysis indicating the poor-
computation is not shown here but results in a speed of 52 mph, est operation.
as shown on the worksheets. Computations again begin with the conversion of all demand
The second ramp is an off-ramp with an adjacent upstream volumes to equivalent passenger cars per hour under ideal condi-
off-ramp. Figure 5-4 suggests that Equation 7 again be used for tions during the peak 15 min of the hour. The driver population
this case: factor is taken to be 1.00, there being no indication that drivers
are unfamiliar with the area. For level terrain, the passenger car
V12 = VR + (VF − VR) PFD
equivalent for trucks is 1.5, which yields a factor of 1/[1 + 0.10(1.5
PFD = 0.760 − 0.000025 (4,864) − 0.000046 (579) = 0.612 − 1)] = 0.952 for 10 percent trucks and 1/[1 + 0.05(1.5 − 1)] =
V12 = 579 + (4,864 − 579) (0.612) = 3,201 pcph 0.976 for 5 percent trucks. Then
Capacity values may now be checked using Table 5-1. The total 5,500
VF = = 6,419 pcph
flow arriving at the diverge (VFI) is 4,864 pcph, well within the 0.90 (0.952) (1.00)
capacity of 6,900 pcph for six-lane freeways with a 60-mph free- 400
VR1 = = 455 pcph
flow speed; the V12 value of 3,201 pcph is also lower than the 0.90 (0.976) (1.00)
capacity of 4,400 pcph for this flow. The off-ramp, with a free- 600
VR2 = = 700 pcph
flow speed of 25 mph, has a capacity of 1,900 pcph (Table 5-6), 0.90 (0.952) (1.00)
which is greater than the ramp flow of 579 pcph. Again, no break-
downs are expected, and the flow is expected to be stable. The freeway flow approaching the second ramp is the sum of the
Density is computed using the off-ramp equation in Table 5-3: freeway flow upstream of Ramp 1 plus the flow entering at Ramp 1,
or 6,419 + 455 = 6,874 pcph. Figure 5-14 shows the worksheets
DR = 4.252 + 0.0086 (3,201) − 0.009 (300) = 29 pc/mi/ln for the two ramps involved in this calculation. From Figure 5-3,
From Table 5-2, this is LOS D. An approximate indication of the Equation 5 should be used to find V12 for the on-ramp of the pair:
average speed within the ramp influence area can be obtained from V12 = VF × PFM
the single-lane off-ramp stable-flow equation in Table 5-4. Not
shown here, the computation results in a speed of 49 mph. PFM = 0.2178 − 0.000125 (455) + 0.01115 (250/30) = 0.254
The section, taken as a whole, is expected to operate within the V12 = 6,419 (0.254) = 1,630 pcph
better range of LOS D. Because the ramp influence areas overlap
somewhat here, the poorer performance predicted for the second It should be noted that this value is considerably lower than might
ramp is likely to dominate operating conditions throughout the be expected. With 1,630 pcph using Lanes 1 and 2, Lanes 3 and
entire section. 4 will carry 6,419 − 1,630 = 4,769 pcph. Such an imbalance is
difficult to accept. Caution suggests that the estimate of V12 may
be unreasonably low.
CALCULATION 3: ON-RAMP–OFF-RAMP PAIR ON EIGHT- Demand flows are now checked against the capacities of Table
LANE FREEWAY
5-1. The total flow downstream of the merge is 6,419 + 455 = 6,874
pcph. This is lower than the capacity for an eight-lane freeway with
Problem a free-flow speed of 60 mph, 9,200 pcph. The flow rate entering
the merge influence area is 1,630 + 455 = 2,085 pcph, which is
Figure 5-13 illustrates the section under consideration in this lower than the capacity of 4,600 pcph for this flow. Thus, no
calculation. The eight-lane freeway section shown runs through breakdowns are expected at the first ramp, and stable flow is
an area of level terrain; all demand flows and other information assumed.
are shown in Figure 5-13. At what level of service would this The density in the merge influence area is computed from the
section be expected to operate? on-ramp equation of Table 5-3:
DR = 5.475 + 0.00734 (455) + 0.0078 (1,637)
− 0.00627 (250) = 20 pc/mi/ln
From Table 5-2, this is right at the boundary of LOS B. The
approximate average speed of vehicles in the ramp influence area
can be estimated using the on-ramp, stable-flow equation of Table
5-4. Though not shown here, this computation results in a speed
of 54 mph.
Figure 5-4 indicates that Equation 10 should be used to compute
V12 for the off-ramp of this pair:
V12 = VR + (VF − VR) PFD
PFD = 0.436
Figure 5-13. Freeway section for Calculation 3. V12 = 700 + (6,874 − 700) (0.436) = 3,392 pcph
5-20 freeways
Figure 5-14a. Worksheet for Calculation 3 (first ramp).

Figure 5-14b. Worksheet for Calculation 3 (second ramp).

5-22 freeways

Comparison of demand flows for the second ramp and capacity in one direction. The total flow entering the merge influence area
values of Table 5-1 reveals no problems. The total flow entering is 1,696 + 1,941 = 3,637 pcph, which is also less than the capacity
the diverge is 6,874 pcph, which is below the capacity for an eight- of 4,600 pcph. Thus, no queueing or breakdowns are expected,
lane freeway (free-flow speed of 60 mph) of 9,200 pcph. The V12 and flow is expected to be stable throughout the subject period of
value of 3,392 pcph is below the associated capacity of 4,400 pcph. analysis.
The two exit legs are also below capacity values. The downstream To compute density for a two-lane on-ramp influence area, the
freeway flow is 6,174 pcph, which is below the capacity of 9,200 normal on-ramp equation in Table 5-3 is used except that the actual
pcph, and the off-ramp flow is 700 pcph, compared with a capacity length of acceleration lane is replaced by the effective length of
(Table 5-6, 30 mph) of 2,000 pcph. the dual acceleration lane:
The off-ramp equation from Table 5-3 is used to compute the
LAeff = 2LA1 + LA2 = 2 (500) + (400) = 1,400 ft
density in the ramp influence area:
DR = 5.457 + 0.00734 (1,941) + 0.0078 (1,796)
DR = 4.252 + 0.0086 (3,392) − 0.009 (250) = 31 pc/mi/ln
− 0.00627 (1,400) = 27 pc/mi/ln
From Table 5-2, this is LOS D. An approximate average vehicle
According to Table 5-2, this is LOS C. The approximate speed of
speed in the ramp influence area is computed from the off-ramp,
vehicles traveling through the ramp influence area can be estimated
stable-flow equation of Table 5-4; the result is 50 mph.
by using the on-ramp, stable-flow equation of Figure 5-4. As with
As initially stated, the influence areas of the two ramps overlap
the density equation, LAeff is used in place of LA. Not shown here,
in the region between the two ramps. The analysis of the first ramp
this computation leads to an average speed of 53 mph.
predicted LOS C operation, but it was noted that the prediction of
V12 was quite possibly too low. Analysis of the second ramp pre-
dicts LOS D operation, which would be expected to prevail CALCULATION 5: OFF-RAMP ON 10-LANE FREEWAY
throughout the section in this case.
Problem
CALCULATION 4: TWO-LANE ON-RAMP
Figure 5-16 illustrates a five-lane (in one direction) segment of
freeway with an off-ramp. All geometric and traffic conditions are
Problem specified. What is the likely level of service under the scenario
shown?
A two-lane on-ramp on a six-lane freeway carries a demand
volume of 1,800 vph with 5 percent trucks. The freeway mainline
Solution
carries 3,000 vph with 5 percent trucks immediately upstream of
the merge area. The free-flow speed of the freeway is 55 mph and
In the Special Applications section, it is indicated that ramps
the free-flow speed of the ramp is 50 mph. Figure 5-15, the work-
on five-lane segments of freeway are treated by estimating V5, the
sheet for this calculation, contains a sketch detailing the section.
flow rate in the left-hand lane. This flow is then deducted from
At what level of service is this ramp expected to operate?
the freeway flow, and an equivalent situation involving an eight-
lane freeway is established.
Solution As in all analyses, demand volumes are converted to passenger
cars per hour under ideal conditions for the peak 15 min of opera-
As in all previous problems, the first computation is the conver- tion during the hour. The driver population is not indicated to be
sion of all demand volumes to equivalent passenger cars per hour unusual, so fp is 1.00, ET for 10 percent trucks in rolling terrain is
under ideal conditions during the peak 15 min of the hour. The 3.0, and fHV = 1/[1 + 0.10(3 − 1)] = 0.833. Then
PHF is given as 0.95. Since no abnormal driver population is cited, 7,200
fp = 1.00. For level terrain, ET = 1.5, and the adjustment factor for VF = = 9,098 pcph
0.95 (0.833) (1.00)
5 percent trucks is 1/[1 + 0.05 (1.5 − 1)] = 0.976. Then
400
3,000 VR = = 506 pcph
VF = = 3,236 pcph 0.95 (0.833) (1.00)
0.95 (0.976) (1.00)
1,800
VR = = 1,941 pcph
0.95 (0.976) (1.00)
In the Special Applications section of this chapter, it is indicated
that in finding V12 for a two-lane on-ramp on a six-lane freeway,
PFM should be set at 0.5550:
V12 = VF × PFM = 3,236 (0.5550) = 1,796 pcph
Although two-lane on-ramps do not affect any of the critical capac-

ity values, the demand flows must be checked against the capacity
values of Table 5-1. The total downstream freeway flow rate is
3,236 + 1,941 = 5,177 pcph, which is less than the capacity of a
six-lane freeway (with a free-flow speed of 55 mph) of 6,750 pcph Figure 5-16. Freeway section for Calculation 5.
5-24 freeways
Table 5-5 indicates that for freeway flows over 7,000 pcph in the
vicinity of an off-ramp, V5 is likely to be 20 percent of VF. Thus,
for an equivalent eight-lane freeway (four lanes in each direction),
VFeff4 = VF − V5 = 9,098 − (9,098)(0.20) = 7,279 pcph
Now, the problem to be analyzed is the equivalent eight-lane free-

way section shown in Figure 5-17. The worksheet for the equiva-
lent section is shown in Figure 5-18.
Since all flow rates have already been converted to passenger Figure 5-17. Equivalent four-lane segment for Calculation 5.
cars per hour under ideal conditions during the peak 15 min of Solution
the hour, capacity comparisons can be made directly. The equiva-
lent four-lane segment carries a total flow approaching the diverge Figure 5-19 shows the worksheet for this calculation, including
of 7,279 pcph, which is below the capacity of 9,200 pcph for a sketch of the section as described.
eight-lane freeways with 60-mph free-flow speed (Table 5-1). The All demand volumes must first be converted to passenger cars
total flow entering the ramp influence area is V12, computed for per hour under ideal conditions during the peak 15 min of the
the equivalent segment using Equation 10 from Figure 5-4: hour. For level terrain, ET is 1.5, and fHV = 1/[1 + 0.15(1.5 − 1)] =
0.93 for 15 percent trucks and 1/[1 + 0.05(1.5 − 1)] = 0.976 for 5
V12 = VR + (VF − VR) PFD = 506 + (7,279 percent trucks. Then
− 506) (0.436) = 3,459 pcph
4,000
VF = = 4,779 pcph
This flow is also below the capacity value of 4,400 pcph (Table 0.90 (0.93) (1.00)
5-1). The off-ramp flow is 506 pcph compared with the ramp 500
VR = = 569 pcph
capacity of 2,100 pcph (Table 5-6). Thus, no breakdowns are 0.90 (0.976) (1.00)
expected, and flow is expected to be stable throughout the study
period. In the Special Applications section, it is suggested that for left-
The density within the ramp influence area is computed using hand ramps on six-lane freeways, flow in the left two lanes is 1.12
the off-ramp equation from Table 5-3: times the flow that would occur in Lanes 1 and 2 if the ramp were
on the right-hand side. Thus, Equation 2 (Figure 5-3) is used to
DR = 4.252 + 0.0086 (3,459) − 0.009 (700) = 28 pc/mi/ln compute V12 as if the ramp were on the right side. This computation
yields 2,853 pcph. Then the flow expected in the left two lanes
From Table 5-2, this is LOS C. With the appropriate equation immediately upstream of the merge is estimated:
from Table 5-3, the average speed of vehicles in the ramp influence
V23 = 2,853 (1.12) = 3,195 pcph
area can be estimated as 53 mph.
The remainder of the problem is solved using V23 in place of V12.
Demand flows are compared with the capacity values of Table 5-1.
CALCULATION 6: LEFT-SIDE ON-RAMP
The total flow leaving the merge area is 4,779 + 569 = 5,348 pcph
(capacity = 7,050 pcph, six-lane freeway, 60 mph). The total flow
entering the merge influence area is 3,195 + 569 = 3,764 pcph (capac-
ity = 4,600 pcph). Thus, there are no capacity problems in this section.
Problem The density is computed using the on-ramp equation of Table
5-3 with V23 instead of V12:
A left-side on-ramp on a six-lane freeway carries a demand DR = 5.457 + 0.00734 (569) + 0.0078 (3,195)
volume of 500 vph with 5 percent trucks. It has a 700-ft accelera-
− 0.00627 (700) = 30 pc/mi/ln
tion lane. The freeway mainline (upstream of the merge) carries
4,000 vph with 15 percent trucks. PHF is 0.90. The terrain is level; From Table 5-2, this is LOS D. An approximate average speed
lane widths, lateral clearances, and driver population are standard; of vehicles in the merge influence area may be computed using
and free-flow speeds are 60 mph for the freeway and 35 mph for the on-ramp, stable-flow equation in Table 5-4; the result is
the ramp. What is the expected level of service for this section? 52 mph.


5-26 freeways

V. REFERENCES
This edition of ramp-freeway terminal analysis procedures results 2. Roess, R.P., and Ulerio, J.M., Capacity and Level of Service
primarily from studies conducted under National Cooperative High- at Ramp-Freeway Junctions. Final Report, NCHRP Project
way Research Program Project 3-37 (2). Some special applications 3-37, Polytechnic University, Brooklyn, N.Y. (1993).
resulted from adaptations of procedures developed by Leisch (3) in 3. Leisch, J., Capacity Analysis Techniques for Design and Oper-
the 1970s. AASHTO policies (1) contain additional material on ation of Freeway Facilities. Federal Highway Administration,
geometric design and geometric design criteria for ramps. Washington, D.C. (1974).
1. A Policy on Geometric Design of Highways and Streets. Ameri- 4. Drew, D., Traffic Flow Theory and Control. McGraw-Hill,
can Association of State Highway and Transportation Officials, Inc., New York, N.Y. (1968).
Washington, D.C. (1990).

chapter 6
FREEWAY SYSTEMS
CONTENTS
i. introduction .......................................................................................................................................................................... 6-2

ii. combined analysis of freeway segments ......................................................................................................................... 6-2
Design Analysis..................................................................................................................................................................... 6-2
Procedures.......................................................................................................................................................................... 6-2
Sample Calculation............................................................................................................................................................ 6-2
Operational Analysis ............................................................................................................................................................. 6-6
Analysis of Breakdown Conditions ...................................................................................................................................... 6-6
iii. freeway surveillance and control ................................................................................................................................. 6-7
Background ............................................................................................................................................................................ 6-7
Control Elements ................................................................................................................................................................... 6-7
Determination of Problems and Control............................................................................................................................... 6-8
Incidents................................................................................................................................................................................. 6-9
iv. capacity of freeway work zones ...................................................................................................................................... 6-9
Observed Work-Zone Capacities .......................................................................................................................................... 6-9
Long-Term Construction Sites .............................................................................................................................................. 6-10
Short-Term Maintenance Sites.............................................................................................................................................. 6-10
Shoulder Use and Traffic Splitting on Three-Lane Segments............................................................................................. 6-10
Lane Narrowing..................................................................................................................................................................... 6-10
Estimating Queue Length and Delay.................................................................................................................................... 6-11
Sample Calculation................................................................................................................................................................ 6-12
v. weather .................................................................................................................................................................................. 6-13
vi. high-occupancy-vehicle lanes on freeways .................................................................................................................. 6-14
Capacity Analysis for HOV Lanes ....................................................................................................................................... 6-14
Effect of HOV Lanes on Freeway Operations..................................................................................................................... 6-14
Sample Calculation................................................................................................................................................................ 6-14
vii. summary .................................................................................................................................................................................. 6-15
viii. references .............................................................................................................................................................................. 6-15

6-2 freeways
I. INTRODUCTION
Chapters 3, 4, and 5 of this manual have treated in detail the present and significantly affect operations. Because of the many
planning, design, and analysis of basic freeway segments, weaving complexities of freeway system operations, these procedures tend
areas, and ramp junctions, respectively. This chapter addresses to be more approximate and less precise than those applied to
how these elements may be combined into a complete freeway specific freeway subsections. They nevertheless provide a basis
design or analysis, and a number of special features that may be for insight and understanding of system effects.
II. COMBINED ANALYSIS OF FREEWAY SEGMENTS
DESIGN ANALYSIS segments may turn out to be either weaving areas or ramp combina-
tions, depending on the final configuration adopted.
Procedures In application, these guidelines lead to fairly straightforward
computations in the following sequence:
In the design use of the procedures in this manual, it is necessary
to consider the kinds of information that generally would be avail- 1. Establish design level of service, demand volume and traffic
able and what results are desired. Capacity analysis is only one of characteristics, horizontal and vertical alignments, and approxi-
several inputs into the design process. Others include geometric mate ramp locations.
standards, safety standards, standards for signing, and so on. 2. Determine the basic number of lanes required for each of
Capacity analysis procedures are used primarily in the design the basic freeway segments identified as previously noted, using
of cross-sectional elements (number of lanes, lane widths, shoul- the procedures detailed in Chapter 3. The basic number of lanes
ders) and in the selection of lane configurations for individual for each ramp may be determined using techniques described in
freeway elements. In general, the following information is required Chapter 5.
for a design analysis: 3. The results of Step 2 will suggest probable configurations
for ramp junctions and potential weaving areas. Analyze each ramp
T Horizontal and vertical alignments.
junction from three points of view: (a) as an isolated ramp, (b)
T Approximate location of ramps and interchanges.
in combination with the adjacent downstream ramp, and (c) in
T Forecast demand volumes.
combination with the adjacent upstream ramp using the procedures
T Forecast demand characteristics, such as, to name a few, the
in Chapter 5. Usually, one or two of these aspects will be invali-
percentages of trucks, buses, and recreational vehicles in the traffic
dated by those procedures, but in other cases, there will be more
stream, and peak-hour factor (PHF).
than one valid analysis. In such cases, the analysis indicating the
The principal problem in coordinating the overall design analy- poorest operations or level of service is taken as the controlling
sis of a freeway facility is the segmenting of the freeway into solution.
component parts for individual consideration using the methods 4. Analyze the weaving areas using the procedures in Chapter
of Chapters 3, 4, and 5. In general, the following guidelines may 4 to determine likely operating conditions. Note that in design, the
be used: case of an on-ramp followed by an off-ramp must be regarded as
both a potential weaving section with an auxiliary lane and a ramp
1. Each section of freeway between ramps or major junctions combination without an auxiliary.
should be considered to be a separate ‘‘basic freeway segment.’’ 5. If the results of Steps 3 and 4 are unsatisfactory, consider
2. Within these basic freeway segments, any grade of more than
1
⁄4 mi (for grades ≥3 percent) or 1⁄2 mi (for grades <3 percent) must T Altering the number and/or location of ramps (which may
be considered a separate basic freeway segment. Any sharp change affect demand distribution).
in terrain, such as from level to rolling terrain, would also necessi- T Changing the design of ramps and/or mainline segments deter-
tate the division of a single segment into separate subsegments. mined in Step 2 to create new configurations.
Long basic segments with no single grade of significance may be T Changing the design of major interchanges to achieve differ-
considered extended segments of level, rolling, or mountainous ent configurations, reduce weaving, and so forth.
terrain, as defined in Chapter 3. Downgrade segments would nor-
mally be considered to be level terrain unless local data allow for Repeat Steps 2 through 4.
more specific treatment (see Chapter 3).
3. Each ramp junction should be considered once in combina-
tion with the adjacent downstream ramp and once in conjunction
Sample Calculation
with the adjacent upstream ramp. Ramps that are clearly part of a
weaving section would not be analyzed using ramp procedures but
would be treated in Step 4. The design indicated in Figure 6-1 illustrates the foregoing pro-
4. Potential weaving and multiple weaving areas should be in- cedures. Note that the given demand volumes are already ex-
vestigated as such. The term ‘‘potential’’ is used because some pressed as peak rates of flow in passenger cars per hour.
freeway systems 6-3
Figure 6-1. Sample design problem.
Step 1—Establish Demand, Alignment, and Ramp Location Ramps A and D could conceivably be considered both isolated
ramps with a simple weaving section in Segment 3 and part of a
Figure 6-1 shows demand, alignment, and ramp locations for
multiple weaving configuration with Segment 3. Both cases would
the sample problem.
be analyzed.
Step 2—Determine Basic Number of Lanes for Open Freeway Ramp A. According to Figure 5-3 for on-ramps, Equation 1
Segments and Ramps applies. Substituting into the equation gives the following results:
The demand on each open freeway segment is shown in Figure V12 = 1,695 pcph VFO = 3,400 pcph VR12 = 2,195 pcph
6-1. Using the criteria in Table 3-1 directly for LOS B, the number Since VFO < 6,900 and VR12 < 4,600, traffic flow is operating below
of lanes in each segment may be found. Because of design deci- capacity, and density and speed can be determined from Tables
sions, 12-ft lanes, adequate lateral clearance, 70-mph free-flow 5-3 and 5-4, respectively. This gives DR = 21 pc/mi/ln and SR =
speed (SFF) on the freeway mainline, and 45-mph free-flow speed 61 mph. From Table 5-2, LOS = C.
(SFR = 45 mph) on the ramps are to be provided. Ramp D. According to Figure 5-4 for off-ramps, Equation 7
applies. Substituting into the equation yields the following results:
No. of Lanes
Segment Flow Rate Required V12 = 2,465 pcph VFO + VR = 3,600 pcph
1 2,900 3 Since VFO + VR < 6,900 and V12 < 4,400, traffic flow is operating
2 3,400 3 to 4 below capacity, and density and speed can be determined from
3 4,000 4 Tables 5-3 and 5-4, respectively. This gives DR = 23 pc/mi/ln and
4 3,600 3 to 4 SR = 61 mph. From Table 5-2, LOS = C.
5 3,300 3 Ramps B and C should not be considered part of the ramp
Table 5-6 may be used to estimate the number of lanes required configuration because the trial design in Figure 6-2 shows them
for each of the ramps. Using the 45-mph free-flow speed criterion, to be in a weaving configuration; as such, they are analyzed in
all of the ramps in Figure 6-1 are single-lane ramps. All accelera- Step 4.
tion and deceleration lanes are 250 ft long (LA and LD = 250 ft). Step 4—Analyze Potential Weaving Areas
On the basis of these results, the design in Figure 6-2 is most
likely to be appropriate. Note that because there is an auxiliary Segments 2 and 3 should be considered a multiple weave. For
lane between Ramps B and C, Segments 2, 3, and 4 make up a the purposes of this analysis, all off-ramp vehicles at C will be
multiple weaving area in this design. assumed to originate from the freeway mainline, a worst-case as-
sumption. Figure 6-3 depicts the resulting flows and weaving
Step 3—Analyze Ramp Junctions diagrams.
Segment 2. Because one of the Segment 2 weaving movements
Given that Ramps B and C are definitely part of a weaving
is made with no lane change and another with one lane change,
section for the trial design in Figure 6-2, the following ramp combi-
this is a Type B section. For Segment 2,
nations remain to be analyzed with ramp procedures:
VR = 900/3,400 = 0.26
T Ramp A, isolated or with a downstream on-ramp (B),
T Ramp D, isolated or with an upstream off-ramp (C). R = 400/900 = 0.44
6-4 freeways
Figure 6-2. Likely design for sample problem.
and
SFF = 70 mph,
v = 3,400 pcph,
N = 3 lanes, and
L = 2,000 ft.
This results in the following estimates of speed for unconstrained

operation:
Sw = 51.0 mph
Snw = 54.9 mph
To determine whether operations are actually unconstrained, the

number of weaving lanes used is now computed by using the
equation given in Table 4-4:
Nw = N[0.085 + 0.703VR + (234.8/L) − 0.018(Snw − Sw)]
where Snw and Sw are as computed above. Substituting the appro-

priate values,
Nw = 0.95 lane
Nnw = 3 − 0.95 = 2.05 lanes
Because Nw is less than the maximum value of 3.50 lanes for

Type B sections (Table 4-4), the section is unconstrained, and the
Figure 6-3. Consideration of multiple weave. original estimates of weaving and nonweaving speeds are taken
to be correct. On the basis of the calculated speeds and effective
lanes for weaving and nonweaving traffic, the corresponding densi-
ties are computed as follows:
Applying Equations 4-2 and 4-3, the speed of weaving and non- Vw 900
Dw = = = 18.6 pc/mi/ln
weaving vehicles is computed: NwSw 0.95 × 51
SFF − 10 Vnw 2,500
Sw or Snw = 15 + Dnw = = = 22.2 pc/mi/ln
1 + a(1 + VR)b(v/N)c/Ld Nnw Snw 2.05 × 54.9
where, from Table 4-3 for unconstrained Type B sections: According to Table 4-6, weaving traffic operates at LOS B,
whereas nonweaving traffic operates at LOS C. A joint measure
Constant Sw Computation Snw Computation for the entire traffic stream can be estimated using the average
a 0.100 0.020 overall speed (weighted by volume), S, where
b 1.20 2.00
c 0.77 1.42 (900 × 51) + (2,500 × 54.9)
S= = 53.9 mph
d 0.50 0.95 900 + 2,500
freeway systems 6-5
and the corresponding density,

3,400
D= = 21 pc/mi/ln
3 × 53.9
yielding an overall LOS C for the weaving section.
Segment 3. This segment should be considered a Type A weav-
ing area because it has an auxiliary lane, as shown in Figure 6-3,
and all weaving vehicles make at least one lane change. Note that
consideration of Segment 3 as a multiple weave is the same as
considering it as a simple weaving section. For Segment 3,
VR = 1,000/4,000 = 0.25
R = 400/1,000 = 0.40
From Table 4-3, for unconstrained Type A weaving areas:
Constant Sw Computation Snw Computation

a 0.226 0.020
b 2.20 4.00
c 1.00 1.30
d 0.90 1.00
and
SFF = 70 mph,
v = 4,000 pcph,
N = 4 lanes, and Figure 6-4. Consideration of multiple weave.
L = 1,500 ft.
assumed that no on-ramp vehicles at B leave that freeway at C or

Then
D (a worst-case assumption).
Sw = 54.7 mph Segment 3, in this case, remains the same as previously, so no
additional analysis is required.
Snw = 62.7 mph Segment 4, however, should be analyzed as a Type B weaving
From Table 4-4, the minimum number of weaving lanes needed section, because one weaving movement is made with no lane
to support unconstrained operation is change and the other requires only one lane change. For Segment 4:
Nw = 2.19NVR0.571LH0.234/Sw0.438 VR = 900/3,600 = 0.25
Nw = 1.30 lanes R = 300/900 = 0.33
Nnw = 4 − 1.30 = 2.70 lanes Constant Sw Computation Snw Computation

a 0.100 0.020
Because Nw is less than the maximum value of 1.4 lanes given in
b 1.20 2.00
Table 4-4, the operation is unconstrained, and the computed speeds
c 0.77 1.42
are correct. Using the same process as that for Segment 2, the
d 0.50 0.95
following speeds and densities are calculated for Segment 3:
and
1,000
Dw = = 14.1 pc/mi/ln
1.3 × 54.7 SFF = 70 mph,
3,000 v = 3,600 pcph,
Dnw = = 17.7 pc/mi/ln N = 3 lanes, and
2.7 × 62.7
L = 2,500 ft.
(1,000 × 54.7) + (3,000 + 62.7)
S= = 59.5 mph
1,000 + 3,000 Then
4,000
D= = 16.8 pc/mi/ln Sw = 52.2 mph
4 × 59.5
Snw = 56.8 mph
A review of the LOS criteria in Table 4-6 indicates that all elements
of Segment 3 operate at LOS B. From Table 4-4, the number of weaving lanes required for un-
Segments 3 and 4. These segments should now be considered constrained operation is
a multiple weaving area, as shown in Figure 6-4. Again, it will be
6-6 freeways
Nw = N[0.085 + 0.703VR + (234.7/L) − 0.018(Snw − Sw)]

Nw = 0.81 lane
Nnw = 3 − 0.81 = 2.19 lanes
Because Nw is less than the maximum allowable value of 3.50
lanes (Table 4-4), the operation is unconstrained, and the computed
speeds are correct. Using the same process as that for Segments
2 and 3, the following densities and speed are calculated for Seg-
ment 4:
900
Dw = = 21.3 pc/mi/ln
0.81 × 52.2
Figure 6-5. Graphic representation of overall level of service.
2,700
Dnw = = 21.7 pc/mi/ln
2.2 × 56.8 has yielded varying results that tend to indicate that the extent of
(900 × 52.2) + (2,700 + 56.8) influence of any individual element can range from as little as
S= = 55.7 mph several hundred feet to more than a mile. Inasmuch as it is not
900 + 2,700
possible to exactly determine the extent of such impacts, weaving
3,600 and ramp junction areas that operate at levels of service poorer
D= = 21.5 pc/mi/ln than adjacennt segments should be viewed with caution because
3 × 55.7
they may affect the operation of upstream sections.
A review of the LOS criteria in Table 4-6 indicates that all elements A graphic technique presented in Figure 6-5 is useful as a tool
on Segment 4 operate at LOS C. to get a pictorial overview of overall operations. Levels of service
Given that all of the weaving areas and ramp junctions meet are plotted for each segment. The illustration shown clearly indi-
the minimum LOS criteria established for the design, the trial cates that the ‘‘bottleneck’’ or limiting segment is the weaving area
design of Figure 6-2 would appear to be acceptable for of segment 4. As long as the indicated operations hold, segment 4
implementation. will operate poorly, at level-of-service E, while other segments
could operate at levels B and C if not prevented from doing so by
spillback from segment 4. As noted previously, the effect of seg-
OPERATIONAL ANALYSIS ment 4 on upstream segments cannot be determined with certainty.
What can be said is that segment 4 should not have an extended
The analysis approach for total freeway evaluation is quite simi- effect as long as it does not break down; in other words, the
lar to the design approach, but is simpler in that there are no demand for segment 4 does not exceed its normal capacity.
alternates to consider. All volumes, geometrics, and traffic condi- If more demand is added, segment 4 would be the first to break
tions are known, and the freeway may be segmented with certain down—and segment 4 is the most susceptible to breakdowns
knowledge of ramp locations, weaving configurations, and other caused by incidents, weather, or other extraneous factors. Once
features. breakdown occurs here, the spatial and time extent of the break-
Once the freeway has been divided into uniform segments ac- down can be estimated using techniques detailed in the next
cording to the guidelines previously noted, the following computa- section.
tional sequence may be followed:
1. Determine the level of service for each potential basic free- ANALYSIS OF BREAKDOWN CONDITIONS
way segment using the procedures of Chapter 3.
2. Determine the level of service for each ramp junction, consid- The behavior of traffic streams during and immediately after
ering each ramp: the occurrence of a breakdown is not well understood. A critical
T As an isolated ramp. issue, however, is the rate at which vehicles can depart from a
T In conjunction with the adjacent downstream ramp. standing queue in an uninterrupted traffic stream. In many cases,
T In conjunction with the adjacent upstream ramp. vehicles are unable to depart from a standing queue at the normal
capacity rate of 2,400 pcphpl (see Figure 3-3). Many researchers
These checks are made using the procedures of Chapter 5. Ramps have noted that the relationships among speed, density, and flow
that are clearly part of a weaving configuration would not be may be discontinuous at the point of capacity and that the maxi-
examined using Chapter 5 procedures. mum rate of flow of vehicles departing from a queue may be less
3. Determine the level of service of each weaving and multiple than capacity under stable flow. Various observations of freeway
weaving segment using the procedures of Chapter 4. queue departure rates range from as low as 1,800 pcphpl to as
Where a given segment falls under several of these analyses, high as 2,400 pcphpl. Local driving characteristics have a major
the analysis resulting in the worst level of service is the controlling effect on this rate, which ranges from a significant reduction in
solution. capacity (compared with 2,400 pcphpl) of up to 25 percent to cases
Once the analysis of segments is complete, the overall interpreta- in which there is virtually no reduction.
tion of results is subject to the exercise of judgment. As was Where standing queues form because of incidents or permanent
presented in Chapter 3, there are general guidelines on the extent bottlenecks, a reduction in lane capacity results, which can have
of influence of weaving areas and ramp junctions. Other research a major effect on the extent of queueing and its dissipation.
freeway systems 6-7
and-go queues because of the 6,325 pcph demand. Capacity further

deteriorates to 3,600 pcph (assuming a drop to 1,800 pcphpl with
two lanes open). Thus, during the first 15 min, 6,325/4 = 1,581
pc arrive, only 3,600/4 = 900 pc are processed, and a queue of
681 pc is formed behind the blockage.
2. After the blockage is removed, capacity improves to 1,800
× 3 = 5,400 pcph because standing queues still exist. Full capacity
cannot be regained until all queues have dissipated. Thus, in the
ensuing 45 min, 6,325 × 3/4 or 4,744 pc arrive and 5,400 × 3/4
or 4,050 pc are processed. The queue continues to build to 681 +
4,744 − 4,050 = 1,375 pc.
3. During the second hour, 5,400 pc arrive and exactly 5,400
pc are processed. The queue is stable, but it does not dissipate.
4. Thereafter, the queue will dissipate because 3,900 pcph arrive
and 5,400 pcph may be processed. The 1,375 queued vehicles
dissipate in 1,375/(5,400 − 3,900) = 0.92 hr, and full capacity is
restored some 2.92 hr after the occurrence of a 15-min blockage.
The queue length (assuming three lanes and 40 ft per vehicle)
reached (1,375/3) × 40 = 18,333 ft, or more than 3 mi at its peak,
which lasted for one full hour.
Figure 6-6 illustrates this analysis in graphic form. The illustra-
tion here is extreme, using the assumed queue discharge rate of
1,800 pcphpl for computational simplicity. In many areas, this
Figure 6-6. Effects of breakdown illustrated.
value will be exceeded. Nevertheless, the expanded time and spa-
tial effects of a breakdown are clearly indicated, as is the need
Consider the case illustrated in Figure 6-6: a three-lane freeway to consider potential incidents in the analysis of freeway system
segment operating under ideal conditions with a demand of 6,325 operation. The value of 40 ft per queued vehicle is approximate
pcph during the peak hour, 5,400 pcph during the hour after the and is based on the assumption of stop-and-go movement within
peak, and 3,900 pcph thereafter. What will occur if an incident the queue.
blocks one lane for 15 min at the beginning of the peak period? This technique is approximate and does not account for many
For illustrative purposes, it is assumed that the formation of a microscopic properties of unstable freeway flows. Nevertheless, it
is useful in estimating the effect of a breakdown in one location
standing queue reduces the lane capacity to 1,800 pcphpl.
on overall operations. However, since the queue discharge rate
The following operational effects should be anticipated:
varies widely depending on local conditions, such an analysis
1. When blockage occurs, capacity immediately drops from should be coordinated with sample field measurements of an ap-
7,200 pcph to 4,800 pcph or lower, which quickly creates stop- propriate discharge rate.
III. FREEWAY SURVEILLANCE AND CONTROL
A complete treatment of this subject is beyond the current scope breakdown (by restricting entries at appropriate locations), or it
of this manual but there are excellent references on the subject. can be controlled so that some specified level of service is main-
The interested reader is referred to a state-of-the-art report by tained. In the latter case, one may specify ramp metering rates in
FHWA (1) and to NCHRP Report 232 (2). anticipation of future growth in demand. Thus, freeway manage-
ment can be used at the planning stage, and not simply as an
BACKGROUND operational correction. It is rare to implement a control scheme
that diverts vehicles from the freeway to maintain a level of service
It is important to recognize that freeway surveillance and control better than E.
is employed relatively commonly and that it has a number of A freeway management system may be planned, or it may be
potential advantages. Some of the key potential advantages are responsive to traffic variations. Further, it may or may not show
an explicit response to incidents.
T Relief of congestion by virtue of exercising control over ex-
cessive entries.
T Decrease in delay, for the same reasons. CONTROL ELEMENTS
T Protection of level of service.
T Response to freeway incidents. The principal elements that are added to the facility because of
a surveillance and control/management effort are
There is an interesting distinction between the first and third items:
a freeway can be controlled with a single objective—to avoid T Vehicular detectors.
6-8 freeways
Figure 6-8. Plot of cumulative ramp demand and output.
serviced curves is the delay per vehicle, and the vertical distance
between the curves is the queue length. From Figure 6-8, the
maximum delay per vehicle would occur at 5:51 PM, and would
Figure 6-7. Illustration of a ramp-metering need.
be approximately 5 min. The queue length at this time would be
about 50 vehicles.
T Ramp metering. It should be noted here, however, that many drivers will be
T Video and/or other observation. unwilling to accept 5-min delays. (In Los Angeles, 1- or 2-min
T Control policies, implemented by central computer or other delays are the average usually observed.) Many of the queued
hardware. vehicles might be expected to seek alternate routings to avoid
T Static and perhaps variable message signing to inform motor- the delay. Thus, a critical consideration in ramp metering is the
ists of alternate routes and/or conditions. availability of alternate routes and the impact of diverted traffic
on those routes.
Of these elements, ramp metering is the most essential because
It should also be noted that some freeway management systems
it is the most positive control action exercised. Chapter 5 addresses
operate on nothing more than application of the above principle
the lack of detailed knowledge on Lane 1 flow effects of metering,
in a consecutive set of freeway segments: the section input is
but its known advantages in control are in smoothing out disruptive
monitored; the segment capacity is known; the ramp input is not
arrival platoons. It is useful to consider an illustration of the ramp
allowed to cause mainline flow to exceed capacity.
and mainline effects of a metered ramp in order to make that
discussion meaningful.
Consider the situation in Figure 6-7: an on-ramp has the demand DETERMINATION OF PROBLEMS AND CONTROL
depicted ranging from 250 to 575 veh/hr (flow rate); the mainline
has 3,500 veh/hr already, with a capacity of 4,000 veh/hr. Clearly, Freeway management is more frequently motivated by opera-
if the ramp demand is allowed to enter, an LOS F situation will tional problems: one or more sections are bottlenecks, with signifi-
occur upstream of the ramp. How may the ramp be metered to cant mainline congestion occurring. The problem is then to
avoid this? What delay and queue will occur at the ramp because alleviate the congestion and to maintain a level of service better
of this? than F. In some cases, the project includes construction at some
The ramp must be metered at 500 veh/hr to avoid exceeding locations to provide additional capacity or includes the incorpora-
capacity on the mainline. This means one vehicle every (3,600/ tion of high-occupancy-vehicle lanes.
500) = 7.2 sec. With a green-red signal at the ramp, this would Although an entire treatment of freeway management is not
usually mean 2 sec of green followed by 5.2 sec of red. This cycle appropriate in the present context, two problem areas deserve spe-
may be implemented in a number of ways, including a conven- cial mention: hidden bottlenecks and origin-destination patterns.
tional electromechanical controller, another local controller (possi- Figure 6-9 depicts a hypothetical freeway with five sections
bly a microprocessor), or a command from a remote computer. and with the input demands shown. Clearly, demand will exceed
From Figure 6-7, ramp demand reaches the 500-veh/hr level at capacity in Segment 3, and LOS F will result. Stop-and-go opera-
approximately 5:09 PM and does not decrease below that level tion can occur in all upstream sections, depending on the duration
again until 5:51 PM. In the interim, a queue will form and continue over which demand exceeds capacity (i.e., over which the conges-
to enlarge, as illustrated in Figure 6-8. tion has a chance to spread).
Figure 6-8 is a plot of ramp vehicles versus time. At any given In practice, the capacities are not computed and one simply
time, the horizontal distance between the demand and vehicles observes severe congestion in Segment 2, caused by Segment 3.
freeway systems 6-9
The congestion may spread to Segment 1 if the peak period is

long enough or if Segment 2 is short.
Assume that some physical reconstruction, perhaps coupled with
a decrease in the 5,300-pcph input via ramp metering further up-
stream, alleviates the problem. Lacking the capacity figures for all
sections, one may overlook the fact that if Segment 3 now outputs
a flow rate higher than 5,200 pcph, a bottleneck will appear at
Segment 5 for the first time. It was always there, but only the
solution of the Segment 3 problem allowed the demand to attain
levels necessary to exhibit it. That is, it was ‘‘hidden’’ by the Figure 6-9. Potential for hidden bottlenecks.
upstream bottleneck in Segment 3.
A complete capacity analysis of the facility should be conducted
to avoid the ‘‘hidden bottleneck’’ problem. In doing so, changes
in flow due to the improvements must be anticipated. For instance,
is the off-ramp in Segment 4 shown at a level of 300 pcph because
it is the true demand or because it is the observed amount that
could get past the original bottleneck? In addition, it must be
recognized that the service flow rates in some sections (e.g., weav- Figure 6-10. Phases of a traffic incident.
ing sections) are functions of the traffic mix, which may change.
Because the flow pattern may be distorted, it is important to
have some knowledge of the origin-destination pattern of traffic. Certainly incident response is desired in order to provide assistance
Further, the origin-destination pattern influences what can be done to the motorists involved (tow, medical, police) as the need arises.
and what should be done. Consider a freeway on which virtually Incident response can also be directed to minimizing the impact
all the outlying ramp entries stay on the facility until it terminates on other vehicles and to recovering use of the facility.
in the downtown area. Consider an identical physical facility on One study (3) showed that an incident removed to the shoulder
which the traffic consists of many short trips, but much outlying on a three-lane facility still reduced capacity by one-third; a single-
traffic exists before another ‘‘layer’’ of traffic enters. The control lane blockage reduced capacity by 50 percent; a two-lane blockage
opportunities and the equity of various control options vary radi- reduced capacity by 79 percent. In addition to the magnitude of
cally between these two extremes. the impact, the duration must also be considered. Refer to Figure
6-10, which identifies four critical phases of an incident. Analo-
gous to the ramp metering illustration under the section ‘‘Control
INCIDENTS
Elements,’’ the effect can persist long after the incident itself is
removed because of the backups created. At one facility (4), it
Incidents occur relatively commonly on traffic facilities, al-
was estimated that peak-period incidents were responsible for more
though it is standard practice to design to a level of service for
delay than recurrent peak period congestion at the location in
the nonincident condition. Clearly, incidents require attention be-
question.
cause they
Incidents may be detected by video observation, audio reports
T Disrupt the level of service being provided. (call boxes, CB), or roadway sensors. Incident response may be
T Reduce capacity radically. by some combination of required assistance, ramp restrictions or
T Present hazards to motorists, particularly those directly closure, and alternate-route advisories. The control actions may be
involved. preplanned or dynamic decisions.
IV. CAPACITY OF FREEWAY WORK ZONES
One of the more frequently occurring disruptions to traffic and observations cited herein must be taken as averages subject
flow on freeways is the required maintenance operations that to some variation.
must take place periodically, either as part of regular maintenance
programs or to correct physical defects of the roadway, roadside,
or supporting structures. Assessment of capacity is a necessary OBSERVED WORK-ZONE CAPACITIES
part of the planning of traffic control strategies during mainte-
nance operations if severe disruptions and delays to traffic are Figure 6-11 shows the range of capacities measured at several
to be avoided. This section details the results of several work- work sites in Texas with an active work crew at the site. The observa-
zone capacity studies that provide considerable insight (5–7). tions are taken to be approximate capacities, because continuous
It should be noted that work-zone capacities will vary depending queues of vehicles were present upstream of the sites included.
on the exact nature of the work being done, the number and size The designation (A,B) is used to identify the various lane closure
of equipment at the site, and the exact location of equipment and situations evaluated. ‘‘A’’ represents the normal number of lanes
crews with respect to moving lanes of traffic. Thus, the criteria in one direction, and ‘‘B’’ represents the number of lanes open
6-10 freeways
observations versus the type of maintenance operations for

informational purposes. Note that flow through the work zone
is also affected by the presence of merging, diverging, or weaving
movements; grades; alignment; trucks; and other factors. The
data in Table 6-2 reflect studies in both Texas and California.
California observations represent peak flow rates, whereas the
Texas data reflect full-hour capacities.
LONG-TERM CONSTRUCTION SITES
Table 6-3 illustrates the results of 10 studies of locations with

long-term, more permanent types of construction operations in prog-
ress. Note that the capacities at such sites are higher than those for
more temporary disruptions primarily because of the use of perma-
nent barriers and other controls and the dissipation of rubbernecking
as drivers become familiar with the site.
SHORT-TERM MAINTENANCE SITES
A study was conducted in Houston, Texas, in which the right

Figure 6-11. Range of observed work-zone capacities—work crew two lanes of a four-lane section were closed to traffic. There was
at site (9). no work activity, however, in the lane immediately adjacent to
moving traffic. In effect, the closure included one full buffer lane
between traffic and maintenance operations. Although capacity
during maintenance operations. Table 6-1 gives the average capac- operations were not observed, capacity of the location was esti-
ity for each closure situation studied. mated to be about 1,800 vphpl, considerably larger than a standard
Average open-lane capacities for (4,2), (3,2), and (4,3) closures (4,2) closure with work activities taking place in the lane adjacent
are approximately 1,500 vphpl. For (5,2) and (2,1) closures, the re- to moving traffic.
ductions are more severe, in the range of 1,350 vphpl. The capacities
of (3,1) closures were the most damaging, averaging only 1,170 SHOULDER USE AND TRAFFIC SPLITTING ON
vphpl. THREE-LANE SEGMENTS
Figure 6-12 shows the cumulative distribution of the observed
work-zone capacities, the function of which is to assist analysts Generally, when work is required on the middle lane of a three-
in identifying the risks in using certain capacity values for given lane section, both the middle and one of the exterior lanes are
lane closures. For example, the 85th-percentile capacity for a closed. Table 6-1 indicates that the average capacity of a single
(3,1) closure is only 1,030 vphpl. The average capacity for this open lane is 1,170 vphpl. Several studies have indicated that this
situation (1,170 vphpl) occurs at the 58th percentile. Thus, on capacity can be increased to 3,000 veh/hr by using a traffic control
the basis of the observed range of values, use of the average approach called ‘‘shifting,’’ in which drivers are encouraged to
value in analysis leads to an overestimate of capacity (and use the shoulder as an additional traffic lane, thereby leaving two
consequently, an underestimate of queues and delays) in 42 effective lanes for traffic movement. Shifting is generally accom-
percent of the cases to which it is applied. Given the variation plished through the use of traffic cones directing drivers onto the
in observed capacities, analysts may wish to use 85th or higher shoulder and adjacent shoulder lane.
percentile values rather than averages to reduce the risk of This same capacity could be achieved using the ‘‘splitting’’
capacity overestimates. approach, in which only the middle lane is closed and traffic is
Because of the limited amount of data available, it is not permitted to move on both sides of the work activity. Since such
possible to statistically correlate capacity with the particular type an operation is often confusing to drivers, a control approach is
of road work taking place. However, Table 6-2 gives individual recommended in which the left lane is closed as much as 1,000
to 1,500 ft upstream of the site. Thus, only two lanes approach
the site. At the maintenance zone, cones are used to direct one
Table 6-1. Measured Average Work-Zone lane to the left and one lane to the right of the closed middle lane.
Capacities (8)
number of lanes number
A B of average capacity LANE NARROWING
normal open studies (veh/hr) (vphpl)
3 1 7 1,170 1,170 A study in Houston considered the effect of lane narrowing with-
2 1 8 1,340 1,340 out closures due to maintenance or construction operations. The sub-
5 2 8 2,740 1,370 ject sites included lane-width reductions to 10 and 11 ft, and portable
4 2 4 2,960 1,480 concrete barriers were used to separate moving traffic from work
3 2 9 2,980 1,490 operations. Capacities in the range of 1,800 vphpl were observed at
4 3 4 4,560 1,520 these sites, which included both three- and four-lane segments.
freeway systems 6-11
Figure 6-12. Cumulative distribution of observed work-zone capacities (9).
Table 6-2. Summary of Observed Capacities for Some Typical Operations

no. of lanes in one directiona
3 2 5 3 or 4 4
type of work 1 1 2 2 3
Median barrier/guardrail — 1,500 — 3,200 4,800
Installation/repair 2,940b 4,570b
Pavement repair 1,050b 1,400 — 3,000 4,500
2,900b
Resurfacing, asphalt removal 1,050b 1,200 2,750b 2,600 4,000
1,300b 2,900b
Striping, slide removal — 1,200 — 2,600 4,000
Pavement markers — 1,100 — 2,400 3,600
Bridge repair 1,350b 1,350b — 2,200 3,400
NOTE: Adapted from paper by Dudek and Richards (9); all values in vehicles per hour.
a
Top row, during normal operation; bottom row, during work-zone operation.
b
Texas data, full-hour capacities; all other data are from California and are expressed as peak flow rates.
ESTIMATING QUEUE LENGTH AND DELAY where
Figure 6-6 illustrates a graphic technique for estimating queue Lt = length of queue, in ft;
buildup and delays for breakdown conditions. This same technique Qt = number of vehicles in queue at time t;
can be applied to work zones where arrival or demand flows exceed N = number of open lanes upstream of the site; and
the capacity of the work zone for some period of time. In particular, ø = average length of vehicle.
the length of the queue may be estimated as
Qt × ø The value of Qt would be found using the graphic technique
Lt = (6-1)
N illustrated in Figure 6-6.
6-12 freeways
SAMPLE CALCULATION In Figure 6-13, work is assumed to begin at 9:00 AM. The
estimated queue length at 1:00 PM, 4 hr after the beginning of
Consider the case of a maintenance operation requiring the closure work, and the time work is assumed to stop is 2.1 mi based on
of the median lane of a three-lane freeway segment. The work will the average capacity of 3,000 veh/hr. This, however, is a 58th-
require 4 hr to complete, including the installation and removal of percentile value. Thus, the queue would be longer than this value
traffic control devices. Data obtained from a nearby traffic counter 42 percent of the time. If the 85th-percentile capacity is used, the
during the previous two weeks were used to estimate the following queue reaches 2.9 mi but would be exceeded only 15 percent of
demand pattern: the time. The 100th-percentile queue length reaches 3.5 mi, which
is not expected to be exceeded under most circumstances.
Time Period Volume Anticipated (veh/hr)
Clearly, such a backup would be most undesirable, and other
9:00 to 10:00 am 2,920
options would be explored in terms of the work-zone operations,
10:00 to 11:00 3,120
including the following:
11:00 to 12:00 noon 3,200
12:00 to 1:00 pm 3,500 1. Work on Saturday or Sunday if volumes are lower then.
1:00 to 2:00 3,830 2. Perform the work at night.
2:00 to 3:00 3,940 3. Reduce the work time or split the work into two shifts.
3:00 to 4:00 4,620 4. Implement additional traffic control strategies.
4:00 to 5:00 5,520
Curves similar to those in Figure 6-13 could be developed for
Referring to Table 6-1 and Figure 6-11, it is seen that the average weekend or night volume conditions. A review of Figure 6-13 also
capacity for a (3,2) work-zone configuration is 1,500 vphpl or indicates that queues could be greatly reduced if the work could
3,000 veh/hr. The 85th-percentile capacity is 1,450 vphpl or 2,900 be accomplished in 3 hr or less. At average capacity, the queue
veh/hr, and the 100th-percentile capacity is 1,420 vphpl or 2,840 after 3 hr would be only 0.8 mi, considerably less than the 2.1-mi
veh/hr. With these assumed capacity values, Figure 6-13 graphs queue that develops after 4 hr. If the work could be divided into
the queue build-up and delays. two 2-hr shifts on two separate days, the queue (at average capac-
ity) would be limited to about 0.5 mi.
Table 6-3. Capacity of Long-Term Construction Sites Other traffic control strategies might include closing of
with Portable Concrete Barriers (8) on-ramps upstream of the site to reduce demand or directing vehi-
cles to use the shoulder past the work zone. The latter strategy
number of number capacity
lanes avg. capacity would add up to 1,500 veh/hr of additional capacity. The issue of
of range ramp closures, however, would have to be carefully considered in
normal open studies (vphpl) veh/hr vphpl
terms of where diverted vehicles would go and what their impact
3 2 7 1,780–2,060 3,720 1,860 on traffic along diversion routes would be. Ramp closures would
2 1 3 — 1,550 1,550 also have to be carefully signed to eliminate driver confusion.

Figure 6-13. Sample calculation—queue analysis for a work zone (9).
V. WEATHER
The capacity of freeway systems is also affected by weather. Another found a typical figure of 8 percent for rain (12), although
The most extreme case is represented by heavy snowfalls that much variation was observed. Indeed, the substantial variations
cause multiple lane closings. However, a variety of weather condi- due to the intensity of the weather condition and the specifics of
tions—rain, snow, fog, glare, and others—affect capacity without the location are entirely rational. It is most important to recognize
such dramatic evidence of their existence. that 10 to 20 percent reductions are typical, and higher percentages
Quantitative information is sparse, but some indications do exist: are quite possible. These effects must be considered in facility
one study found that rain reduced capacity by 14 percent (10,11). design, particularly when adverse conditions are common.

6-14 freeways
VI. HIGH-OCCUPANCY-VEHICLE LANES ON FREEWAYS
The existence of exclusive high-occupancy-vehicle (HOV) freeway a four-lane one, and so on. The effect of cones or other
lanes on freeways raises two issues: (a) what their capacity and dividers may be estimated by treating them as lateral obstructions
what the operating characteristics of such lanes are and (b) what at the roadside edge. Depending on their placement, they may also
effect their presence has on the operation of the remainder of have the effect of narrowing the lane.
the freeway. In contraflow lanes, this latter effect is marked, because vehicles
shy away from the imposing opposite flow of large vehicles at
relatively high speeds. In some instances, an entire adjacent lane
CAPACITY ANALYSIS FOR HOV LANES is taken out of service to act as a buffer zone.
This issue is quite complex. HOV lanes come in many forms,

including SAMPLE CALCULATION
T Exclusive bus lanes. Figure 6-14 illustrates a problem using this estimating technique.
T Exclusive bus and taxi lanes. The problem is to analyze the impact of a proposed contraflow
T Exclusive bus and carpool lanes, with varying occupancy lane on level of service in the direction from which the lane is
restrictions. taken and on the concurrent direction of flow.
T Exclusive bus-taxi-carpool lanes.
In addition, each type may be implemented as a contraflow Before HOV Lane Is Initiated
lane, with the exclusive lane taken from the opposing freeway
lanes, or as a concurrent-flow lane, in which the lane is taken Primary flow is 5,100 + (1.5 × 300) = 5,550 pcph or 1,850
from freeway lanes in the same direction of flow. HOV lanes pcphpl. For 12-ft lanes with no lateral obstructions, the free-flow
are adopted to provide for smooth and speedy flow of passengers speed is estimated according to Equation 3-3 as FFS = 70 − 0 −
in vehicles using the lanes, and they are used to circumvent 0 − 3.0 − 0.0 = 67 mph.
freeway segments operating at or near breakdown conditions. From Figure 3-2, for FFS = 67 mph and flow rate = 1,850
The contrast of high-occupancy vehicles progressing smoothly pcphpl, the passenger-car speed is 65 mph and the density is
while other vehicles are mired in heavy congestion is also 1,850/65 = 28.5 pc/mi/ln (LOS D).
intended to act as an inducement to motorists to abandon their Contraflow is 2,800 pcph in three lanes, or 933 pcphpl. The
car for a bus or carpool. prevailing free-flow speed is 67 mph.
Thus, it is not practical for such a lane to operate at or near From Figure 3-2, the passenger-car speed is 67 mph and the
capacity or at a poor level of service. To do so would defeat density is 933/67 = 13.9 pc/mi/ln (LOS B).
its function and purpose. The issue of the ‘‘capacity’’ of such
lanes is therefore highly speculative, because few (if any) existing
lanes approach this condition at any time. Chapter 12 provides After HOV Lane Is Initiated
guidelines and LOS criteria for HOV lanes, based primarily on
the work of Levinson (13–15). In this section an attempt is Primary flow is 5,100 pcph in three lanes, or 1,700 pcphpl, with
made to provide a general framework for defining the impacts prevailing FFS = 67 mph.
of such a lane on freeway operations. Numerous studies of From Figure 3-2, passenger-car speed is 66 mph and the density
existing operations (16–26) may also be used for general insight is 1,700/67 = 25.4 pc/mi/ln (LOS D).
on the subject. Contraflow = 2,800 pcph in two lanes, or 1,400 pcphpl.
In this case, significant changes have taken place in the off-
peak direction. It is assumed that the average lane width drops
EFFECT OF HOV LANES ON FREEWAY OPERATIONS to 11 ft because of divider placement. Further, the divider is an
obstruction (on the left side) that abuts the lane edge (i.e., there
The existence of an HOV lane on a freeway influences the is zero lateral clearance). Finally, the number of through traffic
operation of remaining freeway lanes in three ways: lanes is reduced to two lanes in that direction. These changes cause
1. A lane is removed from one direction of flow (occasionally a reduction in the prevailing free-flow speed and are reflected in
two are removed, the second being used as a buffer lane). Equation 3-3 as follows: FFS = 70 − 2.0 − 3.6 − 4.5 = 60 mph.
2. Cones or other devices used to demark the lane (where used) From Figure 3-2, with FFS = 60 mph and a flow rate of 1,400
pose lateral obstructions to flow in the adjacent lane if a buffer pcphpl, the passenger-car speed is computed as 60 mph and the
lane is not provided. density is 1,400/60 = 23.3 pc/mi/ln (LOS C).
3. The movement of vehicles into or out of the HOV lane may On the basis of this approximate analysis, the creation of the
be disruptive to other traffic. new lane improves flow in the concurrent direction by removing
buses from the stream. The level of service remains at D, but
Unfortunately, there is no meaningful body of data that has the average running speed increases from 65 to 67 mph, which
quantified these effects. Estimates of the first two effects can, saves each vehicle (5/65 − 5/67)60 = 0.14 min, or 8 sec.
however, be made using techniques presented in Chapter 3. The level of service on the freeway in the reverse direction
The removal of a lane is simply handled by assuming that the decreases from B to C, and speed from 67 to 60 mph, causing
eight-lane freeway becomes a six-lane freeway, and the six-lane each vehicle to lose (5/60 − 5/67)60 = 0.52 min, or 31 sec.
Figure 6-14. Example for analysis of HOV lane impact.
Although not totally definitive, this approximate technique is evaluated in light of the benefits and costs of the HOV lane itself
useful in evaluating the gross effects of HOV-lane implementation and related issues.
on remaining freeway flows. These impacts would have to be
VII. SUMMARY
The freeway is a complex facility made up of many component operational components that may affect overall capacity and level
segments and sections, each having a potential impact on opera- of service. The techniques presented should be considered approxi-
tions in upstream and downstream segments. This chapter has mate and serve primarily to indicate the relative magnitude of
attempted to identify these impacts, as well as various system various operational impacts.
VIII. REFERENCES
1. Everall, P.F., Urban Freeway Surveillance and Control. 7. Abrams, C., and Wang, J., ‘‘Planning and Scheduling Work
FHWA, USGPO Stock No. 5001-00058 (June 1973). Zone Traffic.’’ Report No. DOT-FH-11-9417, Federal High-
2. Blumentritt, C.W., et al. NCHRP Report 232: Guidelines way Administration, Washington, D.C. (Oct. 1980).
for Selection of Ramp Control Systems. (May 1981), 108 pp. 8. Dudek, C., Notes on Work Zone Capacity and Level of Ser-
3. Goolsby, M.E., ‘‘Influence of Incidents on Freeway Quality vice. Texas Transportation Institute, Texas A&M University,
of Service.’’ Presented at 50th TRB Annual Meeting College Station, Texas (1984).
(Jan. 1971). 9. Dudek, C., and Richards, S., ‘‘Traffic Capacity through Ur-
4. McDermott, J.M., ‘‘Automatic Evaluation of Urban Freeway ban Freeway Work Zones in Texas.’’ Transportation Research
Operations.’’ Traffic Engineering (Jan. 1968). Record 869 (1982).
5. Memmott, J., and Dudek, C., ‘‘A Model to Calculate the 10. Jones, E.R., and Goolsby, M.E., ‘‘The Environmental Influ-
Road User Costs at Work Zones.’’ Report No. FHWA/TX- ence of Rain on Freeway Capacity.’’ Highway Research
83/20 + 292-1, Texas Transportation Institute, Texas A&M Record 321, Highway Research Board, Washington, D.C.
University, College Station, Texas (Sept. 1982). (1970).
6. Dudek, C., and Richards, S., ‘‘Traffic Capacity Through 11. Jones, E.R. and Goolsby, M.E., Effect of Rain on Freeway
Work Zones on Urban Freeways.’’ Report No. FHWA/TX- Capacity. Texas Transportation Institute, Research Report No.
81/28 + 228-6. Texas Transportation Institute, Texas A&M 14-23. Texas A&M University (Aug. 1969).
University, College Station, Texas (Apr. 1981).
6-16 freeways
12. Kleitsch and Cleveland, D.E., The Effect of Rainfall on 20. Miller, G.K., and Goodman, K.M., The Shirley Highway
Freeway Capacity. Highway Safety Research Institute, Report Express-Bus-on-Freeway Demonstration Project. Technical
Tr S-6. University of Michigan, Ann Arbor (1971). Analysis Division, National Bureau of Standards, UMTA,
13. ‘‘Transit.’’ TRB Circular 212: Interim Materials on Highway USDOT (1972).
Capacity, Transportation Research Board, Washington, D.C. 21. ‘‘Evaluation of the Shirley Highway Express-Bus-on-Freeway
(Jan. 1980). Demonstration.’’ Final Report, Urban Mass Transportation
14. Levinson, H.S., ET AL., NCHRP Report 143: Bus Use of High- Administration (Aug. 1975).
ways—State of the Art (1973), 406 pp. 22. ‘‘Operation and Management of the Shirley Highway Express-
15. Levinson, H.S., Adams, C.L., and Hoey, W.F., NCHRP Re- Bus-on-Freeway Demonstration Project.’’ North Virginia
port 155: Bus Use of Highways—Planning and Design Guide- Transportation Commission (Sept. 1976).
lines (1975), 161 pp. 23. ‘‘Evaluation of the Kalanianaole Highway Carpool/Bus
16. Interstate 495—Exclusive Bus Lane. Urban Corridor Demon- Lane.’’ Report No. FHWA-RD-77-100 (Aug. 1977).
stration Program, Tri-State Regional Planning Commission, 24. ‘‘Evaluation of the Moanolua Freeway Carpool/Bus Lane.’’
USDOT FH-11-7646 (July 1972). Report No. FHWA-RD-77-99 (Aug. 1977).
17. Report of the Exclusive Bus Lane Demonstration on the South- 25. ‘‘Traffic Control of Carpools and Buses on Priority Lanes on
east Expressway. Bureau of Traffic Operations, Massachusetts
Interstate 95 in Miami.’’ Draft Final Report, Federal Highway
Department of Public Works (1971).
Administration (Aug. 1977).
18. Exclusive Bus Lane Study. Report on Second Phase of Field
26. Various reports by the Texas Transportation Institute on
Tests for the 195-Rte 3 Bus Lane to the Lincoln Tunnel, Port
HOVL projects. Report Nos.: TTI-2-10-74-205-4; TTI-2-10-
of New York Authority (Mar. 1966).
74-205-5; TTI-2-10-74-205-1.
19. Vuchic, V.R., and Stranger, R.M., ‘‘Lindenwold Line, Shir-
ley Busway: A Comparison.’’ Highway Research Record 459,
Highway Research Board, Washington, D.C. (1973).

chapter 7
MULTILANE RURAL AND SUBURBAN HIGHWAYS
CONTENTS
i. introduction .......................................................................................................................................................................... 7-2

Characteristics of Multilane Highways................................................................................................................................. 7-2
Relationship Between Highway Types................................................................................................................................. 7-2
Free-Flow Speed.................................................................................................................................................................... 7-3
Speed-Flow and Density-Flow Relationships....................................................................................................................... 7-3
Ideal Conditions..................................................................................................................................................................... 7-4
ii. free-flow speed adjustments ............................................................................................................................................. 7-5
Speed Enforcement................................................................................................................................................................ 7-5
Design Speed ......................................................................................................................................................................... 7-5
Speed Limit............................................................................................................................................................................ 7-6
Lane Width and Lateral Clearance ....................................................................................................................................... 7-6
Median Type.......................................................................................................................................................................... 7-6
Access Points......................................................................................................................................................................... 7-6
Other Adjustments................................................................................................................................................................. 7-6
iii. volume factors ..................................................................................................................................................................... 7-7
Peak-Hour Factor................................................................................................................................................................... 7-7
Heavy-Vehicle Factor............................................................................................................................................................ 7-7
iv. methodology.......................................................................................................................................................................... 7-7
Free-Flow Speed.................................................................................................................................................................... 7-7
Determination of Free-Flow Speed....................................................................................................................................... 7-9
Field Measurement........................................................................................................................................................... 7-9
Guidelines for Estimating Free-Flow Speed ................................................................................................................... 7-9
Determination of Flow Rate.................................................................................................................................................. 7-11
Peak-Hour Factor ............................................................................................................................................................. 7-12
Adjustment for Presence of Heavy Vehicles .................................................................................................................. 7-12
Volume Distribution by Lane ............................................................................................................................................... 7-14
Determination of Level of Service ....................................................................................................................................... 7-14
Segmenting the Roadway...................................................................................................................................................... 7-14
v. procedures for application ................................................................................................................................................ 7-15
Data Requirements ........................................................................................................................................................... 7-15
Segmenting the Roadway ................................................................................................................................................ 7-15
Computational Steps ........................................................................................................................................................ 7-16
Interpretation of Results................................................................................................................................................... 7-16
Design Analysis..................................................................................................................................................................... 7-16
Relationship to AASHTO Design Criteria...................................................................................................................... 7-18
Segmenting the Roadway ................................................................................................................................................ 7-18
Planning Analysis .................................................................................................................................................................. 7-18
Signalized Intersections on Multilane Highways ................................................................................................................. 7-19
Three-Lane Highways with Permanently Assigned Third Lanes ........................................................................................ 7-19

7-2 rural and suburban highways
vi. sample calculations ............................................................................................................................................................ 7-21

Calculation 1—Operational Analysis of an Undivided Highway........................................................................................ 7-21
Calculation 2—Operational Analysis of a Divided Highway.............................................................................................. 7-21
Calculation 3—Design of a Multilane Highway.................................................................................................................. 7-24
Calculation 4—Design Analysis of an Existing Multilane Roadway ................................................................................. 7-24
Calculation 5—Planning Analysis for a New Roadway...................................................................................................... 7-30
vii. acknowledgments................................................................................................................................................................. 7-30
viii. references .............................................................................................................................................................................. 7-30
APPENDIX I. Figures and Worksheets for Use in the Analysis of Multilane Highways ............................................................. 7-31
ADDENDUM. Adjustment for Driver Population ............................................................................................................................ 7-37
I. INTRODUCTION
The procedures in this chapter are designed to analyze the capac- multilane suburban and rural highways from freeways are as
ity, level of service, lane requirements, and impacts of traffic and follows:
design features of rural and suburban multilane highways. The
procedures are not applicable at points along highways at which T Vehicles may enter and leave the highway at intersections
there are traffic signals, but can be used to analyze sections of the and driveways, and through the median at selected points.
highway between widely spaced signals, where platooning from T Traffic signals may be located on this type of facility.
the upstream signal does not affect flow conditions. T The general design standards of multilane highways tend to
be lower than those found on freeways, although an ideal multilane
highway approaches freeway conditions as access points and turn-
CHARACTERISTICS OF MULTILANE HIGHWAYS ing volumes approach zero.
T The visual setting and developed frontage along multilane
Multilane highways, which are analyzed using procedures in highways have more impact on drivers than do the development
this chapter, generally have posted speed limits of between 40 and and location of such features along freeways.
55 mph. They usually have four or six lanes, often with physical
medians or two-way left-turn-lane (TWLTL) medians, although When compared with urban arterials, the multilane highway is
they may also be undivided. Multilane highways are typically similar in many respects but lacks the regularity of traffic signals
located in suburban communities leading to central cities or along and tends to have greater control on the number of access points
high-volume rural corridors that connect two cities or significant
per mile. Also, design standards are generally higher than those
activities generating a substantial number of daily trips. Traffic
used on urban arterials. Speed limits on multilane highways are
signals may be found along such highways, although traffic signals
often 5 to 15 mph higher than speed limits on urban arterials.
spaced at 2.0 mi or less typically create urban arterial conditions.
Pedestrian activity, as well as parking, is minimal when compared
Traffic volumes on multilane highways widely vary but might
with that on urban arterials.
typically range from 15,000 to 40,000 vehicles per day (vpd). In
Multilane highways differ substantially from two-lane high-
some cases, volumes as high as 100,000 vpd have been observed
ways, principally because of the ability of a driver on a multilane
when access across the median is severely restricted and when all
major crossings are grade separated. highway to pass slower-moving vehicles without using lanes desig-
Illustrations 7-1 through 7-4 show typical multilane highways. nated for oncoming traffic. Multilane highways also tend to be
located adjacent to urban areas or to connect urban areas and
often have better design features, such as horizontal and vertical
RELATIONSHIP BETWEEN HIGHWAY TYPES curvature, compared with two-lane highways.
The methodology described in this chapter is intended solely
Multilane highways in suburban and rural settings have different for uninterrupted-flow segments with access at cross streets and
operational characteristics than do freeways, urban arterials, and with direct access from adjacent properties. Chapter 11 contains
two-lane highways. Multilane highways are not completely access the methodology suited to facilities distinguished by one or more
controlled. At-grade intersections and, occasionally, traffic signals of the following factors:
are found along these highways. In addition, friction created by
opposing vehicles on undivided multilane highways and the impact T Any segment influenced significantly by upstream and down-
of access to roadside development contribute to a different opera- stream signals (this is indirectly indicated by a signal spacing of
tional setting than that found on freeways. Multilane highways 2.0 mi or less),
span the range between the uninterrupted-flow conditions found T Any significant presence of on-street parking,
on freeways and the flow conditions on urban arterials, which are T Presence of bus stops that have significant use, or
frequently interrupted by signals. The key factors that distinguish T Significant pedestrian activity.
multilane rural and suburban highways 7-3
Illustration 7-1. Divided multilane highway in a rural Illustration 7-2. Divided multilane highway in a suburban
environment. environment.
Illustration 7-4. Undivided multilane highway in a suburban

environment.
Illustration 7-3. Undivided multilane highway in a rural
environment. of capacity and level of service under uninterrupted-flow condi-
tions using the procedures in this chapter.
Several methods of determining free-flow speed are available
to the highway and transportation analyst. Field determination of
the free-flow speed is accomplished by performing travel time
Uninterrupted-flow facilities that allow access solely through a
studies during periods of low volume. In comparing free-flow
system of on-ramps and off-ramps from grade separations or ser-
speed with other measures of speed, it should be noted that op-
vice roads are considered freeways and should be evaluated using
erating speed, as defined in previous capacity manuals and in
the methodology of Chapter 3.
AASHTO documents, is similar to free-flow speed when taken
under low-volume conditions. For the analyses made using this
FREE-FLOW SPEED chapter, the upper limit for low-volume conditions is 1,400 passen-
ger cars per hour per lane (pcphpl).
An important characteristic of multilane highways is the free-
flow speed of vehicles. Free-flow speed is the theoretical speed SPEED-FLOW AND DENSITY-FLOW RELATIONSHIPS
of traffic as density approaches zero. Practically, it is the speed
at which drivers feel comfortable traveling under the physical, Figures 7-1 and 7-2 show the speed-flow and density-flow rela-
environmental, and traffic control conditions existing on an uncon- tionships for a typical uninterrupted-flow segment on a multilane
gested section of multilane highway. This is similar to the average highway under either ideal or nonideal conditions in which free-
desired speed of all drivers on an arterial segment used in Chapter flow speed is known. The capacity and level of service for a
11. Free-flow speeds will be lower on sections of highway with multilane highway may be slightly lower than those of a freeway,
restricted vertical or horizontal alignments. Free-flow speeds tend even under ideal conditions, because drivers on multilane high-
to be lower when posted speed limits are lower. The importance ways allow for potential conflicts with turning traffic, even when
of free-flow speed is that it is the starting point for the analyses there are no access points in the vicinity.
Figure 7-1. Speed-flow relationships on multilane highways.
Figure 7-1 indicates that the speed of traffic on a multilane relationships and adjustments to flow. For the multilane highways
highway is insensitive to traffic volume up to a flow rate of 1,400 to be analyzed in this chapter, those conditions (from which adjust-
pcphpl. Figure 7-1 also indicates that the capacity of a multilane ments will be made) are as follows:
highway under ideal conditions is 2,200 pcphpl for highways with
a 60-mph, free-flow speed. At flow rates between 1,400 and 2,200
T Level terrain, with grades no greater than 1 to 2 percent.
pcphpl, the speed on a multilane highway with a 60-mph, free-flow
T 12-ft lane widths.
speed drops 5 mph.
T A minimum of 12 ft of total lateral clearance in the direction
Figure 7-2 shows that density varies continuously throughout the
of travel. This total represents the lateral clearances from the edge
full range of flow rates. Figures 7-1 and 7-2 are indicative of op-
of the traveled lanes (shoulders included) to obstructions along the
erating conditions for multilane highways with a given free-flow
edge of the road and in the median. Lateral clearances greater than
speed.
6 ft are considered in computations to be equal to 6 ft.
The capacity value of 2,200 pcphpl is representative of the maxi-
T No direct access points along the roadway.
mum 15-min flow rate that can be accommodated under ideal condi-
T A divided highway.
tions for 60-mph, free-flow highways; however, actual capacities on
T Only passenger cars in the traffic stream.
specific multilane highway sections may vary from this. Neverthe-
T A free-flow speed of 60 mph or more.
less, for the purpose of determining the level of service of a multilane
highway section, the capacity values presented in this chapter should
be used unless local measurements suggest otherwise. These ideal conditions represent the highest type of multilane
rural and suburban highways. The procedures in this chapter deter-
IDEAL CONDITIONS mine the reduction in travel speed that occurs for less-than-ideal
conditions. It should be noted that these conditions are ideal only
Studies of the flow characteristics of multilane highways have from the point of view of capacity and level of service and do not
defined a set of ideal conditions as the basis for developing flow relate to safety or other factors.

Figure 7-2. Density-flow relationships on multilane highways.
II. FREE-FLOW SPEED ADJUSTMENTS
The free-flow speed for multilane highways is found by using several studies (1–3) have found the enforcement effects to be
the mean speed of passenger cars measured under low to moderate temporary, both temporally and spatially. The degree to which the
traffic flow conditions. Level of service is based on density. Den- enforcement effect is sustained depends upon the type and duration
sity may be calculated by dividing per-lane flow by speed. The of technique used. In general, it is likely that a stationary enforce-
research for this chapter showed that on multilane highways speed ment activity will affect no more than an 8- to 9-mi length of
is fairly insensitive to flow over a wide range of flows. It also roadway. The effect has been found to decay exponentially down-
demonstrated statistically significant effects of geometric factors stream from the enforcement site. The speeds at the site may be
(lane width, median type, access points) on free-flow speed. affected over time as much as 2 to 3 days after removal of the
The capacity of a multilane highway is the maximum sustained activity.
hourly rate of flow at which vehicles can be reasonably expected If the roadway in question is located in a community in which
to traverse a uniform segment of roadway under prevailing road- speed enforcement is anticipated to affect operation often, the user
way and traffic conditions. The time period used for analysis is may make local measurements to calibrate the relationship between
15 min. This maximum is a function of density and therefore speed. 85th-percentile speed and free-flow speed. If this is done, the data
Because free-flow speed is a direct input to the density calcula- should be taken at times when the average anticipated enforcement
tion, adjustments relating to geometric factors are made to free- effects will be operating.
flow speed and not to capacity. Factors that characterize traffic
conditions are discussed in Section III, Volume Factors.
Several traffic control, physical, and traffic conditions affect DESIGN SPEED
the free-flow speed along a given highway. These conditions are
described in the following sections. The design speed of the principal physical elements of a multi-
lane roadway can affect travel speeds. In particular, the horizontal
and vertical alignments of a highway may contribute to the actual
SPEED ENFORCEMENT speeds of vehicles. Because design speed is difficult to assess in
the field, several alternative methods are described in this chapter
Vehicular speeds, and the proportion of vehicles exceeding the that allow the analyst to estimate free-flow speed along a given
speed limit, are affected by speed enforcement activities. However, section of multilane highway. The design speed, and particularly
the horizontal and vertical alignments, is implicitly included in raised or depressed median or with a median at least 10 ft wide
these methods. If a multilane highway is characterized by some- (including a TWLTL) should be considered as divided. A raised
what extreme horizontal or vertical conditions, the analyst should curb in the median, even when interrupted with openings on a
determine free-flow speed from field observation and field study. regular basis, should be treated as a raised median. Short sections
of raised or flush median (less than 500 ft long), however, should
not be considered as a raised median. Travel speeds on an undi-
SPEED LIMIT
vided multilane highway are somewhat lower across the entire
range of volume conditions than speeds on a divided highway of
Posted speed limits normally have an effect on the free-flow similar design. Thus, when the free-flow speed is determined, the
speed of passenger cars (4). Typically, the mean speed of passenger median type (undivided or divided) is an influence that must be
cars is above the posted speed limit for multilane highways. Data accounted for.
from multilane highways indicate that the posted speed limit has For divided multilane highways with different median designs,
a significant correlation with the speed at which vehicles move there is no significant difference in speed conditions. These designs
along the highway. The procedures in this chapter allow the analyst include the TWLTL and the barrier, curb, and landscaped designs.
to estimate free-flow speed on the basis of posted speed limit when Thus, for the purposes of this chapter, all highways with barrier
no other estimate is available for free-flow speed. These guidelines designs (a 10-ft-or-wider TWLTL or a curb and landscaped design)
were developed from field data taken at sites throughout the United are considered divided. This would include those with a raised
States. Enforcement levels were considered typical, and there was median of any width, as long as the raised curb acted as a physical
no evidence of speed enforcement activity in the sections studied barrier.
during the time of data collection.
ACCESS POINTS
LANE WIDTH AND LATERAL CLEARANCE
An important influence on free-flow speed is the number of
Two adjustments are used in this chapter to predict the effect access points along the right side of the roadway. The data base
of a constricted cross section on free-flow speeds—the average used to establish the procedures in this chapter indicated that the
width of lanes and the combined lateral clearance along the right number of access points was the critical element in reducing free-
side and the median of a multilane highway. Lane widths that are flow speeds along a section of multilane highway. Although the
less than the ideal of 12 ft reduce travel speeds, but lane widths amount of activity at each point also contributes to changes in
of more than 12 ft are not considered to increase speed above the travel speed, it is apparent that drivers adjust their travel speed
ideal situation. For lateral clearance, a total clearance (left side not only on the basis of entrances and exits at such points but also
plus right side along a roadway in one direction of travel) of 12 on the mere existence of access points. As expected, the addition
ft or more is considered to be the ideal condition. When less than of intersections or driveways along a multilane highway will re-
12 ft, combined lateral clearance has a negative effect on travel duce travel speeds. The procedures of this chapter show that for
speeds. For both lane width and lateral clearance, the procedures every 10 access points per mile that affect a given direction of
of this chapter reflect a change in travel speed for flow conditions travel on a multilane highway, travel speed will be reduced by
throughout the range of flows from nearly zero to 2,200 pcphpl. 2.5 mph.
Where the free-flow speed is reduced, capacity is also reduced. Note that this procedure takes into account only those access
Considerable judgment must be used in determining whether points on one side of the roadway and not those on the opposite
roadside and median objects and barriers present true obstructions. side of the roadway or openings in the median. If access points
Such obstructions may be continuous, such as a retaining wall, or on the opposite side of the roadway or median openings for U-
may not be continuous, such as light supports or bridge abutments. turns are expected to have a significant effect on traffic flow in
In some cases, drivers may become accustomed to certain types the direction of interest, these intersections, driveways, or openings
of obstructions; as a result, their effect on traffic flow becomes may be included in the determination of access-point density.
negligible. Certain common types of traffic barrier, for example,
may have little or no impact on traffic, even when closer than 6
ft from the travel lanes. These include the reinforced-concrete OTHER ADJUSTMENTS
traffic barriers and the W-beam barriers often used on highways.
This chapter is based on traffic flow data taken throughout
the United States in the late 1980s. The procedures developed
MEDIAN TYPE from these data contain several adjustments that are used to
predict the free-flow speed when a section of multilane highway
Typically, there are three types of median along multilane rural is analyzed. Although data on the effect of driver population
and suburban highways—an undivided median composed of a are few and undocumented, it is generally assumed that the
striped centerline; a two-way left-turn-lane (TWLTL) median com- commuter or frequent driver may travel at a higher speed than
posed of a full-width lane used for left turns in either direction; other drivers. No data were available to suggest a significant
and a median composed of a raised curb, barrier, or natural terrain and consistent relationship between changes in the type of driver
or landscaping. The data used to establish the procedures of this using the facility and changes in traffic flow characteristics such
chapter were arrayed in two groups, those for multilane highways as travel speed.
with no median (that is, undivided) and those that had a median Another factor found in earlier capacity documents and not in-
including a TWLTL or any type of raised or barrier configuration, cluded in this chapter is the difference between rural and suburban
including a paved flush median. Any multilane highway with a driving environments. In previous methods, this distinction was
made on the basis of the difference in amount of access provided in a direct manner. Thus, the suburban-rural factor is not part of
to the roadway. Rather than differentiate between these two envi- this procedure.
ronments, the procedures of this chapter account for access points
III. VOLUME FACTORS
The previous section described the conditions that affect travel as the ratio of total hourly volume to the maximum 15-min rate
speed along a multilane highway. Flow is the second component of flow within the hour.
of the density calculation. The estimate of volume is adjusted by
factors relating to both the composition and fluctuation of the HEAVY-VEHICLE FACTOR
traffic so that all roadways may be compared with an equivalent
measure, passenger cars per hour per lane (pcphpl). These factors The second factor used to adjust the volume is heavy vehicles.
are identified in the following sections. A factor that converts trucks, buses, and recreational vehicles
(RVs) into an equivalent number of passenger cars is used. In this
chapter, only two categories of heavy vehicles are used, trucks
PEAK-HOUR FACTOR and RVs. For analysis purposes, buses moving along multilane
highways should be considered as trucks. The impact of converting
The basis for traffic volumes used in this chapter is a 15-min heavy vehicles to equivalent passenger cars is especially important
peak-period flow, which typically occurs during a peak hour of for sections of highway with vertical grades. For level terrain and
the day. Thus, the analyst will be concerned with volumes in two especially for conditions near capacity, trucks, buses, and RVs
time periods, a peak-hour volume and the flow rate within the tend to operate like passenger cars, and the equivalency factors
peak 15 min of the peak hour. The tables and charts used in the begin to approach unity. Once the analysis volumes for heavy
procedures are based on flow rates. Conversion from the hourly vehicles have been converted to equivalent passenger cars, the
volume to a flow rate is accomplished by dividing the hourly procedures of this chapter are based on a flow rate expressed in
volume by the peak-hour factor. The peak-hour factor is defined pcphpl.
IV. METHODOLOGY
FREE-FLOW SPEED LOS boundaries are represented in Figure 7-3, by sloped lines,
each corresponding to a constant value of density. Complete LOS
The definition of ideal conditions for multilane highways speci- criteria are given in Table 7-1; the user should note that these
fies a free-flow speed of at least 60 mph. In the field, the free- criteria are based on the typical speed-flow–density-flow relation-
flow speed is the mean speed of passenger cars measured under ships shown in Figures 7-1 and 7-2. The criteria reflect the shape
low to moderate flow conditions (up to 1,400 pcphpl) where speeds of those curves, particularly that speed remains relatively constant
are insensitive to flow rates, as shown in Figure 7-1. This essen- across LOS A to D but is reduced as capacity is approached. For
tially represents the average desired speed at which a driver would average free-flow speeds of 60, 55, 50, and 45 mph, Table 7-1
like to travel. A variety of conditions, including grades, horizontal gives the average travel speed, the maximum value of v/c, and the
curves, speed limits, speed enforcement, vehicle operating charac- corresponding maximum service flow rate (MSF) for each level
teristics, and driver preferences, typically limits free-flow speeds of service. Under ideal conditions, the speeds, v/c ratios, and MSF
to a range of 40 to 60 mph. tabulated are expected to exist in traffic streams operating at the
Any multilane highway segment can be characterized by a free- densities defined for each level of service.
flow speed curve similar to those shown in Figure 7-1. The recom- LOS criteria depend on the free-flow speed of the highway
mended methodology to determine the free-flow speed for a partic- element being studied—an isolated geometric element, such as a
ular multilane highway segment is discussed in a following section. curve or grade of significant length that operates with a reduced
speed, or a series of such geometric elements that affect the opera-
tion of a longer segment of highway.
LEVEL-OF-SERVICE CRITERIA LOS A describes completely free-flow conditions. The operation
of vehicles is virtually unaffected by the presence of other vehicles,
Level-of-service (LOS) criteria for multilane highways are de- and operations are constrained only by the geometric features of
fined in terms of density. Density is a measure that quantifies the the highway and by driver preferences. Vehicles are spaced at an
proximity of vehicles to each other within the traffic stream and average of 440 ft at a maximum density of 12 passenger cars per
indicates the degree of maneuverability within the traffic stream. mile per lane (pc/mi/ln). Maneuverability within the traffic stream
Various levels of service are applied to the speed-flow curves is good. Minor disruptions to flow are easily absorbed at this level
presented in Figure 7-1 to give density boundary values. These without a change in travel speed.
Table 7-1. Level-of-Service Criteria for Multilane Highways

free-flow speed
60 mph 55 mph 50 mph 45 mph
max max max max
service service service service
level max average flow max average flow max average flow max average flow
of density speed max rate density speed max rate density speed max rate density speed max rate
service (pc/mi/ln) (mph) v /c (pcphpl) (pc/mi/ln) (mph) v /c (pcphpl) (pc/mi/ln) (mph) v /c (pcphpl) (pc/mi/ln) (mph) v /c (pcphpl)
A 12 60 0.33 720 12 55 0.31 660 12 50 0.30 600 12 45 0.28 540
B 20 60 0.55 1,200 20 55 0.52 1,100 20 50 0.50 1,000 20 45 0.47 900
C 28 59 0.75 1,650 28 54 0.72 1,510 28 50 0.70 1,400 28 45 0.66 1,260
D 34 57 0.89 1,940 34 53 0.86 1,800 34 49 0.84 1,670 34 44 0.79 1,500
E 40 55 1.00 2,200 41 51 1.00 2,100 43 47 1.00 2,000 45 42 1.00 1,900
NOTE: The exact mathematical relationship between density and v / c has not always been maintained at LOS boundaries because of the use of rounded values.
Density is the primary determinant of LOS. LOS F is characterized by highly unstable and variable traffic flow. Prediction of accurate flow rate, density, and
speed at LOS F is difficult.
LOS B is also indicative of free flow, although the presence of LOS D represents a range in which ability to maneuver is se-
other vehicles begins to be noticeable. Average travel speeds are verely restricted because of traffic congestion. Travel speed begins
the same as in LOS A, but drivers have slightly less freedom to to be reduced by increasing volumes. The average spacing of
maneuver. Vehicles are spaced at an average of approximately 264 vehicles is 155 ft at a maximum density of 34 pc/mi/ln. Only
ft at a maximum density of 20 pc/mi/ln. Minor disruptions are still minor disruptions can be absorbed without the formation of exten-
easily absorbed at this level, although localized deterioration in sive queues and the deterioration of service to LOS E and LOS F.
level of service will be more obvious. LOS E represents operations at or near capacity and is quite
LOS C represents a range in which the influence of traffic den- unstable. The densities at LOS E vary depending upon the free-
sity on operations becomes marked. The ability to maneuver within flow speed. At LOS E, vehicles are operating with the minimum
the traffic stream is now clearly affected by the presence of other spacing at which uniform flow can be maintained. Thus, as the
vehicles. Average travel speeds begin to show some reduction for limits for the level of service are approached, disruptions cannot
multilane highways with free-flow speeds over 50 mph. The aver- be damped or readily dissipated, and most disruptions will cause
age spacing of vehicles is reduced to approximately 189 ft at a queues to form and service to deteriorate to LOS F. For the major-
maximum density of 28 pc/mi/ln. Minor disruptions may be ex- ity of multilane highways with free-flow speeds between 45 and
pected to cause serious local deterioration in service, and queues 60 mph, passenger-car speeds at capacity range from 40 to 55 mph
may form behind any significant traffic disruption. but are highly variable and unpredictable within that range.
Figure 7-3. Speed-flow curves with LOS criteria. *Maximum density for respective levels of service. **Maximum densities for LOS E
occur at volume-to-capacity ratio of 1.0. They are 40, 41, 43, and 45 pc/mi/ln at free-flow speeds of 60, 55, 50, and 45 mph, respectively.
LOS F represents forced or breakdown flow. It occurs either ies may be used. It is recommended that the field study be con-
at a point where vehicles arrive at a rate greater than the rate ducted in the more stable regime of low to moderate flow
at which they are discharged or at a point on a planned facility conditions (up to 1,400 pcphpl). Off-peak hours are generally good
where forecast demand exceeds computed capacity. Although times to observe low flow rates. If the speed study must be con-
operations at such points (and on sections immediately down- ducted at a flow rate of more than 1,400 pcphpl, the free-flow
stream) will appear to be at capacity, queues will form behind speed can be found by using the model speed-flow curve, assuming
these breakdowns. Operations within queues are highly unstable, that data on traffic volumes were taken at the same time. The
with vehicles experiencing brief periods of movement followed y-intercept of a curve that depicts the observed speed at the mea-
by stoppages. Average travel speeds with queues are generally sured flow rate is the free-flow speed. The speed study should
less than 30 mph. Note that the term ‘‘LOS F’’ may be used measure the speeds of all passenger cars or a systematic sample
to characterize both the point of the breakdown and the operating of passenger cars (e.g., every 10th passenger car). The speed study
condition within the queue. It must be remembered, however, should not only measure speeds for unimpeded vehicles but should
that it is the point of breakdown that causes the queue to form also include representative numbers of impeded vehicles. A sample
and that operations within the queue are generally not related of at least 100 passenger-car speeds should be obtained. Further
to defects along the highway segment over which the queue guidance on the conduct of speed studies is found in standard
extends. Chapters 3 and 6 contain more detailed discussions of traffic engineering publications such as the Manual of Traffic Engi-
the use and application of LOS F and of the analysis of neering Studies, published by the Institute of Transportation
breakdown conditions. Engineers.
As with other LOS criteria, the MSFs of Table 7-1 are stated The average of all passenger-car speeds measured in the field
in terms of rates of flow based on the peak 15 min. Demand or study under low-volume conditions can be used directly as the
forecast volumes are generally divided by the peak-hour factor to free-flow speed if such measurements were taken at flow rates at
reflect a maximum hourly flow rate before they are compared with or below 1,400 pcphpl. This speed reflects the net effects of all
the criteria of Table 7-1. conditions at the site that influence speed, including those con-
The prediction of level of service for a multilane highway gener-
tained in this procedure (lane width, lateral clearance, type of
ally involves three steps:
median, and access points), as well as others such as speed limit
1. Determination of free-flow speed, and vertical and horizontal alignment.
2. Adjustment of volume, and If field measurement of the roadway under study is not feasible,
3. Determination of level of service. data taken at a similar facility may be used. The surrogate roadway
should be similar with respect to the variables that are identified
These steps are described in this order in subsequent sections. in this chapter as affecting free-flow speed.
Highway agencies with ongoing speed-monitoring programs or
with existing speed data on file may prefer to use those data rather
DETERMINATION OF FREE-FLOW SPEED
than conduct a new speed study or use an indirect speed estimation
technique. Such data can be used directly if collected in accordance
The first step in the assessment of the level of service on a
with the procedures presented above. Data that include both pas-
multilane highway is to determine the free-flow speed for the
senger-car and heavy-vehicle speeds can probably be used for level
highway. The free-flow speed is measured using the mean speed
terrain or moderate downgrades but should not be used for rolling
of passenger cars under low to moderate flow conditions (up to
or mountainous terrain. Recent data for multilane highways indi-
1,400 pcphpl). Average speed is virtually constant across this range
cate that mean free-flow speed under ideal conditions ranges from
of flow rates.
approximately 1 mph lower than the 85th-percentile speed when
Two general methods can be used to determine the free-flow
the latter is 40 mph to 3 mph lower when the 85th-percentile speed
speed for the roadway: (a) field measurement and (b) estimation
with guidelines provided in this chapter. The field measurement is 60 mph.
procedure is provided for those who prefer to gather these data When it is not possible to use data from an existing similar
directly or to incorporate these measurements in an existing speed- roadway, an estimate may be necessary based on available data,
monitoring program. However, field measurements are not neces- experience, and the consideration of the variety of factors that
sary to apply the procedure. The free-flow speed can be estimated have an identified effect on free-flow speed. Speed limit is one
with guidelines based on field data and user knowledge of condi- factor that affects free-flow speed. Recent research suggests that
tions on the roadway in question. free-flow speed under ideal conditions is approximately 7 mph
higher than the speed limit for 40- and 45-mph speed limits
and 5 mph higher for 50- and 55-mph speed limits. Analyses
Field Measurement based upon these rules of thumb should be used with caution.
The free-flow speed of a highway can be determined directly

from a speed study conducted in the field. If field-measured data Guidelines for Estimating Free-Flow Speed
are used, no subsequent adjustments are made to free-flow speed.
The speed study should be conducted at a representative location
within the highway segment being evaluated; for example, a seg- The free-flow speed can be estimated indirectly when field data
ment on an upgrade should not be selected within a site that is are not available. The free-flow speed is estimated as
generally level. Any speed measurement technique that has been
found acceptable for other types of traffic engineering speed stud- FFS = FFSI − FM − FLW − FLC − FA (7-1)

where: free-flow speed for narrower lanes. The data in Table 7-3 indicate
that multilane highways with 11-ft lanes have free-flow speeds
FFS = estimated free-flow speed (mph),
that are 1.9 mph lower than highways with 12-ft lanes, whereas
FFSI = estimated free-flow speed (mph) for ideal conditions,
highways with 10-ft lanes have free-flow speeds 6.6 mph less
FM = adjustment for median type (from Table 7-2),
than highways with 12-ft lanes. To use Table 7-3, lane widths
FLW = adjustment for lane width (from Table 7-3),
greater than 12 ft are considered as 12 ft. No data exist for
FLC = adjustment for lateral clearance (from Table 7-4), and
lane widths less than 10 ft.
FA = adjustment for access points (from Table 7-5).
Table 7-4 presents the adjustment for lateral clearance to fixed
The first adjustment that is used to modify the estimated free- obstructions on the roadside or in the median. Fixed obstructions
flow speed relates to median type. The data in Table 7-2 indicate whose lateral clearance effects should be considered include light
that the average free-flow speed should be decreased by 1.6 mph standards, signs, trees, abutments, bridge rails, traffic barriers, and
for undivided roadways to account for the friction caused by oppos- retaining walls. Standard raised curbs are not considered obstruc-
ing traffic in an adjacent lane. The analyst should consider dividing tions. Table 7-4 shows the appropriate reduction in free-flow speed
the highway section into homogeneous sections that reflect based on the total lateral clearance, which is defined as
changes in median characteristics. TLC = LCR + LCL (7-2)
Ideal conditions for multilane highways are based on 12-ft
lanes. Table 7-3 presents the adjustment to modify the estimated where:
TLC = total lateral clearance (ft),
Table 7-2. Adjustment for Median Type LCR = lateral clearance (ft) from the right edge of the travel
lanes to roadside obstructions (if greater than 6 ft, use
reduction in
free-flow speed 6 ft), and
median type (mph) LCL = lateral clearance (ft) from the left edge of the travel lanes
to obstructions in the roadway median (if the lateral clear-
Undivided Highways 1.6
Divided Highways (including TWLTLs) 0.0 ance is greater than 6 ft, use 6 ft). For undivided roadways,
there is no adjustment for left-side lateral clearance. The
undivided design itself is taken into account by the median
Table 7-3. Adjustment for Lane Width adjustment. Therefore, in order to use Table 7-4 for undi-
vided highways, the lateral clearance on the left edge is
lane reduction always 6 ft. Lateral clearance in the median of roadways
width in free-flow
(ft) speed (mph) with TWLTLs is considered to be 6 ft.
10 6.6 Thus, a total lateral clearance of 12 ft is used for a completely
11 1.9 unobstructed roadside and median, whereas the actual value is
12 0.0 used when obstructions are located closer to the roadway. The
adjustment for lateral clearance on six-lane highways is slightly
less than for four-lane highways because lateral obstructions have
Table 7-4. Adjustment for Lateral Clearance a minimal effect on traffic operations in the center lane of a three-
four-lane highways six-lane highways lane roadway. Illustrations 7-5 through 7-8 show various types of
roadside and median treatments that can affect the flow on multi-
reduction in reduction in
total lateral free-flow total lateral free-flow lane highways.
clearancea speed clearancea speed Table 7-5 presents the adjustment for various levels of access-
(ft) (mph) (ft) (mph) point density. The data indicate that each access point per mile
12 0.0 12 0.0 decreases the estimated free-flow speed by approximately 0.25
10 0.4 10 0.4 mph, regardless of the type of median. The access-point density
8 0.9 8 0.9 on a divided roadway is found by dividing the total number of
6 1.3 6 1.3 access points (intersections and driveways) on the right side of
4 1.8 4 1.7 the roadway in the direction of travel being studied by the
2 3.6 2 2.8
length of the section in miles. An intersection or driveway
0 5.4 0 3.9
a
should only be included by the analyst if it is considered to
Total lateral clearance is the sum of the lateral clearances of the median (if
greater than 6 ft, use 6 ft) and shoulder (if greater than 6 ft, use 6 ft). There- have a significant influence on traffic flow. Access points that
fore, for analysis purposes, total lateral clearance cannot exceed 12 ft. are difficult to identify by the driver or where there is little
activity should not be included in the determination of access-
point density. Such access points might include private driveways
Table 7-5. Access-Point Density Adjustment to individual residences or service driveways at commercial sites.
access points reduction in Although development of the access-point adjustments did not
per mile free-flow speed (mph) include data for one-way multilane highways, at the analyst’s
0 0.0 discretion it may be appropriate to include intersections and
10 2.5 driveways on both sides of a one-way roadway to determine
20 5.0 the total number of access points per mile.
30 7.5 When data on the number of access points on a highway section
40 or more 10.0 are unavailable (e.g., when the highway has not yet been con-
Illustration 7-5. Bridge pier in center of normally undivided Illustration 7-6. Inadequate shoulder and obstructions on
suburban multilane highway. roadway.
Illustration 7-7. Ideal divided multilane highway. Illustration 7-8. Undivided multilane highway with no
obstructions.
structed), the guidelines presented in Table 7-6 may be used. The

analyst must use some judgment when applying these guidelines mate the ideal free-flow speed of a highway section by overstating
to determine access-point density. the effect of a given driver population.
The data on which the free-flow speed relationships presented
in this chapter are based include both commuter and noncommuter
DETERMINATION OF FLOW RATE
traffic conditions. No significant differences between the two
were detected. However, it is frequently held that commuters
Two adjustments must be made to hourly volume counts or
or other frequent users of a highway will use the facility more
estimates to arrive at the equivalent passenger-car flow rate used
efficiently than do recreational users or other occasional drivers.
in LOS analyses. These adjustments are the peak-hour factor
If the effect of a particular driver population is a concern, the
and the heavy-vehicle adjustment factor. The number of lanes
analyst is encouraged to measure free-flow speed in the field.
is also used so the flow rate can be expressed on a per-lane
If field measurements cannot be made, the analyst may select
basis. These adjustments are applied in the following manner:
a free-flow speed that reflects the anticipated effect of the
particular driver population. Care should be taken not to underesti- V
vp = (7-3)
(N) (PHF) (fHV)
where:
Table 7-6. Number of Access Points for General
Development Environments vp = service flow rate (pcphpl),
access points per mile V = volume (number of vehicles passing a point in 1 hr),
type of development (one side of roadway) N = number of lanes,
PHF = peak-hour factor, and
Rural 0–10
Low-Density Suburban 11–20 fHV = heavy-vehicle adjustment factor.
High-Density Suburban 21 or more PHF and fHV are discussed in the following sections.
Peak-Hour Factor 1. Level terrain—Any combination of horizontal and vertical

alignment permitting heavy vehicles to maintain approximately
PHF represents the temporal variation in traffic flow within an the same speed as passenger cars; this generally includes short
hour. Observation of traffic flow consistently indicate that the flow grades of no more than 1 to 2 percent.
rates found in the peak 15-min period within an hour are not 2. Rolling terrain—Any combination of horizontal and vertical
sustained throughout the entire hour. Application of PHF in Equa- alignment causing heavy vehicles to reduce their speeds substan-
tion 7-3 accounts for this. tially below those of passenger cars, but not causing heavy vehicles
The PHFs for multilane highways have been observed to range to operate at crawl speeds for any significant length of time or at
from 0.76 to 0.99. Lower PHFs are typical of rural or off-peak frequent intervals.
conditions, whereas higher factors are typical of urban and sub- 3. Mountainous terrain—Any combination of horizontal and
urban peak-hour conditions. Users are encouraged to use their own vertical alignment causing heavy vehicles to operate at crawl
data to develop PHFs applicable to their conditions. Where local speeds for significant distances or at frequent intervals.
data are not available, 0.85 is a reasonable estimate of the PHF
for rural multilane highways and 0.92 for suburban multilane For all such general highway segments, values of ET and ER are
selected from Table 7-7.
highways.
Any grade of 3 percent or less that is longer than 1 mi or
any grade greater than 3 percent that is longer than 1⁄2 mi should
Adjustment for Presence of Heavy Vehicles be treated as an isolated significant grade. The upgrade and
downgrade must be treated separately because the impact of
The presence of heavy vehicles in the traffic stream decreases heavy vehicles varies substantially for these two conditions.
the free-flow speed because under ideal conditions the traffic Tables 7-8 and 7-9 give passenger-car equivalents for trucks
stream is composed only of passenger cars. Rarely will such an and buses (ET) and for RVs (ER), respectively, on uniform upgrades.
ideal traffic stream exist on multilane highways. Therefore, traffic Table 7-8 is based on an average weight-to-horsepower ratio of
volumes must be adjusted to an equivalent flow rate expressed in 167 lb/hp, which is typical of the truck population currently found
passenger cars per hour per lane. This is accomplished through on multilane highways.
application of the factor fHV. When several consecutive grades of different steepness form
Adjustment for the presence of heavy vehicles in the traffic a composite grade, an average uniform grade is computed and
stream applies for three types of vehicles: trucks, RVs, and buses. used to enter the tables. The average grade is commonly computed
There is no evidence to indicate any distinct differences in perform- as the total rise from the beginning of the grade divided by the
ance characteristics between the truck and bus populations on mul- total horizontal distance over which the rise was accomplished.
tilane highways, and thus buses are considered as trucks in this Consider the following example. Three consecutive upgrades
procedure. Finding the heavy-vehicle adjustment factor requires are to be analyzed:
two steps:
1. 3 percent grade—1,000 ft long,
1. Find the equivalent factor for trucks and buses (ET) and for 2. 4 percent grade—2,000 ft long, and
RVs (ER) for the prevailing operating conditions. 3. 2 percent grade—1,000 ft long.
2. Using the values found in Step 1, compute an adjustment
The total rise of the 4,000-ft grade may be computed as
factor that corrects for all heavy vehicles in the traffic stream.
1,000 × 0.03 = 30 ft
Finding Passenger-Car Equivalents 2,000 × 0.04 = 80 ft
1,000 × 0.02 = 20 ft
Passenger-car equivalents can be selected for two conditions:
extended general highway segments and specific grades. Values 130 ft
of passenger-car equivalents are selected from Tables 7-7 through The average grade may now be expressed as follows:
7-10 for a variety of basic conditions. For long segments of high-
way over which no single grade has a significant impact on opera- Average grade = (130/4,000) × 100 = 3.25 percent
tions, Table 7-7 is used to select passenger-car equivalent values Passenger-car equivalents would then be selected for a 4,000-ft
for trucks and buses, ET, and RVs, ER. A long multilane highway grade of 3.25 percent.
segment may be classified as a general segment if no grades ex- The average grade technique is reasonably accurate for grades
ceeding 3 percent are longer than 1⁄2 mi or grades of 3 percent or of 4,000 ft or less, or no greater than 4 percent. For steeper and
less do not exceed 1 mi. Such segments should be categorized as longer grades, a more exact technique is described in Appendix I
follows: of Chapter 3. When applying the composite grade technique where
a large change in grade occurs for a significant length, the user
should consider segmenting the roadway. The composite grade
technique may need to be supplemented by a specific grade analy-
Table 7-7. Passenger-Car Equivalents on Extended sis when a single steep grade creates a critical effect that might
General Multilane Highway Segments not otherwise be identified when it is included in a length of
type of terrain highway with otherwise flat conditions.
Downgrade conditions for trucks and buses are handled with
factor level rolling mountainous
Table 7-10. For all downgrades less than 4 percent and for steeper
ET (Trucks and Buses) 1.5 3.0 6.0 downgrades less than or equal to 2 mi long, use the passenger-car
ER (RVs) 1.2 2.0 4.0 equivalents for trucks and buses in level terrain given in
Table 7-8. Passenger-Car Equivalents for Trucks and Buses on Uniform Upgrades
grade length
(%) (mi) ETa
percent trucks and
buses 2 4 5 6 8 10 15 20 25
<2 All 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5
2 0–1⁄4 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5
1
⁄4–1⁄2 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5
1
⁄2–3⁄4 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5
3
⁄4–1 2.5 2.0 2.0 2.0 1.5 1.5 1.5 1.5 1.5
1–11⁄2 4.0 3.0 3.0 3.0 2.5 2.5 2.0 2.0 2.0
>11⁄2 4.5 3.5 3.0 3.0 2.5 2.5 2.0 2.0 2.0
3 0–1⁄4 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5
1
⁄4–1⁄2 3.0 2.5 2.5 2.0 2.0 2.0 2.0 1.5 1.5
1
⁄2–3⁄4 6.0 4.0 4.0 3.5 3.5 3.0 2.5 2.5 2.0
3
⁄4–1 7.5 5.5 5.0 4.5 4.0 4.0 3.5 3.0 3.0
1–11⁄2 8.0 6.0 5.5 5.0 4.5 4.0 4.0 3.5 3.0
>11⁄2 8.5 6.0 5.5 5.0 4.5 4.5 4.0 3.5 3.0
4 0–1⁄4 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5
1
⁄4–1⁄2 5.5 4.0 4.0 3.5 3.0 3.0 3.0 2.5 2.5
1
⁄2–3⁄4 9.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5
3
⁄4–1 10.5 8.0 7.0 6.5 6.0 5.5 5.0 4.5 4.0
>1 11.0 8.0 7.5 7.0 6.0 6.0 5.0 5.0 4.5
5 0–1⁄4 2.0 2.0 1.5 1.5 1.5 1.5 1.5 1.5 1.5
1
⁄4–1⁄3 6.0 4.5 4.0 4.0 3.5 3.0 3.0 2.5 2.0
1
⁄3–1⁄2 9.0 7.0 6.0 6.0 5.5 5.0 4.5 4.0 3.5
1
⁄2–3⁄4 12.5 9.0 8.5 8.0 7.0 7.0 6.0 6.0 5.0
3
⁄4–1 13.0 9.5 9.0 8.0 7.5 7.0 6.5 6.0 5.5
>1 13.0 9.5 9.0 8.0 7.5 7.0 6.5 6.0 5.5
6 0–1⁄4 4.5 3.5 3.0 3.0 3.0 2.5 2.5 2.0 2.0
1
⁄4–1⁄3 9.0 6.5 6.0 6.0 5.0 5.0 4.0 3.5 3.0
1
⁄3–1⁄2 12.5 9.5 8.5 8.0 7.0 6.5 6.0 6.0 5.5
1
⁄2–3⁄4 15.0 11.0 10.0 9.5 9.0 8.0 8.0 7.5 6.5
3
⁄4–1 15.0 11.0 10.0 9.5 9.0 8.5 8.0 7.5 6.5
>1 15.0 11.0 10.0 9.5 9.0 8.5 8.0 7.5 6.5
NOTE: If a length of grade falls on a boundary condition, the equivalent from the longer-grade category is used.
a
Four- or six-lane highways.
Table 7-9. Passenger-Car Equivalents for Recreational Vehicles on Uniform Upgrades

grade length
(%) (mi) ERa
percent rvs 2 4 5 6 8 10 15 20 25
≤2 All 1.2 1.2 1.2 1.2 1.2 1.2 1.2 1.2 1.2
3 0–1⁄2 1.2 1.2 1.2 1.2 1.2 1.2 1.2 1.2 1.2
>1⁄2 2.0 1.5 1.5 1.5 1.5 1.5 1.2 1.2 1.2
4 0–1⁄4 1.2 1.2 1.2 1.2 1.2 1.2 1.2 1.2 1.2
1
⁄4–1⁄2 2.5 2.5 2.0 2.0 2.0 2.0 1.5 1.5 1.5
>1⁄2 3.0 2.5 2.5 2.0 2.0 2.0 2.0 1.5 1.5
5 0–1⁄4 2.5 2.0 2.0 2.0 1.5 1.5 1.5 1.5 1.5
1
⁄4–1⁄2 4.0 3.0 3.0 3.0 2.5 2.5 2.0 2.0 2.0
>1⁄2 4.5 3.5 3.0 3.0 3.0 2.5 2.5 2.0 2.0
6 0–1⁄4 4.0 3.0 2.5 2.5 2.5 2.0 2.0 2.0 1.5
1
⁄4–1⁄2 6.0 4.0 4.0 3.5 3.0 3.0 2.5 2.5 2.0
>1⁄2 6.0 4.5 4.0 4.0 3.5 3.0 3.0 2.5 2.0
NOTE: If a length of grade falls on a boundary condition, the equivalent from the longer-grade category is used.
a

Table 7-10. Passenger-Car Equivalents for Trucks on Step 3. Find the point on the horizontal axis corresponding to
Downgrades the appropriate flow rate (vp) in pcphpl.
downgrade length Step 4. Read up to the FFS curve identified in Step 2 and
(%) (mi) ETa determine the average passenger-car travel speed corresponding
percent trucks 5 10 15 20 to that point.
Step 5. Determine the level of service by determining the density
<4 All 1.5 1.5 1.5 1.5
4 ≤4 1.5 1.5 1.5 1.5 region within which the point on the FFS curve falls. These regions
4 >4 2.0 2.0 2.0 1.5 are labeled in Figure 7-3. The density can also be computed as
5 ≤4 1.5 1.5 1.5 1.5
5 >4 5.5 4.0 4.0 3.0 D = vp /S (7-5)
6 ≤2 1.5 1.5 1.5 1.5
6 >2 7.5 6.0 5.5 4.5 where:
a
D = density (pc/mi/ln),
vp = service flow rate (pcphpl), and
S = average passenger-car travel speed (mph).
Table 7-7. For grades of at least 4 percent and longer than 2 mi,
use the specific values shown in Table 7-10. For RVs on down- The level of service can then be determined from the density ranges
grades, use the passenger-car equivalents for level terrain given in shown in Table 7-1.
Table 7-7 in all cases. A graphic example of these steps is given in Figure 7-4 for a
multilane highway with a free-flow speed equal to 52 mph and a
Computing the Heavy-Vehicle Adjustment Factor flow rate equal to 1,700 pcphpl. A free-flow speed curve for 52
mph has been shown with a dashed line; at this speed, capacity is
Once values for ET and ER have been determined, the adjustment
approximately 2,040 pcphpl. Reading up to this curve from a flow
factor for heavy vehicles may be computed as follows:
rate of 1,700 pcphpl, one finds that the average travel speed is
fHV = 1/[1 + PT (ET − 1) + PR(ER − 1)] (7-4) approximately 50 mph and that the highway is operating at LOS D.
If the user does not need to know the average travel speed but
where: only the level of service, this can be accomplished by determining
fHV = adjustment factor for the presence of heavy vehicles in from Table 7-1 the MSF for each level of service for the appro-
the traffic stream, priate FFS (or by interpolation between adjacent FFS columns, if
ET, ER = passenger-car equivalents for trucks and buses and for necessary) and then the level of service within which the flow
RVs, respectively, and rate falls.
PT, PR = proportion of trucks and buses and RVs, respectively,
in the traffic stream (expressed as a decimal).
SEGMENTING THE ROADWAY
VOLUME DISTRIBUTION BY LANE
The procedures described in this chapter are best applied to
It is not necessary to know the volume distribution by lane to homogeneous sections of roadway where the variables that affect
determine the capacity and level of service of a multilane highway; travel speeds are constant. Therefore, it may sometimes be neces-
nevertheless, there may be situations in which this lane distribution sary for the analyst to divide a roadway into separate sections for
would be useful. Field data show that under high-volume condi- analysis. The following conditions should generally indicate that
tions (equivalent to LOS D and E), the right lane of a multilane segmenting of the roadway is necessary:
highway is underutilized. For one direction of flow on a four-lane
highway, the user can expect 40 percent of the traffic to be moving T A change in the basic number of travel lanes along the
in the right lane and 60 percent in the left lane. With three lanes highway,
in one direction, the user can expect 25 percent of the traffic in T A change in the median treatment along the highway,
the right lane, 37.5 percent in the center lane, and 37.5 percent in T A change of grade of 2 percent or more or a constant upgrade
the left lane. over 4,000 ft long,
T The presence of a traffic signal along the multilane highway,
DETERMINATION OF LEVEL OF SERVICE
T A significant change in the density of access points within a
defined area on the route,
The level of service on a multilane highway can be determined T Different speed limits along the highway, and
directly from Figure 7-3 on the basis of the free-flow speed (FFS) T The presence of a bottleneck condition.
and the service flow rate (vp) in pcphpl. The procedure is as
Some judgment must be applied when a roadway is segmented
follows:
for analysis. In general, the minimum length of a study section
Step 1. Define and segment the highway as appropriate. should be 2,500 ft. Also, the limits of study sections should be no
Step 2. On the basis of the actual free-flow speed on a highway closer than 1⁄4 mi to a signalized intersection. The procedures in
segment, an appropriate speed-flow curve of the same shape as this chapter are based on average conditions observed over an
the typical curves in Figure 7-3 is drawn. The curve should inter- extended highway section with generally consistent physical
cept the y-axis at the free-flow speed. characteristics.
Figure 7-4. Example of graphic solution using speed-flow curves.
V. PROCEDURES FOR APPLICATION
The procedures for capacity analysis of multilane rural and sub- (c) lateral clearances, (d) grades, (e) length of grades, and (f) type
urban highways are divided into three analysis types: operational, of terrain.
design, and planning. 2. Volume—The existing traffic volume or the projected future
volume must be known in vehicles per hour (vph) for the hour of
interest (usually the peak hour).
OPERATIONAL ANALYSIS 3. Speed—The existing free-flow speed for passenger cars ei-
ther measured directly or estimated must be known for the hour
For these applications, traffic and geometric conditions must be of interest.
known for an existing highway or estimated for a future highway. 4. Traffic Characteristics—Detailed traffic characteristics are
The analysis focuses on the determination of a level of service needed for operational analysis, including (a) the PHF, (b) percent
and on estimates of travel speed and density of the traffic stream of trucks and buses, and (c) percent of RVs.
along the highway. 5. Roadway Environment—The multilane highway must be
The typical situations that can be resolved through this type of classified as either divided or undivided and the total number of
analysis include the comparison of flow conditions for different access points (driveways plus unsignalized intersections) along
volume levels and number of lanes. This type of analysis might each side of the highway should be known.
also be used to establish the impacts of a change in the number of
access points along a given section of multilane highway. Another
typical application of an operational analysis might be to develop Segmenting the Roadway
several alternative packages that would be used to improve the
level of service or travel speed along a multilane highway.
As described in Section IV, Methodology, the analysis proce-
dures are best applied to multilane highway segments of relatively
Data Requirements uniform characteristics. Therefore, significant changes in roadway
or traffic characteristics require the roadway to be divided into
The following information must be available as inputs to the separate segments for analysis. Signalized intersections should be
operational analysis procedure: located at least 1⁄4 mi from the segment ends because they can
affect speed and volume. Signalized intersections cannot be ana-
1. Geometrics—The geometrics of the highway should be spec- lyzed using the procedure in this chapter; the procedure contained
ified in detail, including (a) number of lanes, (b) lane widths, in Chapter 9 should be used instead. It is inappropriate to classify
a long segment at a single level of service when various subseg- is built into the access-point adjustment. The ability of drivers to
ments are experiencing different levels of service and different change lanes on multilane facilities and the presence of shoulders
operating conditions. Careful division of the roadway into uniform and turn lanes play a large role in minimizing the delay associated
analysis segments will avoid this inadequacy. with turning vehicles. On multilane highway segments, in which
turning volumes significantly affect traffic flow, either because of
poor geometric conditions or unusually high turning movement
Computational Steps volumes, a field study should be performed to determine an appro-
priate speed adjustment attributable to a particular access point or
The general approach taken in operational analysis for which points. However, any adjustments should replace those already
field-measured speeds are unavailable is to use Equations 7-1 and in the procedure so as not to double-count the effect of turning
7-3 to solve for the free-flow speed (FFS) and the service flow movements.
rate (vp). The resulting values are used to find the density and level Use of posted speed limits to determine free-flow speeds should
of service in Figures 7-2 and 7-3 and in Table 7-1. be undertaken with extreme care. The research on which these
The following computational steps are used. analysis procedures are based indicates that speed limits are gener-
1. The segment’s free-flow speed is determined by either direct ally lower than 85th-percentile speeds. Speed limits, which are set
field measurements of passenger-car speeds or by using data from unusually low, should not be relied on to determine free-flow
a similar roadway. If the free-flow speed is estimated, Equation speed. It also must be realized that a change in speed limit will
7-1 must be used to convert the ideal free-flow speed to an actual not necessarily cause a change in free-flow speeds or level of
free-flow speed. The adjustments needed can be found in the ap- service.
propriate tables: When the analysis of a segment suggests the existence of LOS
F conditions, it may be useful to estimate the propagation of queues
FM, median type (Table 7-2) upstream of the breakdown. A detailed technique for such analysis
FLW, lane width (Equation 7-2 and Table 7-3) is included in Chapter 6.
FLC, lateral clearance (Table 7-4)
FA, access-point density (Table 7-5 or 7-6)
2. The hourly flow rate in pcphpl is calculated for each direction DESIGN ANALYSIS
of flow by using Equation 7-3. The heavy-vehicle adjustment fac-
tor is calculated by using Equation 7-4 and Tables 7-7 through
To use the procedures in this chapter for design, a forecast
7-10.
of future traffic volumes has to be made and the general geometric
3. Figure 7-3 is used to set a speed-flow curve at the appropriate
and traffic control conditions, such as speed limits, must be
free-flow speed. Then the travel speed and level of service can be
estimated. With these data and a threshold level of service, an
determined by reading up from the flow rate (pcphpl).
estimate of the number of lanes required for each direction of
4. Density is determined by using either Table 7-1 and Figure
travel can be made. Another application for design involves
7-2 or more accurately by using Equation 7-5.
5. The maximum service flow rate (Mvp), maximum v/c ratio, identification of the level of service achievable if the design
and maximum density for a given level of service can be deter- used a minimum number of lanes (rather than increased the
mined by using Table 7-1. lanes to maintain the objective level of service). The design
application might also be used in conjunction with Chapter 8
A worksheet for operational analysis is shown in Figure 7-5. to assess the traffic flow characteristics of a four-lane road
It provides a useful format for the organization and display of compared with a two-lane cross section.
computations and can also be used for design applications as de-
scribed in the following section.
Data Requirements
Interpretation of Results
The design analysis requires less-detailed data than does the
Operational analysis results in an estimate of the operating char- operational analysis. Data are required on general geometric condi-
acteristics of the traffic stream for the road segment under study. tions, future traffic volumes, and roadway environment.
The densities and travel speeds estimated on the basis of Figures
7-2 and 7-3 represent average U.S. conditions; local conditions 1. Geometric Conditions: (a) lane width, (b) lateral clearance
may vary somewhat from these values. and median type, (c) type of terrain, (d) grade, and (e) grade length.
The densities drawn from Figure 7-2 are expressed in passenger 2. Volume: (a) directional design-hour volume, (b) traffic com-
cars per mile per lane. When field measurements of density are position, and (c) PHF.
used to determine level of service, data expressed in vehicles per 3. Roadway Environment: development environment, rural or
mile per lane must be converted to passenger cars per mile per suburban.
lane using heavy-vehicle equivalencies before they are compared
with the density criteria in Table 7-1. The average travel speeds The desired or expected level of service is normally input into
drawn from Figure 7-3 are based on passenger cars in the traffic the design analysis. Many of the preceding adjustments can be
stream. controlled in the design process, and the varying geometric or
Although the effect of turning volumes is not explicitly consid- environmental conditions may affect the number of lanes necessary
ered in the analysis procedures, some effect of turning movements to reach a desired level of service.
Figure 7-5. Worksheet for operational and design analysis.

Relationship to AASHTO Design Criteria quences. Although there are no set guidelines for such decisions,
designers should perform an operational analysis on the possible
The LOS values used in the AASHTO Policy on Geometric choices for N to determine the level of service and approximate
Design of Highways and Streets (1990) are not directly applicable speed and density that would result. This allows such decisions
to this chapter. The AASHTO design criteria for multilane high- to be made with some knowledge of the operational impacts—
ways may, however, be related to using the following v/c values: knowledge that must be weighed against the relative costs
involved.
1. Rural design—0.50 (i.e., LOS B, 1,000 pcphpl maximum).
The decision on number of lanes in a specific segment of a
2. Suburban design—0.75 (i.e., LOS C, 1,500 pcphpl
multilane highway also depends on their continuity with lanes in
maximum).
adjacent segments and with the rest of the highway system. Fre-
quent adding or dropping of lanes along a highway is not practical,
Segmenting the Roadway although either may be considered at critical locations.
On specific grades, a larger number of lanes may be required
The roadway undergoing design must be separated into uniform on the upgrade than on the downgrade. This is a clear indication
segments. Changes in terrain, significant grades, major junctions at that a climbing lane is required. For a more precise treatment of
which demand volume changes significantly, changes in the devel- such cases, Chapter 3 contains a detailed procedure for the design
opment environment, and similar conditions would indicate the need and evaluation of climbing lanes.
to begin a new segment for design analysis. Along significant grades,
the upgrade and downgrade must be considered separately.
PLANNING ANALYSIS
Computational Steps A planning analysis is directed toward estimating the number

of lanes required to accommodate given traffic conditions. It differs
The general approach to design analysis involves determination from the design application in that the analyst usually has available
of the number of lanes, N, required to carry the expected traffic a value for annual average daily traffic (AADT) and a minimal
volumes at a desired level of service. The following computational definition of the facility being planned.
steps are used: In the planning stage, details of specific grades and other geo-
1. An ideal free-flow speed is estimated, based on either local metric features usually do not exist. Further, traffic forecasts are
conditions or an anticipated speed limit. As stated throughout these not precise. Thus, at the planning level, capacity analysis is approx-
procedures, however, care must be exercised when speed limit is imate and serves to give a general idea of the highway geometrics
used as a basis for free-flow speed estimation. required.
2. By using Equation 7-1, the actual free-flow speed is deter-
mined on the basis of anticipated geometric and environmental
conditions. The adjustments necessary are Data Requirements
FM, median type (Table 7-2)

The planning methodology assumes that ideal geometrics exist
FLW, lane width (Table 7-3)
and that traffic streams consist of only passenger cars and trucks.
FLC, lateral clearance (Equation 7-2 and Table 7-4)
The access-point density assumed is applicable to fringe urban and
FA, access-point density (Table 7-5 or 7-6)
suburban conditions. The required input data include
3. By using Figure 7-3, the appropriate speed-flow curve is
1. A forecast of the AADT for the design year,
drawn corresponding with the actual free-flow speed.
2. A forecast percent of trucks,
4. By using Figure 7-3, the service flow rate necessary to reach
3. The anticipated ideal free-flow speed of the roadway seg-
the desired level of service is determined. This flow rate may be
ment, and
maximum, minimum, or at some midpoint within the LOS range
4. A general classification of terrain type.
depending on the design goals.
5. By using Equation 7-3, the number of lanes, N, necessary to Table 7-11 was developed using free-flow speeds of 60 and 50
reach the desired level of service is calculated. The heavy-vehicle mph under ideal conditions. Using Equation 7-1, the free-flow
adjustment factor is determined by using Equation 7-4 and Tables speed was adjusted for 20 access points per mile (all other condi-
7-7 through 7-10. tions are ideal). The limiting flows (in pcphpl) for LOS A through
Figure 7-5 is a worksheet that may be used for design as well E were found from Figure 7-3 and multiplied by the peak-hour
as operational analysis. It is a useful form for performing and factor (0.9). These are the values in vehicles per hour per lane when
summarizing the results of design computations. there are no trucks. The vehicle volumes with varying percents of
trucks were found by applying the appropriate heavy-vehicle
factor.
Interpretation of Results The AADT is a necessary input for the planning of any highway
and will generally be available for capacity analysis. Vertical align-
Design computation for N generally results in a fraction. Be- ment and truck presence may only be estimates on the part of the
cause the number of lanes must be an integer value, the designer analyst, based on the general terrain conditions of the area through
is faced with deciding whether to reduce or increase the computed which the highway will pass and on the anticipated character of
value to the nearest integer, a decision with economic conse- traffic it is intended to serve.
Table 7-11. Service Flow Rates in Vehicles per Lane for Use in Planning Analysis
type of level of
terrain service free-flow speed (ideal conditions) = 60 mph free-flow speed (ideal conditions) = 50 mph
percent trucks 0 5 10 15 20 0 5 10 15 20
Level A 590 580 570 550 540 490 470 460 450 440
B 990 970 940 920 900 810 790 770 750 740
C 1,360 1,330 1,290 1,260 1,240 1,130 1,110 1,080 1,050 1,030
D 1,620 1,580 1,540 1,510 1,470 1,350 1,320 1,290 1,260 1,230
E 1,890 1,840 1,800 1,760 1,720 1,710 1,670 1,630 1,590 1,550
Rolling A 590 540 500 460 420 490 440 410 370 350
B 990 900 830 760 710 810 740 680 620 580
C 1,360 1,240 1,130 1,050 970 1,130 1,030 950 870 810
D 1,620 1,470 1,350 1,250 1,160 1,350 1,230 1,130 1,040 960
E 1,890 1,720 1,580 1,450 1,350 1,710 1,550 1,430 1,320 1,220
Mountainous A 590 480 400 340 300 490 390 320 280 240
B 990 790 660 570 500 810 650 540 460 410
C 1,360 1,090 910 780 680 1,130 910 760 650 570
D 1,620 1,300 1,080 930 810 1,350 1,080 900 770 680
E 1,890 1,510 1,260 1,080 950 1,710 1,370 1,140 980 860
NOTE: Lane widths are 12 ft. Shoulder width is 6 ft. PHF = 0.9. Number of access points = 20 per mi. Divided highway.
Computational Steps direction offer a good indication that a multilane highway may be
inappropriate for the anticipated conditions and that a limited-
The following steps are involved in conducting a planning access highway should be considered.
analysis:
1. AADT is converted to DDHV using the following equation: SIGNALIZED INTERSECTIONS ON MULTILANE
HIGHWAYS
DDHV = AADT × K × D (7-6)
where: Multilane highways will generally have signalized intersections
at widely spaced intervals, occurring at major junction points
AADT = forecast average annual daily traffic (vpd),
that are not grade separated. These intersections may be subjected
DDHV = directional design hourly volume (vph),
to a detailed analysis using the methodology of Chapter 9.
K = percent of AADT occurring in the peak hour, and
D = percent of peak-hour traffic in the heaviest direction.
THREE-LANE HIGHWAYS WITH PERMANENTLY ASSIGNED
Values of K and D should be based on local or regional characteris- THIRD LANES
tics. Further discussion of these parameters appears in Section VI,
Sample Calculations, Calculation 5. Use of three-lane highways, which declined in the late 1960s,
2. An appropriate maximum service flow rate in vehicles, Mvv, has recently increased. Three-lane highways may be operated in
is selected from Table 7-11 for the prevailing truck percent and a number of ways, the most common of which include
terrain and for the desired level of service.
3. The number of lanes that would be required in each direction 1. Use of the center lane as a continuous left-turn lane (more
of the highway is computed using the following equation: common in suburban settings).
2. Alternate assignment of the center lane to one direction, then
N = DDHV/(Mvv) (7-7) to the other, providing alternating, exclusive passing lanes for each
direction of flow.
Figure 7-6 is a worksheet that may be used for planning analysis.
3. Permanent operation of a long segment of three-lane highway
with two lanes in one direction and one in the other.
The added lane in the preferred direction on a three-lane highway
Planning analysis results in an estimate of N, the number of is generally less efficient than a full four-lane facility, because the
lanes required in each direction, for the multilane highway in added lane exists for distances of less than 1 to 2 mi. The added lane
question. This estimate is based on general information, and plan- is used primarily to pass slower-moving vehicles (particularly on
ning computations must be refined during the design phase of the long upgrades) and to make left turns. This added lane increases the
project. capacity of the two-lane highway by providing for more efficient
Multilane highways with more than three lanes in each direction passing and reducing left-turn conflicts but such a three-lane high-
are rare, and those with more than four lanes are virtually nonexis- way would not approach the capacity of a four-lane highway even
tent. Computations resulting in more than four lanes in each in the preferred direction.
Figure 7-6. Worksheet for planning analysis.

VI. SAMPLE CALCULATIONS
CALCULATION 1—OPERATIONAL ANALYSIS OF On the free-flow speed curve, the average speed is still 46.0 mph
AN UNDIVIDED HIGHWAY with LOS C and a density of 26.4 pc/mi/ln.
For the 2.5 percent downgrade over the same distance, a
1. Description—A 3.25-mi section of undivided highway in
different value of fHV results. Because the grade is less than 4
generally level terrain has a free-flow speed of 46 mph measured
percent, the values of ET and ER are those used in the general
at a flow rate of 1,000 pcphpl. At a distance of 6,000 ft from one
terrain analysis for level terrain. The heavy-vehicle factor, fHV,
end of the section is a 3,200-ft segment with a 2.5 percent grade.
has been determined above as 0.94, and vp as 1,123 pcphpl.
There are 14 driveways on each side of the road in the 3,200-ft
The average travel speed on the downgrade is 46.0 mph with
segment and approximately 21 driveways per mile throughout the
LOS C and a density of 24.4 pc/mi/ln. Figure 7-8 shows the
rest of the section. The cross section for the entire roadway is
solution for the downgrade.
composed of four 11-ft lanes with a 4-ft shoulder on either side.
The afternoon peak-hour traffic volume throughout is 1,900 vph
in each direction with 8 percent trucks, 3 percent buses, and 2 CALCULATION 2—OPERATIONAL ANALYSIS OF A
percent RVs. The peak-hour factor is 0.90. DIVIDED HIGHWAY
2. Objective—Determine the overall average travel speed for
passenger cars, the density of traffic, and the level of service for 1. Description—An east-west multilane highway has a five-
each direction of flow. Are there any potential problem areas? lane cross section and is composed of two travel lanes in each
3. Solution—Free-flow speed has been measured under low- direction separated by a two-way left-turn lane (TWLTL). Lane
volume conditions for this road. No adjustments to free-flow speed widths are 12 ft and there is sufficient lateral clearance on each
will be made. side of the roadway. The study area is about 2 mi long and contains
a 6,000-ft-long, 4 percent upgrade westbound, followed by level
terrain for 5,000 ft. The section has a volume of 1,500 vph in each
The value for free-flow-speed is 46 mph. A 46-mph curve is direction with 4 percent trucks and 2 percent buses. The north side
sketched onto a copy of Figure 7-3 (as shown in Figure 7-7). The of the highway has 27 access points evenly distributed approxi-
only calculation required is for service flow rate. An initial solution mately 400 ft apart throughout the section. The south side has only
is obtained using a general level terrain segment. 10 access points, all located in the level terrain section. Data are
Equation 7-3 is used to calculate the service flow rate: available on current travel speeds along the roadway. The 85th-
percentile speed of passenger cars on the upgrade (westbound) is
V 48 mph and 54 mph on the downgrade (eastbound). On the level
vp =
(N)(PHF)(fHV) section, the 85th-percentile speed is 52 mph in both directions.
The PHF is 0.90.
The heavy-vehicle adjustment factor, fHV, is calculated using Equa- 2. Objective—Determine the level of service in the study area.
tion 7-4. 3. Solution—Proper analysis requires the 2-mi study area to be
separated into two segments, the level segment and the segment
fHV = 1/[1 + PT (ET − 1) + PR (ER − 1)] with sustained grade. Note that because the number of access
points varies by direction, the level segment must be evaluated in
ET, the passenger-car equivalent for trucks and buses, is determined both directions. Figures 7-9 and 7-10 present the worksheets used
to be 1.5 from Table 7-7 on the basis of an extended level terrain for this problem.
segment. Similarly, from Table 7-7, ER is 1.2. The resulting heavy-
vehicle factor, fHV, is 0.94. The first step in determining the level of service for each seg-
The service flow rate is calculated as ment is to calculate the free-flow speed for ideal conditions for
each segment on the basis of the rule of thumb for 85th-percentile
vp = (1,900 vph)/(2 × 0.90 × 0.94) = 1,123 pcphpl speeds. These speeds are
Level segment (both directions): 50 mph;
As shown in Figure 7-4, the average travel speed on the total Sustained-grade segment:
section is 46.0 mph and the level of service is C. The density is Westbound (upgrade), 46 mph;
calculated as 24.4 pc/mi/ln. Figure 7-7 shows this part of the Eastbound (downgrade), 52 mph.
solution and the calculations as worked on the operational analysis
worksheet. The only segment that might have a different result is The free-flow speed can then be calculated using Equation 7-1:
the 3,200-ft 2.5 percent grade, analyzed as an upgrade in one FFS = FFSI − FM − FLW − FLC − FA
direction and a downgrade in the other.
For the 2.5 percent upgrade over approximately 0.6 mi, the FM, the adjustment for median type, is determined using Table
value of ET is 2.2 from Table 7-8 by interpolation for 11 percent 7-2. Because a TWLTL is considered to have the same effect as
trucks and buses on an upgrade between 2 and 3 percent over 1⁄2 a median, FM = 0.0. FLW, the lane width adjustment, and FLC, the
to 3⁄4 mi. ER is 2.0 from Table 7-9 for 2 percent RVs on a grade lateral clearance adjustment, are both 0.0, as determined from
between 2 and 3 percent and more than 1⁄2 mi long. By using Tables 7-3 and 7-4, respectively. FA, the adjustment for access-
Equation 7-5, fHV is determined to be 0.87. The service flow rate point density, is determined using Table 7-5. The access-point
on the upgrade is vp = (1,900 vph)/(2 × 0.90 × 0.87) = 1,213 pcphpl. densities for each segment are
Figure 7-7. Illustration of solution to Calculation 1—general segment.

Figure 7-8. Illustration of solution to Calculation 1—grade segment.

Level segment: By using Equation 7-1, the free-flow speed is calculated as 52.5.
Westbound, 13 access points per mile; A speed-flow curve is drawn at a free-flow speed of 52.5 mph.
Eastbound, 10 access points per mile. From the speed-flow curve, at LOS D, the maximum service flow
Sustained-grade segment: rate for an ideal facility is determined to be approximately 1,750
Westbound (upgrade), 13 access points per mile; pcphpl.
Eastbound (downgrade), 0 access points per mile. Applying the percentages for design-hour volume and direc-
tional distribution produces a volume of 3,300 vph in the peak
From Table 7-5, the access-point adjustments for the level seg- direction. By using Equation 7-3 and a heavy-vehicle adjustment
ment are 3.3 and 2.5 mph in the westbound and eastbound direc- factor based on Table 7-7, N, the minimum number of lanes re-
tions, respectively, or a 0.25-mph reduction per access point. On quired per direction, is 2.4. Because essentially ideal geometric
the grade segment, FA for the upgrade (westbound) is 3.3 mph and conditions were assumed, three lanes per direction, or a six-lane
0.0 mph for the downgrade (eastbound). cross section, are needed.
The free-flow speeds that result after appropriate adjustments If it is assumed that design criteria require a 12-ft median to
are 46.7 and 47.5 in the westbound and eastbound directions of allow for turning bays as needed, 12-ft lanes, and 10-ft shoulders,
the level segment, 42.7 on the upgrade (westbound), and 52.0 on a six-lane cross section would require 104 ft of right-of-way. If
the downgrade (eastbound). The appropriate free-flow speed the existing 90-ft right-of-way is the maximum allowable, one
curves can then be drawn on Figure 7-3. solution would be to reduce the lane widths to 11 ft, allow for
Next the service flow rate is calculated for each segment by only a 6-ft raised median and 4-ft shoulders, and provide a 5-ft
direction using Equation 7-3. The heavy-vehicle adjustment factor offset to the right-of-way. The resulting total lateral clearance is
is calculated using Equation 7-4. For the level segment, ET is 1.5 at least 10 ft in each direction of travel. The access-point density
from Table 7-7. In the upgrade direction, Table 7-8 provides an would be unchanged. This particular design would provide free-
ET of 7.0, and Table 7-10 provides an ET of 1.5 in the downgrade flow speed of 50.2 mph. A six-lane cross section of this reduced
direction. The service flow rates are calculated as design would provide an LOS D (almost LOS C) for the estimated
Level segment (both directions): 859 pcphpl; service flow rate of 1,343 pcphpl. The anticipated average travel
Sustained-grade segment: speed would be 50.2 mph for this six-lane cross section. Thus, a
Westbound (upgrade), 1,126 pcphpl; six-lane roadway with less than ideal geometry would satisfy the
Eastbound (downgrade), 859 pcphpl. design requirements of LOS D operation. This design is only one of
several, however, that would produce the required level of service.
On the basis of these flow rates and the free-flow speeds, it is Furthermore, the analysis does not address safety considerations,
determined from the speed-flow curves that the level segment which may be overriding.
operates at LOS B in both directions. On the sustained-grade sec-
tion, the upgrade operates at LOS C and the downgrade section
CALCULATION 4—DESIGN ANALYSIS OF AN
at LOS B. EXISTING MULTILANE ROADWAY
1. Description—A six-lane divided roadway located in an urban

CALCULATION 3—DESIGN OF A MULTILANE setting is the subject of a rehabilitation program aimed at improv-
HIGHWAY ing traffic operations. The 2.5-mi section located in level terrain
has signalized intersections at either end and one signal installation
1. Description—A 2-mi section of a multilane highway must
in the middle. This last signal is being replaced by a grade-
be designed to carry an average daily traffic volume of 60,000
separated roadway at the same location. The current peak-hour
vehicles at LOS D. On the basis of local data from other multilane
flow rate is 1,400 pcphpl, and the current average travel time
highways, the design-hour volume should be 10 percent with a
through the section is 3.0 min. At-grade access is provided only
55/45 directional split, a PHF of 0.9, and 5 percent trucks. The
at the signalized intersections. The current roadway has 11-ft lanes
highway is expected to have a 50-mph speed limit and approxi-
separated by a 16-ft raised median. The shoulder on each side of
mately 10 access points per mile and to be located in rolling terrain.
the roadway measures 4 ft.
Existing right-of-way consists of a 90-ft corridor.
2. Objective—Determine the expected travel speed of the im-
2. Objective—Determine the cross section that must be pro-
proved roadway when the grade separation is complete. How much
vided to meet the design criteria and find the expected travel speed
additional traffic can be added and maintain the improved level of
for passenger cars on this highway.
service?
3. Solution—A design analysis allows the designer to deter-
3. Solution—The new free-flow speed as a result of the change
mine, by trial and error, the appropriate geometrics that are neces-
in travel time must be found. The existence of the grade separation
sary to provide a given level of service. Initially, it is assumed
is not a reason for segmenting the roadway. Figure 7-12 shows
that an ideal facility will be provided. The resulting cross section
the results of the analysis.
will have 12-ft lanes, a raised median, and shoulders 6 ft or more
in each direction. The restrictions imposed by available right-
of-way can then be investigated. Figure 7-11 shows the worksheet Before removal of the signal, the average travel time through
for this problem. the 2.5-mi section was 180 sec under free-flow conditions, or an
average travel speed of 50 mph. At a flow rate of 1,400 pcphpl,
LOS is D. Removing the traffic signal is estimated to reduce the
In the absence of better information, the free-flow speed under average travel time by 30 to 150 sec. This corresponds to a free-
ideal conditions is estimated using the rule of thumb in Section flow speed of approximately 60 mph. (No increase in traffic vol-
IV, Determination of Free-Flow Speed. ume is expected as a result of constructing the grade separation.)
Figure 7-9. Illustration of solution to Calculation 2—level segment.

Figure 7-10. Illustration of solution to Calculation 2—grade segment.

Figure 7-11. Illustration of solution to Calculation 3.



Sketching the new free-flow speed on the graph indicates that the percent of the AADT expected in the design hour. The value
at a flow rate of 1,400 pcphpl, the expected average travel speed for D reflects the directional distribution of traffic in that hour.
is 60 mph under LOS C conditions. By reading to the right to the Historically, these values have been referred to as the K-factor and
density curve marking the boundary between LOS C and D, the the D-factor. For suburban areas, these values have traditionally
maximum flow rate for LOS C at 60 mph is estimated to be 1,620 been 0.10 and 60/40, respectively. In this example, these values
pcphpl. Thus, an additional 220 passenger cars per lane could be result in a DDHV of 2,520 vph. The table at the bottom of the
added and still maintain LOS C. worksheet in Figure 7-13 indicates that a six-lane facility with a
free-flow speed of 50 mph under ideal conditions can handle 2,850
CALCULATION 5—PLANNING ANALYSIS FOR A to 3,090 vph and still operate at LOS C. The calculations for this
NEW ROADWAY planning analysis are displayed on the worksheet in Figure 7-13.
1. Description—A new corridor is to be developed on the out- The value for K should represent local practice for the design
skirts of a metropolitan area. The highway is to be constructed in hourly volume. It may be the ratio between the 30th, 50th, or some
approximately 10 years. Future traffic projections indicate that the other hour of the year and AADT. This information is generally
highway should be designed to carry 42,000 vehicles per day with available to the user. The data for similar facilities in the general
5 to 10 percent trucks. At the current time it is anticipated that area may have been collected for other analyses or as a part of
the corridor will have ideal design conditions through rolling ter- the modeling process for developing a traffic forecast. The value
rain. From similar multilane highways, it is expected that the free- for K, which is dependent on the environment around the roadway,
flow speed on this highway will be 50 mph. has been observed to increase with distance from urban areas. The
2. Objective—Determine the number of lanes needed to provide value for D, the directional distribution, varies as a function of
LOS C operation. route type and distance from activity centers. Usually the high-
3. Solution—In the absence of more specific information on volume direction is evaluated.
the proposed highway and forecast traffic, the planning approach Use of default values can lead to results that differ significantly
may be considered. An estimate will be made of the directional from those using more specific local data. It is strongly recom-
design hourly volume (DDHV), which will be compared with the mended that when default values are used in applying the planning
values for Table 7-11. methodology, local values for defaults be determined. Further-
more, the planning method is more useful for scaling roadways
The first step is to calculate the DDHV using Equation 7-7. on a systemwide basis than for making design decisions for a
AADT reflects a year-round average. The value for K represents specific roadway.
VII. ACKNOWLEDGMENTS
This chapter is based upon research performed for the National the Committee on Highway Capacity and Quality of Service. The
Cooperative Highway Research Program by JHK & Associates members of the subcommittee are Ulrich Brannolte; Barbara
and the Midwest Research Institute. The principal researchers were Ostrom; Ronald Pfefer, Chairman; William Reilly; and Fred
William R. Reilly, Douglas W. Harwood, James M. Schoen, and Rooney.
Michael F. Holling.
The chapter is a revision of the research product and was devel-
oped by members of the Subcommittee on Multilane Highways of
VIII. REFERENCES
1. Hool, J.N., Maghsoodloo, S., Veren, A.D., and Brown, 3. Hauer, E., and Ahlin, F.J., ‘‘Speed Enforcement and Speed
D.D., ‘‘Analysis of Selective Enforcement Strategy Effects on Choice.’’ Accident Analysis and Prevention, Vol. 14, No. 4
Rural Alabama Traffic Speeds.’’ Transportation Research (1982), pp. 267–278.
Record 910, Transportation Research Board, Washington, D.C. 4. Tignor, S.C., ‘‘Driver Speed Behavior on U.S. Streets and
(1983), pp. 74–81. Highways.’’ Compendium of Technical Papers, Institute of
2. Armour, M., ‘‘The Effect of Police Presence on Urban Driving Transportation Engineers (August 5–8, 1990).
Speeds.’’ Australian Road Research, Vol. 14, No. 3 (Sept.
1984), pp. 142–148.

APPENDIX I
FIGURES AND WORKSHEETS FOR USE IN THE ANALYSIS OF

MULTILANE HIGHWAYS
FIGURES
Figure 7-1 Speed-flow relationships on multilane highways......................................................................................................... 7-32

Figure 7-2 Density-flow relationships on multilane highways ...................................................................................................... 7-33
Figure 7-3 Speed-flow curves with LOS criteria. *Maximum density for respective levels of service. **Maximum densities
for LOS E occur at volume-to-capacity ratio of 1.0. They are 40, 41, 43, and 45 pc/mi/ln at free-flow speeds of
60, 55, 50, and 45 mph, respectively ........................................................................................................................... 7-34
WORKSHEETS
Operational and Design Analysis Worksheet.......................................................................................................................................... 7-35
Planning Analysis Worksheet .................................................................................................................................................................. 7-36

7-32

rural and suburban highways
Figure 7-1. Speed-flow relationships on multilane highways.

Figure 7-2. Density-flow relationships on multilane highways.

7-34

rural and suburban highways
Figure 7-3. Speed-flow curves with LOS criteria. *Maximum density for respective levels of service. **Maximum densities for LOS E occur at volume-to-capacity ratio of
1.0. They are 40, 41, 43, and 45 pc/mi/ln at free-flow speeds of 60, 55, 50, and 45 mph, respectively.

Planning Analysis Worksheet

ADDENDUM: ADJUSTMENT FOR DRIVER POPULATION
As described under Other Adjustments (p. 7-6), the methodology that a lesser value reflecting more recreational or weekend traffic
in Chapter 7 for the analysis of traffic flow on multilane highways characteristics should be applied. Where greater accuracy is
is based on traffic data collected across the United States. These needed, comparative field studies of weekday and weekend traffic
data did not specifically focus on the possible variation in traffic flow and speeds are recommended.
flow for differing driver populations (i.e., commuters or frequent In the analysis of multilane highway level of service, the driver
drivers versus infrequent drivers). Although little documented data population effect is reflected as an adjustment to the hourly service
are currently available and the effect of driver population on traffic flow rate, vp . Thus, to determine the service flow rate, the analyst
flow is not well understood, it is generally accepted that traffic should replace Equation 7-3 (p. 7-11) with the following equation:
streams with different characteristics from those consisting of com- V
muters and frequent drivers (i.e., those who drive on weekends, vp =
(N) (PHF) (fHV) (fp)
for recreation, and perhaps even at mid-day) use roadways less
efficiently than do those who drive frequently. A limited number where:
of studies on uninterrupted-flow roadways have reported lower vp = service flow rate (pcphpl),
capacities on weekends, particularly in recreational areas. V = volume (number of vehicles passing a point in 1 hr),
The adjustment factor fp is used to reflect this effect. The values N = number of lanes,
for fp range from 1.0 to 0.85. Typically, the analyst should select PHF = peak-hour factor,
1.0, which reflects weekday commuter traffic (i.e., familiar users), fHV = heavy-vehicle adjustment factor, and
unless there is sufficient evidence or it is the analyst’s judgment fp = driver population adustment factor.

chapter 8
TWO-LANE HIGHWAYS
CONTENTS
i. introduction .......................................................................................................................................................................... 8-2

Levels of Analysis................................................................................................................................................................. 8-2
Operational Characteristics.................................................................................................................................................... 8-2
Ideal Conditions..................................................................................................................................................................... 8-4
ii. methodology.......................................................................................................................................................................... 8-5
Levels of Service................................................................................................................................................................... 8-5
Use of the Peak Hour Factor........................................................................................................................................... 8-7
Analysis of General Terrain Segments............................................................................................................................ 8-7
Analysis of Specific Grades ............................................................................................................................................ 8-8
Highway System Planning .................................................................................................................................................... 8-13
Operational Analysis of General Terrain Segments............................................................................................................. 8-14
Operational Analysis of Specific Grades.............................................................................................................................. 8-15
Planning ................................................................................................................................................................................. 8-17
iv. design and operational treatments................................................................................................................................. 8-17
Passing Sight Distance .......................................................................................................................................................... 8-18
Paved Shoulders .................................................................................................................................................................... 8-18
Three-Lane Highways ........................................................................................................................................................... 8-18
Passing Lanes ................................................................................................................................................................... 8-18
Continuous Two-Way Median Left-Turn Lanes............................................................................................................. 8-20
Reversible Lane................................................................................................................................................................ 8-20
Intersection Treatments .................................................................................................................................................... 8-20
Climbing Lanes ................................................................................................................................................................ 8-20
Turnouts............................................................................................................................................................................ 8-20
Short Four-Lane Sections...................................................................................................................................................... 8-21
v. sample calculations ............................................................................................................................................................ 8-21
Calculation 1—Finding Service Flow Rates for a General Terrain Segment..................................................................... 8-21
Calculation 2—Finding Level of Service for a General Terrain Segment ......................................................................... 8-23
Calculation 3—Finding Service Flow Rates for a Specific Grade...................................................................................... 8-23
Calculation 4—Finding Level of Service and Capacity of a Specific Grade..................................................................... 8-24
Calculation 5—Consideration of a Climbing Lane.............................................................................................................. 8-26
Calculation 6—Planning Application 1 ................................................................................................................................ 8-26
vi. references .............................................................................................................................................................................. 8-27
appendix I. Figures and Worksheets for Use in Analysis of Two-Lane Highways............................................................. 8-28
8-1
8-2 rural highways
I. INTRODUCTION
A two-lane highway may be defined as a two-lane roadway The utilization of capacity reflects the access function, and is
having one lane for use by traffic in each direction. Passing of defined as the ratio of the demand flow rate to the capacity of the
slower vehicles requires the use of the opposing lane where sight facility.
distance and gaps in the opposing traffic stream permit. As Level-of-service criteria utilize all three of the parameters noted
volumes and/or geometric restrictions increase, the ability to above, with percent time delay being the primary measure of ser-
pass decreases, resulting in the formation of platoons in the vice quality. Speed and capacity utilization are secondary
traffic stream. Motorists in these platoons are subject to delay measures.
because of the inability to pass. This chapter provides specific definitions and methodologies
Two-lane highways compose the predominant mileage of most for the estimation of level of service for all types of two-lane
national highway systems. They are used for a variety of highways. Subsequent sections provide a descriptive list of
functions, are located in all geographic areas, and serve a wide treatments for alleviating both spot and section design and/or
range of traffic requirements. Consideration of operating quality operational problems that may arise because of high volume
must account for these disparate traffic functions. and/or geometric restrictions. A set of example calculations is
Efficient mobility is the principal function of major two-lane provided to illustrate the use and application of procedures. A
highways used as primary arteries connecting major traffic genera- complete set of worksheets for all levels of analysis is also
tors or as primary links in state and national highway networks. provided. Illustration 8-1 shows typical views of two-lane, two-
Such routes tend to serve long-distance commercial and recre- way rural highways.
ational travelers, and may have sections of many miles through
rural environments without traffic control interruptions. Consistent
high-speed operations and infrequent passing delays are desirable LEVELS OF ANALYSIS
for these facilities.
Many paved, two-lane rural roads basically serve an accessibility
function. They provide all-weather accessibility to an area, often This chapter is based on a comprehensive study of two-lane
for relatively low traffic volumes. The provision of cost-effective highway operation (1,2). Microscopic simulation combined with
access is the dominant policy consideration. High speed, while additional field data (3) and theoretical considerations were used
beneficial, is not the principal concern. Delay, as indicated by the to develop the methodology. Analysis is provided at two levels:
formation of platoons, and the utilization of capacity are more
1. Operational analysis—This application is intended to deter-
relevant measures of service quality.
mine the level of service for an existing two-lane highway with
Two-lane roads also serve scenic and recreational areas where
existing traffic and roadway conditions, or for projected future
the vista and environment are to be experienced and enjoyed with-
conditions; operational analysis applications are presented for gen-
out traffic interruption or delay. A safe roadway is desired, but
eral terrain segments and for specific grades.
high-speed operation is neither expected nor desired.
2. System planning—This application enables planners to
Short sections of high-volume two-lane roads sometimes serve
quickly determine the AADT volumes which can be accommo-
as short connections between two major multilane roadways or
dated on two-lane highways for various levels of service and terrain
urban centers. For such short links, traffic conditions tend to be
conditions.
better than might be expected for longer two-lane segments, and
the expectations of motorists regarding service quality are gener- Design computations cannot be readily performed for two-lane
ally higher than for longer sections. highways because the number of lanes is fixed. Modifications
For these reasons, three parameters are used to describe service to grade and alignment, however, could improve the operational
quality for two-lane highways: efficiency of a two-lane facility. For other design options, proce-
dures for the appropriate types of facilities would be consulted.
1. Average travel speed. Procedures of Chapter 3, ‘‘Basic Freeway Segments,’’ and Chapter
2. Percent time delay. 7, ‘‘Multilane Highways,’’ would often be useful in investigating
3. Capacity utilization. design alternatives.
The selection of an appropriate level of analysis is based on the
objectives of the analysis, the available data base, and the accuracy
Average travel speed reflects the mobility function of two- requirements.
lane highways, and is the length of the highway segment under
consideration divided by the average travel time of all vehicles
traversing the segment in both directions over some designated
OPERATIONAL CHARACTERISTICS
time interval.
Percent time delay reflects both mobility and access functions,
and is defined as the average percent of time that all vehicles Traffic operations on two-lane, two-way highways are unique.
are delayed while traveling in platoons due to the inability to Lane-changing and passing are possible only in the face of on-
pass. ‘‘Percent time delay’’ is difficult to measure directly in the coming traffic in the opposing lane. Passing demand increases
field. The percent of vehicles traveling at headways less than 5 rapidly as traffic volumes increase, while passing capacity in
sec can be used as a surrogate measure in field studies. the opposing lane declines as volumes increase. Thus, unlike
two-lane highways 8-3
Illustration 8-1. Typical views of two-lane

two-way highways in rural environments.
other types of uninterrupted flow facilities, on two-lane highways, ‘‘Percent time delay’’ is the average percent of the total travel
normal traffic flow in one direction influences flow in the other time that all motorists are delayed in platoons while traveling a
direction. Motorists are forced to adjust their individual travel given section of highway. Motorists are defined to be delayed
speed as volume increases and the ability to pass declines. Two when traveling behind a platoon leader at speeds less than their
traffic stream characteristics, average travel speed and percent time desired speed and at headways less than 5 sec. For field measure-
delay, are used as operational measures describing the quality of ment purposes, percent time delay in a section is approximately
service provided to motorists on a two-lane highway. the same as the percentage of all vehicles traveling in platoons at
A relatively high running speed has become an accepted cri- headways less than 5 sec (2,5).
terion for primary highway design. Mean speeds of traffic flow Percent time delay reflects the changing service quality per-
are frequently observed above 55 mph on primary rural high- ceived by motorists under a wide range of geometric and traffic
ways. Research has shown that speed is fairly insensitive to conditions. At low traffic volumes, motorists are almost never
volume on two-lane highways without significant grades or turn- delayed because demand for passing is low, average headways
ing traffic (4). Consequently, average speeds of less than 50 mph are high, and the ability to pass is high. The percent time delay
are judged undesirable for primary two-lane highways in level for such conditions is near 0 percent. As volumes approach
terrain because of the high percentage of time motorists would capacity, passing demand greatly exceeds passing capacity, ma-
be delayed. jor platoons of traffic exist, and motorists are delayed almost
8-4 rural highways
a. Relationship between average speed and flow on two-lane b. Relationship between percent time delay and flow on two-lane
highways. highways.
Figure 8-1. Speed-flow and percent time delay-flow relationships for two-lane rural highways (ideal conditions).
100 percent of the time. Even though speeds may be relatively reflects the impact of opposing vehicles on passing opportunities,
high near capacity (40 mph or more), driver frustration would be and therefore on the ability to efficiently fill gaps in the traffic
excessive if these conditions routinely existed for long periods stream. This phenomenom restricts capacity to a lower value than
of time. the 2,000 pcphpl which may be accommodated on multilane unin-
The basic relationships between average travel speed, percent terrupted flow facilities.
time delay, and flow are shown in Figure 8-1. These curves Directional distribution is defined to be 50/50 for ideal condi-
assume ideal traffic and roadway conditions. The average speed tions. Most directional distribution factors observed on rural
represents the average travel or space mean speed of all traffic two-lane highways range from 55/45 to 70/30. On recreational
traveling in both directions over the section of highway in routes, the directional distribution may be as high as 80/20 or
question. Percent time delay is the average for all vehicles in more during holiday or other peak periods. Some variation in
the traffic stream. speed and percent time delay occurs by direction with changing
directional distribution factors and volume levels. Minor changes
IDEAL CONDITIONS in average traffic stream characteristics will also occur with
directional distribution.
Ideal conditions for two-lane highways are defined as no restric- The frequency of no passing zones along a two-lane highway
tive geometric, traffic, or environmental conditions. Specifically, is used to characterize roadway design and to define expected
they include: traffic conditions. A no passing zone is defined as any marked no
passing zone or, as a surrogate, any section of road wherein the
1. Design speed greater than or equal to 60 mph. passing sight distance is 1,500 ft or less. The average percentage
2. Lane widths greater than or equal to 12 ft. of no passing zones in both directions along a section is used in
3. Clear shoulders wider than or equal to 6 ft. the procedures.
4. No ‘‘no passing zones’’ on the highway. The typical percentage of no passing zones found on rural
5. All passenger cars in the traffic stream. two-lane highways ranges from 20 percent to 50 percent. Values
6. A 50/50 directional split of traffic. approaching 100 percent can be found on sections of winding
7. No impediments to through traffic due to traffic control or
mountainous roads. No passing zones have a greater effect in
turning vehicles.
mountainous terrain than on level or rolling highway segments.
8. Level terrain.
Heavy platoon formation along a highway section also may
The capacity of two-lane rural highways under these ideal cause greater than expected operational problems on an adjacent
conditions is 2,800 pcph, total, in both directions. This capacity downstream section having restricted passing opportunities.
II. METHODOLOGY
LEVELS OF SERVICE Table 8-2 gives level-of-service criteria for specific grade seg-
ments. These criteria relate the average travel speed of upgrade vehi-
As noted previously, level-of-service criteria for two-lane cles to level of service. Operations on sustained two-lane grades are
highways address both mobility and accessibility concerns. The substantially different from extended segments of general terrain.
primary measure of service quality is percent time delay, with The speed of upgrade vehicles is seriously impacted, as the formation
speed and capacity utilization used as secondary measures. Level- of platoons behind slow-moving vehicles intensifies and passing ma-
of-service criteria are defined for peak 15-min flow periods, and neuvers generally become more difficult. Further, unlike general ter-
are intended for application to segments of significant length. rain segments, where the approximate average travel speed at which
Level-of-service criteria for general terrain segments are given in capacity occurs can be identified, the capacity speed for a specific
Table 8-1. For each level of service, the percent time delay is shown. grade depends on the steepness and length of the grade and volume.
Average travel speed is also shown, with values varying slightly by Because of this, estimation of capacity is complex. Thus, Table 8-2
type of terrain. The body of the table includes maximum values of defines separate level-of-service criteria for specific grade segments.
v/c ratio for the various terrain categories and levels of service A In addition, this chapter includes special computational procedures
through F. The v/c ratios shown in Table 8-1 are somewhat different for sustained grades on two-lane highways.
from those used in other chapters. For two-lane highways, the values Downgrade operations are not specifically addressed by these
given represent the ratio of flow rate to ‘‘ideal capacity,’’ where procedures. Downgrade operations on gentle grades (less than 3
ideal capacity is 2,800 pcph for a level terrain segment with ideal percent) are generally comparable to those on a level roadway.
geometrics and 0 percent no passing zones. Two-lane highways are On more severe grades, downgrade operations are about midway
quite complex, and capacities vary depending on terrain and the de- between those experienced on a level roadway and those experi-
gree of passing restrictions. To simplify computational procedures, enced on an upgrade of equivalent traffic and roadway characteris-
v/c ratios are given in terms of the constant ‘‘ideal capacity’’ of 2,800 tics. The principal concern on steep downgrades is the potential
pcph, total in both directions of flow. for ‘‘runaway’’ trucks.
The level-of-service criteria of Table 8-1 are for extended The highest quality of traffic service occurs when motorists
segments of two-lane rural highways where efficient mobility is are able to drive at their desired speed. Without strict enforce-
the primary objective of the facility. Where speeds have been ment, this highest quality, representative of level-of-service A,
restricted by an agency, such as through a town or village, the would result in average speeds approaching 60 mph on two-
percentage of time delay and capacity utilization are the only lane highways. The passing frequency required to maintain these
meaningful indicators of level of service. speeds has not reached a demanding level. Passing demand is
Table 8-1. Level-of-Service for General Two-Lane Highway Segments

v/c ratioa
level terrain rolling terrain mountainous terrain
percent
time avgb percent no passing zones avgb percent no passing zones avgb percent no passing zones
los delay speed 0 20 40 60 80 100 speed 0 20 40 60 80 100 speed 0 20 40 60 80 100
A ≤ 30 ≥ 58 0.15 0.12 0.09 0.07 0.05 0.04 ≥ 57 0.15 0.10 0.07 0.05 0.04 0.03 ≥ 56 0.14 0.09 0.07 0.04 0.02 0.01
B ≤ 45 ≥ 55 0.27 0.24 0.21 0.19 0.17 0.16 ≥ 54 0.26 0.23 0.19 0.17 0.15 0.13 ≥ 54 0.25 0.20 0.16 0.13 0.12 0.10
C ≤ 60 ≥ 52 0.43 0.39 0.36 0.34 0.33 0.32 ≥ 51 0.42 0.39 0.35 0.32 0.30 0.28 ≥ 49 0.39 0.33 0.28 0.23 0.20 0.16
D ≤ 75 ≥ 50 0.64 0.62 0.60 0.59 0.58 0.57 ≥ 49 0.62 0.57 0.52 0.48 0.46 0.43 ≥ 45 0.58 0.50 0.45 0.40 0.37 0.33
E > 75 ≥ 45 1.00 1.00 1.00 1.00 1.00 1.00 ≥ 40 0.97 0.94 0.92 0.91 0.90 0.90 ≥ 35 0.91 0.87 0.84 0.82 0.80 0.78
F 100 < 45 — — — — — — < 40 — — — — — — < 35 — — — — — —
a
Ratio of flow rate to an ideal capacity of 2,800 pcph in both directions.
b
These speeds are provided for information only and apply to roads with design speeds of 60 mph or higher.
8-6 rural highways
Table 8-2. Level-of-Service Criteria for Specific Passing is virtually impossible under level-of-service E condi-
Grades tions, and platooning becomes intense when slower vehicles or
level of average upgrade other interruptions are encountered.
service speed (mph) The highest volume attainable under level-of-service E defines
A ≥ 55 the capacity of the highway. Under ideal conditions, capacity is
B ≥ 50 2,800 pcph, total in both directions. For other conditions, capacity
C ≥ 45 is lower. Note that the v/c ratios of Table 8-1 are not all 1.00 at
D ≥ 40 capacity. This is because the ratios are relative to ‘‘ideal capacity’’
E ≥ 25–40a as discussed. Operating conditions at capacity are unstable and
F < 25–40a
difficult to predict. Traffic operations are seldom observed near
a
The exact speed at which capacity occurs varies with the percentage and capacity on rural highways, primarily because of a lack of demand.
length of grade, traffic compositions, and volume; computational procedures
are provided to find this value. Capacity of two-lane highways is affected by the directional
split of traffic. As directional split moves away from the 50/50
‘‘ideal’’ condition, total two-way capacity is reduced, as follows:
Directional Total Ratio of Capacity to

well below passing capacity, and almost no platoons of three or Split Capacity (pcph) Ideal Capacity
more vehicles are observed. Drivers would be delayed no more
than 30 percent of the time by slow-moving vehicles. A maximum 50/50 2,800 1.00
flow rate of 420 pcph, total in both directions, may be achieved 60/40 2,650 0.94
under ideal conditions. 70/30 2,500 0.89
Level-of-service B characterizes the region of traffic flow 80/20 2,300 0.83
wherein speeds of 55 mph or slightly higher are expected on level 90/10 2,100 0.75
terrain. Passing demand needed to maintain desired speeds be- 100/0 2,000 0.71
comes significant and approximately equals the passing capacity
at the lower boundary of level-of-service B. Drivers are delayed For short lengths of two-lane road, such as tunnels or bridges,
up to 45 percent of the time on the average. Service flow rates of opposing traffic interactions may have only a minor effect on
750 pcph, total in both directions, can be achieved under ideal capacity. The capacity in each direction may approximate that of
conditions. Above this flow rate, the number of platoons forming a fully loaded single lane, given appropriate adjustments for the
in the traffic stream begins to increase dramatically. lane width and shoulder width (5).
Further increases in flow characterize level-of-service C, re- As with other highway types, level-of-service F represents heav-
sulting in noticeable increases in platoon formation, platoon size, ily congested flow with traffic demand exceeding capacity. Vol-
and frequency of passing impediment. Average speed still exceeds umes are lower than capacity, and speeds are below capacity speed.
52 mph on level terrain, even though unrestricted passing demand Level-of-service E is seldom attained over extended sections on
exceeds passing capacity. At higher volume levels, chaining of level terrain as more than a transient condition; most often, pertur-
platoons and significant reductions in passing capacity begin to bations in traffic flow as level E is approached cause a rapid
occur. While traffic flow is stable, it is becoming susceptible to transition to level-of-service F.
congestion due to turning traffic and slow-moving vehicles. Per-
cent time delays are up to 60 percent. A service flow rate of up
to 1,200 pcph, total in both directions, can be accommodated under
ideal conditions.
Unstable traffic flow is approached as traffic flows enter level- OPERATIONAL ANALYSIS
of-service D. The two opposing traffic streams essentially begin
to operate separately at higher volume levels, as passing becomes
extremely difficult. Passing demand is very high, while passing This section presents the methodology for operational analysis
capacity approaches zero. Mean platoon sizes of 5 to 10 vehicles of general terrain segments and specific grades on two-lane high-
are common, although speeds of 50 mph can still be maintained ways. Separate procedures for general highway segments and
under ideal conditions. The fraction of no passing zones along the grades are used, because the dynamics of traffic interaction on
roadway section usually has little influence on passing. Turning sustained two-lane grades differ from those on general terrain seg-
vehicles and/or roadside distractions cause major shock-waves in ments. Grades of less than 3 percent or shorter than 1/2 mile may
the traffic stream. The percentage of time motorists are delayed be included in general terrain analysis. Grades both longer and
approaches 75 percent. Maximum service flow rates of 1,800 pcph, steeper than these values should generally be treated as specific
total in both directions, can be maintained under ideal conditions. grades. Level, rolling, and mountainous terrain are as defined in
This is the highest flow rate that can be maintained for any length Chapters 1 and 3.
of time over an extended section of level terrain without a high The length of grade is taken to be the tangent length of grade
probability of breakdown. plus a portion of the vertical curves at the beginning and end
Level-of-service E is defined as traffic flow conditions on two- of the grade. About one-fourth of the length of vertical curves
lane highways having a percent time delay of greater than 75 at the beginning and end of a grade are included in the grade
percent. Under ideal conditions, speeds will drop below 50 mph. length. Where two grades (in the same direction) are joined by
Average travel speeds on highways with less than ideal condi- a vertical curve, one-half the length of the curve is included in
tions will be slower, as low as 25 mph on sustained upgrades. each grade segment.
The objective of operational analysis is generally the deter- The decision to use flow rates or full-hour volumes in an analysis
mination of level of service for an existing or projected facility is related to whether or not peaking characteristics will cause sub-
operating under existing or projected traffic demand. Operational stantial fluctuation in operating conditions within the peak hour,
analysis may also be used to determine the capacity of a two- and whether the impact of such fluctuations will impact design
lane highway segment, or the service flow rate which can and/or operational policy decisions. In general, where the peak
be accommodated at any given level of service. hour factor is less than 0.85, operating conditions will vary substan-
tially within the hour.
Where the peak hour factor can be determined from local field
Use of the Peak Hour Factor
data, this should be done. Where field data are not available, the
factors tabulated in Table 8-3 may be used. These are based solely
As for other facility types, two-lane highway analysis is based on the assumption of random flow and may be somewhat higher
on flow rates for a peak 15-min period within the hour of interest, than those obtained from field studies. When level of service is
which is usually the peak hour. The criteria of Table 8-1 refer to be determined for a given traffic volume, a value appropriate
to equivalent hourly flow rates based on the peak 15 min of to the volume level on the subject segment is selected from the
flow. upper portion of the table. When a service flow rate is to be
These criteria are used to compute service flow rates, SF, which computed, a value is selected from the lower portion of the table,
are compared to existing or projected flow rates to determine level because volume is unknown.
of service. Thus, full-hour demand volumes must be converted to
flow rates for the peak 15 min, as follows:
Analysis of General Terrain Segments
v = V/PHF
The general terrain methodology estimates average traffic opera-
where: tional measures along a section of highway based on average ter-
v = flow rate for the peak 15 min, total for both directions rain, geometric, and traffic conditions. Terrain is classified as level,
of flow, in vph; rolling, or mountainous, as described previously. The general ter-
V = full-hour volume total for both directions of flow, in rain procedure is usually applied to highway sections of at least 2
vph; and miles in length.
PHF = peak hour factor. Highway geometric features include a general description of
longitudinal section characteristics and specific roadway cross-
When criteria are compared to flow rates, the predicted op- section information. Longitudinal section characteristics are de-
erating characteristics are expected to prevail for the 15-min scribed by the average percent of the highway having no passing
period for which the flow rate applies. For many rural condi- zones. The average for both directions is used. The percentage
tions, the analyst may wish to examine average conditions over of roadway along which sight distance is less than 1,500 ft may
a peak hour. Full-hour volumes, unadjusted for the PHF, are be used as a surrogate for no passing zone data. Roadway cross-
compared to criteria directly for these cases. It should be noted, section data include lane width and usable shoulder width. Geo-
however, that prediction of an average level-of-service C during metric data on design speed and specific grades are not used
a full hour may include portions of the hour operating at level directly, but are reflected in the other geometric factors dis-
D or E, while other portions operate at A or B. cussed.
Table 8-3. Peak Hour Factors for Two-Lane Highways Based on Random Flow
A. Level-of-Service Determinations
total 2-way peak hour total 2-way peak hour
hourly volume factor hourly volume factor
(vph) (phf) (vph) (phf)
100 0.83 1,000 0.93
200 0.87 1,100 0.94
300 0.90 1,200 0.94
400 0.91 1,300 0.94
500 0.91 1,400 0.94
600 0.92 1,500 0.95
700 0.92 1,600 0.95
800 0.93 1,700 0.95
900 0.93 1,800 0.95
≥ 1,900 0.96
B. Service Flow-Rate Determinations
Level of Service A B C D E
Peak Hour Factor 0.91 0.92 0.94 0.95 1.00
8-8 rural highways
Traffic data needed to apply the general terrain methodology magnitude of passing restrictions. Note that v/c ratios at ca-
include the two-way hourly volume, a peak hour factor, and the pacity are not equal to 1.00 for rolling or mountainous terrain. This
directional distribution of traffic flow. Peak hour factors may be is because the ratios are based on an ideal capacity of 2,800 pcph,
computed from field data, or appropriate default values may be which cannot be achieved on severe terrains. Further, as the forma-
selected from Table 8-3. Traffic data also include the proportion tion of platoons is more frequent where terrain is rolling or mountain-
of trucks, recreational vehicles (RV’s), and buses in the traffic ous, passing restrictions have a greater effect on capacity and service
stream. When estimates of the traffic mix are not available, the flow rate than on level terrain.
following default values for these fractions may be used for pri- 3. Adjustment for directional distribution—All of the v/c values
mary routes: in Table 8-1 are for a 50/50 directional distribution of traffic on
a two-lane highway. For other directional distributions, the factors
shown in Table 8-4 must be applied to Table 8-1 values.
T PT = 0.14 (trucks)
4. Adjustment for narrow lanes and restricted shoulder width—
T PR = 0.04 (RV’s)
Narrow lanes force motorists to drive closer to vehicles in the
T PD = 0.00 (buses) opposing lane than they would normally desire. Restricted or nar-
row shoulders have much the same effect, as drivers shy away
Recreational routes would typically have a higher proportion of from roadside objects or point restrictions perceived to be close
recreational vehicles than shown for primary rural routes. enough to the roadway to pose a hazard. Motorists compensate
1. General relationship—The general relationship describing for driving closer to opposing vehicles by slowing down and/or
traffic operations on general terrain segments is as follows: by leaving larger headways between vehicles in the same lane.
Both reactions result in lower flow rates being sustained at any
given speed.
SFi = 2,800 × (v/c)i × fd × fw × fHV (8-1)
Factors reflecting this behavior are shown in Table 8-5, and are
applied to v/c values taken from Table 8-1. Factors at capacity are
where: higher than those for other levels of service, as the impact of
narrow lanes and restricted shoulder widths is less deleterious
SFi = total service flow rate in both directions for prevailing when vehicles are already traveling at reduced speeds which pre-
roadway and traffic conditions, for level of service i, vail under capacity operation.
in vph; 5. Adjustment for heavy vehicles in the traffic stream—The
(v/c)i = ratio of flow rate to ideal capacity for level of service i, v/c ratios of Table 8-1 are based on a traffic stream consisting of
obtained from Table 8-1; only passenger cars. All vehicles having only four wheels con-
fd = adjustment factor for directional distribution of traffic, tacting the pavement may be considered to be passenger cars. This
obtained from Table 8-4; includes light vans and pick-up trucks.
fw = adjustment factor for narrow lanes and restricted shoulder ‘‘Heavy vehicles’’ are categorized as trucks, recreational vehi-
width, obtained from Table 8-5; cles, or buses, and the traffic stream is characterized by the propor-
fHV = adjustment factor for the presence of heavy vehicles in tion of such vehicles in the traffic mix. The adjustment factor for
the traffic stream, computed as: heavy vehicles, fHV, is computed using Eq. 8-2 and the passenger-
fHV = 1/[1 + PT (ET − 1) + PR(ER − 1) + PB(EB − 1)] (8-2) car equivalents given in Table 8-6.
A wide range in the proportions of trucks and RV’s in the
traffic stream are found on rural highways. Equation 8-2 will
where: yield an adjustment factor for any given mix. In addition, there
is some variation in the weight distribution between heavy
PT = proportion of trucks in the traffic stream, expressed as a
(>35,000 lb) and medium-duty (≤35,000 lb) trucks. The equiva-
decimal;
lents of Table 8-6 assume a 50/50 distribution between heavy
PR = proportion of RV’s in the traffic stream, expressed as a
and medium-duty trucks. Two-lane highways serving unusually
decimal;
high proportions of heavy trucks, such as in coal, gravel, or
PB = proportion of buses in the traffic stream, expressed as a timber operations, particularly those in mountainous terrain,
decimal; would have higher values of ET than those shown in the table.
ET = passenger-car equivalent for trucks, obtained from Table The deleterious impact of heavy vehicles on two-lane highways
8-6; increases markedly as terrain becomes more severe. As heavy
ER = passenger-car equivalent for RV’s, obtained from Table vehicles slow on steeper grades, platoon formation becomes more
8-6; and frequent and severe. This effect is compounded by passing sight
EB = passenger-car equivalent for buses, obtained from Table 8-6. distance restrictions often accompanying severe terrain and leads
to serious deterioration of traffic flow.
Equation 8-1 takes an ideal capacity of 2,800 pcph, and adjusts
it to reflect a v/c ratio appropriate for the desired level of service,
directional distributions other than 50/50, lane width restrictions Analysis of Specific Grades
and narrow shoulders, and heavy vehicles in the traffic stream.
2. Adjustment for v/c ratio—The v/c ratios given in Table The analysis of extended specific grades on two-lane highways
8-1 reflect a complex relationship among speed, flow, delay, and is more complex than for general terrain segments. The analysis
geometric parameters for two-lane highways. Specifically, v/c procedures assume that the approach to the grade is level. On
values vary with level-of-service criteria, terrain type, and the such grades, the operation of upgrade vehicles is substantially
Table 8-4. Adjustment Factors for Directional Distribution on General Terrain Segments
Directional Distribution 100/0 90/10 80/20 70/30 60/40 50/50
Adjustment Factor, fd 0.71 0.75 0.83 0.89 0.94 1.00
Table 8-5. Adjustment Factors for the Combined Effect of Narrow Lanes and Restricted Shoulder Width, fw
usablea 12-ft 11-ft 10-ft 9-ft

shoulder lanesb lanesb lanesb lanesb
width los los los los los los los los
(ft) a–d e a–d e a–d e a–d e
≥6 1.00 1.00 0.93 0.94 0.84 0.87 0.70 0.76
4 0.92 0.97 0.85 0.92 0.77 0.85 0.65 0.74
2 0.81 0.93 0.75 0.88 0.68 0.81 0.57 0.70
0 0.70 0.88 0.65 0.82 0.58 0.75 0.49 0.66
a
Where shoulder width is different on each side of the roadway, use the average shoulder width.
b
For analysis of specific grades, use LOS E factors for all speeds less than 45 mph.
impacted, while downgrade vehicles experience far less impact. Table 8-6. Average Passenger-Car Equivalents for
As a result, level-of-service criteria presented in Table 8-2 are Trucks, RV’s, and Buses on Two-Lane Highways Over
based on the average upgrade travel speed. This speed is the General Terrain Segments
average speed of all vehicles traveling up the grade. type of terrain
vehicle level of
Where composite grades are present, the average grade is type service level rolling mountainous
used in analysis. The average grade is the total rise, in feet, of
Trucks, ET A 2.0 4.0 7.0
the composite grade divided by the horizontal length of the
B and C 2.2 5.0 10.0
grade, in feet, multiplied by 100 to adjust from a decimal to D and E 2.0 5.0 12.0
a percentage.
RV’s, ER A 2.2 3.2 5.0
The average upgrade speed at which capacity occurs varies be- B and C 2.5 3.9 5.2
tween 25 and 40 mph, depending upon the percent grade, the D and E 1.6 3.3 5.2
percentage of no passing zones, and other factors. Because op-
Buses, EB A 1.8 3.0 5.7
erating conditions at capacity vary for each grade, the finding of B and C 2.0 3.4 6.0
capacity is not as straightforward as service flow rate computations D and E 1.6 2.9 6.5
for levels-of-service A through D, where speed is established using source: Ref. 6
the criteria of Table 8-2.
Research has found that grades on two-lane highways have a
more significant impact on operations than similar grades on multi-
lane highways. Platoons forming behind slow-moving vehicles can
be broken up or dissipated only by passing maneuvers using the
opposing lane. On two-lane highways, the same geometric features
causing platoons to form also tend to restrict passing opportunities
as well. It has also been found that most passenger cars, even in
The service flow rate for any given average upgrade speed is
the absence of heavy vehicles, are affected by extended grades,
given by the following relationship:
and will operate less efficiently than on level terrain. Additional
operational problems due to vehicle stalls, accidents, or other inci-
SFi = 2,800 × (v/c)i × fd × fw × fg × fHV (8-3)
dents are not accounted for in the procedure. The effects of rain,
snow, ice, and other negative environmental factors are also not
considered. where:
1. Relationship between speed and service flow rate on specific SFi = service flow rate for level-of-service i, or speed i, total
grades—Average upgrade speeds on two-lane highways may be vph for both directions, for prevailing roadway and traffic
estimated for specific grades of a given percent and length of conditions.
grade, assuming a level approach to the grade. Two-way service (v/c)i = v/c ratio for level-of-service i or speed i, obtained from
flow rates, SF, may be calculated for a specific level of service, Table 8-7;
or correspondingly, for any designated average upgrade speed. fd = adjustment factor for directional distribution, obtained
The need to provide a climbing lane based on AASHTO’s safety from Table 8-8;
warrant is not part of the procedure, but sample calculation 5 fw = adjustment factor for narrow lanes and restricted
illustrates the evaluation of a potential climbing lane. shoulder width, obtained from Table 8-5;
8-10 rural highways
fg = adjustment factor for the operational effects of passenger-car operation. This is an important point, because a
grades on passenger cars, computed as described be- v/c ratio of 1.00 in Table 8-7 DOES NOT necessarily signify
low; and capacity. The solution for capacity of an extended grade is dis-
fHV = adjustment factor for the presence of heavy vehicles in cussed later. However, solutions for capacity or service flow rate
the upgrade traffic stream, computed as described exceeding 2,000 vph total indicate that the specific grade is not
subsequently. affecting operations and that the general terrain methodology
should be used.
Values of v/c approaching or equal to 0.00 mean that the
This relationship for specific grades is generally not applied to associated average upgrade speed is difficult or impossible to
grades of less than 3 percent or shorter than 1/2 mile. achieve for the percent grade and percent no passing zones
2. Adjustment for v/c ratio—Table 8-7 shows values of v/c indicated.
ratio related to percent grade, average upgrade speed, and per- 3. Adjustment for directional distribution—On extended
cent no passing zones. The values shown are the ratio of flow grades, the directional distribution can be a critical factor af-
rate to an ideal capacity of 2,800 pcph, and assume that pas- fecting operations. Table 8-8 contains adjustment factors for a
senger cars are unaffected by extended grades. Another adjust- range of directional distributions with a significant upgrade com-
ment is applied to account for the impacts of grades on ponent.
Table 8-7. Value of v/c Ratioa vs. Speed, Percent Grade, and Percent No Passing Zones for Specific Grades
average upgrade percent no passing zones

percent grade speed (mph) 0 20 40 60 80 100
3 55 0.27 0.23 0.19 0.17 0.14 0.12
52.5 0.42 0.38 0.33 0.31 0.29 0.27
50 0.64 0.59 0.55 0.52 0.49 0.47
45 1.00 0.95 0.91 0.88 0.86 0.84
42.5 1.00 0.98 0.97 0.96 0.95 0.94
40 1.00 1.00 1.00 1.00 1.00 1.00
4 55 0.25 0.21 0.18 0.16 0.13 0.11
52.5 0.40 0.36 0.31 0.29 0.27 0.25
50 0.61 0.56 0.52 0.49 0.47 0.45
45 0.97 0.92 0.88 0.85 0.83 0.81
42.5 0.99 0.96 0.95 0.94 0.93 0.92
40 1.00 1.00 1.00 1.00 1.00 1.00
5 55 0.21 0.17 0.14 0.12 0.10 0.08
52.5 0.36 0.31 0.27 0.24 0.22 0.20
50 0.57 0.49 0.45 0.41 0.39 0.37
45 0.93 0.84 0.79 0.75 0.72 0.70
42.5 0.97 0.90 0.87 0.85 0.83 0.82
40 0.98 0.96 0.95 0.94 0.93 0.92
35 1.00 1.00 1.00 1.00 1.00 1.00
6 55 0.12 0.10 0.08 0.06 0.05 0.04
52.5 0.27 0.22 0.18 0.16 0.14 0.13
50 0.48 0.40 0.35 0.31 0.28 0.26
45 0.85 0.76 0.68 0.63 0.59 0.55
42.5 0.93 0.84 0.78 0.74 0.70 0.67
40 0.97 0.91 0.87 0.83 0.81 0.78
35 1.00 0.96 0.95 0.93 0.91 0.90
30 1.00 0.99 0.99 0.98 0.98 0.98
7 55 0.00 0.00 0.00 0.00 0.00 0.00
52.5 0.13 0.10 0.08 0.07 0.05 0.04
50 0.34 0.27 0.22 0.18 0.15 0.12
45 0.77 0.65 0.55 0.46 0.40 0.35
42.5 0.86 0.75 0.67 0.60 0.54 0.48
40 0.93 0.82 0.75 0.69 0.64 0.59
35 1.00 0.91 0.87 0.82 0.79 0.76
30 1.00 0.95 0.92 0.90 0.88 0.86
a
Ratio of flow rate to ideal capacity of 2,800 pcph, assuming passenger-car operation is unaffected by grade.
NOTE: Interpolate for intermediate values of ‘‘Percent No Passing Zone’’; round ‘‘Percent Grade’’ to the next higher integer value.
Table 8-8. Adjustment Factor for Directional PHV = total proportion of heavy vehicles (trucks + RV’s
Distribution on Specific Grades, fd + buses) in the upgrade traffic stream;
percent of traffic EHV = passenger-car equivalent for specific mix of heavy
on upgrade adjustment factor vehicles present in the upgrade traffic stream, com-
100 0.58 puted as:
90 0.64
80 0.70 EHV = 1 + (0.25 + PT/HV) (E − 1) (8-7)
70 0.78
60 0.87
50 1.00
40 1.20 PT/HV = proportion of trucks among heavy vehicles, i.e., the
≤ 30 1.50 proportion of trucks in the traffic stream divided by
the total proportion of heavy vehicles in the traffic
stream; and
E = base passenger-car equivalent for a given percent
grade, length of grade, and speed, selected from
Table 8-9.
4. Adjustment for narrow lanes and/or restricted shoulder
The passenger-car equivalents presented in Table 8-9 represent
width—The impact of narrow lanes and/or restricted shoulder
an average mix of trucks, recreational vehicles, and buses in the
widths on grades is the same as for general terrain segments. The
traffic stream. This average mix is for 14 percent trucks, 4 percent
appropriate factor is selected from Table 8-5, presented previously.
RV’s, and no buses. The values of EHV computed by this procedure
5. Adjustment for passenger cars on grades—The v/c ratios of
yield equivalent volumes which travel at the same average overall
Table 8-7 assume that passenger cars will maintain their speed
speed as the actual mixed traffic stream under stable flow condi-
on grades if unimpeded. Recent studies (1,2) have indicated that
tions. Any tendency of vehicles to stall or perform sluggishly at
passenger-car operation is affected by grades, even where heavy
high volume levels and power requirements is not accounted for
vehicles are not present in the traffic stream. The factor fg adjusts
in these procedures.
the v/c ratios of Table 8-7 to account for this effect. The factor is
The existence of heavy vehicles on two-lane highway grades is
computed as:
a particularly difficult problem, because an increase in formation
of platoons is caused at the same time as passing restrictions usu-
fg = 1/[1 + (PPIP)] (8-4) ally also increase. Thus, the decision of whether to provide a
climbing lane for heavy vehicles is often a critical one for extended
where: grades on two-lane highways. A common criterion sometimes used
in the design of grades is to include a climbing lane where the
fg = adjustment factor for the operation of passenger cars on
operating speed of trucks falls 10 mph or more (11). Figures 8-2
grades;
and 8-3 show speed reduction curves for a 200-lb/hp truck and a
PP = proportion of passenger cars in the upgrade traffic stream,
300-lb/hp truck. The former is considered indicative of a represen-
expressed as a decimal;
tative truck for the average mix of trucks occurring on two-lane
IP = impedance factor for passenger cars, computed as:
highways. The latter is representative of a ‘‘heavy’’ truck, such
IP = 0.02 (E − Eo) (8-5) as heavily loaded farm vehicles, coal carriers, gravel carriers, or
log carriers. The choice of which type of truck should be used is
E = base passenger-car equivalent for a given percent grade, based on safety considerations. Speed reduction is related to the
length of grade, and speed, selected from Table 8-9; and steepness and length of the grade in Figures 8-2 and 8-3. For a
Eo = base passenger-car equivalent for 0 percent grade and a given more detailed depiction of the operating characteristics of trucks
speed, selected from Table 8-9. on extended upgrades, the truck performance curves included in
Appendix I of Chapter 3 may be consulted.
The passenger-car equivalents of Table 8-9 are used for both In addition to the 10-mph speed reduction criterion, a climbing
the passenger-car and heavy vehicle adjustment factors. The pas- lane might be considered wherever a level-of-service analysis indi-
senger-car factor adjusts from the base v/c ratios, which assume cates a serious deterioration in operating quality on an extended
no operational impact of grades on cars, to prevailing conditions grade when compared to the adjacent approach segment of the
of grade. The heavy vehicle adjustment factor is based on passen- same highway.
ger-car equivalents related to passenger cars operating on the grade Heavy vehicles in the traffic stream on extended grades also
specified. cause delay to other vehicles. Delay can be evaluated as the differ-
6. Adjustment for heavy vehicles in the traffic stream—The ence in travel time between what vehicles could achieve if unim-
adjustment factor for heavy vehicles is computed as follows: peded by heavy vehicles and the travel time actually experienced
in the mixed traffic stream. Sample calculations illustrate the com-
fHV = 1/[1 + PHV (EHV − 1)] (8-6)
putation of this delay.
7. Capacity of specific grade segments—Sections 1 through
where:
6 above describe the computation of service flow rates on specific
fHV = adjustment factor for the presence of heavy vehicles two-lane highway grades. For levels-of-service A through D,
in the upgrade traffic stream; this is a simple process. The speed relating to the desired LOS
8-12 rural highways
Table 8-9. Passenger-Car Equivalents for Specific Grades of Two-Lane Rural Highways, E and EO
length average upgrade speed (mph)
of
grade grade
(%) (mi) 55.0 52.5 50.0 45.0 40.0 30.0
0 All 2.1 1.8 1.6 1.4 1.3 1.3
1
3 ⁄4 2.9 2.3 2.0 1.7 1.6 1.5
1
⁄2 3.7 2.9 2.4 2.0 1.8 1.7
3
⁄4 4.8 3.6 2.9 2.3 2.0 1.9
1 6.5 4.6 3.5 2.6 2.3 2.1
1 1⁄ 2 11.2 6.6 5.1 3.4 2.9 2.5
2 19.8 9.3 6.7 4.6 3.7 2.9
3 71.0 21.0 10.8 7.3 5.6 3.8
a
4 48.0 20.5 11.3 7.7 4.9
1
4 ⁄4 3.2 2.5 2.2 1.8 1.7 1.6
1
⁄2 4.4 3.4 2.8 2.2 2.0 1.9
3
⁄4 6.3 4.4 3.5 2.7 2.3 2.1
1 9.6 6.3 4.5 3.2 2.7 2.4
1 1⁄ 2 19.5 10.3 7.4 4.7 3.8 3.1
2 43.0 16.1 10.8 6.9 5.3 3.8
a
3 48.0 20.0 12.5 9.0 5.5
a a
4 51.0 22.8 13.8 7.4
1
5 ⁄4 3.6 2.8 2.3 2.0 1.8 1.7
1
⁄2 5.4 3.9 3.2 2.5 2.2 2.0
3
⁄4 8.3 5.7 4.3 3.1 2.7 2.4
1 14.1 8.4 5.9 4.0 3.3 2.8
1 1⁄ 2 34.0 16.0 10.8 6.3 4.9 3.8
2 91.0 28.3 17.4 10.2 7.5 4.8
a a
3 37.0 22.0 14.6 7.8
a a a
4 55.0 25.0 11.5
1
6 ⁄4 4.0 3.1 2.5 2.1 1.9 1.8
1
⁄2 6.5 4.8 3.7 2.8 2.4 2.2
3
⁄4 11.0 7.2 5.2 3.7 3.1 2.7
1 20.4 11.7 7.8 4.9 4.0 3.3
1 1⁄ 2 60.0 25.2 16.0 8.5 6.4 4.7
a
2 50.0 28.2 15.3 10.7 6.3
a a
3 70.0 38.0 23.9 11.3
a a a
4 90.0 45.0 18.1
1
7 ⁄4 4.5 3.4 2.7 2.2 2.0 1.9
1
⁄2 7.9 5.7 4.2 3.2 2.7 2.4
3
⁄4 14.5 9.1 6.3 4.3 3.6 3.0
1 31.4 16.0 10.0 6.1 4.8 3.8
1 1⁄ 2 a
39.5 23.5 11.5 8.4 5.8
a
2 88.0 46.0 22.8 15.4 8.2
a a a
3 66.0 38.5 16.1
a a a a a
4 28.0
a
Speed not attainable on grade specified.
NOTE: Round ‘‘Percent Grade’’ to next higher integer value.
is selected from Table 8-2, and appropriate adjustment factors where:

are selected for use in Eq. 8-3.
Sc = speed at which capacity occurs, in mph; and
The service flow rate at capacity, i.e., SFE, is not as easily
determined, because the speed at which it occurs varies depending vc = flow rate at capacity, in mixed vph.
on the percent and length of the grade in question. For the normal
For convenience, the equation predicts upgrade speeds based
range of grades, i.e., 3 to 7 percent up to 4 miles long, capacity
on total two-way flow rates. The equation is valid for speed up to
may occur at speeds ranging from 25 to 40 mph. The speed at
40 mph.
which capacity occurs is related to the flow rate at capacity by the
If the service flow rates computed for various speeds using
following equation:
Eq. 8-3 and the capacity speed vs. capacity flow rate relationship
Sc = 25 + 3.75(vc /1000)2 (8-8) of Eq. 8-8 are plotted, the two curves will intersect. The inter-
Figure 8-2. Speed reduction curve for a 200-lb/hp truck. Figure 8-3. Speed reduction curve for a 300-lb/hp truck.
section defines both the speed at capacity and the flow rate at The AADT’s presented in Table 8-10 illustrate a wide range
capacity for the grade in question. This procedure for deter- of conditions, and were computed from service flow rates as
mining capacity is illustrated in the sample calculations. follows:
AADTi = SFi × PHF/K (8-9)

HIGHWAY SYSTEM PLANNING
The planning procedure enables highway operating agencies to where:

perform very general planning and policy studies of a rural two-
lane highway system. Traffic, geometric, and terrain data would AADTi = the maximum AADT for level-of-service i, based on
be only generally classified, with traffic demand expressed in terms the assumed conditions described above, in vpd;
of an average annual daily traffic (AADT), perhaps of some future SFi = maximum service flow rate for level-of-service i, com-
forecast year. puted from Eq. 8-3, based on the assumed conditions
Table 8-10 presents estimated maximum AADT’s for two-lane described above, in vph;
highways as related to: PHF = peak hour factor, selected from Table 8-3 for the indi-
cated level of service; and
1. Level of service. K = design hour factor, i.e., the proportion of AADT ex-
2. Type of terrain. pected to occur in the design hour.
3. Design hour factor, K.
The K-factor is normally expressed in design problems as
The levels of service refer to operating conditions within the DHV = AADT × K, where the DHV is the total two-way
peak 15-min period of the day. In constructing Table 8-10, the design hour volume, and K is estimated from the ratio of the
default values of the peak hour factor (PHF) shown in Table 30th HV to the AADT from a similar site. The 30th HV is the
8-3 were assumed. For each level of service, the related percent 30th highest hourly volume during the year and is often used
time delay criteria were applied across all three types of terrain. as a design volume for rural highways. Since the DHV should
The planning criteria also assume a typical traffic mix of 14 be less than SFi for the selected level of service, the actual AADT
percent trucks, 4 percent RV’s, and no buses. A 60/40 directional for a road should be less than the maximum value shown in
split is used, along with percent no passing zone values of 20 Table 8-10. Traffic conditions occurring during the highest
percent, 40 percent, and 60 percent for level, rolling, and mountain- hourly volume of the year (1st HV) would usually be no worse
ous terrain, respectively. Ideal geometrics of 12-ft lanes, 6-ft shoul- than one level of service less than that existing for the 30th HV
ders, and 60-mph design speed were used. for most rural highways.
8-14 rural highways
Table 8-10. Maximum AADT’s vs. Level of Service and Type of Terrain for Two-Lane Rural Highways
level of service
k-factor a b c d e
Level Terrain
0.10 2,400 4,800 7,900 13,500 22,900
0.11 2,200 4,400 7,200 12,200 20,800
0.12 2,000 4,000 6,600 11,200 19,000
0.13 1,900 3,700 6,100 10,400 17,600
0.14 1,700 3,400 5,700 9,600 16,300
0.15 1,600 3,200 5,300 9,000 15,200
Rolling Terrain
0.10 1,100 2,800 5,200 8,000 14,800
0.11 1,000 2,500 4,700 7,200 13,500
0.12 900 2,300 4,400 6,600 12,300
0.13 900 2,100 4,000 6,100 11,400
0.14 800 2,000 3,700 5,700 10,600
0.15 700 1,800 3,500 5,300 9,900
Mountainous Terrain
0.10 500 1,300 2,400 3,700 8,100
0.11 400 1,200 2,200 3,400 7,300
0.12 400 1,100 2,000 3,100 6,700
0.13 400 1,000 1,800 2,900 6,200
0.14 300 900 1,700 2,700 5,800
0.15 300 900 1,600 2,500 5,400
NOTE: All values rounded to the nearest 100 vpd. Assumed conditions include 60/40 directional split, 14 percent trucks, 4 percent RV’s, no buses, and PHF values
from Table 8-3. For level terrain, 20 percent no passing zones were assumed; for rolling terrain, 40 percent no passing zones; for mountainous terrain, 60 percent
no passing zones.
The methodology described in the previous section is generally which are hypothesized and/or forecast. The general approach
applied in either the operational analysis or planning mode. will be to compute service flow rates for each level of service and
Design computations, as used in this manual, focus on the deter- compare these values with the existing flow rate on the facility.
mination of the number of lanes required for a given facility. Such This is done using Eq. 8-1:
computations have little significance for two-lane highways, where
the number of lanes is fixed. Such design features as horizontal
SFi = 2,800 × (v/c) i × fd × fw × fHV
and vertical alignment, however, have a significant impact on oper-
ations. Operational analyses can be performed for alternative de-
signs to document this impact. Where computations indicate that where all terms are as previously defined. A service flow rate
a two-lane highway is not adequate for existing or projected de- for each LOS is computed because the heavy vehicle factor varies
mands, various multilane options may be considered and analyzed with LOS, and a direct solution of the equation for v/c ratio
using other chapters of this manual. would be iterative. Users preferring to solve for v/c may do so,
A separate section of this chapter deals with operational and but must iterate until the assumed LOS used in computing the
design measures for two-lane highways, short of reconstructing heavy vehicle factor is the same as that indicated by the v/c
the entire highway as a multilane facility. This material should be ratio found.
consulted where a two-lane facility presently has or is expected In general, the following computational steps are used. Compu-
to experience operational difficulties. tations may be conveniently performed on the worksheet illustrated
in Figure 8-4.
OPERATIONAL ANALYSIS OF GENERAL TERRAIN

1. Summarize all input data on traffic and roadway conditions,
SEGMENTS including:
T Existing or forecast peak hour volume, in vph.

The objective in operational analysis is to determine the level T Peak hour factor, PHF, from local data or default value
of service for a given segment or segments of roadway for a selected from Table 8-3.
known existing set of conditions, or for a future set of conditions T Traffic composition (% trucks, % RV’s, % buses).
Figure 8-4. Worksheet for operational analysis of general

terrain segments.
T Directional distribution of traffic. alleviation measures presented in the next section should be
T Terrain type. considered, as well as the expansion of the facility to four or more
T Lane and usable shoulder widths, in ft. lanes. Expansion to a multilane facility should be examined using
T Design speed, in mph. the methodology presented in Chapter 7.
2. Select appropriate values of the following factors for each
LOS from the tables indicated: OPERATIONAL ANALYSIS OF SPECIFIC GRADES
T The v/c ratio from Table 8-1.
T The directional distribution factor, fd, from Table 8-4. The operational analysis of specific grades is similar to the
procedure for general terrain segments. The level of service for
T The lane width and shoulder width factor, fw, from Table
the upgrade direction is sought, and is found by comparing an
8-5.
actual two-way flow rate to the service flow rates for the various
T Passenger-car equivalents, ET, ER, and EB, for trucks, RV’s,
levels of service. As noted in the ‘‘Methodology’’ section, how-
and buses, from Table 8-6.
ever, the determination of capacity for specific grades requires the
3. Compute the heavy vehicle factor, fHV, for each LOS from: plotting of a service flow rate-speed curve, and a curve represent-
ing the relationship of speed at capacity to flow rate at capacity.
fHV = 1/[1 + PT (ET − 1) + PR(ER − 1) + PB(EB − 1)] The worksheet shown in Figure 8-5 is used to simplify the follow-
ing computational steps.
4. Compute the service flow rate, SF, for each LOS from:
1. Summarize all required input data on traffic and roadway
SFi = 2,800 × (v/c)i × fd × fw × fHV conditions, including:
5. Convert the existing or forecast volume to an equivalent flow T Existing or forecast peak hour volume, in vph.
rate, as follows: v = V/PHF. T Peak hour factor, PHF, from local data or default value
6. Compare the actual flow rate of step 5 with the service flow from Table 8-3.
rate of step 4 to determine the level of service. T Traffic composition (% trucks, % RV’s, % buses, %
passenger cars).
Where the level of service is found to be inadequate, the T Directional distribution of traffic.
8-16
rural highways
Figure 8-5(a). Worksheet for operational analysis of specific grades on two-lane Figure 8-5(b). Worksheet for operational analysis of specific grades on two-lane
highways (page 1). highways (page 2).
T Percent grade. 6. Plot the service flow rates vs. speeds resulting from the
T Percent no passing zones. computations of steps 2–5 on the grid included in the worksheet
T Length of grade, in miles. of Figure 8-5. Note that the curve for speed at capacity vs. flow
T Lane and usable shoulder width, in ft. rate at capacity is already drawn on this grid.
T Design speed, in mph. 7. Find the speed at capacity and the service flow rate at
capacity from the intersection of the two curves on the plot of
2. Select values of the following factors from the indicated step 6.
tables for the following average speeds: 55 mph (LOS A), 52.5 8. Summarize the service flow rates for each level of service
mph, 50 mph (LOS B), 45 mph (LOS C), 40 mph (LOS D), on the worksheet as indicated.
and 30 mph. This range of speeds will allow the plotting of a 9. Convert the actual or forecast volume to a flow rate, as
service flow rate vs. speed curve to find capacity and the speed follows: v = V/PHF.
at capacity. 10. Compare the actual flow rate of step 9 with the service flow
T The v/c ratio from Table 8-7. rates of step 8 to determine the level of service.
T The directional distribution factor, fd, from Table 8-8. As with general terrain segments, a two-lane highway grade
T The lane and shoulder width factor, fw, from Table 8-5. displaying unacceptable operating conditions would be considered
T The passenger-car equivalent, E, for the percent and length for improvement. If heavy vehicles on the upgrade are the principal
of grade, from Table 8-9. difficulty, the addition of a truck climbing lane should be consid-
T The passenger-car equivalent, Eo, for a 0 percent grade, ered. If operational problems are more broad-based, any of the
from Table 8-9. alleviation techniques discussed in the next section could be con-
3. Compute the grade factor, fg, as follows: sidered, as well as expansion of the facility to four or more lanes.
Again, the multilane option would be examined using procedures
fg = 1/[1 + PpIp] in Chapter 7.
Ip = 0.02(E − Eo)
PLANNING
where all values are as previously defined.
4. Compute the heavy vehicle factor, fHV, for each of the speeds The highway system planning technique described in the
noted in step 2 as follows: ‘‘Methodology’’ section is easily applied. Table 8-10 may be en-
tered with a known or forecast AADT to determine expected level
fHV = 1/[1 + PHV (EHV − 1)]
of service during the peak 15 min of flow, or with a known LOS
EHV = 1 + (0.25 + PT/HV)(E − 1)
to find the maximum allowable AADT. No computations are
PT/HV = PT /[PT + PR + PB]
needed to use this table, although users are cautioned that any
where all values are as previously defined. conditions varying widely from those noted in the footnotes to
5. Compute the service flow rate, SF, for each of the speeds Table 8-10 will indicate the need to conduct an operational analysis
noted in step 2 as follows: for the facility in question.
Users may also find Table 8-10 useful in making preliminary
SFi = 2,800 × (v/c)i × fd × fw × fg × fHV estimates of LOS in general terrain segment analysis.
IV. DESIGN AND OPERATIONAL TREATMENTS
Addressing those operational problems that may exist on rural severe operational and safety problems due to a variety of traffic,
two-lane highways requires an understanding of the nature of two- geometric, and environmental causes. Special treatments for
lane highway systems. Only about 30 percent of all travel in the such highways may be needed before capacity levels are ap-
United States occurs on rural two-lane roads, even though this proached. In some areas, the two-lane rural arterial system
network comprises 80 percent of all paved rural highways. For carries a disproportionately large share of rural traffic, including
the most part, two-lane highways carry light traffic and experience significant components involved in interstate commerce. Many
few operational problems. Highway agencies are typically more of these highways are located near major urban areas and are experi-
concerned with pavement maintenance and roadside safety issues encing rapid growth in traffic. Heavy turning movements
on such highways. to roadside developments can block through traffic and increase
Some two-lane highways, however, periodically experience delay.
8-18 rural highways
As much as 60 percent of all two-lane highway mileage is highway design and, as illustrated by Tables 8-1 and 8-7, has
located in terrain classified as rolling or mountainous. This, cou- a critical impact on capacity and service flow rate. Where long
pled with occasionally high opposing volumes, is not favorable to queues are likely to form because of severe passing restrictions,
either passing or turning maneuvers. When these and other rural every effort should be made to continuously and completely dis-
highways experience increased recreational travel, major opera- perse the platoon once significant passing sight distance is re-
tional problems may arise. Large numbers of recreational and other gained. In these passing sections, short segments with passing sight
heavy vehicles in the traffic stream increase the demand for pass- distance restrictions should be avoided where possible. Inclusion
ing, while at the same time making such maneuvers more difficult. of periodic passing lanes for each direction should be considered
Two-lane highways serving as major routes to recreational areas where the distance between segments with passing sight distance
may operate at or near capacity on weekends in peak seasons. available is long and queuing extensive.
When any of the foregoing situations exist, the frequent result
is a reduced level of service, increased platooning, increased delay,
PAVED SHOULDERS
an increase in questionable passing maneuvers, and generally frus-
trated drivers. Nevertheless, many such situations do not justify
the reconstruction of the two-lane highway to a full multilane A roadway that is constructed with structurally adequate paved
facility. In these cases, one or more of the special design and/or shoulders can be used to assist in dispersal and breakup of platoons.
operational treatments discussed in this section may be useful. Slower moving vehicles may temporarily use the shoulder to per-
A wide range of design and operational solutions are needed to mit faster vehicles to pass, returning to the travel lane when passing
address the variety of problems encountered on two-lane highways. maneuvers have been completed. In Texas and Canada, where
The operational and/or safety problems on a particular section may some agencies construct wide shoulders for a total roadway width
be so severe as to call for an expansion of the facility to four or of 40 to 44 ft, a high percentage of the driving population uses
more lanes. However, limited reconstruction funds, difficult ter- the shoulder in this manner—particularly in western Canada where
rain, and other problems may not always permit full reconstruction long distance recreational travel is heavy during the summer. Illus-
of a two-lane facility as a multilane highway. Less costly and less tration 8-2 presents a typical use of paved shoulders as described
previously.
environmentally disruptive solutions may be required. Highways
Five states allow the use of shoulders for slow-moving vehicles
experiencing less severe operational and/or safety problems, to-
at all times. An additional ten states permit such use under specified
gether with those experiencing site-specific reductions in level of
conditions.
service, may be candidates for treatment with one or more of the
following alleviation techniques:
THREE-LANE HIGHWAYS
1. Realignment to improve passing sight distance.
2. Use of paved shoulders.
3. Three-lane roadways with two lanes designated for travel in Three-lane roadways are a rational intermediate solution to four-
one direction (passing prohibited or permitted in opposing lane expansions for two-lane highways experiencing operational
direction). problems. Because of funding and terrain constraints, three-lane
4. Three-lane road sections with continuous two-way median roadways may be considered for spot and segment improvements.
left-turn lanes. There are numerous methods for using the third travel lane on
such segments.
5. Three-lane roadway with reversible center lane.
In the 1940’s and 1950’s, the third (center) lane was used for
6. Special intersection treatments.
passing by vehicles in either direction—the first vehicle to occupy
7. Truck or heavy vehicle climbing lanes.
the center lane had the right-of-way. This condition was found to
8. Turnouts.
be hazardous, particularly in hilly terrain. This use of three-lane
9. Short four-lane segments.
highways in the United States has been generally discontinued.
Selection of the appropriate treatment requires identification of Other three-lane highway treatments are being safely and effi-
the probable causes of the operational and safety problems existing, ciently applied, including the use of passing lanes, turning lanes,
and the determination of cost-effectiveness of the design alterna- and climbing lanes.
tives for a given set of highway geometric, traffic, and system
constraints. The following discussions address the use of allevia-
Passing Lanes
tion measures on two-lane highways. They are intended to provide
the user with general information, and should not be construed as
This three-lane roadway design assigns the third (center) lane
firm guidelines or criteria.
to one direction of travel for a short distance (approximately 1
mile), then alternates the assignment of the passing lane to the
other direction. This cyclic process may be continued along an
entire highway section, or may be combined in an urban fringe area
with two-way left-turn lanes and/or specific intersection turning
PASSING SIGHT DISTANCE
treatments.
In a rural setting, intermittently spaced passing lane sections
The opportunity to pass, given a constant volume, is a function have been successfully used to break up platoons and reduce
of the availability of passing sight distance. Provision of passing delay. Two lanes are provided for unimpeded passing in one
sight distance is an important component in basic two-lane direction for 1 to 2 miles followed by a transition to two lanes
Illustration 8-2. Slow-moving vehicle

uses the shoulder of a two-lane rural
highway, permitting faster vehicles to
pass.
of similar design for the opposing flow. Advance signing advises may be prohibited. Figure 8-6 depicts these markings, and var-
motorists of the next upcoming passing lane to reduce driver ious methods of providing for the transition when the direction
anxiety and frustration. Two operational markings are practiced: of the passing lane is changed. Permissive passing for the one-
passing in the single-lane direction may be permitted if passing lane direction is not used by some agencies when the AADT
sight distance is available, or passing in the single-lane direction exceeds about 3,000 vpd.
Figure 8-6. Use of third lane for passing lanes.

8-20 rural highways
Table 8-11. Spacing of Passing Lanes on Two-Lane Detailed analysis of intersections may be performed using the
Highways procedures of Chapter 9, ‘‘Signalized Intersections,’’ and Chap-
Two-Way Peak Hourly Volume (vph) 400 300 200 ter 10, ‘‘Unsignalized Intersections.’’
Distance to Next Passing Lane (miles) 5 6.5 9
Climbing Lanes
Traditional climbing lanes also form three-lane cross sections

An analytic study of passing lane requirements was conducted
when used in conjunction with two-lane highways. They are gener-
in Ontario, Canada (7). This study recommended that passing lanes
ally applied as a spot improvement, most often on steep, sustained
should consistently be from 1.0 to 1.25 miles long. This length was
grades which cause heavy vehicles, particularly heavy trucks, to
found to be adequate to disperse most platoons, to provide for addi-
travel at slow speeds. This reduces capacity, creates platoons, and
tional transition zones, and yet not be too long to change drivers’
increases delay. Additionally, safety problems may arise when the
expectations about the true nature of the highway. Table 8-11 gives
reduction in speed of heavy trucks exceeds 10 mph along the grade.
the recommended spacing between passing lanes in a given direction
which resulted from the study. Estimated operating speed characteristics of trucks are illustrated
in Figures I.3-1, I.3-2, and I.3-3 in Appendix I of Chapter 3.
Resulting lengths of grade producing 10-mph speed reductions are
Continuous Two-Way Median Left-Turn Lanes plotted in Figures 8-2 and 8-3, presented earlier in this chapter.
AASHTO presently warrants a climbing lane wherever the speed
On two-lane highways having sizable left-turn traffic, a single of a 300-lb/hp truck is reduced by 10 mph or more and the volume
travel lane in each direction often becomes subject to long delays and percentage of heavy trucks justify the added cost. One set of
as vehicles await opportunities to complete left turns. By providing criteria that might be applied to reflect the economic considera-
a continuous refuge area for left-turning traffic, the two-way left- tions is:
turn lane can help to maintain through traffic capacity, with the
added benefit of separating opposing flows. The ability to pass, 1. Upgrade traffic flow rate exceeds 200 vph.
however, is eliminated. 2. Upgrade truck flow rate exceeds 20 vph.
Two-way left-turn lanes are not usually used where speeds are 3. One of the following conditions exists:
less than 25 mph or more than 50 mph, and are most often used
in urban fringe areas or on a major route passing through a small T Level-of-service E or F exists on the grade.
town or village. T A reduction of two or more levels of service is experienced
when moving from the approach segment to the grade.
T A 10-mph or greater speed reduction is expected for a
Reversible Lane
typical heavy truck.
This is another use of the third (center) lane of a three-lane These general guides for the consideration of climbing lanes on
highway which is most applicable where travel demands are of grades would apply only to climbing lanes on two-lane highways
a tidal nature—that is, extreme directional splits occur. The and should not be used in conjunction with consideration of climb-
center lane is reversed by time of day to match the peak flow. ing lanes on multilane highways.
The center lane is controlled by overhead signs or traffic signals
indicating the direction of travel assigned at the time. Passing
is not permitted in this application in the direction of the
single lane. Turnouts
The reversible lane technique is most applicable to routes joining
residential areas and high-employment centers, and for many recre- The use of turnouts for improving the level of service on two-
ational routes. lane, two-way highways is more prevalent in the rolling and moun-
tainous terrain of the western United States. Turnouts are short
Intersection Treatments segments of a third lane added to one side of the highway or the
other which permit slow vehicles at the head of platoons to pull
Conventional analysis of two-lane highways assumes unin- off the main roadway, allowing faster vehicles to pass. Turnouts
terrupted flow, which is normally representative of rural con- are used satisfactorily on both upgrades and downgrades, as well
ditions. With increasing development occurring in some rural as on level terrain, to improve traffic flow. Impeding motorists
areas, and in suburban fringe areas, the demand for high-volume are legally required to use turnouts where provided under certain
access and egress can grow. Major intersections along two-lane prescribed conditions, which vary by state.
highways become more common and important to the overall A recent study of operational characteristics revealed that few
quality of flow on main routes. Adequate protected turning lanes drivers actually stop at turnouts (8). Several additional conclusions
for both left and right turns are useful in minimizing disruption drawn from this study included:
to through traffic. Bypass lanes for through traffic may be con-
sidered where a protected left-turn lane is not feasible, partic- 1. Turnouts are safe when properly used.
ularly where paved shoulders are provided and/or where T- 2. A series of turnouts at regular intervals can provide con-
intersections are involved. siderable delay reduction.
3. Turnouts are not a substitute for a passing or climbing Table 8-12. Length of Turnouts on Two-Lane
lane of adequate length. Highways
4. About 10 percent of all platoon leaders use properly desig- Approach Speed 25 30 40 50 55 60
nated turnouts. (mph)
5. Large trucks tend to avoid turnouts.
Minimum 200 200 250 375 450 535
Turnouts are a short but functional treatment of irritating causes Length of
of operational delay. A western state recommends that the length Turnout (ft)
of turnouts vary with approach speed according to the criteria of
Table 8-12 (9).
Approach speeds of potential turnout-users vary with prevailing
traffic and roadway conditions, and differ between upgrades and both directions. The decision to use a short four-lane segment,
downgrades. Turnout lengths of more than 500 ft are only used as compared to using a three-lane option, may be based on long-
on downgrades exceeding 3 percent where high approach speeds range planning objectives for the facility, availability of rights-of-
are expected to exist. Lengths greater than 600 ft are never de- way, existing cross section, topography, and on the desire to reduce
signed, as drivers may mistakenly attempt to use them as pass- platooning and passing problems.
ing lanes. The transition from a two-lane to a four-lane roadway should
be designed to provide sufficient sight distance for passing. For
the length of four-lane segments, AASHTO suggests that they
SHORT FOUR-LANE SECTIONS be sufficiently long to permit several vehicles in line behind a
slow-moving vehicle to pass before reaching the normal section
Short sections of four-lane cross section may be constructed of two-lane highway. Four-lane sections of 1.0 to 1.5 miles
along a primarily two-lane highway to break up platoons, to should be sufficiently long to dissipate most queues formed,
provide the desired frequency of safe passing zones, and to depending on volume and terrain conditions. Further, it is noted
eliminate interference from low-speed vehicles. Such sections that sections of four-lane highway, particularly divided sections,
are particularly advantageous in rolling terrain, or where the longer than 2 miles may cause drivers to lose their sense of
alignment is winding or the profile includes critical grades from awareness that the road is basically a two-lane facility.
V. SAMPLE CALCULATIONS
CALCULATION 1—FINDING SERVICE FLOW

RATES FOR A GENERAL TERRAIN SEGMENT
1. Description—A segment of rural two-lane highway is ex- (v/c)C = 0.43 (Table 8-1, level terrain 0 percent no passing zones,
pected to have the following characteristics: LOS C);
(v/c)E = 1.00 (Table 8-1, level terrain, 0 percent no passing zones,
a. Roadway characteristics—70-mph design speed; 12-ft LOS E);
lanes; 10-ft paved shoulders; level terrain; 0 percent no fd = 0.89 (Table 8-4, 70/30 split);
passing zones; length = 5 miles. fw = 1.00 (Table 8-5, 12-ft lanes, >6-ft shoulders);
b. Traffic characteristics—70/30 directional split; 10 percent ET = 2.2 for LOS C, 2.0 for LOS E (Table 8-6, level terrain);
trucks; 5 percent recreational vehicles; 1 percent buses; 84 ER = 2.5 for LOS C, 1.6 for LOS E (Table 8-6, level terrain);
percent passenger cars. EB = 2.0 for LOS C, 1.6 for LOS E (Table 8-6, level terrain);
PT = 0.10 (Given);
What is the capacity of the section? What is the maximum flow
PR = 0.05 (Given); and
rate which can be accommodated at level-of-service C?
PB = 0.01 (Given).
2. Solution—The solution to this problem is found by comput-
ing the service flow rates for levels-of-service C and E (capacity),
using Eq. 8.1:
Then:
SFi = 2,800 × (v/c)i × fd × fw × fHV
where fHV(LOS C) = 1/[1 + 0.10(2.2 − 1) + 0.05(2.5 − 1) +

0.01(2.0 − 1)]
fHV = 1/[1 + PT (ET − 1) + PR(ER − 1) + PB(EB − 1)] = 0.83
fHV(LOS E) = 1/[1 + 0.10(2.0 − 1) + 0.05(1.6 − 1) +
The following values are selected for use in these computa- 0.01(1.6 − 1)]
tions: = 0.88
8-22
rural highways
Figure 8-7. Worksheet summarizing solution to Calculation 1. Figure 8-8. Worksheet summarizing solution to Calculation 2.
and: PT = 0.05 (Given); and

PR = 0.10 (Given).
SFC = 2,800 × 0.43 × 0.89 × 1.00 × 0.83 = 889 vph
Then:
SFE = 2,800 × 1.00 × 0.89 × 1.00 × 0.88 = 2,193 vph
fHV (LOS A) = 1/[1 + 0.05(7 − 1) + 0.10(5.0 − 1)] = 0.588
Thus, the highway will have an expected capacity of 2,193 vph, (LOS B, C) = 1/[1 + 0.05(10 − 1) + 0.10 (5.2 − 1)] = 0.535
total in both directions, and can accommodate a flow rate of up (LOS D, E) = 1/[1 + 0.05(12 − 1) + 0.10(5.2 − 1)] = 0.508
to 889 vph at level-of-service C. The worksheet for general terrain
sections may be used to perform these computations, as shown in and:
Figure 8-7.
SFA = 2,800 × 0.02 × 0.94 × 0.75 × 0.588 = 23 vph
SFB = 2,800 × 0.12 × 0.94 × 0.75 × 0.535 = 127 vph

CALCULATION 2—FINDING LEVEL OF SERVICE FOR A
GENERAL TERRAIN SEGMENT
SFC = 2,800 × 0.20 × 0.94 × 0.75 × 0.535 = 211 vph
1. Description—A two-lane rural highway carries a peak hour SFD = 2,800 × 0.37 × 0.94 × 0.75 × 0.508 = 371 vph
volume of 180 vph and has the following characteristics:
SFE = 2,800 × 0.80 × 0.94 × 0.88 × 0.508 = 941 vph
a. Roadway characteristics—60-mph design speed; 11-ft
lanes; 2-ft shoulders; mountainous terrain; 80 percent no If the actual flow rate of 207 vph (which represents the flow
passing zones; length = 10 miles. rate during the peak 15 min of flow) is compared to these values,
b. Traffic characteristics—60/40 directional split; 5 percent it is seen that it is higher than the service flow rate for LOS B
trucks; 10 percent recreational vehicles; no buses; 85 per- (127 vph), but is less than the service flow rate for LOS C (211
cent passenger cars. vph). Therefore, the level of service for the highway is C for the
conditions described.
This problem illustrates several points. On severe terrain, such
At what level of service will the highway operate during peak as the situation for this problem, ‘‘good’’ operating conditions can
periods? be sustained only at low flow rates. The capacity of the roadway
2. Solution—The solution is found by comparing the actual is also severely limited, reaching only 941 vph, which is approxi-
flow rate to service flow rates computed for each LOS. The actual mately one-third of the ideal capacity of 2,800 vph. Note that the
flow rate is found as: v/c ratio used in the computation of capacity is only 0.80. This is
because all v/c ratios in the two-lane methodology are referenced
v = V/PHF to the ideal capacity of 2,800 vph, which cannot be achieved in
severe terrain with passing sight distance restrictions.
where: This solution may be summarized or done on the general terrain
section worksheet, as shown in Figure 8-8.
V = 180 vph (Given)
PHF = 0.87 (Default value, Table 8-3, 200 vph)
and: CALCULATION 3—FINDING SERVICE FLOW RATES FOR A

SPECIFIC GRADE
v = 180/0.87 = 207 vph
Service flow rates are computed from Eq. 8-1: 1. Description—A rural two-lane highway in mountainous ter-
rain has a 6 percent grade of 2 miles. Other relevant characteristics
SFi = 2,800 × (v/c)i × fd × fw × fHV include:
fHV = 1/[1 + PT (ET − 1) + PR(ER − 1) + PB(EB − 1)] a. Roadway characteristics—12-ft lanes; 8-ft shoulders; 60
percent no passing zones.
where: b. Traffic characteristics—70/30 directional split; 12 percent
trucks; 7 percent recreational vehicles; 1 percent buses, 80
v/c = 0.02 for LOS A, 0.12 for LOS B, 0.20 for LOS C, 0.37 percent passenger cars; PHF = 0.85.
for LOS D, 0.80 for LOS E (Table 8-1, mountainous terrain,
80 percent no passing zones);
fd = 0.94 (Table 8-4, 60/40 split); What is the maximum volume which can be accommodated on
fw = 0.75 for LOS A through D, 0.88 for LOS E (Table 8-5, the grade at a speed of 40 mph (LOS D, Table 8-2)?
11-ft lanes, 2-ft shoulders); 2. Solution—Service flow rate on specific grades is computed
ET = 7 for LOS A, 10 for LOS B, C, 12 for LOS D, E, (Table using Eq. 8-3, as follows:
8-6, mountainous terrain);
ER = 5.0 for LOS A, 5.2 for LOS B-E (Table 8-6, moun-
tainous terrain); SFi = 2,800 × (v/c)i × fd × fw × fg × fHV
8-24 rural highways
where: At what level of service does the grade operate? What upgrade
speed can be expected during the peak 15 min of flow? What is
fg = 1/[1 + Pp Ip] from Eq. 8-4 the capacity of the grade? If the approach speed to the grade is
55 mph, what delay is incurred by vehicles climbing the grade?
Ip = 0.02 (E − Eo) from Eq. 8-5 2. Solution—The finding of capacity for a specific grade re-
quires plotting of the service flow rate vs. speed curve which
and: results from Eq. 8-3:
fHV = 1/[1 + PHV(EHV − 1)] from Eq. 8-6 SFi = 2,800 × (v/c)i × fd × fw × fg × fHV
EHV = 1 + (0.25 + PT/HV)(E − 1) from Eq. 8-7 where:
The following values are used in these computations: fg = 1/[1 + Pp Ip]

(v/c)D = 0.83 (Table 8-7, 40 mph, 6 percent grade, 60 percent Ip = 0.02 (E − Eo)
no passing zones);
fd = 0.78 (Table 8-8, 70/30 split, 70 percent upgrade); and:
fw = 1.00 (Table 8-5, 12-ft lanes, >6-ft shoulders);
E = 10.7 (Table 8-9, 40 mph, 6 percent for 2-mile grade);
fHV = 1/[1 + PHV(EHV − 1)]
Eo = 1.3 (Table 8-9, 40 mph, 0 percent grade);
PHV = PT + PR + PB = 0.12 + 0.07 + 0.01 = 0.20; and EHV = 1 + (0.25 + PT/HV) (E − 1)
PT/HV = PT /PHV = 0.12/0.20 = 0.60.
Capacity is found at the point where this curve intersects the
Then, computing factors fg and fHV: speed at capacity vs. flow rate at capacity curve on the specific
grade worksheet. The upgrade speed is found by entering this
Ip = 0.02 (10.7 − 1.3) = 0.188 curve with the actual flow rate.
fg = 1/[1 + (0.80 × 0.188)] = 0.87 To plot the curve, the procedure recommends computing service
EHV = 1 + (0.25 + 0.60) (10.7 − 1) = 9.25 flow rate points for the following speeds: 55 mph (LOS A), 52.5
fHV = 1/[1 + 0.20(9.25 − 1)] = 0.38 mph, 50 mph (LOS B), 45 mph (LOS C), 40 mph (LOS D),
and 30 mph. These points would be plotted on the specific grade
worksheet of Figure 8-5, and a smooth curve constructed. Once
The service flow rate for the peak 15 min is now computed
capacity is determined, the service flow rates for every LOS will
using Eq. 8-3:
be known, and the actual LOS can be determined by comparing
the actual flow rate to the computed values.
SFD = 2,800 × 0.83 × 0.78 × 1.00 × 0.87 × 0.38 = 599 vph
The following values are used in these computations:
v/c = 0.00 for 55 mph 0.05 for 52.5 mph
Since the question asks for a maximum volume, rather than a 0.15 for 50 mph 0.40 for 45 mph
flow rate, the service flow rate is converted to a full hour volume 0.64 for 40 mph 0.88 for 30 mph
as follows: (Table 8-7, 7 percent grade, 80 percent no passing zones);
fd = 0.70 (Table 8-8, 80/20 split);
V = SF × PHF = 599 × 0.85 = 509 vph fw = 0.85 for 55–45 mph
0.92 for 45–30 mph
Thus, the maximum full-hour volume which can be accommo- (Table 8-5, 11-ft lanes, 4-ft shoulders);
dated at 40 mph, or LOS D, on the grade described is 509 vph. E = 88.0 for 52.5 mph 46.0 for 50 mph
The maximum flow rate is 599 vph. 22.8 for 45 mph 15.4 for 40 mph
8.2 for 30 mph
(Table 8-9, 7 percent grade, 2 miles, no value given for
CALCULATION 4—FINDING LEVEL OF SERVICE AND 55 mph);
CAPACITY OF A SPECIFIC GRADE Eo = 1.8 for 52.5 mph 1.6 for 50 mph
1.4 for 45 mph 1.3 for 40 mph, 30 mph
1. Description—A rural two-lane highway in mountainous ter- (Table 8-9, 0 percent grade);
rain has a grade of 7 percent, 2 miles long. It currently carries Pp = 0.84 (Given);
a peak hour volume of 500 vph. Other relevant characteristics PHV = PT + PR + PB = 0.04 + 0.10 + 0.02 = 0.16; and
include: PT/HV = PT /PHV = 0.04/0.16 = 0.25.
a. Roadway characteristics—60-mph design speed; 11-ft
lanes; 4-ft shoulders; 80 percent no passing zones. Values of fg may now be computed as follows:
b. Traffic characteristics—80/20 directional split; 4 percent
trucks; 10 percent recreational vehicles; 2 percent buses; Ip(52.5) = 0.02(88.0 − 1.8) = 1.724
84 percent passenger cars; PHF = 0.85. (50.0) = 0.02(46.0 − 1.6) = 0.888
(45.0) = 0.02(22.8 − 1.4) = 0.428 Note that the low or zero service flow rates for 55.0 and 52.5
(40.0) = 0.02(15.4 − 1.3) = 0.282 mph indicate that these average upgrade speeds are virtually im-
(30.0) = 0.02(8.2 − 1.3) = 0.138 possible to maintain on the upgrade described in this problem.
fg(52.5) = 1/[1 + 0.84(1.724)] = 0.41 These computations are summarized on the specific grade work-
(50.0) = 1/[1 + 0.84(0.888)] = 0.57 sheet shown in Figure 8-9. The curve defined by these points is
(45.0) = 1/[1 + 0.84(0.428)] = 0.74 also plotted on the worksheet. The intersection of the plotted curve
(40.0) = 1/[1 + 0.84(0.282)] = 0.81 with the speed at capacity vs. flow rate at capacity curve indicates
(30.0) = 1/[1 + 0.84(0.138)] = 0.90 that capacity is 950 vph, total in both directions, which occurs at
an average upgrade speed of 28.0 mph.
Values of fHV are also computed: To find the existing level of service, the volume of 500 vph is
EHV(52.5) = 1 + (0.25 + 0.25)(88.0 − 1) = 44.5 converted to a flow rate for the peak 15-min period:
(50.0) = 1 + (0.25 + 0.25)(46.0 − 1) = 23.5 v = V/PHF = 500/0.85 = 588 vph
(45.0) = 1 + (0.25 + 0.25)(22.8 − 1) = 11.9
The plotted curve is entered on the worksheet with 588 vph,
(40.0) = 1 + (0.25 + 0.25)(15.4 − 1) = 8.2
and the upgrade speed is found to be 37 mph. Because this speed
(30.0) = 1 + (0.25 + 0.25)(8.2 − 1) = 4.6
is less than 40 mph, the minimum value for LOS D (Table 8-2),
fHV(52.5) = 1/[1 + 0.16(44.5 − 1)] = 0.13
but greater than the speed at capacity (28 mph), the level of service
(50.0) = 1/[1 + 0.16(23.6 − 1)] = 0.22 is E. This can also be determined by comparing the actual flow
(45.0) = 1/[1 + 0.16(11.9 − 1)] = 0.36 rate of 588 vph with the service flow rate for LOS D (40 mph)
(40.0) = 1/[1 + 0.16(8.2 − 1)] = 0.46 of 430 vph and capacity (950 vph).
(30.0) = 1/[1 + 0.16(4.6 − 1)] = 0.63 The last part of this problem asks to find the delay incurred by
Having computed all relevant factors, the total two-way service vehicles traveling up the grade. ‘‘Delay’’ is defined as the differ-
flow rates for the designated speeds may be computed: ence in travel time experienced by vehicles traversing the upgrade
at the existing speed and the travel time which would be experi-
speed 2,800 × v/c × fd × fw × fg × fHV = SF enced if they were able to maintain their approach speed on the
grade. Thus:
55.0 2,800 0.00 0.70 0.85 — — 0 vph
52.5 2,800 0.05 0.70 0.85 0.41 0.13 4 vph Travel time at 55.0 mph = (2 miles/55 mph) × 3600 sec/hour
50.0 2,800 0.15 0.70 0.85 0.57 0.22 31 vph = 130.9 sec/veh
45.0 2,800 0.40 0.70 0.85 0.74 0.36 178 vph Travel time at 37.0 mph = (2 miles/37 mph) × 3600 sec/hour
40.0 2,800 0.64 0.70 0.92 0.81 0.46 430 vph = 194.6 sec/veh
30.0 2,800 0.88 0.70 0.92 0.90 0.63 900 vph Delay = 194.6 − 130.9 = 63.7 sec/veh
Figure 8-9. Worksheet for Calculation 4 (pages 1 and 2).

8-26 rural highways
CALCULATION 5—CONSIDERATION OF A PT/HV = 0.15/0.21 = 0.71.

CLIMBING LANE
Using these values to compute the service flow rate at level-
1. Description—A rural two-lane highway has a 4 percent up- of-service D:
grade of 11⁄2 miles, and has the following other characteristics:
Ip = 0.02(3.8 − 1.3) = 0.05
a. Roadway characteristics—level terrain approach; 12-ft fg = 1/[1 + (0.79 × 0.05)] = 0.96
lanes; 8-ft shoulders; 40 percent no passing zones. EHV = 1 + (0.25 + 0.71)(3.8 − 1) = 3.69
b. Traffic characteristics—DHV = 400 vph; 15 percent fHV = 1/[1 + 0.21(3.69 − 1)] = 0.64
trucks; 5 percent recreational vehicles; 1 percent buses; 79 SFD = 2,800 × 1.00 × 0.87 × 1.00 × 0.96 × 0.64 = 1,497 vph
percent passenger cars; 60/40 directional split; PHF = 0.85.
The actual flow rate is the DHV divided by the PHF, or 400/
0.85 = 471 vph. As this is clearly less than the service flow rate
Is the addition of a climbing lane justified at this location? for LOS D, the existing LOS is not E, and this condition is not met.
2. Solution—It is assumed that a climbing lane on a two-lane The next condition to investigate is whether a 10-mph speed
highway is generally justified when the following conditions are met: reduction of heavy trucks would exist on the grade described.
Based on the assumption that the typical truck on this grade has
1. Upgrade flow rate is greater than 200 vph. a weight/horsepower ratio of 200 lb/hp, Figure 8-2 is used to
2. Upgrade truck flow rate is greater than 20 vph. estimate the speed reduction experienced as shown below:
3. One of the following occurs:
a. The grade operates at LOS E or F.
b. The typical heavy truck reduces its speed by more than
10 mph on the grade.
c. The LOS on the grade is two or more levels poorer
than on the approach to the grade.
Each of these conditions should be checked to justify the con-

struction of the climbing lane:
Upgrade flow rate = 400 × 0.60/0.85 = 282 vph > 200 vph OK
Upgrade trucks = 400 × 0.15 × 0.60/0.85 = 42 vph > 20
vph OK
To justify a climbing lane, only one of the conditions specified

in item 3 must be demonstrated. The LOS will be E or worse if
the actual flow rate exceeds the service flow rate for LOS D. This
value is computed using Eq. 8-3:
SFD = 2,800 × (v/c)D × fd × fw × fg × fHV
where:
It can be seen that the speed reduction will be well in excess
fg = 1/[1 + Pp Ip] of 20 mph, which is greater than 10 mph, fulfilling the last required
condition for justifying a climbing lane. Note that because only
Ip = 0.02 (E − Eo) one of the conditions in item 3 needs to be satisfied, it is not
necessary to investigate the third condition.
and: It can be concluded that a climbing lane is justified on the basis
of the stated criteria.
fHV = 1/[1 + PHV(EHV − 1)]
EHV = 1 + (0.25 + PT/HV)(E − 1)

CALCULATION 6—PLANNING APPLICATION 1
The following values are used:
(v/c)D = 1.00 (Table 8-7, 4 percent grade, 40 mph, 40 percent 1. Description—A rural two-lane highway in mountainous ter-
no passing zones); rain is located in an area where the design hour factor, K, is 0.14.
fd = 0.87 (Table 8-8, 60/40 directional split); What is the maximum AADT which can be accommodated without
fw = 1.00 (Table 8-5); the LOS falling below D during the peak 15-min flow period?
E = 3.8 (Table 8-9, 4 percent, 11⁄2-mile grade, 40 mph); 2. Solution—The solution is simply found by entering Table
Eo = 1.3 (Table 8-9, 0 percent grade, 40 mph); 8-10 with mountainous terrain, LOS D, and a K-factor of 0.14.
PHV = 0.15 + 0.05 + 0.01 = 0.21; and The maximum permissible AADT is found to be 2,700 vpd.
CALCULATION 7—PLANNING APPLICATION 2 percent per year. The responsible highway agency’s policy is to
expand two-lane highways to four lanes before the level of service
1. Description—A rural two-lane highway is located in rolling becomes E during peak periods. In how many years will expansion
terrain in an area where the design hour factor, K, is 0.12. Its of the facility have to be completed under this policy? If it will
current AADT is 5,000 vpd. What is the likely LOS during the take 7 years to construct a four-lane highway, how long will it be
peak 15 min of flow? before the construction project should begin?
2. Solution—Again, the solution is straightforward using Table 2. Solution—The policy requires that expansion of the highway
8-10. The maximum AADT’s for the various levels of service are be completed before the AADT exceeds the maximum allowable
found for rolling terrain and a K-factor of 0.12. The 5,000 AADT value for LOS D. From Table 8-10, the maximum AADT for LOS
is seen to fall between the maximum values for LOS C (4,400 D, for level terrain and a K-factor of 0.12, is 11,200 vpd.
vpd) and LOS D (6,600 vpd). The LOS is therefore expected to The question now becomes: How many years will it take an
be D during the peak 15 min of flow. AADT of 6,600 vpd to grow to 11,200 vpd at a rate of 5 percent
per year? Therefore:
CALCULATION 8—PLANNING APPLICATION 3 11,200 = 6,600(1 + 0.05)n

1. Description—A two-lane highway carrying an AADT of
n = 10.9 years
6,600 vpd is located in level terrain in an area where the design
hour factor, K, is 0.12. The area has a traffic growth rate of 5 Construction should begin in 10.9 − 7 years, or in 3.9 years.
VI. REFERENCES
1. Messer, C.J., ‘‘Two-Lane, Two-Way Rural Highway Level cies of Trucks, Buses, and Recreational Vehicles for Two-
of Service and Capacity Procedures.’’ Project report, NCHRP Lane Rural Highways.’’ Transportation Research Record 615
Project 3-28A, Texas Transportation Institute, College Station, (1976).
Tex. (Feb. 1983). 7. Development of Passing Lane Criteria. Ontario Ministry of
2. Messer, C.J., ‘‘Two-Lane, Two-Way Rural Highway Capac- Transportation and Communications, Downsview, Canada
ity.’’ Final report, NCHRP Project 3-28A, Texas Transporta- (1975).
tion Institute, College Station, Tex. (Feb. 1983). 8. Rooney, F., Turnouts: Traffic Operational Report No. 2. Of-
3. Krummins, I., ‘‘Capacity and Level of Service of Two-Lane fice of Traffic, California Department of Transportation, Sac-
Rural Highways in Alberta.’’ Thesis, University of Calgary, ramento, Calif. (1976).
Calgary, Alberta, Canada (Sept. 1981). 9. Theoretical Analysis: Slow Moving Vehicle Turnouts. Oregon
4. Yagar, S., ‘‘Capacity and Level of Service for 2-Lane Department of Transportation (1978).
Rural Highways.’’ Report to the Ontario Ministry of Trans- 10. St. John, A.D. and Kobett, D.R., ‘‘Grade Effects on Traffic
portation and Communications, Downsview, Ontario, Canada Flow Stability and Capacity.’’ NCHRP Report 185 (1978)
(1980). 110 pp.
5. Traffic Capacity of Major Routes. Organization for Economic 11. A Policy on Geometric Design of Highways and Streets. Amer-
Development, Paris (Jan. 1983). ican Association of State Highway and Transportation Offi-
6. Werner, A., and Morrall, J.F., ‘‘Passenger Car Equivalen- cials, Washington, D.C. (1984).
8-28 rural highways
APPENDIX I
FIGURES AND WORKSHEETS FOR USE IN ANALYSIS OF TWO-LANE

HIGHWAYS
FIGURES PAGE
Figure 8-1. Speed-flow and percent time delay-flow relationships for two-lane rural highways (ideal conditions) ................... 8-28
Figure 8-2. Speed reduction curve for a 200-lb/hp truck................................................................................................................ 8-29
Figure 8-3. Speed reduction curve for a 300-lb/hp truck................................................................................................................ 8-29
WORKSHEETS
Worksheet for General Terrain Segments............................................................................................................................................... 8-30
Worksheet for Specific Grades (Page 1)................................................................................................................................................. 8-31
Worksheet for Specific Grades (Page 2)................................................................................................................................................. 8-32
two-lane highways
Figure 8-1. Speed-flow and percent time delay-flow relationships for two-lane rural highways (ideal conditions).
8-29
8-30
rural highways
Figure 8-2. Speed reduction curve for a 200-lb/hp truck. Figure 8-3. Speed reduction curve for a 300-lb/hp truck.
8-32 rural highways
chapter 9
SIGNALIZED INTERSECTIONS
CONTENTS
i. introduction .......................................................................................................................................................................... 9-2

Traffic Signals ....................................................................................................................................................................... 9-2
Green Time, Effective Green Time, and Lost Times in Signal Cycles .............................................................................. 9-4
Capacity and Level of Service.............................................................................................................................................. 9-5
Capacity of Signalized Intersections ............................................................................................................................... 9-5
Level of Service for Signalized Intersections ................................................................................................................. 9-6
Relating Capacity and Level of Service.......................................................................................................................... 9-7
Computational Alternatives for Delay and Level of Service ......................................................................................... 9-7
Levels of Analysis............................................................................................................................................................ 9-8
Suitability of Operational Configurations ....................................................................................................................... 9-8
ii. methodology.......................................................................................................................................................................... 9-8
Input Module .................................................................................................................................................................... 9-9
Volume Adjustment Module............................................................................................................................................ 9-12
Saturation Flow Rate Module.......................................................................................................................................... 9-14
Capacity Analysis Module............................................................................................................................................... 9-22
LOS Module..................................................................................................................................................................... 9-27
Overview of Planning Method ........................................................................................................................................ 9-32
Field Data Requirements.................................................................................................................................................. 9-32
Default Values.................................................................................................................................................................. 9-33
Synthesis of Signal Operation ......................................................................................................................................... 9-33
Other Analyses ................................................................................................................................................................. 9-34
Input Module .................................................................................................................................................................... 9-34
Volume Adjustment Module............................................................................................................................................ 9-38
Saturation Flow Rate Module.......................................................................................................................................... 9-38
Capacity Analysis Module............................................................................................................................................... 9-43
LOS Module..................................................................................................................................................................... 9-46
Supplemental Uniform Delay Worksheet........................................................................................................................ 9-48
Worksheet Operations ...................................................................................................................................................... 9-50
Computational Requirements........................................................................................................................................... 9-50
Lane Volume Worksheet ................................................................................................................................................. 9-50
Signal Operations Worksheet .......................................................................................................................................... 9-56
Limitations of Planning Method...................................................................................................................................... 9-57
Procedures for Other Analyses ............................................................................................................................................. 9-58
Calculation 1: Operational Analysis of Existing Pretimed, Two-Phase Signal .................................................................. 9-60
Calculation 2: Operational Analysis of Three-Phase, Pretimed Signal............................................................................... 9-69
Calculation 3: Operational Analysis of Multiphase Actuated Signal.................................................................................. 9-78
Calculation 4: Planning Analysis of Intersection with Multilane Approaches ................................................................... 9-84
Calculation 5: Planning Analysis of Intersection with Single-Lane Approaches ............................................................... 9-86
Calculation 6: Determining v/c and Service Flow Rates—An Alternative Use of Operational Analysis Procedure ....... 9-88

9-2 urban streets
v. references .............................................................................................................................................................................. 9-96

appendix i. Intersection Geometrics—Suggestions for Estimating Design Elements .......................................................... 9-97
appendix ii. Suggestions for Establishing Signal Design in Analysis .................................................................................. 9-98
appendix iii. Measurement of Intersection Control Delay in the Field ................................................................................ 9-117
appendix iv. Direct Measurement of Prevailing Saturation Flow Rates .............................................................................. 9-121
appendix v. Worksheets for Use in Analysis ........................................................................................................................ 9-124
appendix vi. Extension of Signal Delay Models To Incorporate Effect of Initial Queue ................................................... 9-139
I. INTRODUCTION
This chapter contains procedures for the analysis of signalized on capacity are outlined briefly. The following terms are com-
intersection capacity and level of service. The signalized intersec- monly used to describe traffic signal operation:
tion is one of the most complex locations in a traffic system.
Signalized intersection analysis must consider a wide variety of Cycle—any complete sequence of signal indications;
prevailing conditions including the amount and distribution of traf- Cycle length—the total time for the signal to complete one cycle,
fic movements, traffic composition, geometric characteristics, and stated in seconds and given the symbol C;
the details of intersection signalization. The methodology of this Interval—a period of time during which all signal indications
chapter focuses on the determination of level of service for known remain constant;
or projected prevailing conditions but presents computational alter- Phase—the part of a cycle allocated to any combination of
natives for determining other variables using an assumed or desired traffic movements receiving the right-of-way simultaneously dur-
level of service. ing one or more intervals;
In other chapters of this manual, the capacity of a highway is Change-and-clearance interval—the yellow plus all-red inter-
related primarily to the geometric characteristics of the facility as vals that occur between phases to provide for clearance of the
well as to the composition of the traffic stream on the facility. intersection before conflicting movements are released, stated in
Geometrics is a fixed, or nonvarying, characteristic of a facility. seconds and given the symbol Y;
Thus, allowing for some variation in traffic composition over time, Green time—the time within a given phase during which the
the capacity of a facility is generally a stable value that can be green indication is shown, stated in seconds and given the sym-
significantly improved only by initiating geometric improvements. bol G;
At the signalized intersection, an additional element is intro- Lost time—time during which the intersection is not effectively
duced into the concept of capacity: time allocation. A traffic signal used by any movement, which occurs during the change-and-clear-
essentially allocates time among conflicting traffic movements that ance intervals (when the intersection is cleared) and at the begin-
seek use of the same physical space. The way in which time is ning of each phase as the first few vehicles in a standing queue
allocated has a significant impact on the operation and the capacity experience start-up delays, given the symbol L;
of the intersection and its approaches. Effective green time—the time that is effectively available to a
The methodology presented here addresses the capacity and movement, generally taken to be the green time plus the change-
level of service of intersection approaches and the level of service and-clearance interval minus the lost time for the designated move-
of the intersection as a whole. Capacity is evaluated in terms of ment, stated in seconds and given the symbol gi;
the ratio of demand flow rate (volume) to capacity (v/c ratio), Effective green ratio—the ratio of effective green time to the
whereas level of service is evaluated on the basis of average control cycle length, given the symbol, gi /C;
delay per vehicle (in seconds per vehicle). The capacity of the Effective red time—the time during which a given movement
intersection as a whole is not addressed, because both the design or set of movements is effectively not permitted to occur, the cycle
and the signalization of intersections focus on the accommodation length minus the effective green time, stated in seconds and given
of major movements in the intersection and on its approaches. the symbol ri.
Capacity is therefore only meaningful as applied to these major
movements and approaches. Traffic engineering textbooks describe three types of traffic sig-
nal controllers:
1. Pretimed controllers: A preset sequence of phases is dis-

TRAFFIC SIGNALS
played in repetitive order. Each phase has a fixed green time and
a change-and-clearance interval that are repeated in each cycle to
Modern traffic signals allocate time in a variety of ways, from produce a constant cycle length.
the simplest two-phase pretimed mode to the most complex multi- 2. Fully actuated controllers: The timing on all of the ap-
phase actuated mode. The basic terminology of traffic signals is proaches to an intersection is influenced by vehicle detectors. Each
described and the various types of signal operation and their impact phase is subject to a minimum and a maximum green time, and
signalized intersections 9-3
some phases may be skipped if no demand is detected. The cycle is not intended to design timing plans but to estimate the average
length for fully actuated control will vary from cycle to cycle. value of the cycle length and phase times that will result from a
3. Semiactuated controllers: Some approaches (typically on the specified combination of traffic conditions and controller settings.
minor street) have detectors, and some do not. The earliest form This procedure may be applied to controllers with coordinated
of semiactuated control was designed to confine the green indica- phases in addition to those that operate in an isolated mode. It is
tion to the major street in the absence of a minor-street actuation. necessary to know the details of the actuated controller settings
Once actuated, the minor-street green is displayed for a period just for each phase because these settings will have a significant impact
long enough to accommodate the traffic demand. on the resulting timing plan. Although this procedure fairly repre-
sents the traffic signal timing that can be expected at an intersection
Although these equipment-based definitions have persisted in traf- with actuated control given the stated variables, it does not repre-
fic engineering terminology, the evolution of traffic control tech- sent the minimum delay cycle or optimum operation.
nology has complicated their function from the analyst’s These two signal timing estimation procedures provide a useful
perspective. For purposes of capacity and level of service (LOS) computational resource. However, local policies or methods should
analysis, it is no longer sufficient to consider the controller type
also be consulted when traffic signal operation is determined. The
as a global descriptor of the intersection operation. Instead, an
timing plan estimation methodology in this chapter is provided to
expanded set of these definitions must be applied individually to
assist in capacity analysis and should not be construed to suggest
each lane group.
nationally accepted standards, criteria, or guidelines for traffic sig-
Each traffic movement may be served by a phase that is either
nal operation.
actuated or nonactuated. Nonactuated phases may be coordinated
It is not only the allocation of green time that has a significant
with neighboring signals on the same route, or they may function
impact on capacity and operations at a signalized intersection but
in an isolated mode without any influence from other signals.
also the manner in which turning movements are accommodated
Nonactuated phases generally operate with fixed minimum green
within the phase sequence. Signal phasing can provide for pro-
times, which may be extended by reassigning unused green time
tected, permitted, or not opposed turning movements.
from actuated phases with low demand, if such phases exist.
Actuated phases, on the other hand, may be used for intersec- A permitted turning movement is made through a conflicting
tions at which other phases are coordinated, but they may not, for pedestrian flow or opposing-vehicle flow. Thus, a left-turn move-
the purposes of this chapter, be coordinated themselves. Actuated ment that is made at the same time as the opposing through move-
phases are subject to being shortened on cycles with low demand. ment is considered to be permitted, as is a right-turn movement
On cycles with no demand, they may be skipped entirely, or they made at the same time as pedestrian crossings in a conflicting
may be displayed for their minimum duration. With systems in crosswalk.
which the nonactuated phases are coordinated, the actuated phases Protected turns are those made without these conflicts, such as
are also subject to early termination (force-off) to accommodate turns made during an exclusive left-turn phase or a right-turn phase
the progression design for the system. during which conflicting pedestrian movements are prohibited.
The capacity analysis procedures in this chapter are based on Permitted turns experience the friction of selecting and passing
known or projected signalization plans. Two alternative procedures through gaps in a conflicting vehicle or pedestrian flow. Thus, a
are provided to assist the analyst in establishing signalization plans. single permitted turn often consumes more of the available green
The first is the planning method, which produces estimates of the time than a single protected turn. Either permitted or protected
cycle length and green times that could be considered to constitute turning phases may be more efficient in a given situation, de-
a ‘‘reasonable and effective’’ signal timing plan. The planning pending on the turning and opposing volumes, intersection geome-
method requires minimal field data and relies instead on default try, and other factors.
values for the required traffic and control parameters. Although Turning movements that are not opposed do not receive a dedi-
intended primarily for planning purposes, this method may be used cated left-turn phase (i.e., a green arrow), but because of the nature
to design initial timing plans for pretimed signals. Timing plans of the intersection, they are never in conflict with through traffic.
produced by the planning method will not generally be optimal This condition occurs on one-way streets, at T-intersections, and
with respect to intersection performance, and they may not even with signal phasing plans that provide complete separation between
be implementable because certain practical considerations such as all movements in opposite directions (i.e., split-phase operation).
minimum phase times are ignored. Their main purpose will be to Such movements must be treated differently in some cases because
support the analysis of capacity and level of service at signalized they may be accommodated in shared lanes without impeding the
intersections using the methodology in this chapter. through traffic. It is important to distinguish between left turns
A more detailed procedure is provided in Appendix II for esti- that are not opposed at any time and those that may be unopposed
mating the timing plan at both pretimed and traffic-actuated sig- during some part of the signal cycle and opposed during another
nals. The pretimed procedure provides the basis for the design of part. Left turns that are opposed during any part of the sequence
signal timing plans that equalize the degree of saturation on the will impede through traffic in shared lanes.
critical approaches for each phase of the signal sequence. This The preceding discussion emphasizes this primary concept: the
procedure does not provide for optimum operation. The planning capacity of an intersection is highly dependent on the signalization
method builds on this procedure by adding some assumptions and present. Given the range of potential signal control schemes, inter-
approximations to produce a complete worksheet for timing plan section capacity is far more variable than that on other types of
estimation. facilities, where capacity is mainly dependent upon the physical
Controllers with traffic-actuated phases will respond to detector geometry of the roadway. In effect, signalization, which can be
inputs to generate different timing plans on each cycle of operation. changed frequently and quickly, allows considerable latitude in
Therefore the traffic-actuated procedure contained in Appendix II the management of the physical capacity of the intersection space
9-4 urban streets
Figure 9-1. Relationship among actual green, lost-time elements, extension of effective green, and effective green.
and geometry. Thus, the concept of intersection capacity is some- the two lost times and apply both at the beginning of a particular
what different from that discussed in previous chapters. traffic movement, particularly for protected-plus-permitted phas-
ing analysis. Thus the following relationships exist for typical
conditions, and the relationship among actual green, lost time,
GREEN TIME, EFFECTIVE GREEN TIME, AND LOST TIMES
extension of effective green, and effective green is illustrated in
IN SIGNAL CYCLES
Figure 9-1.
For any given lane group at a signalized intersection, only three l1 = 2 (typical)
signal indications are seen: green, yellow, and red. The red indica- l2 = Y i − e
tion usually includes a short period during which all indications
are red, referred to as an ‘‘all-red’’ interval, which with the yellow where e = 2 (typical; e may be higher in congested conditions);
indication forms the change-and-clearance interval between two
green phases. tL = l1 + l2 = l1 + Yi − e = 2 + Yi − 2 = Yi
For analysis purposes, it is convenient to divide the signal cycle (typical; tL may be less in congested conditions).
for a given lane group into two simplified components: effective As shown, the total lost time for the movement is deducted from
green and effective red. Effective green time for a given lane group the beginning of the actual green phase. Thus, a small portion of
is the time that may be used by vehicles in the subject lane group Gi becomes part of the effective red, ri. This portion is equal to
at the saturation flow rate. Effective red is defined as the cycle the lost time for the movement, tL. Because all of the lost time for
length minus the effective green. the movement is deducted at the beginning of the green, effective
It is important that the relationship between the actual green, green can be assumed to run through the end of the yellow-plus-
yellow, and red times shown on signal faces and the effective all-red change-and-clearance interval, Yi. Thus, for any given
green and red times be well understood. Each time a movement movement:
is started and stopped, two ‘‘lost times’’ are experienced. At the
beginning of movement, the first several vehicles in the queue gi = Gi + Yi − tL (9-1)
experience start-up losses that result in their movement at less than
the saturation flow rate (Figure 9-1). At the end of a movement, ri = Ri + tL (9-2)
there is a portion of the change-and-clearance interval (yellow and The simplified concept of applying all of the lost time at the
all-red) that is not used for vehicular movement. beginning of a movement makes it easier to analyze more complex
The following definitions apply to these variables: signalizations involving protected-plus-permitted left-turn
Gi = actual green time allotted to lane group i, sec; treatments.
Yi = sum of actual yellow change time plus all-red clearance As a general rule, a lost time tL is applied each time a movement
time allotted to lane group i, sec; starts. Thus, where a given movement starts in a protected phase
Ri = actual red time exclusive of the all-red clearance time allot- and continues through a permitted phase (or vice versa), only one
ted to lane group i, sec; lost time is deducted. No lost time is assumed to occur at the
gi = effective green time for lane group i, sec; boundary between the permitted and protected phases for continu-
ri = effective red time for lane group i, sec; ing movements.
l1 = start-up lost time, sec; Figure 9-2 diagrams a more complex phasing involving a pro-
l2 = clearance lost time, sec; tected-plus-permitted left-turn movement, a classic lead-lag phas-
e = extension of effective green (the amount of change-and- ing scheme in which left turns are protected in Phase 1a [east-
clearance time usable as effective green), sec; and bound (EB)] and Phase 1c [westbound (WB)] and permitted during
tL = total lost time for the lane group (the sum of l1 and l2), sec. the common Phase 1b. The question of how many lost times are
included in such a phase sequence is an important one. Using the
Research has shown that start-up lost time (l1) is normally about general rule that the entire lost time for a movement is applied at
2 sec. It has also shown that the extension of effective green the time the movement begins, the following may be determined:
(e) is normally about 2 sec (sometimes longer under congested
conditions). The remainder of the change-and-clearance time is T In Phase 1a, the EB through and left-turn movements begin.
the clearance lost time (l2). It is analytically convenient to combine Thus, a lost time tL is applied to both movements.
Figure 9-2. Protected-plus-permitted signal phasing.
T In Phase 1b, the EB through and left-turn movements con- The v/c ratio is a measure of capacity sufficiency, that is,
tinue. No lost times are assigned to the continuing movements in whether or not the physical geometry and signal design provide
this phase. The WB through and left-turn movements begin in sufficient capacity for the subject movement or movements. Delay
this phase, however, and a lost time tL must be applied to these is a measure of quality of service to the road user. Both must be
movements. analyzed to fully understand the anticipated operational character-
T In Phase 1c, only the WB through and left turns continue. istics of the intersection, and neither can be substituted for the
Because these movements did not start in this phase, no lost time other. As a practical matter, however, it must be recognized that an
is applied here. Further, because no movements begin in Phase intersection cannot operate beyond its capacity indefinitely without
1c, no lost time is applied to any movement in Phase 1c. experiencing excessive delay.
T In Phase 2, northbound (NB) and southbound (SB) move- For planning purposes, it may be more appropriate to consider
ments begin, and a lost time tL must be applied. the provision of adequate future capacity as related to geometric
design features. Delay may be less of a concern, because it may
The total lost time in the signal cycle, L, is also important. This
be improved significantly through coordination of signals and im-
is the total lost time involved in the critical path through the signal
proved signal design. In the analysis of existing problem locations,
cycle. The determination of the critical path and the finding of L
delay may be a more significant consideration when improved
are discussed later in this chapter.
controls are considered. Both of these important concepts are dis-
cussed in more detail in the sections that follow.
CAPACITY AND LEVEL OF SERVICE
The concepts of capacity and level of service are central to the Capacity of Signalized Intersections
analysis of intersections, as they are for all types of facilities. In
intersection analysis, however, the two concepts are not as strongly Capacity at intersections is defined for each lane group. The
correlated as they are for other facility types. In previous chapters, lane group capacity is the maximum rate of flow for the subject
the same analysis results yielded a determination of both the capac- lane group that may pass through the intersection under prevailing
ity and the level of service of the facility. For signalized intersec- traffic, roadway, and signalization conditions. The rate of flow is
tions, the two are analyzed separately and are not related in a generally measured or projected for a 15-min period, and capacity
simple way to each other. It is critical to note at the outset, however, is stated in vehicles per hour (vph).
that both capacity and level of service must be fully considered Traffic conditions include volumes on each approach, the distri-
to evaluate the overall operation of a signalized intersection. bution of vehicles by movement (left, through, right), the vehicle
A separate capacity is computed for each lane group ap- type distribution within each movement, the location of and use
proaching an intersection. A lane group is defined as one or more of bus stops within the intersection area, pedestrian crossing flows,
lanes that accommodate traffic and have a common stop line and and parking movements near the intersection area.
capacity shared by all vehicles. Capacity analysis results in the Roadway conditions include the basic geometrics of the intersec-
computation of volume-to-capacity (v/c) ratios for each lane group. tion, including the number and width of lanes, grades, and lane
The v/c ratio is the actual or projected rate of flow on a designated use allocations (including parking lanes).
lane group during a 15-min interval divided by the capacity of the Signalization conditions include a full definition of the signal
lane group. Although the capacity of the entire intersection is not phasing, timing, and type of control, and an evaluation of signal
defined, a composite v/c ratio for the sum of the critical lane groups progression for each lane group.
within the intersection is computed as an indication of the overall The capacity of designated lane groups within an approach is
intersection sufficiency. evaluated and determined using the procedures in this chapter.
Level of service is based on the average control delay per vehicle This may be done to isolate lanes serving a particular movement
for various movements within the intersection. Although v/c affects or movements, such as an exclusive right- or left-turn lane. Lanes
delay, there are other parameters that more strongly affect it, such so designated for separate analysis were defined earlier as lane
as the quality of progression, length of green phases, cycle lengths, groups. The procedure in this chapter contains guidelines for when
and others. Thus, for any given v/c ratio, a range of delay values and how separate lane groups should be designated on an approach.
may result, and vice versa. For this reason, both the capacity and Capacity at signalized intersections is based on the concept of
level of service of the intersection must be carefully examined. saturation flow and saturation flow rate, defined as the maximum
9-6 urban streets
rate of flow that can pass through a given lane group under prevail- where
ing traffic and roadway conditions, assuming that the lane group Xc = critical v/c ratio for the intersection;
has 100 percent of real time available as effective green time. ∑(v/s)ci = summation of flow ratios for all critical lane groups, i;
Saturation flow rate is given the symbol s and is expressed in units C = cycle length, sec; and
of vehicles per hour of effective green time (vphg) for a given L = total lost time per cycle, computed as the sum of the
lane group. lost time tL for all critical lane groups, i.
The flow ratio for a given lane group is defined as the ratio of
the actual or projected demand flow rate for the lane group (vi) to Equation 9-5 is useful in evaluating the overall intersection with
the saturation flow rate (si). The flow ratio is given the symbol respect to the geometrics and total cycle length provided and also
(v/s)i (for lane group i). in estimating signal timings when they are unknown or not speci-
The capacity of a given lane group may be stated as fied by local policies or procedures. It gives the v/c ratio for all
critical movements, assuming that green time has been allocated
ci = si (gi /C) (9-3) in proportion to the v/s values. It is therefore possible to have a
critical v/c ratio of less than 1.0 and still have individual move-
where ments oversaturated within the signal cycle. A critical v/c ratio
ci = capacity of lane group i, vph; less than 1.0, however, does indicate that all movements in the
si = saturation flow rate for lane group i, vphg; and intersection can be accommodated within the defined cycle length
gi /C = effective green ratio for lane group i. and phase sequence by proportionally allocating green time. In
essence, the total available green time in the phase sequence is
The ratio of flow rate to capacity (v/c), often called the volume- adequate to handle all movements if allocated solely on the basis
to-capacity ratio, is given the symbol X in intersection analysis. of v/s.
This new symbol is introduced in this chapter to emphasize the The Xc value can, however, be misleading when used as an
strong relationship of capacity to signalization conditions and for indicator of the overall sufficiency of the intersection geometrics,
consistency with the literature, which also refers to this variable as is often required in planning applications. The problem is that
as the degree of saturation. low flow rates dictate the need for short cycle lengths to minimize
For a given lane group i: delay. Inspection of Equation 9-5 suggests that shorter cycle
lengths produce a higher Xc, for a specified level of traffic demand.
Xi = (v/c)i = vi /(sigi /C) = viC/(sigi) (9-4) Furthermore, many signal timing methods, including the planning
where method described later in this chapter, are based on a fixed target
value of Xc. This tends to make Xc independent of the demand
Xi = (v/c)i = ratio for lane group i; volumes.
vi = actual or projected demand flow rate for lane group A broader indicator of the overall sufficiency of the intersection
i, vph; is therefore obtained by substituting the maximum cycle length
si = saturation flow rate for lane group i, vphg; acceptable to the agency responsible for the signal operation in
gi = effective green time for lane group i, sec; and place of the actual cycle length in Equation 9-5:
C = cycle length, sec.
Xcm = o (v/s) ci [Cmax/(Cmax − L)] (9-6)
Sustainable values of Xi range from 1.0 when the flow rate
equals capacity to zero when the flow rate is zero. Values above where
1.0 indicate an excess of demand over capacity. Xcm = critical v/c ratio based on the maximum acceptable cycle
The capacity of the full intersection is not a significant concept length, and
and is not specifically defined here. Rarely do all movements at Cmax = maximum acceptable cycle length, sec.
an intersection become saturated at the same time of day. It is the
ability of individual movements to move through the intersection For planning purposes, Xcm offers a more appropriate indicator
with some efficiency that is the critical concern. of the proportion of the actual capacity of the intersection that is
Another capacity concept of utility in the analysis of signalized being used by the specified traffic volumes. In the balance of
intersections, however, is the critical v/c ratio Xc, which is the this chapter, Xc will be used to represent the critical v/c ratio for
v/c ratio for the intersection as a whole, considering only the lane operational analyses and Xcm will be used for planning analysis.
groups that have the highest flow ratio (v/s) for a given signal The analysis of capacity in this chapter focuses on the computa-
phase. For example, in a two-phase signal, opposing lane groups tion of saturation flow rates, capacities, and v/c ratios for various
move during the same green time. Generally, one of these two lane groups of the intersection. Procedures for these computations
lane groups will require more green time than the other (i.e., it are described in greater detail in Sections II, Methodology, and
will have a higher flow ratio). This would be the ‘‘critical’’ lane III, Procedures for Application.
group for the subject signal phase. Each signal phase will have a
critical lane group that determines the green-time requirements for
the phase. When signal phases overlap, the identification of these Level of Service for Signalized Intersections
critical lane groups becomes somewhat complex; this situation is
discussed in Section II, Methodology. Level of service for signalized intersections is defined in terms
The critical v/c ratio for the intersection is defined in terms of of delay, which is a measure of driver discomfort, frustration, fuel
critical lane groups or approaches: consumption, and lost travel time. The delay experienced by a
motorist is made up of a number of factors that relate to control,
Xc = o (v/s) ci [C/(C − L)] (9-5) geometrics, traffic, and incidents. Total delay is the difference
Table 9-1. Level-of-Service Criteria for Signalized high v/c ratios below 1.0 with many individual cycle failures. Poor
Intersections progression and long cycle lengths may also be major contributing
level of service control delay per vehicle (sec) factors to such delay levels.
A ≤10
B >10 and ≤20 Relating Capacity and Level of Service
C >20 and ≤35
D >35 and ≤55
E >55 and ≤80 Because delay is a complex measure, its relationship to capacity
F >80 is also complex. The levels of service in Table 9-1 were established
on the basis of the acceptability of various amounts of delay to
drivers. Although local standards may vary, LOS C may be re-
garded as a desirable design objective. It is important to note
that this concept is not related to capacity in a simple one-to-one
between the travel time actually experienced and the reference
fashion.
travel time that would result during ideal conditions: in the absence
In previous chapters, the lower bound of LOS E was defined
of traffic control, in the absence of geometric delay, in the absence
to be capacity; that is, the v/c ratio is by definition 1.0. This is not
of any incidents, and when there are no other vehicles on the road.
the case for the procedures in this chapter. It is possible, for exam-
In Chapters 9, 10, and 11, only the portion of total delay attributed
ple, to have delays in the range of LOS F (unacceptable) while
to the control facility is quantified. This delay is called control
the v/c ratio is below 1.0, perhaps as low as 0.75 to 0.85. Very
delay. Control delay includes initial deceleration delay, queue
long delays can occur at such v/c ratios when some combination
move-up time, stopped delay, and final acceleration delay. In con-
of the following conditions exists: (a) the cycle length is long, (b)
trast, in previous versions of this chapter of the HCM (1994 and
the lane group in question is disadvantaged by the signal timing
earlier), delay included only stopped delay. In this chapter, control
(has a long red time), and (c) the signal progression for the subject
delay may also be referred to as signal delay.
movements is poor.
Specifically, LOS criteria for traffic signals are stated in terms
The reverse is also possible: a saturated lane group (i.e., v/c
of the average control delay per vehicle, typically for a 15-min
ratio greater than 1.0) may have short delays if (a) the cycle length
analysis period. The criteria are given in Table 9-1. Delay may be
is short or (b) the signal progression is favorable for the subject
measured in the field or estimated using procedures presented later
lane group, or both.
in this chapter. Delay is a complex measure and is dependent on
Thus, the designation LOS F does not automatically imply that
a number of variables, including the quality of progression, the
the intersection, approach, or lane group is over capacity, nor does
cycle length, the green ratio, and the v/c ratio for the lane group
a level of service better than E automatically imply that unused
in question.
capacity is available.
LOS A describes operations with very low control delay, up to
The procedures and methods in this chapter require the analysis
10 sec per vehicle. This level of service occurs when progression
of both capacity and LOS conditions to fully evaluate the operation
is extremely favorable and most vehicles arrive during the green
of a signalized intersection. It is imperative that the analyst recog-
phase. Most vehicles do not stop at all. Short cycle lengths may
nize the unique relationship of these two concepts as they apply
also contribute to low delay.
to signalized intersections.
LOS B describes operations with control delay greater than 10
and up to 20 sec per vehicle. This level generally occurs with good
progression, short cycle lengths, or both. More vehicles stop than Computational Alternatives for Delay and Level of Service
with LOS A, causing higher levels of average delay.
LOS C describes operations with control delay greater than 20 This chapter defines the level of service at a signalized intersec-
and up to 35 sec per vehicle. These higher delays may result from tion in terms of average control delay per vehicle. It also establishes
fair progression, longer cycle lengths, or both. Individual cycle threshold delay values for the various levels of service and presents
failures may begin to appear at this level. The number of vehicles a detailed computational methodology for estimating delay. The
stopping is significant at this level, though many still pass through methodology prescribed here represents a broad accumulation of
the intersection without stopping. professional knowledge, experience, and research. As such, it of-
LOS D describes operations with control delay greater than 35 fers a consistent and impartial means of assessing the level of
and up to 55 sec per vehicle. At level D, the influence of congestion service at signalized intersections under a full range of operating
becomes more noticeable. Longer delays may result from some conditions.
combination of unfavorable progression, long cycle lengths, or It must be recognized, however, that delay is a quantity that may
high v/c ratios. Many vehicles stop, and the proportion of vehicles be directly measured in the field. The results of the computations
not stopping declines. Individual cycle failures are noticeable. described in this chapter cannot be expected to supersede the re-
LOS E describes operations with control delay greater than 55 sults of properly executed field studies that measure delay. Further-
and up to 80 sec per vehicle. This level is considered by many more, the literature contains a variety of models that offer delay
agencies to be the limit of acceptable delay. These high delay estimation techniques based on complex software algorithms, some
values generally indicate poor progression, long cycle lengths, and of which require additional field data. Some of these models are
high v/c ratios. Individual cycle failures are frequent occurrences. designed to deal explicitly with unusual situations of geometrics,
LOS F describes operations with control delay in excess of 80 signal operation, driver behavior, and so forth. It is not, therefore,
sec per vehicle. This level, considered to be unacceptable to most possible to argue the superiority of the macroscopic model con-
drivers, often occurs with oversaturation, that is, when arrival flow tained in this chapter over all of the more microscopic methods
rates exceed the capacity of the intersection. It may also occur at under all conditions, as described in Chapter 1.
9-8 urban streets
Although worksheets are provided for all the calculations, the analysis into the level of operational analysis. The accuracy of the
use of computerized versions of these procedures is a universal computed level of service will depend on the degree of effort
practice. The primary purpose of the worksheets is to explain the applied to the development of the data items that are represented
computational methodology in a clear and understandable manner. by default values (e.g., lane widths, truck proportions) and on the
Productivity considerations dictate the need for automation, and quality of the approximated signal timing plan. Thus, planning
fortunately there is an excellent choice of software products that and operational analyses may be viewed as two applications that
implement the methodology of this chapter. It is not the purpose of represent the extremes of a continuous range of possibilities.
this manual to endorse, compare, or even mention these products; The operational analysis methodology considers the full details
however, their importance to the analysis of signalized intersection of each of four components: demand flow rates at the intersection,
capacity and level of service cannot be ignored, nor can the profes- signalization of the intersection, geometric design or characteristics
sional responsibility of the analyst for the final results. The soft- of the intersection, and the delay or level of service that results
ware must be viewed as a supplement to this manual to be used from these. The methodology is capable of treating any of these
with a thorough understanding of the procedures, and not simply four components as an unknown to be determined knowing the
as a time-saving alternative. For example, the interpretation of details of the other three. Thus the method can be used to
the level of precision available from computations performed by
1. Solve for level of service, knowing details of intersection
software should be compatible with the accuracy of the input data.
flows, signalization, and geometrics;
To simplify the description of the computational process, certain
2. Solve for allowable service flow rates for selected levels of
parameters may be assumed to apply globally to all movements.
service, knowing the details of signalization and geometrics;
This is common practice in signalized intersection analysis and is
3. Solve for signal timing (for an assumed phase plan), knowing
reflected in all the worksheets and sample calculations. It is not,
the desired level of service and the details of flows and geomet-
however, the intent in this chapter to preclude the use of move-
rics; and
ment-specific values when such information is available and a
4. Solve for basic geometrics (number or allocation of lanes),
higher level of precision is required.
knowing the desired level of service and the details of flows and
signalization.
Levels of Analysis Although the methodology is capable of computations in all
four modes, the specific procedures and worksheets are designed
Two levels of analysis are presented. The primary methodology for the first of these, that is, a solution for level of service. In
used is the operational analysis. At this level, detailed information developing alternative signal and geometric designs, it is often
on all prevailing traffic, roadway, and signalization characteristics necessary to consider changes simultaneously in both. Rarely can
must be provided. The method provides for a full analysis of signalization be considered in isolation from geometric design and
capacity and level of service and can be used to evaluate alternative vice versa. Thus, the most frequent type of analysis would consider
traffic demands, geometric designs, or signal plans, or all three. such alternatives on a trial-and-error basis and would not attempt
A second method is provided for planning analysis. At this to hold one constant and solve for the other. Sample calculations,
level, only capacity is addressed because it is not necessary, nor however, illustrate alternative uses of the methodology.
is it practical, to perform detailed calculations of delay given the
accuracy of the data that are generally available for planning pur- Suitability of Operational Configurations
poses. Basic information on intersection geometrics, lane utiliza-
tion, and movement-specific traffic volumes is required, along The methodology presented in this chapter covers a wide range
with the manner in which each of the left turns is accommodated of operational configurations, including combinations of phase
(protected, permitted, etc.) and the presence or absence of parking plans, lane utilization, and left-turn treatment alternatives. It is
on each approach. important to note that some of these configurations may be consid-
The planning method generates two important products: (a) a ered unacceptable from a traffic safety point of view by some
projection of the status of the intersection with respect to its capac- operating agencies. The safety aspect of signalized intersections
ity and (b) an approximation of a signal timing plan. Combining cannot be ignored, and the provision in this chapter of an analysis
this approximation with appropriate values for other parameters methodology for a specific operational configuration does not im-
used in the operational analysis, it is possible to extend the planning ply an endorsement of its suitability for application at all locations.
II. METHODOLOGY
OPERATIONAL ANALYSIS or projected for future situations. Because the operational analysis
of signalized intersections is complex, it is divided into five distinct
Operational analysis results in the determination of capacity and modules, as follows:
level of service for each lane group as well as the level of service
for the intersection as a whole. It requires that detailed information 1. Input Module: All required information upon which subse-
be provided concerning geometric, traffic, and signalization condi- quent computations are based is defined. The module includes all
tions at the intersection. These may be known for existing cases necessary data on intersection geometry, traffic volumes and
Figure 9-3. Operational analysis procedure.
conditions, and signalization. It is used to provide a convenient The methodology in this chapter provides formulas and look-
summary for the remainder of the analysis. up tables for all factors that are to be used. In all cases, the tables
2. Volume Adjustment Module: Demand volumes should be pro- provide entries for the extreme limits that are allowed by the
vided in terms of the average flow rates (vph) for the 15-min method; in no case should the tabulated values be extrapolated
analysis period, in which case peak-hour factor values of 1.0 should beyond these limits except when extrapolation is explicitly recom-
be used. Demand volumes may also be stated in terms of average mended (e.g., for lane width factors). Interpolation between tabu-
hourly volumes (vph), in which case the Volume Adjustment Mod- lated values is suggested to avoid the discontinuities that can occur
ule uses the peak-hour factors provided to convert these to flow without interpolation, but the recommended practice in all cases
rates for the 15-min analysis period. In special cases, analysis is to use the formulas that are provided to completely avoid the
periods other than 15 min may be used, in which case average issues of both interpolation and extrapolation. All the examples
flow rates (vph) should be provided for the analysis period and presented later in this chapter are based on the formulas.
peak-hour factors of 1.0 should be used. The definition of lane
groups for analysis also takes place in this module.
3. Saturation Flow Rate Module: The saturation flow rate is Input Module
computed for each of the lane groups established for analysis. The
flow rate is based on the adjustment of an ‘‘ideal’’ saturation flow Figure 9-4 provides a summary of the input information required
rate to reflect a variety of prevailing conditions. to conduct an operational analysis. This information forms the
4. Capacity Analysis Module: Volumes and saturation flow basis for selecting computational values and procedures in the
rates are manipulated to compute the capacity and v/c ratios for modules that follow. The data needed are detailed and varied,
each lane group and the critical v/c ratio for the intersection. and fall into four main categories; geometric conditions, traffic
5. LOS Module: Delay is estimated for each lane group estab- conditions, signalization conditions, and default values.
lished for analysis. Delay measures are aggregated for approaches
and for the intersection as a whole, and levels of service are Geometric Conditions
determined.
Intersection geometry is generally presented in diagrammatic
Figure 9-3 is a diagram of the modules and the analysis proce- form and must include all of the relevant information, including
dure. Each module is discussed in detail in the sections that follow. approach grades, the number and width of lanes, and parking con-
9-10 urban streets
ple-period analysis using the procedures in Appendix VI should

be performed using each of these subperiods individually. The
length of the subperiods would normally be, but not be limited to,
15 min each.
Vehicle type distribution is quantified as the percent of heavy
vehicles (%HV) in each movement, where heavy vehicles are de-
fined as those with more than four wheels touching the pavement.
The number of local buses on each approach should also be identi-
fied, including only those buses making stops to pick up or dis-
charge passengers at the intersection (on either the approach or
departure side). Buses not making such stops are considered to be
heavy vehicles.
Pedestrian flows are needed, because these will interfere with
permitted right turns. The pedestrian flow for a given vehicular
approach is the flow in the crosswalk interfering with right turns
from the approach. Thus, for a westbound approach, the pedestrian
flow in the north crosswalk would be used; for an eastbound ap-
proach, the south-crosswalk flow; for a northbound approach, the
east-crosswalk flow; and for a southbound approach, the west-
crosswalk flow.
One of the most critical traffic characteristics that must be quan-
tified to complete an operational analysis of a signalized intersec-
tion is the quality of the progression. The parameter that best
describes this characteristic is the arrival type (AT) for each lane
group. This parameter is a general categorization that represents
the quality of progression in an approximate manner. Six arrival
types are defined for the dominant arrival flow as follows:
Arrival Type I: Dense platoon, containing over 80 percent of
the lane group volume, arriving at the start of the red phase. This
AT is representative of network links that may experience very
poor progression quality as a result of conditions such as overall
Figure 9-4. Input data needs for each analysis lane group. network signal optimization.
Arrival Type 2: Moderately dense platoon arriving in the middle
of the red phase or dispersed platoon, containing 40 to 80 percent of
ditions. The existence of exclusive left- or right-turn lanes should the lane group volume, arriving throughout the red phase. This AT
be noted, along with the storage lengths of such lanes. is representative of unfavorable progression on two-way arterials.
When the specifics of geometry are to be designed, these fea- Arrival Type 3: Random arrivals in which the main platoon
tures must be assumed for the analysis to continue. State or local contains less than 40 percent of the lane group volume. This AT
policies and guidelines should be used in establishing the trial is representative of operations at isolated and noninterconnected
design. When these are not readily available, Appendix I of this signalized intersections characterized by highly dispersed platoons.
chapter contains suggestions for geometric design that may be It may also be used to represent coordinated operation in which
useful in preparing a preliminary design for analysis. the benefits of progression are minimal.
Arrival Type 4: Moderately dense platoon arriving in the middle
Traffic Conditions of the green phase or dispersed platoon, containing 40 to 80 percent
of the lane group volume, arriving throughout the green phase.
Traffic volumes for the intersection must be specified for each
This AT is representative of favorable progression quality on a
movement on each approach. These volumes are the flow rates in
two-way arterial.
vehicles per hour for the 15-min analysis period, which is the
Arrival Type 5: Dense to moderately dense platoon, containing
normal analysis period length (T = 0.25). If the 15-min data are
over 80 percent of the lane group volume, arriving at the start of
not known, they may be estimated using hourly volumes and peak-
the green phase. This AT is representative of highly favorable
hour factors. In situations where v/c is greater than about 0.9,
progression quality, which may occur on routes with low to moder-
control delay is significantly affected by the length of the analysis
ate side-street entries and which receive high-priority treatment in
period. In these cases, if the 15-min flow rate remains relatively
the signal timing plan design.
constant for more than 15 min, the length of time the flow is
Arrival Type 6: This arrival type is reserved for exceptional
constant should be used as the analysis period, T, in hours.
progression quality on routes with near-ideal progression charac-
If v/c exceeds 1.0 during the analysis period, the length of the
teristics. It is representative of very dense platoons progressing
analysis period should be extended to cover the period of oversatur-
over a number of closely spaced intersections with minimal or
ation in the same fashion, as long as the average flow during that
negligible side-street entries.
period is relatively constant. If the resulting analysis period is
longer than 15 min and different flow rates can be identified during The arrival type is best observed in the field but could be approx-
equal-length subperiods within the longer analysis period, a multi- imated by examining time-space diagrams for the arterial or street
Table 9-2. Relationship Between Arrival Type and than the average pedestrian walking speed of 4.5 fps cited in
Platoon Ratio (Rp) Chapter 13, Pedestrians. The lower value is intended to accommo-
range of date crossing pedestrians who walk at speeds slower than the aver-
arrival platoon ratio default value progression age. Where local policy uses different criteria for estimating
type (Rp) (Rp) quality minimum pedestrian crossing requirements, these criteria should
1 ≤0.50 0.333 Very poor be used in lieu of Equation 9-8.
2 >0.50 and ≤0.85 0.667 Unfavorable When signal phases are actuated, the cycle length and green times
3 >0.85 and ≤1.15 1.000 Random arrivals will vary from cycle to cycle in response to demand. To establish
4 >1.15 and ≤1.50 1.333 Favorable values for analysis, the operation of the signal should be observed in
5 >1.50 and ≤2.00 1.667 Highly favorable the field during the same period that volumes are observed. Average
6 >2.00 2.000 Exceptional
values of cycle length and green time may then be used.
When signalization is to be established as part of the analysis,
state or local policies and procedures should be applied where
in question. The arrival type should be determined as accurately appropriate in designing the signalization for analysis. Appendix
as possible because it will have a significant impact on delay II and the planning method presented later in this chapter contain
estimates and LOS determination. Although there are no definitive suggestions for the design of a trial signalization that may also be
parameters to precisely quantify arrival type, the following ratio useful. These should not be construed to be standards or criteria
is a useful value: for signal design. It should be noted that a trial signalization cannot
be designed until the Volume Adjustment and Saturation Flow
Rp = P(C/gi) (9-7) Rate modules have been completed. In some cases, the computa-
where tions will be iterative, because left-turn adjustments for permitted
turns used in the Saturation Flow Rate Module depend on signal
Rp = platoon ratio; timing. Appendix II also contains suggestions for estimating the
P = proportion of all vehicles in movement arriving during the timing of an actuated signal if field observations are unavailable.
green phase; It should be noted that an operational analysis requires the speci-
C = cycle length; and fication of a signal timing plan for the intersection under study.
gi = effective green time for the movement or lane group. The planning level analysis presented later in this chapter offers
a method for establishing a ‘‘reasonable and effective’’ signal
P may be estimated or observed in the field, whereas gi and C are
timing plan. The planning procedure is based on the methodology
computed from the signal timing. Note that when P is estimated, its
presented in Appendix II to determine an appropriate cycle length
value may not exceed 1.0. The approximate ranges of Rp are related
and green time allocation for pretimed control. This procedure is
to arrival type as shown in Table 9-2, and default values are
recommended only for the estimation of level of service and not
suggested for use in subsequent computations.
for the design of an implementable signal timing plan. The signal
Another traffic condition of interest is the activity in parking
timing design process is more complicated and involves, for exam-
lanes adjacent to analysis lane groups. Parking activity is measured
ple, iterative checks for minimum green time violations. When
in terms of the number of parking maneuvers per hour within 250
one or more phases are traffic actuated, the timing plan will differ
ft upstream of the stop line (Nm). Each vehicle entering or leaving
on each cycle. The traffic-actuated procedure presented in Appen-
a parking place is considered to be a parking maneuver.
dix II may be used to estimate the average cycle lengths and phase
Signalization Conditions times under these conditions, provided that the controller settings
are available.
Complete information regarding signalization is needed. This The design of an implementable timing plan is a complex and
includes a phase diagram illustrating the phase plan, cycle length, iterative process that may be carried out with the assistance of
green times, and change-and-clearance intervals. Actuated lane computer software. Although the methodology presented here is
groups must be identified, including the existence of pushbutton oriented toward the estimation of delay at traffic signals, it was
pedestrian-actuated phases. suggested in Section I of this chapter that the computations could
If pedestrian timing requirements exist, the minimum green time be applied iteratively to develop a signal timing plan. Some of the
for the phase should be indicated and must be provided for in available signal timing software products employ the methodology
the signal timing. The minimum green time for a phase may be of this chapter at least in part.
estimated as There are, however, several aspects of signal timing design that
are beyond the scope of this manual. One such aspect is the choice
Gp = 7.0 + (W/4.0) − Yi (9-8)
of the timing strategy itself. At intersections with traffic-actuated
where phases, the signal timing plan is determined on each cycle by the
instantaneous traffic demand and the controller settings. When all
Gp = minimum green time, sec;
of the phases are pretimed, a timing plan design must be developed.
W = distance from the curb to the center of the farthest travel
Timing plan design and estimation are covered in detail in Appen-
lane on the street being crossed or to the nearest pedestrian
dix II.
refuge island if the pedestrian crossing is to be made over
two signal cycles, ft; and
Yi = change-and-clearance interval (yellow + all-red time), sec. Default Values
It is assumed that the 15th-percentile walking speed of pedestri- Occasionally, some of the field data noted in Figure 9-4 will
ans crossing a street is 4.0 fps in this computation. This is lower not be available. When critical variables are missing, it may be
9-12 urban streets
Table 9-3. Default Values for Use in Operational Volume Adjustment Module
and Planning Analyses
characteristic default value Two major analytic steps are performed in the Volume Adjust-
ment Module: (a) movement volumes are adjusted to flow rates
Traffic
for a 15-min period of analysis, if necessary, and (b) lane groups
Ideal saturation flow rate 1,900 pcphgpl for analysis are established.
Conflicting pedestrian volume None: 0 peds/hr
(assume none unless field data Low: 50 peds/hr Adjustment of Movement Volumes To Reflect Peak Flow Rates
indicate otherwise) Moderate: 200 peds/hr
High: 400 peds/hr As with other chapters and procedures in this manual, the initial
Percent heavy vehicles 2
computational process is to convert any demands stated as hourly
Grade (percent) 0
Number of stopping buses 0/hr volumes to flow rates for the 15-min analysis period within the
Parking conditions No parking hour. This is done by dividing the movement volumes by an appro-
Parking maneuvers 20/hr where parking exists priate peak-hour factor (PHF), which may be defined for the inter-
Arrival type section as a whole, for each approach, or for each movement.
Lane groups with through 3 if isolated
movements 4 if coordinateda vp = V/PHF (9-9)
Lane groups without through 3
movements where
Peak-hour factor 0.90
Lane utilization adjustment factor See Table 9-4 vp = flow rate during 15-min analysis period, vph;
V = hourly volume, vph; and
Facility and Traffic Signal PHF = peak-hour factor.
Signal type Pretimed
Because not all intersection movements may peak at the same
Cycle length range 60–120 sec
Start-up lost time 2.0 sec time, it is valuable to observe 15-min flows directly and select
Extension of effective green time 2.0 sec critical periods for analysis. The conversion of hourly volumes to
Yellow plus all-red change-and- 4.0 sec/phase (for peak flow rates using the PHF assumes that all movements peak
clearance interval planning) during the same 15-min period, and it is therefore a conservative
Unit extension 3.0 secb approach. It is particularly conservative if different PHF values
Area type Non-CBD are assumed for each movement. It should be noted also that
Lane width 12 ft
statistically valid surveys of the PHF for individual movements
a
Better arrival types are often possible with favorable progression design.
b are difficult to obtain during a single peak hour.
Unit extensions may vary significantly based on local conditions.
Determination of Lane Groups for Analysis

Table 9-4. Default Lane Utilization Adjustment The operational analysis procedure is disaggregate; that is, it is
Factors designed to consider individual intersection approaches and indi-
percent of vidual lane groups (as defined in Section I) within approaches. It
traffic in lane utilization is therefore necessary to determine appropriate lane groups for
lane group no. of lanes most heavily adjustment analysis.
movements in lane group traveled lane factor (f LU) Segmenting the intersection into lane groups is generally a rela-
5
Through or 1 100.0 1.00 tively obvious process that considers both the geometry of the
shared 2 52.5 0.95 intersection and the distribution of traffic movements. In general,
3a 36.7 0.91
the smallest number of lane groups is used that adequately de-
52
Exclusive left 1 100.0 1.00 scribes the operation of the intersection. The following guidelines
a
turn 51.5 0.97 may be applied:
52
Exclusive 1 100.0 1.00
right turn a
56.5 0.88 1. An exclusive left-turn lane or lanes should normally be desig-
a
nated as a separate lane group unless there is also a shared left-
If lane group has more lanes than number shown in this table, it is recom-
mended that surveys be made or the smallest fLU shown for that type of lane through lane present, in which case the proper lane grouping will
group be used. depend on the distribution of traffic volume between the move-
ments. The same is true of an exclusive right-turn lane.
2. On approaches with exclusive left-turn or right-turn lanes,
necessary to conduct a planning analysis. However, default values or both, all other lanes on the approach would generally be included
may be used for some of the variables without seriously compro- in a single lane group.
mising computations. Caution should be used when such values 3. When an approach with more than one lane includes a lane
are applied, and it must be recognized that results become more that may be used by both left-turning vehicles and through vehi-
approximate as more default values are used. cles, it is necessary to determine whether conditions permit equilib-
Tables 9-3 and 9-4 summarize default values for use when field rium conditions to exist or whether there are so many left turns
data are not available. Use of many of these defaults generates no that the lane essentially acts as an exclusive left-turn lane, which
adjustments to the base, ideal conditions, but this is not true for is referred to as a de facto left-turn lane.
every default, as in the case of percent heavy vehicles, peak-hour De facto left-turn lanes cannot be identified effectively until the
factor, and lane utilization adjustment factor. proportion of left turns in the shared lane has been computed. A
Figure 9-5. Typical lane groups for analysis.
procedure for estimating this quantity will be presented later. If block so many of the available gaps as to leave insufficient capacity
the computed proportion of left turns in the shared lane equals or for left-turning vehicles. The interaction of all these mechanisms
exceeds 1.0 (i.e., 100 percent), the shared lane must be considered results in vehicles’ establishing an equilibrium through their selec-
a de facto left-turn lane. tion of lanes. The procedures in this chapter attempt to address this
When two or more lanes are included in a lane group for analysis equilibrium state and allow approaches containing shared left-turn
purposes, all subsequent computations treat these lanes as a single and through lanes to be analyzed as a single lane group.
entity. Figure 9-5 shows some common lane group schemes for
analysis. Adjustment for Right Turn on Red (RTOR)
The operation of a shared left-turn and through lane with permit-
ted left-turn phasing is quite complex. Left-turning vehicles exe- When RTOR is permitted, the right-turn volume may be reduced
cute their turning maneuvers through gaps in the opposing traffic by the volume of right-turning vehicles moving on the red phase.
stream. The first gap, however, does not appear until the queue of This is generally done on the basis of hourly volumes before
opposing vehicles clears the intersection. If a left-turner arrives converting to flow rates.
during the interval in which the opposing queue is clearing, it The number of vehicles able to turn right on a red phase is a
effectively blocks the lane for both through and turning vehicles function of several complex factors:
until the first gap appears. Thereafter, left-turning vehicles may T Approach lane allocation (shared or exclusive right-turn lane),
move through gaps in the opposing traffic stream until the green T Demand for right-turn movements,
phase terminates, at which time as many as two left-turning vehi- T Sight distance at the intersection approach,
cles may be able to execute turns during the change interval. Any T Degree of saturation of the conflicting through movement,
lane blockages or congestion in the shared lane will influence lane T Arrival patterns over the signal cycle,
distribution as vehicles move to adjacent lanes to avoid turbulence T Left-turn signal phasing on the conflicting street, and
and delays. T Conflicts with pedestrians.
Another factor also influences lane distribution. If a through vehi-
cle arrives at the intersection at the time that a gap appears in the For an existing intersection, it is appropriate to consider the
opposing traffic stream, no left-turning vehicle will be able to use right turns on red that actually occur. For both the shared lane and
the gap. A large number of through vehicles in the shared lane may the exclusive right-turn lane conditions, the number of right turns
9-14 urban streets
Table 9-5. Adjustment Factor for Average Lane Width (fw) Table 9-6. Adjustment Factor for Heavy Vehicles (fHV)
average lane lane width percent heavy vehicles, %hv heavy vehicle factor, fHV
width, W (ft) factor, fw 0 1.000
8 0.867 2 0.980
9 0.900 4 0.962
10 0.933 6 0.943
11 0.967 8 0.926
12 1.000 10 0.909
13 1.033 15 0.870
14 1.067 20 0.833
15 1.100 25 0.800
16 1.133 30 0.769
W − 12 35 0.741
Note: fw = 1 + for W ≥ 8 (if W > 16, a two-lane analysis may be 40 0.714
30
considered). 45 0.690
50 0.667
75 0.571
100 0.500
on red may be subtracted from the right-turn volume before the 100
Note: fHV = for 0 ≤ %HV ≤ 100, where ET = 2.0
analysis of lane group capacity or level of service. At an existing 100 + %HV (ET − 1)
intersection, the number of right turns on red should be determined passenger cars per heavy vehicle.
by field observations.
If the analysis is dealing with future conditions or if the RTOR Table 9-7. Adjustment Factor for Approach Grade (fg)
volume is not known from field data, it is necessary to estimate
the number of vehicles that will turn right on the red. This is a grade, %g
very difficult quantity to estimate because of the complexity of type percent grade factor, fg
the process and variations in driver behavior. In the absence of Downhill −6 or less 1.030
field data, it is preferable for most purposes to utilize the right- −4 1.020
turn volumes directly without a reduction for the number of right −2 1.010
turns on red except when an exclusive right-turn lane movement Level 0 1.000
is ‘‘shadowed’’ by a protected left-turn phase from the cross street. Uphill +2 0.990
+4 0.980
For example, the westbound left turn will shadow the northbound +6 0.970
right turn. In this case the shadowing left-turn volume per lane +8 0.960
may be removed from the total right-turn volume as right turns +10 or more 0.950
on red. Free-flowing right turns that are not under signal control %G
Note: fg = 1 − for − 6 ≤ %G ≤ +10.
should be removed from the analysis. 200
Saturation Flow Rate Module fp = adjustment factor for the existence of a parking lane adja-
cent to the lane group and the parking activity in that lane,
In the Saturation Flow Rate Module, a saturation flow rate for given in Table 9-8;
each lane group is computed. The saturation flow rate is the flow fbb = adjustment factor for the blocking effect of local buses that
in vehicles per hour that could be accommodated by the lane group stop within the intersection area, given in Table 9-9;
assuming that the green phase was always available to the lane fa = adjustment factor for area type, given in Table 9-10;
group, that is, that the green ratio (g/C) was 1.0. Computations fLU = adjustment factor for lane utilization, computed as de-
begin with the selection of an ‘‘ideal’’ saturation flow rate, usually scribed in the following sections;
1,900 passenger cars per hour of green time per lane (pcphgpl), fRT = adjustment factor for right turns in the lane group, given
and this value is adjusted for a variety of prevailing conditions in Table 9-11; and
that are not ideal. All the adjustment factors are given in Tables fLT = adjustment factor for left turns in the lane group, given
9-4 through 9-12. in Table 9-12 or computed as described in the following
sections.
s = so N fw fHV fg f p fbb fa fLU fRT fLT (9-10)
where Measured values of the prevailing saturation flow rate will pro-
duce more accurate results than the estimation procedure described
s = saturation flow rate for the subject lane group, expressed
here and can be used directly without further adjustment. Appendix
as a total for all lanes in the lane group under prevailing
IV gives a procedure for measuring the prevailing saturation flow
conditions, vphg;
rate directly.
so = ideal saturation flow rate per lane, usually 1,900 pcphgpl;
N = number of lanes in the lane group; Adjustment Factors
fw = adjustment factor for lane width (12-ft lanes are standard),
given in Table 9-5; The use of adjustment factors is similar to that in previous
fHV = adjustment factor for heavy vehicles in the traffic stream, chapters. Each factor accounts for the impact of one or several
given in Table 9-6; prevailing conditions that are different from the ideal conditions
fg = adjustment factor for approach grade, given in Table 9-7; for which the ideal saturation flow rate applies.
Table 9-8. Adjustment Factor for Parking (fp)

no. of lanes no. of parking maneuvers per hour, Nm
in lane group,
N no. parking 0 10 20 30 40a
1 1.000 0.900 0.850 0.800 0.750 0.700
2 1.000 0.950 0.925 0.900 0.875 0.850
3a 1.000 0.967 0.950 0.933 0.917 0.900
N − 0.1 − 18Nm/3600
Note: fp = for 0 ≤ Nm ≤ 180, fp ≥ 0.05. a
Use formula for more than 3 lanes or more than 40 maneuvers per hour.
N
Table 9-9. Adjustment Factor for Bus Blockage (fbb)

no. of no. of buses stopping per hour, NB
lanes in
lane group,
N 0 10 20 30 40a
1 1.000 0.960 0.920 0.880 0.840
2 1.000 0.980 0.960 0.940 0.920
3a 1.000 0.987 0.973 0.960 0.947
N − 14.4NB/3600
Note: fbb = for 0 ≤ NB ≤ 250, fbb ≥ 0.05. a
Use formula for more than 3 lanes or more than 40 buses stopping per hour.
N
Table 9-10. Adjustment Factor for Area Type (fa)

type of area area type factor, fa
CBD or similar 0.90
All other areas 1.00
Table 9-11a. Adjustment Factor for Right Turns ( fRT): Formulas

Cases 1–6: Exclusive/Shared Lanes and Protected/Permitted Phasing
fRT = 1.0 − PRT [0.15 + (PEDS/2100) (1 − PRTA)]
0.0 ≤ PRT ≤ 1.0 Proportion of RT in lane group = 1.00 for excl. RT lane (Cases 1–3);
<1.00 for shared lane (Cases 4–6).
0.0 ≤ PRTA ≤ 1.0 Proportion of RT using protected phase = 1.00 for complete protection—no peds;
< 1.00 for permitted with conflicting peds.
0 ≤ PEDS ≤ 1700 Volume (peds/hr) of peds conflicting with RT (if PEDS > 1700, use 1700).
fRT ≥ 0.05
Case 7: Single-Lane Approach (all traffic on approach in a single lane, as defined in Figure 9-5)
fRT = 0.90 − PRT [0.135 + (PEDS/2100)]
0 ≤ PRT ≤ 1.0 Proportion of RT in lane group.
0 ≤ PEDS ≤ 1700 Volume (peds/hr) of peds conflicting with RT (use 0 if RT is completely protected).
fRT = 1.00 if PRT = 0.0
fRT ≥ 0.05
range of variable values

case PRT PRTA peds simplified formula
1 Excl. RT lane; prot. RT phase 1.0 1.0 0 0.85
2 Excl. RT lane; perm. RT phase 1.0 0.0 0–1700 0.85 − (PEDS/2100)
3 Excl. RT lane; prot. + perm. RT phase 1.0 0–1.0 0–1700 0.85 − (PEDS/2100) (1 − PRTA)
4 Shared RT lane; prot. RT phase 0–1.0 1.0 0 1.0 − PRT [0.15]
5 Shared RT lane; perm. RT phase 0–1.0 0.0 0–1700 1.0 − PRT [0.15 + (PEDS/2100)]
6 Shared RT lane; prot. + perm. RT phase 0–1.0 0–1.0 0–1700 1.0 − PRT [0.15 + (PEDS/2100)(1 − PRTA)]
7 Single-lane approach 0–1.0 — 0–1700 0.9 − PRT [0.135 + (PEDS/2100)]

9-16 urban streets
Table 9-11b. Adjustment Factor for Right Turns: Factors

proportion of rt’s in lane group, PRT
cases
cases 4, 5, 6 1, 2, 3
case PRTA peds 0 .2 .4 .6 .8 1.0
2 0 0 1.00 .970 .940 .910 .880 .850
and 50 (Low) 1.00 .965 .930 .896 .861 .826
5 100 1.00 .960 .921 .881 .842 .802
200 (Mod.) 1.00 .951 .902 .853 .804 .755
400 (High) 1.00 .932 .864 .796 .728 .660
800 1.00 .894 .788 .681 .575 .469
1200 1.00 .856 .711 .567 .423 .279
≥1700 1.00 .808 .616 .424 .232 .050
.20 0 1.00 .970 .940 .910 .880 .850
50 (Low) 1.00 .966 .932 .899 .865 .831
100 1.00 .962 .925 .887 .850 .812
200 (Mod.) 1.00 .955 .910 .864 .819 .774
400 (High) 1.00 .940 .879 .819 .758 .698
800 1.00 .909 .818 .727 .636 .545
1200 1.00 .879 .757 .636 .514 .393
≥1700 1.00 .840 .681 .521 .362 .202
.40 0 1.00 .970 .940 .910 .880 .850
50 (Low) 1.00 .967 .934 .901 .869 .836
3 100 1.00 .964 .929 .893 .857 .821
and 200 (Mod.) 1.00 .959 .917 .876 .834 .793
6 400 (High) 1.00 .947 .894 .841 .789 .736
800 1.00 .924 .849 .773 .697 .621
1200 1.00 .901 .803 .704 .606 .507
≥1700 1.00 .873 .746 .619 .491 .364
.60 0 1.00 .970 .940 .910 .880 .850
50 (Low) 1.00 .968 .936 .904 .872 .840
100 1.00 .966 .932 .899 .865 .831
200 (Mod.) 1.00 .962 .925 .887 .850 .812
400 (High) 1.00 .955 .910 .864 .819 .774
800 1.00 .940 .879 .819 .758 .698
1200 1.00 .924 .849 .773 .697 .621
≥1700 1.00 .905 .810 .716 .621 .526
.80 0 1.00 .970 .940 .910 .880 .850
50 (Low) 1.00 .969 .938 .907 .876 .845
100 1.00 .968 .936 .904 .872 .840
200 (Mod.) 1.00 .966 .932 .899 .865 .831
400 (High) 1.00 .962 .925 .887 .850 .812
800 1.00 .955 .910 .864 .819 .774
1200 1.00 .947 .894 .841 .789 .736
≥1700 1.00 .938 .875 .813 .750 .688
1
and 1.00 0 1.00 .970 .940 .910 .880 .850
4
7 — 0 1.00 .873 .846 .819 .792 .765
50 (Low) 1.00 .868 .836 .805 .773 .741
100 1.00 .863 .827 .790 .754 .717
200 (Mod.) 1.00 .854 .808 .762 .716 .670
400 (High) 1.00 .835 .770 .705 .640 .575
800 1.00 .797 .694 .590 .487 .384
1200 1.00 .759 .617 .476 .335 .194
≥1700 1.00 .711 .522 .333 .144 .050

Table 9-12. Adjustment Factor for Left Turns (fLT)

case type of lane group left-turn factor, fLT
1 Exclusive LT Lane; Protected 0.95
Phasing
2 Exclusive LT Lane; Permitted Special procedure; see worksheet in Fig. 9-17 or 9-18
Phasing
3 Exclusive LT Lane; Apply Case 1 to protected phase
Protected-Plus-Permitted Phasing Apply Case 2 to permitted phase
4 Shared LT Lane; fLT = 1.0/(1.0 + 0.05 PLT)
Protected Phasing Proportion of Left Turns, PLT
0.00 0.20 0.40 0.60 0.80 1.00
Factor 1.00 0.99 0.98 0.97 0.96 0.95
5 Shared LT Lane; Permitted Special procedure; see worksheet in Fig. 9-17 or 9-18
Phasing
6 Shared LT Lane; fLT = (1,400 − vo′)/[(1,400 − vo′) + (235 + 0.435 vo′)PLT] vo′ ≤ 1,220 vph
Protected-Plus-Permitted Phasing
fLT = 1/[1 + 4.525 PLT] vo′ > 1,220 vph
where vo′ = vo /fLUo
Opposing Proportion of Left Turns, PLT
Volume
vo′ 0.00 0.20 0.40 0.60 0.80 1.00
0 1.00 0.97 0.94 0.91 0.88 0.86
200 1.00 0.95 0.90 0.86 0.82 0.78
400 1.00 0.92 0.85 0.80 0.75 0.70
600 1.00 0.88 0.79 0.72 0.66 0.61
800 1.00 0.83 0.71 0.62 0.55 0.49
1,000 1.00 0.74 0.58 0.48 0.41 0.36
1,200 1.00 0.55 0.38 0.29 0.24 0.20
≥1,220 1.00 0.52 0.36 0.27 0.22 0.18
Lane Width Adjustment Factor. The lane width adjustment per hour in parking areas directly adjacent to the lane group and
factor, fw, accounts for the deleterious impact of narrow lanes on within 250 ft upstream from the stop line. If more than 180 maneu-
saturation flow rate and allows for an increased flow on wide vers per hour exist, a practical limit of 180 should be used. If the
lanes. Twelve-foot lanes are the standard. The lane width factor parking is adjacent to an exclusive-turn-lane group, the factor only
may be calculated with caution for lane widths greater than 16 ft, applies to that lane group. On a one-way street, parking on the
or an analysis using two narrow lanes may be conducted. Note left side will affect the leftmost lane group. If parking is on both
that use of two lanes will always result in a higher saturation flow sides of a single-lane group, as in a one-way street with no exclu-
rate than a single wide lane, but in either case the analysis should sive-turn lanes, the number of maneuvers used is the total for both
reflect the way in which the width is actually used or expected to sides of the lane group. Note that parking conditions with zero
be used. In no case should the lane width factor be calculated for maneuvers are not the same as no parking.
lane widths less than 8 ft.
Bus Blockage Adjustment Factor. The bus blockage adjust-
Heavy Vehicle and Grade Adjustment Factors. The effects of ment factor, fbb, accounts for the impacts of local transit buses that
heavy vehicles and grades are treated by separate factors, fHV and stop to discharge or pick up passengers at a near-side or far-side
fg, respectively. Their separate treatment recognizes that passenger bus stop within 250 ft of the stop line (upstream or downstream).
cars are affected by approach grades, as are heavy vehicles. The This factor should only be used when stopping buses block traffic
heavy vehicle factor accounts for the additional space occupied flow in the subject lane group. If more than 250 buses per hour
by these vehicles and for the differential in the operating capabili- exist, a practical limit of 250 should be used. When local transit
ties of heavy vehicles with respect to passenger cars. The passenger buses are believed to be a major factor in intersection performance,
car equivalent (ET) used for each heavy vehicle is 2.0 passenger Chapter 12, Transit Capacity, may be consulted for a more precise
car units (pcu) and is reflected in the formula. The grade factor method of quantifying this effect. The factor used here assumes
accounts for the effect of grades on the operation of all vehicles. an average blockage time of 14.4 sec during a green indication.
Parking Adjustment Factor. The parking adjustment factor, fp, Area Type Adjustment Factor. The area type adjustment fac-
accounts for the frictional effect of a parking lane on flow in an tor, fa, accounts for the relative inefficiency of business area inter-
adjacent lane group, as well as for the occasional blocking of an sections in comparison with those in other locations, primarily
adjacent lane by vehicles moving into and out of parking spaces. because of the complexity and general congestion in the business
Each maneuver (either in or out) is assumed to block traffic in the environment.
lane next to the parking maneuver for an average of 18 sec. The Application of the area type adjustment factor reduction is typi-
number of parking maneuvers used is the number of maneuvers cally appropriate in areas that exhibit many central business district
9-18 urban streets
(CBD) characteristics. These characteristics include narrow street Item 5 should be determined by field observation, but a gross
rights-of-way; narrow sidewalks; frequent parking maneuvers; ve- estimate can be made from the signal timing by assuming that the
hicle blockages; abundant taxi or bus activity, or both; small-radius proportion of right-turning vehicles using the protected phase is
turns; limited use of exclusive-turn lanes; high pedestrian activity; approximately equal to the proportion of the turning phase that is
dense population; mid-block curb cuts; and so forth. Use of this protected. If PRTA = 1.0—that is, the right turn is completely pro-
factor should be determined on a case-by-case basis. This factor tected from conflicting pedestrians—a pedestrian volume of zero
is not limited to designated CBD areas, nor will this factor need should be used. The right-turn factor is 1.0 if the lane group does
to be used for all CBD areas. Instead, it should be used in areas not include any right turns.
where the geometric design and the traffic or pedestrian flows, or When RTOR is permitted, the right-turn volume may be reduced
both, are such that the vehicle headways are significantly increased as described in the discussion of the Volume Adjustment Module.
to the point where the capacity of the intersection is affected.
Left-Turn Adjustment Factor. The left-turn adjustment factor,
Lane Utilization Adjustment Factor. The lane utilization adjust- fLT, is based on variables similar to those for the right-turn adjust-
ment factor accounts for the unequal distribution of traffic on each ment factor, including
lane in a lane group when more than one lane exists and provides an
adjustment to the saturation flow rate to reflect the rate at which 1. Whether left turns are made from exclusive or shared lanes,
vehicles are discharged from a lane group when variations in lane 2. Type of phasing (protected, permitted, or protected plus
use exist. The adjustment factor is calculated on the basis of the flow permitted),
in the lane with the highest volume, as follows: 3. Proportion of left-turning vehicles using a shared lane
group, and
fLU = vg /(vg1N) (9-11)
4. Opposing flow rate when permitted left turns are made.
where
The left-turn adjustment factor is 1.0 if the lane group does not
fLU = lane utilization adjustment factor; include any left turns. When a left turn is not opposed at any
vg = unadjusted demand flow rate for the lane group, vph; time by through vehicles but encounters conflicting pedestrian
vg1 = unadjusted demand flow rate on the single lane in the lane movements, the left turn should be treated using the adjustment
group with the highest volume; and procedure for right turns. If no conflicting pedestrian movements
N = number of lanes in the lane group. are present, a normal protected left-turn adjustment should be
The saturation flow rate is normally adjusted for lane utilization performed.
to account for the effect of unbalanced lane usage on lane group Basically, turn factors account for the fact that these movements
delay. This adjustment can be used to account for the variation in cannot be made at the same saturation flow rates as through move-
traffic flow on the individual lanes in a lane group caused by ments. They consume more of the available green time and conse-
changes in upstream or downstream roadway characteristics such quently more of the lane group’s available capacity.
as the number of lanes available or flow characteristics such as The turn adjustment factors in Tables 9-11 and 9-12 reflect
the prepositioning of traffic within a lane group due to heavy seven different conditions under which turns may be made, as
turning movements within a short distance from an intersection. follows:
Actual lane volume distributions observed in the field, if known,
should be used in the computation of the lane utilization adjustment Case 1: Exclusive lane with protected phasing,
factor. A lane utilization factor of 1.0 may be used when uniform Case 2: Exclusive lane with permitted phasing,
traffic distribution can be assumed across all lanes in the lane group Case 3: Exclusive lane with protected-plus-permitted phasing,
or when a lane group is composed of a single lane. When average Case 4: Shared lane with protected phasing,
conditions exist or traffic distribution on a lane group is not known, Case 5: Shared lane with permitted phasing,
the default values summarized in Table 9-4 may be used. Case 6: Shared lane with protected-plus-permitted phasing,
Case 7: Single-lane approaches (right-turn factors only).
Right-Turn Adjustment Factor. Turning factors depend on a
number of parameters. The most important characteristic is the Special Procedure: Left-Turn Adjustment Factor for Permitted
manner in which turns are accommodated in the intersection. Turns Phasing
may operate out of exclusive or shared lanes, with protected or
permitted signal phasing, or with some combination of these condi- When permitted left turns exist, either from shared lanes or from
tions. The impact of turns on saturation flow rates is very much exclusive lanes, their impact on intersection operations is quite
dependent on the mode of turning operations. complicated. The procedure outlined in this section is applied to
The right-turn adjustment factor, fRT, depends on a number of Cases 2, 3, and 5 above.
variables, including Basic Case: Permitted Left Turns. The basic case for which
1. Whether the right turn is made from an exclusive or this model was developed is one in which there are simple permit-
shared lane; ted left turns from either exclusive or shared lanes. This case does
2. Type of signal phasing (protected, permitted, or protected not consider the complications of protected-plus-permitted phasing
plus permitted)—a protected right-turn phase has no conflicting nor cases in which an opposing leading phase may exist. These
pedestrian movements and a permitted phase has conflicting pedes- complications are discussed later.
trian movements; Consider Figure 9-6, which shows a permitted left turn being
3. Volume of pedestrians using the conflicting crosswalk; made from a shared lane group. When the green is initiated, the
4. Proportion of right-turning vehicles in the shared lane; and opposing queue begins to move. While the opposing queue clears,
5. Proportion of right turns using the protected portion of a left turns from the subject direction are effectively blocked. The
protected-plus-permitted phase. portion of effective green blocked by the clearance of an opposing
Basic Model for Multilane Approaches and Exclusive-Permitted

Left-Turn Lanes. On the basis of this conception of permitted left-
turn operations, the left-turn adjustment factor for the lane from
which permitted left turns are made can be stated as
1 g 2(1.0) + 3 4 1 2 31 + P (E 4
gf g q − gf g 1
fm = (0.0) + u
g g L L1 − 1)
(9-13)
1 g 2 + 1 g 2 31 + P (E 4
gf gu 1
fm = (9-13a)
L L1 − 1)
Note that there is no term in this formulation to account for
‘‘sneakers,’’ that is, vehicles completing left turns during the effec-
tive-red portion of the change-and-clearance interval. This is be-
cause in saturation flow rate measurements, vehicles are counted
Figure 9-6. Permitted left turn. (Source: W. McShane and R. P.
when they enter the intersection, not when they leave it. However,
Roess, Traffic Engineering, Prentice-Hall, Englewood Cliffs,
N.J.,1990, Fig. 21-8, p. 434.) there is a practical minimum number of left turns that will be made
on any phase, defined by sneakers.
To account for this, a practical minimum value must be imposed
queue of vehicles is designated gq. During this time, the shared on fm. One sneaker per cycle may be assumed as a minimum. The
lane from which subject left turns are made is blocked when a left- probability that a second sneaker will be in position at the end of
turning vehicle arrives. Until the first left-turning vehicle arrives, the green phase will be equal to the proportion of left turns in the
however, the shared lane is unaffected by left-turners. The portion shared lane, PL. The estimated number of sneakers per cycle may
of effective green until the arrival of the first left-turning vehicle therefore be computed as (1 + PL). Assuming an approximate
is designated gf. average headway of 2 sec per vehicle in an exclusive lane on a
Once the opposing queue of vehicles clears, subject left-turning protected phase, the practical minimum value of fm may be esti-
vehicles filter through an unsaturated opposing flow at a rate af- mated as 2(1 + PL)/g.
fected by the magnitude of the opposing flow. The portion of the For multilane groups, the impact of left turns on a shared
effective green during which left turns filter through the opposing lane must be extended to include their impact on the entire lane
flow is designated gu. group. One might simply assume that the factor for the shared
This portioning of the effective green phase for permitted left lane is fm and that the factor for each other lane in the group
turns creates up to three distinct periods for which the impact is 1.0. This assumes, however, that left turns affect only the
of left turns on a shared or exclusive left-turn lane must be lane from which they are made. This is an incorrect assumption,
considered: because vehicles maneuver from lane to lane to avoid left-
T gf: Until the arrival of the first left-turning vehicle, a shared turn congestion. Regression studies suggest that the following
lane is unaffected by left turns. During this period of time, the relationship is more realistic:
effective left-turn adjustment factor is logically 1.0, because no fLT = [fm + 0.91(N − 1)]/N (9-14)
left turns are present. By definition, gf = 0.0 sec for exclusive-
permitted left-turn lanes, because it is assumed that a queue of where
left-turners is present at the beginning of the phase. fLT = left-turn adjustment factor applied to a total lane group from
T gq − gf: If the first left-turning vehicle arrives before the oppos- which left turns are made, and
ing queue clears, it waits until the opposing queue clears, blocking fm = left-turn adjustment factor applied only to the lane from
the shared lane, and then seeks a gap in the unsaturated opposing which left turns are made.
flow that follows. During this period of time, there is effectively
no movement in the shared lane, and the left-turn adjustment factor When a single (or double) exclusive-permitted left-turn lane is
(fLT) applied to the shared lane is logically 0.0. involved, fLT = fm.
When the first left-turning vehicle arrives after the opposing To implement this model, it is necessary to estimate the subpor-
queue clears, this period of time does not exist; that is, gq − gf has tions of the effective green phase, gf, gq, and gu. Regression rela-
a practical minimum value of zero. The value of gq has a practical tionships have been developed to permit this, as follows:
range of 0.0 to g. 1. Compute gf:
T gu: After the opposing queue clears, left-turning vehicles
select gaps through the unsaturated opposing flow. This occurs gf = G exp (−0.882LTC 0.717) − tL (9-15)
at a reduced rate because of the interference of opposing vehicles (shared-permitted left-turn lanes)
and the effect this has on other vehicles in the shared lane from gf = 0.0 (exclusive-permitted left-turn lanes) 0 ≤ gf ≤ g
which left turns are made. During this period, Figure 9-7 assigns
where
EL1 through-car equivalents for each left-turning vehicle. From
this, an adjustment factor can be computed for this period: G = actual green time for the permitted phase, sec;
LTC = left turns per cycle, vpc, computed as vLTC/3600;
1/[1.0 + PL (ELl − 1)] (9-12)
vLT = adjusted left-turn flow rate, vph;
where PL is the proportion of left-turning vehicles in the shared C = cycle length, sec; and
lane. For exclusive-permitted left-turn lanes, PL = 1.0. tL = lost time for subject left-turn lane group, sec.
9-20 urban streets
a
Use formula for more than 1,200 effective opposing flow; vo must be greater than zero.
EL1 = sTH/sLT (exclusive)
EL1 = sTH/sLT − 1 (shared)
sLT = [vo′ exp (−vo′tc / 3,600]/[(1 − exp (−vo′tf / 3,600)]
where
EL1 = through-car equivalent for permitted left turns;
sTH = saturation flow of through traffic, vphgpl = 1900 vphgpl;
sLT = filter saturation flow of permitted left turns, vphgpl;
tc = critical gap, sec = 4.5 sec; and
tf = follow-up headway, sec = 2.5 sec (exclusive), 4.5 sec (shared).
Figure 9-7. Through-car equivalents, EL1, for permitted left turns (1).
2. Compute gq: where

volc qro PLT = proportion of left turns in the lane group, and
gq = − tL (9-16)
0.5 − [volc (1 − qro)/go] N = number of lanes in the lane group.
volc (1 − qro)/go ≤ 0.49 Note: When an exclusive-permitted left-turn lane is involved,
0.0 ≤ gq ≤ g PL = PLT = 1.0.
where 6. Compute fm using Equation 9-13.
volc = adjusted opposing flow rate per lane per cycle, computed 7. Compute fLT using Equation 9-14.
as voC/(3600No fLUo), vplpc;
vo = adjusted opposing flow rate, vph; Basic Model for Single-Lane Approaches Opposed by Single-
fLUo = lane utilization adjustment factor for opposing flow Lane Approaches. The case of a single-lane approach opposed by
No = number of opposing lanes; another single-lane approach has a number of unique features that
qro = opposing queue ratio, that is, the proportion of opposing must be reflected in the model. The most critical of these is the
flow rate originating in opposing queues, computed as effect of opposing left turns. An opposing left-turning vehicle in
1 − Rpo(go /C), qro ≥ 0; effect creates a gap in the opposing flow through which a subject
Rpo = platoon ratio for the opposing flow, obtained from Table left turn may be made. This can occur during the clearance of the
9-2 on the basis of opposing arrival type; opposing queue as well as during the unsaturated portion of the
go = effective green for the opposing flow, sec; and green phase.
tL = lost time for opposing lane group. Thus, the assumption in the multilane model that there is no
flow during the period gq − gf (where gq > gf) is not applicable to
3. Compute gu: opposing single-lane approaches, on which there is flow during
gu = g − g q when gq ≥ gf this period at a reduced rate reflecting the blocking effect of left-
gu = g − g f when gq < gf turning vehicles as they await an opposing left turn. Left-turning
vehicles during the period gq − gf are assigned a ‘‘through-car
where g = effective green time for subject permitted left turn, sec. equivalent’’ value, EL2, based upon simple queueing analysis,
Note: When gq < gf, that is, when the first left-turning vehicle which can be converted to an adjustment factor for application
does not arrive until after the opposing queue clears, an effective during this period of the green.
adjustment factor of 1.0 is applied throughout gf and a factor based Since vehicles do not have the flexibility to choose lanes on a
upon EL1 thereafter. single-lane approach, regression relationships for predicting gf and
4. Select the appropriate value of EL1 from Figure 9-7 on the gq are also different from those for the multilane case. Further, for
basis of the opposing flow rate, vo, and the lane utilization adjust- a single-lane approach, fLT = fm, and PL = PLT. As in the multilane
ment factor of the opposing flow, fLUo. For the purposes of de- case, the opposing single-lane model has no term to account for
termining vo, opposing right and left turns from exclusive lanes sneakers but has a practical minimum value of fLT = 2(1 + PLT)/g.
are not included in vo. The basic model for opposing single-lane approaches is
5. Compute PL (proportion of left turns in shared lane): therefore
3 24
12 1 23 4
gf g 1
1
fLT = fm = (1.0) + diff
(N − 1) g g g 1 + PLT (EL2 − 1)
PL = PLT 1 + (9-17)
1 23 4
gu g 1
gf + + 4.24 + u (9-18)
EL1 g 1 + PLT (EL1 − 1)
1 g 2 + 1 g 2 31 + P 4 1 2 31 + P 4
gf gdiff 1 gu 1 Special Cases for Permitted Left Turns. Two special cases for
fLT = +
LT (EL2 − 1) g (EL1 − 1)
LT
fully permitted left turns must be addressed: a single-lane approach
(9-18a) opposed by a multilane approach, and vice versa.
where gdiff = max (gq − gf, 0). Note that when no opposing left When the subject lane in these cases is the single-lane approach,
turns are present, the value of gdiff is to be set to zero. it is opposed by a multilane opposing flow. Even if the opposing
To implement this model, it is again necessary to estimate the approach is a single through lane and an exclusive left-turn lane,
subportions of the effective green phase, gf, gq, and gu, as follows: opposing left turns will not open gaps in the opposing flow. Thus,
the special structure of the single-lane model does not apply. The
1. Compute gf : multilane model is applied, except that fLT = fm. The value of gf,
gf = G exp (−0.860LTC 0.629) − tL, 0 ≤ gf ≤ g (9-19) however, should be computed using the single-lane equation, gf =
G exp(−0.860LTC 0.629) − tL.
where When the multilane approach is considered, the reverse is true.
G = actual green time for the permitted phase, sec; The opposing flow is in a single lane, and opposing left turns
LTC = left turns per cycle, vpc, computed as vLTC/3600; could conceivably open gaps for subject left-turners. The single-
vLT = adjusted left-turn flow rate, vph; lane model may be applied, with several notable revisions:
C = cycle length, sec; and
tL = lost time for subject left-turn lane group, sec. T gf should be computed using the multilane equation:
2. Compute gq : gf = G exp(−0.882LTC 0.717) − tL
gq = 4.943vo/c0.762 qro1.061 − tL, 0.0 ≤ gq ≤ g (9-20) T PL must be estimated and substituted for PLT in the single-
lane model. PL may be estimated from PLT using the multilane
where equation:
31 24
volc = adjusted opposing flow rate per lane per cycle, computed
as voC/(3600fLUo) vplpc; (N − 1) g
PL = PLT 1 + (9-17)
1 2
vo = adjusted opposing flow rate, vph; g
gf + u + 4.24
fLUo = lane utilization adjustment factor for opposing flow EL1
qro = opposing queue ratio, that is, the proportion of opposing
flow rate originating in opposing queues, computed as T fLT does not equal fm. Thus, the conversion must be made using
1 − Rpo(go /C), qro ≥ 0; the multilane equation, except when the subject approach is a dual
Rpo = platoon ratio for the opposing flow, obtained from Table left-turn lane.
9-2 on the basis of opposing arrival type; and
fLT = [fm + 0.91 (N − 1)]/N
go = effective green for the opposing flow, sec.
tL = lost time for opposing lane group Worksheets that may be used to assist in implementing the
special models for permitted left-turn movements are presented in
3. Compute gu:
Section III of this chapter. These worksheets do not account for the
g u = g − gq when gq ≥ gf modifications that must be made to analyze single-lane approaches
g u = g − gf when gq < gf opposed by multilane approaches, and vice versa.
where g = effective green time for subject permitted left turn, sec. More Complex Phasing with Permitted Left Turns. The models
Note: When gq < gf, that is, when the first left-turning vehicle and worksheets presented in the previous section apply directly to
does not arrive until after the opposing queue clears, an effective situations in which left turns are made only on permitted phases
adjustment factor of 1.0 is applied throughout gf and a factor based (without protection) and in which no protected phases or opposing
upon EL1 thereafter. leading green phases exist. The models may, however, be applied
to these more complex cases with some modifications.
4. Select the appropriate value of EL1 from Figure 9-7 on the In general, protected-plus-permitted phases for exclusive lanes
basis of the opposing flow rate, vo, and the lane utilization adjust- are analyzed by separating the portions of the phase into two lane
ment factor of the opposing flow, fLUo. groups for the sake of analysis. Each portion of the phase is then
handled as it would be normally if the other were not present. The
5. Compute EL2: protected portion of the phase is treated as a protected phase, and
EL2 = (1 − PnTHo)/PLTo, EL2 ≥ 1.0 (9-21) a left-turn adjustment factor appropriate to a protected phase is
selected. The permitted portion of the phase is treated as a permit-
where ted phase, and the special procedures outlined here are used to
estimate a left-turn adjustment factor (with modifications as de-
PLTo = proportion of left turns in opposing single-lane approach; fined in this section).
PTHo = proportion of through and right-turning vehicles in oppos- By doing this, separate saturation flow rates may be computed
ing single-lane approach, computed as 1 − PLTo; and for each portion of the phase. A method for estimating delays
in such cases is described later in this chapter. This method
n = maximum number of opposing vehicles that could arrive
does not require that the demand volume for the protected-plus-
during gq − gf, computed as (gq − gf)/2. Note that n is
permitted movement be divided between the two portions of
subject to a minimum value of zero.
the phase. However, the computation of the critical v/c ratio,
6. Compute fLT using Equation 9-18. Xc, does require this apportionment. The following is a reasonable
9-22 urban streets
and conservative approach to apportioning the volumes for For the NB phase, gq is referenced to the beginning of the
purposes of computing Xc: opposing (SB) effective green. Again, the value needed is the
portion of the NB g* blocked by the clearance of the opposing
T The first portion of the phase, whether protected or permitted,
queue. Because the NB effective green (g*) does not account for
is assumed to be fully utilized, that is, to have a v/c of 1.0, unless
lost time, gq* = gq + tL. For the SB phase, the usual computation
total demand is insufficient to use the capacity of that portion of
of gq is indexed to the start of the opposing (NB) flow, which
the phase.
begins in the leading phase. For analysis of the permitted phase,
T Any remaining demand not handled by the first portion of
however, only the portion that blocks the SB permitted effective
the phase is assigned to the second portion of the phase, whether
green is of interest. Thus, gq* = gq − G1 − G/Y1.
protected or permitted.
The foregoing discussion is illustrative. The relationship be-
This approach assumes that when the movement is initiated, a tween the normal calculations of g, G, gf, and gq and their adjusted
queue is available that uses available capacity in the initial portion counterparts, g*, G*, gf*, and gq*, is best illustrated by Figure 9-
of the phase. In cases of a failed cycle, the unserved queue will 8, which may be used in conjunction with the standard worksheets
exist after the end of the second portion of the phase, with those to arrive at the appropriate left-turn adjustment factor for the per-
vehicles queued and ready to use the initial portion of the phase mitted portion of a protected-plus-permitted phase plan. Obviously,
on the next cycle. In this sense, the initial portion of the movement ‘‘north’’ and ‘‘south’’ can be reversed or replaced by ‘‘east’’ and
can never operate at a v/c of more than 1.0. ‘‘west’’ without any change in the equation shown.
In the analysis of the permitted portion of such phases, as well
as those with opposing leading protected left-turn phases, the basic
Capacity Analysis Module
models described previously may be applied. The difficulty is in
selecting values of G, g, gf, gq, and gu for use in these models. The
In the Capacity Analysis Module, computational results of previ-
equation for gf is indexed to the beginning of effective green in
ous modules are manipulated to compute key capacity variables,
the subject direction, and gq is indexed to the beginning of the
including
effective green for the opposing flow. When leading or lagging
phasing or protected-plus-permitted phasing exists, these equations 1. Flow ratio for each lane group.
must be modified to account for shifts in the initiation and overlap 2. Capacity of each lane group.
of various green times. 3. Volume-to-capacity ratio of each lane group, and
Some common examples are shown in Figure 9-8. The following 4. Critical v/c ratio for the overall intersection.
notation is used: G, g, gf, and gq are computed as shown in the
models and worksheets. These values are modified as shown and Flow ratios are computed by dividing the adjusted demand flow,
replaced on the worksheets with G*, g*, gf*, and gq* for the permit- v, computed in the Volume Adjustment Module by the adjusted
ted portion of protected-plus-permitted phasing. This extended no- saturation flow rate, s, computed in the Saturation Flow Rate
tation is required to cover the general case of complex left-turn Module.
phasing. In most practical cases, it will not be necessary to use all The capacity of each lane group is computed from Equation 9-3:
the superscripted terms. ci = si(gi /C)
The standard case is shown in Figure 9-8(a) as a starting point.
Case 2 is a leading green phase. The equations shown are valid If the signal timing is not known, a timing plan will have to be
for either exclusive-lane or shared-lane operation, except that gf is estimated or assumed to make these computations. Appendix II
zero by definition for the exclusive-lane case. For exclusive-lane contains suggestions for making these estimates, but state or local
operation, the leading green, G1, is followed by G/Y1, a period policies and guidelines should also be consulted whenever possi-
during which the left-turn change-and-clearance interval is dis- ble. The planning method described later also offers a procedure
played, and the through movement continues with a green indica- for the synthesis of timing plans based on the concepts presented
tion. G2 has a green indication for both the through and left-turn in Appendix II.
movements, followed by a full change-and-clearance interval for The v/c ratio for each lane group is computed directly by divid-
all north-south movements, Y2. ing the adjusted flows by the capacities computed above, as in
The effective green time for the permitted phase, g*, is equal Equation 9-4:
to G2 + Y2 for the NB direction and G2 + Y2 − tL for the SB direction. Xi = vi /ci
Note that there is no lost time for NB movements, since both were
initiated in the leading phase, and the lost time is assessed there. The final capacity parameter of interest is the critical v/c ratio,
Thus, the NB and SB effective green times that must be used are Xc, for the intersection. It is computed from Equation 9-5 as
not equal. follows:
For the NB phase, gf is computed using the total green time for
Xc = ∑(v/s)ci C/(C − L)
NB left-turn movement, G1 + G/Y1 + G2. The computed value,
however, begins with the leading-phase effective green, as shown. This ratio indicates the proportion of available capacity that
The value that needs to be applied to the permitted phase, however, could be utilized by vehicles in critical lane groups. If this ratio
is that portion of gf that overlaps g*, which results in gf* = gf − exceeds 1.0, one or more of the critical lane groups will be over-
G1 − G/Y1 + tL. This computation would be done for a shared lane, saturated. A ratio over 1.0 is an indication that the intersection
and the result, gf*, would have to be a value between 0 and g*. design, cycle length, or phase plan is inadequate, or all three are
For an exclusive-lane case, gf and gf* are by definition zero. For inadequate, for the given demand. A ratio of less than 1.0 indicates
the SB phase, gf as normally computed is the same as gf*, and no that the design, cycle length, and phase plan are adequate to handle
adjustment is necessary. all critical flows without having demand exceed capacity, assum-
Figure 9-8. Green time adjustments for protected-plus-permitted phasing: (a) standard case and Case 2, (b) Cases 3 and 4,
and (c) Case 5. (Continued on next page.)
ing that green times are proportionally assigned. When phase splits ward: in each discrete phase, the lane group with the highest v/s
are not proportional to the v/s ratios, some movement demands ratio is critical.
may exceed movement capacities even where the critical v/c ratio Thus, when phases do not overlap,
is less than 1.0.
1. There is one critical lane group for each signal phase,
The computation of the critical v/c ratio, Xc, requires that critical
2. In each phase, the critical lane group is the one with the
lane groups be identified. During each signal phase, one or more
highest v/s ratio among the lane groups moving in that phase, and
lane groups are given the green. One lane group will have the most
3. The critical lane group v/s ratios are summed for use in
intense demand and will be the one that determines the amount of
computing Xc.
green time needed. This lane group would be the critical lane
group for the phase in question. The critical lane group for each Overlapping phases are more difficult to analyze, because vari-
signal phase in effect controls the required signal timing, or, given ous lane groups may move in several phases of the signal, and
the signal timing, the critical lane group is the one most constrained some left-turn movements may operate on a protected-and-permit-
by it. ted basis in various portions of the cycle. In such cases, it is
The normalized measure of demand intensity on any lane group necessary to find the critical path through the signal cycle. The
is given by the v/s ratio for the lane group. When there are no path having the highest sum of v/s ratios is the critical path.
overlapping phases in the signal design, such as in a simple two- When phases overlap, the critical path must conform to the
phase signal, the determination of critical lane groups is straightfor- following rules:
9-24 urban streets
Figure 9-8 (continued).
1. Excluding lost times, one critical lane group must be moving protected-and-permitted movements from a shared lane under left-
at all times during the signal cycle, turn Case 6). In such cases, the user may need to allocate volumes
2. At no time in the signal cycle may more than one critical in the most logical manner possible or simply omit the critical
lane group be moving, and v/c determination from the analysis.
3. The critical path has the highest sum of v/s ratios.
These rules are more easily explained by example. Consider the
In some complex phasing situations, it may not be possible to case of a leading and lagging green phase plan on an arterial with
identify critical movements using the guidelines stated above (e.g., exclusive left-turn lanes, as shown in Figure 9-9.
Figure 9-8 (continued).
Figure 9-9. Critical lane group determination: leading and lagging green phase plan with exclusive left-turn lanes.
9-26 urban streets
Phase 1 is discrete, with NB and SB lane groups moving

simultaneously. The critical lane group for Phase 1 would
therefore be chosen on the basis of the highest v/s ratio. As
the v/s ratio for the NB lane group is 0.30 and that for the SB
lane group is 0.25, the NB lane group is the critical lane group
for this phase.
Phase 2 involves overlapping leading and lagging green phases.
There are two possible paths through Phase 2 that conform to the
rule stated above, that is, that (except for lost times) there must
be only one critical lane group moving at all times. The EB through
and right-turn (T/R) lane group moves through Phases 2A and 2B
with a v/s ratio of 0.30. The WB left-turn lane group moves only
in Phase 2C with a v/s ratio of 0.15. The total v/s ratio for this
path is therefore 0.30 + 0.15 or 0.45. The only alternative path
involves the EB left-turn lane group, which moves only in Phase
2A (v/s = 0.25), and the WB T/R lane group, which moves in
Phases 2B and 2C (v/s = 0.25). Because the sum of the v/s ratios
for this path is 0.25 + 0.25 = 0.50, which is higher than the v/s
ratio for the alternative, this is the critical path through Phase 2.
Thus, the sum of critical v/s ratios for the cycle is 0.30 for Phase
1 plus 0.50 for Phase 2, for a total of 0.80.
The solution for Xc also requires that the lost time for the critical Figure 9-10. Critical lane group determination: leading and lag-
ging green phase plan with addition of permitted left turn in
path (L) through the signal be determined. Using the general rule
Phase 2B.
that a movement’s lost time tL is applied when a movement is
initiated, the following conclusions are reached:
the option without permitted left turns in Phase 2B, which is an
T The critical NB movement is initiated in Phase 1, and its lost expected result.
time is applied. Again, the lost time for the critical path is determined as follows:
T The critical EB left-turn movement is initiated in Phase 2A,
and its lost time is applied. T The NB critical flow begins in Phase 1, and its lost time is
T The critical WB T/L movement is initiated in Phase 2B, and applied.
its lost time is applied. T The critical EB left turn (protected) is initiated in Phase 2A,
T No critical movement is initiated in Phase 2C. Therefore, no and its lost time is applied.
lost time is applied to the critical path here. Although the WB left- T The critical WB left turn (permitted) is initiated in Phase 2B,
turn movement is initiated in this phase, it is not a critical move- and its lost time is applied.
ment, and its lost time is not included in L. T The critical WB left turn (protected) is a continuation of the
T Therefore, for this case, L = 3tL, assuming that each movement WB left turn (permitted). Because the left-turn movement is al-
has the same lost time, tL. ready moving when Phase 2C is initiated, no lost time is applied
here.
This problem may be altered significantly by adding a permitted T Thus, for this case, L = 3tL, assuming that each movement
left turn in both directions to Phase 2B. This is shown in Figure has the same lost time, tL.
9-10, with the v/s ratios resulting. Note that in this case, a separate
This is the same result obtained previously.
v/s ratio is computed for the protected and permitted portions of
Figure 9-11 shows another complex case with actuated control
the EB and WB left-turn movements. In essence, the protected
and a typical eight-phase plan. Although eight phases are provided
and permitted portions of these movements are treated as separate
on the controller, the path through the cycle cannot include more
lane groups.
than six of these phases, as shown. The leading phases (1B and
The analysis of Phase 1 does not change, because it is discrete.
2B) will be chosen on the basis of which left-turn movements
The NB lane group is still critical, with a v/s ratio of 0.30. There
have higher demands on a cycle-by-cycle basis.
are now, however, four different potential paths through Phase 2
The possible critical paths through Phase 1 are as follows:
that conform to the rules for determining critical paths:
EB left turn (protected) + EB left turn (permitted),
WB T/R + EB left turn (protected) = 0.25 + 0.20 = 0.45. EB left turn (protected) + WB left turn (permitted),
EB T/R + WB left turn (protected) = 0.30 + 0.05 = 0.35. EB left turn (protected) + WB T/R,
EB left turn (protected) + EB left turn (permitted) + WB left WB left turn (protected) + WB left turn (permitted),
turn (protected) = 0.20 + 0.15 + 0.05 = 0.40. WB left turn (protected) + EB left turn (permitted), and
EB left turn (protected) + WB left turn (permitted) + WB left WB left turn (protected) + EB T/R.
turn (protected) = 0.20 + 0.22 + 0.05 = 0.47.
Again, the combination with the highest v/s ratio would be chosen
The critical path through Phase 2 is the one with the highest as the critical path. A similar set of choices exists for Phase 2,
total v/s ratio. This is the last choice, and yields a v/s ratio of 0.47, with NB replacing EB and SB replacing WB.
which when added to the 0.30 for Phase 1 results in a sum of The most interesting aspect of this problem is the number of
critical v/s ratio of 0.77. Note that this is a smaller total than for lost times that must be included in L for each of these paths. The
Figure 9-11. Critical lane group determination: complex multiphase signal.
paths involving EB left turn (protected) + EB left turn (permitted) to the control delay value, as shown in Table 9-1. The values
and WB left turn (protected) + WB left turn (permitted) involve derived from the formulas represent the average control delay
only one application of tL each, because the turning movement in experienced by all vehicles that arrive in the analysis period, in-
question moves continuously throughout the three subphases. All cluding delays incurred beyond the analysis period when the lane
other paths involve two applications of tL, because each critical group is oversaturated.
movement is initiated in a portion of the phase. Note that the left The average control delay per vehicle for a given lane group is
turn that does not continue in Phase 1B or 2B is a discontinuous
d = d1PF + d2 + d3 (9-22)
movement; that is, it moves as a protected turn in Phase 1A or
2A, stops in Phase 1B or 2B, and moves again as a permitted turn where
in Phase 1C or 2C.
d1 = uniform control delay component assuming uniform arriv-
Thus, for this complex phasing, the lost time through each major
als, sec/veh;
phase could have one or two lost times applied, on the basis of
PF = uniform delay progression adjustment factor that accounts
the critical path. Therefore, for the total cycle, two to four lost
for the effects of signal progression on delay;
times will be applied, again depending on the critical path.
d2 = incremental delay component to account for the effect of
In general terms, up to n lost times are to be applied in the
random and oversaturation queues, adjusted for the duration
calculation of the total lost time per cycle, where n is the number
of the analysis period and the type of signal control; this
of movements in the critical path through the signal cycle. For the
delay component assumes that there is no residual demand
purposes of determining n, a protected-plus-permitted movement
for the lane group at the start of the analysis period, sec/
is considered to be one movement if the protected and permitted
veh; and
phases are adjacent.
d3 = residual demand delay to account for oversaturation queues
that may have existed before the analysis period, sec/veh;
LOS Module this component is detailed in Appendix VI.
Uniform Delay, d1
In the LOS Module, the average control delay per vehicle is
estimated for each lane group and averaged for all approaches and Equation 9-23 gives an estimate of delay assuming perfectly
for the intersection as a whole. Level of service is directly related uniform arrivals and stable flow. It is based on the first term of
9-28 urban streets
Webster’s delay formulation and is widely accepted as an accurate Rpgi /C. Arrival Type 3 should be assumed for all uncoordinated
depiction of delay for the idealized case of uniform arrivals. Note lane groups.
that values of X beyond 1.0 are not used in the computation of d1. Movements made from exclusive left-turn lanes on protected
phases are not usually provided with good progression. Thus, Ar-
0.50C(1 − g/C)2 rival Type 3 is usually assumed for coordinated left turns. When
d1 = (9-23)
1 − Min(1,X)g/C the actual arrival type is known, it should be used. When the
coordinated left turn is part of a protected-permitted phasing, only
where the effective green for the protected phase should be used to deter-
C = cycle length, sec [cycle length used in pretimed signal mine PF since the protected phase is normally the phase associated
control, or average cycle length for actuated control (see with platooned coordination. When a lane group contains move-
Appendix II for signal timing estimation of actuated ments that have different levels of coordination, a flow-weighted
control parameters)]; average of P should be used in determining the PF.
g = effective green time for lane group, sec [green time used in
pretimed signal control, or average green time for actuated Incremental Delay d2
control (see Appendix II for signal timing estimation of Equation 9-25 estimates the incremental delay due to nonuni-
actuated control parameters)]; and form arrivals and temporary cycle failures (random delay) as well
X = v/c ratio or degree of saturation for lane group. as that caused by sustained periods of oversaturation (oversatura-
Progression Adjustment Factor, PF. Good signal progression tion delay). It is sensitive to the degree of saturation of the lane
will result in a high proportion of vehicles arriving on the green. group (X), the duration of the analysis period of interest (T), the
Poor signal progression will have a low percentage of vehicles capacity of the lane group (c), and the type of signal control as
arriving on the green. The progression adjustment factor, PF, ap- reflected by the control parameter (k). The formula assumes that
plies to all coordinated lane groups, including both pretimed con- there is no unmet demand causing residual queues at the start of
trol and nonactuated lane groups in semiactuated control systems. the analysis period (T). Should that not be the case, the reader may
In circumstances where coordinated control is explicitly provided consult Appendix VI for additional procedures that can account for
for actuated lane groups, PF may also be applied to these lane the effect of a nonzero initial queue on signal delay. Finally, the
groups. Progression primarily affects uniform delay, and for this incremental delay term is valid for all values of X, including highly
reason, the adjustment is applied only to d1. The value of PF may oversaturated lane groups. The expression for d2 is
be determined by
3 !(X − 1) + 4
2
8kIX
d2 = 900T (X − 1) + (9-25)
(1 − P)fP cT
PF = (9-24)
1 − (g/C) where
where T = duration of analysis period, hours;
k = incremental delay factor that is dependent on controller
P = proportion of vehicles arriving on the green,
settings;
g/C = proportion of green time available, and
I = upstream filtering/metering adjustment factor;
fP = supplemental adjustment factor for when the platoon
c = lane group capacity, vph; and
arrives during the green.
X = lane group v/c ratio, or degree of saturation.
The default values for fP are 0.93 for Arrival Type 2, 1.15 for
Arrival Type 4, and 1.0 for all other arrival types. Incremental Delay Calibration Term (k)
The value of P may be measured in the field or estimated from
the arrival type. If field measurements are carried out, P should The calibration term (k) is included in Equation 9-25 to incorpo-
be determined as the proportion of vehicles in the cycle that arrive rate the effect of controller type on delay. For pretimed signals, a
at the stop line or join the queue (stationary or moving) while the value of k = 0.50 is used throughout. This value is based on a
green phase is displayed. PF may be computed from measured queueing process with random arrivals and uniform service time
values of P using the default values for fP. Alternatively, Table 9- equivalent to the lane group capacity. Actuated controllers, on the
13 may be used to determine PF as a function of the arrival type other hand, have the ability to tailor the green time to the cyclic
based on the default values for P (i.e., Rpgi /C) and fP associated demand, thus reducing the overall incremental delay component.
with each arrival type. If PF is estimated by Equation 9-24, its The delay reduction depends in part on the controller’s unit exten-
calculated value may exceed 1.0 for Arrival Type 4 with extremely sion and the prevailing v/c ratio. Recent research indicates that
low values of g/C. As a practical matter, PF should be assigned lower unit extensions result in lower values of k and d2. However,
a maximum value of 1.0 for Arrival Type 4. This has already been when v/c approaches 1.0, an actuated controller will behave in a
taken into consideration in Table 9-13. similar manner to a pretimed controller at the maximum settings.
Application of the adjustment factor for progression requires Thus, the k parameter will converge to the pretimed value of 0.50
detailed knowledge of offsets, travel speeds, and intersection sig- at X ≥ 1.0. The recommended k values for pretimed and actuated
nalization. When delay for future situations involving coordination lane groups are given in Table 9-14.
is estimated, particularly when alternatives are analyzed, it is advis- For unit extension values other than those listed in Table 9-
able to assume Arrival Type 4 as a base condition for coordinated 14, k values may be interpolated. If the formula in Table 9-14
lane groups (except left turns), in which case P may be estimated is used the kmin value (the k value for X = 0.5) should first be
using the Rp default values from Table 9-2 and Equation 9-7 as interpolated for the given unit extension and then the formula
Table 9-13. Progression Adjustment Factor (PF)

progression adjustment factor (PF)
PF = (1 − P)fP/(1 − g/C) (see Note)
arrival type (AT)
green ratio (g/C) AT-1 AT-2 AT-3 AT-4 AT-5 AT-6
3
0.20 1.167 1.007 1.000 1.000 0.833 0.750
0.30 1.286 1.063 1.000 0.986 0.714 0.571
0.40 1.445 1.136 1.000 0.895 0.555 0.333
0.50 1.667 1.240 1.000 0.767 0.333 0.000
0.60 2.001 1.395 1.000 0.576 0.000 0.000
0.70 2.556 1.653 1.000 0.256 0.000 0.000
Default, fP 1.00 0.93 1.00 1.15 1.00 1.00
Default, RP 0.333 0.667 1.000 1.333 1.667 2.000
Note: 1. Tabulation is based on default values of fP and RP.
2. P = RPg/C (may not exceed 1.0).
3. PF may not exceed 1.0 for AT-3 through AT-6.
Table 9-14. Recommended k Values for Lane Groups Under Actuated and Pretimed Control
UNIT degree of saturation (X)
EXTENSION
(sec) ≤0.50 0.60 0.70 0.80 0.90 ≥1.0
≤2.0 0.04 0.13 0.22 0.32 0.41 0.50
2.5 0.08 0.16 0.25 0.33 0.42 0.50
3.0 0.11 0.19 0.27 0.34 0.42 0.50
3.5 0.13 0.20 0.28 0.35 0.43 0.50
4.0 0.15 0.22 0.29 0.36 0.43 0.50
4.5 0.19 0.25 0.31 0.38 0.44 0.50
5.01 0.23 0.28 0.34 0.39 0.45 0.50
Pretimed or 0.50 0.50 0.50 0.50 0.50 0.50
nonactuated
movement
Note: For a given UE and its kmin value at X = 0.5: k = (1 − 2kmin)(X − 0.5) + kmin, k ≥ kmin, k ≤ 0.5.
1
For UE > 5.0, extrapolate to find k, keeping k ≤ 0.5.
should be used. Table 9-14 may be extrapolated for unit extension proach and for the intersection as a whole. In general, this is done
values beyond 5.0 sec, but in no case should the extrapolated by computing weighted averages, where the lane group delays are
k value exceed 0.5. weighted by the adjusted flows in the lane groups.
Thus, the delay for an approach is computed as
Upstream Filtering/Metering Adjustment Factor, I
dA = o dv
i i
(9-26)
The incremental delay adjustment factor, I, incorporates the effects
of metering arrivals from upstream signals, as described in Chapter ov i
11. In an isolated signal analysis in this chapter, an I value of 1.0 where

is used.
dA = delay for approach A, sec/veh;
Residual Demand Delay, d3 di = delay for lane group i (on approach A), sec/veh; and
vi = adjusted flow for lane group i, vph.
When a residual demand from a previous time period causes a
residual queue to occur at the start of the analysis period (T), addi- Approach control delays can then be further averaged to provide
tional delay is experienced by the vehicles arriving in the period, the average delay for the intersection:
since the residual queues must first clear the intersection. A proce-
dure for determining this supplemental delay is described in detail dI = odv
A A
(9-27)
in Appendix VI. If this is not the case, a d3 value of zero is used. ov A
This procedure is also extended to analyze delay over multiple time
where
periods, each having a duration (T) in which a residual demand may
be carried from one time period to the next. dl = average delay per vehicle for the intersection, sec/veh, and
vA = adjusted flow for approach A, vph.
Aggregating Delay Estimates
LOS Determination
The procedure for delay estimation yields the average control
delay per vehicle for each lane group. It is also desirable to aggre- Intersection level of service is directly related to the average
gate these values to provide average delay for an intersection ap- control delay per vehicle. Once delays have been estimated for
9-30 urban streets
each lane group and aggregated for each approach and the intersec- ratio is not greater than 1.0. This adjustment is also necessary for
tion as a whole, Table 9-1 is consulted, and the appropriate levels the analysis of simple protected operation as described previously.
of service are determined for each component. If the v/c ratio is greater than 1.0, the area contained by the polygon
will not be defined. The effect of v/c ratios greater than 1.0 is
Special Procedure for Uniform Delay with expressed by the second term of the delay equation.
Protected-Plus-Permitted Operation It is first necessary to distinguish between protected-plus-permit-
ted (leading left-turn) phasing and permitted-plus-protected (lag-
The delay estimation model just presented is based on a well- ging left-turn) phasing. Three of the five cases shown in Figure
established formulation originally proposed by Webster and subse- 9-12 are associated with leading left-turn phases and the other two
quently modified by others. The delay per vehicle is expressed as are associated with lagging left-turn phases. The five cases are
the sum of two terms. The first term represents the delay that would identified as follows:
result from completely uniform arrival of vehicles over the signal
cycle. The second term recognizes the tendency for occasional Case 1—leading left-turn phase: no queue remains at the end
overflow to occur as a result of random arrivals. of the protected or permitted phase.
The first term is easily derived as a function of the area contained Case 2—leading left-turn phase: a queue remains at the end of
within the plot of queue storage as a function of time. With a the protected phase but not at the end of the permitted phase.
single green phase per cycle, this plot assumes a triangular shape; Case 3—leading left-turn phase: a queue remains at the end of
that is, the queue size increases linearly on the red phase and the permitted phase but not at the end of the protected phase. Note
decreases linearly on the green. The peak storage occurs at the that it is not possible to have a queue at the end of both the
end of the red phase. The geometry of the triangle depends on the protected and permitted phases if the v/c ratio is not allowed to
arrival flow rate, the queue discharge rate, and the length of the exceed 1.0 for purposes of the uniform delay term.
red and green signal phases. Case 4—lagging left-turn phase: no queue remains at the end
This simple triangle becomes a more complex polygon when of the permitted phase. In this case there will be no queue at the
left turns are allowed to proceed on both protected and permitted end of the protected phase either, because the protected phase
phases. However, the area of this polygon, which determines the follows immediately after the permitted phase and will therefore
uniform delay, is still relatively easy to compute given the proper accommodate all of its arrivals without further delay.
values for the arrival and discharge rates during the various inter- Case 5—lagging left-turn phase: a queue remains at the end of
vals of the cycle, along with the interval lengths that determine the permitted phase. If the v/c ratio is kept below 1.0 as just
its shape. In the following discussion, the protected phase is re- discussed, this queue will be fully served during the protected
ferred to as the primary phase and the permitted phase is referred phase.
to as the secondary phase. Some intermediate computations are required to provide a con-
Specifically, the following quantities must be known to evaluate sistent framework for dealing with all of these cases. Three queue
the uniform delay: lengths may be determined at various transition points within the
T The arrival rate, qa (veh/sec), presumed to be uniform over cycle. These values are defined as follows:
the entire cycles; T Queue size Qa (veh) at the beginning of the green arrow,
T The saturation flow rate sp (veh/sec) for the primary phase; T Queue size Qu (veh) at the beginning of the unsaturated inter-
T The saturation flow rate ss (veh/sec) for the unsaturated por- val of the permitted green phase, and
tion of the secondary phase (the unsaturated portion begins when T Residual queue size Qr (veh) at the end of either the permitted
the queue of opposing vehicles has been served); or the protected phase.
T The effective green time, g (sec), for the primary phase in
which a green arrow is displayed to the left turns; These queue sizes dictate the shape of the polygon whose area
T The green time gq (sec) during the secondary phase when the determines the value of uniform delay. Separate formulas will be
opposing through movement blocks the permitted left turns (this given for computing each of the queue sizes for the five cases just
interval begins at the start of the permitted green and continues described. Formulas will be provided for computing the uniform
until the queue of opposing through vehicles has been fully delay as a function of the queue sizes.
discharged);
T The green time gu (sec) that is available for left-turning vehi-
cles to filter through gaps in the oncoming traffic [this interval
begins when the queue of opposing through vehicles has been
The results of an operational analysis will yield two key values:
satisfied (i.e., at the end of gq) and continues until the end of the
permitted green phase]; and 1. Volume-to-capacity ratios for each lane group and for all of
T The red time r (sec) during which the signal is effectively the critical lane groups within the intersection as a whole, and
red for the left turn. 2. Average control delays for each lane group and approach
and for the intersection as a whole and the corresponding levels
The input-output relationships that determine the shape and area
of service.
of the polygon are shown in Figure 9-12. Note that the queueing
polygon may assume five different shapes depending on the rela- Any v/c ratio greater than 1.0 is an indication of actual or poten-
tionship of arrivals and departures. Slightly different mathematical tial breakdown and a condition requiring amelioration. When the
formulas must be applied to determine the area for each of the overall intersection v/c ratio is less than 1.0 but some critical lane
different shapes. In all cases, the arrival rate must be adjusted to groups have v/c ratios greater than 1.0, the green time is generally
ensure that, for purposes of uniform delay computation, the v/c not appropriately apportioned, and a retiming using the existing
Figure 9-12. Queue accumulation polygons.
phasing should be attempted. Appendix II may be consulted for In some cases, delay will be high even when v/c ratios are low.
suggestions in this regard. In these situations, poor progression or an inappropriately long
A critical v/c ratio greater than 1.0 indicates that the overall cycle length (or both), is generally present. The following point
signal and geometric design provides inadequate capacity for the must be emphasized: unacceptable delay can exist where capacity
given flows. Improvements that might be considered include any is a problem as well as in cases in which it is adequate. Further,
or all of the following: acceptable delay levels do not automatically ensure that capacity
is sufficient. The analysis must consider the results of both the
1. Basic changes in intersection geometry (number and use of
Capacity Analysis Module and the LOS Module to obtain a com-
lanes),
plete picture of existing or projected intersection operations.
2. Increases in the signal cycle length if it is determined to be
Because of the complexity of this methodology, detailed work-
too short, or
sheets are provided for the computations of each analysis module.
3. Changes in the signal phase plan.
These are presented and discussed in Section III, Procedures for
Appendixes I and II may be consulted for suggestions with Application.
regard to these improvements. Existing state and local policies or
standards should also be consulted in the development of potential
improvements. PLANNING ANALYSIS
It should also be noted that v/c ratios near 1.0 represent situations
with little available capacity to absorb demand increases. Particu- The operational analysis method for signalized intersections pre-
larly when projected volumes are being used, normal inaccuracies sented in this chapter provides an extremely detailed treatment of
in such projections can cause an intersection projected to operate the operation of a traffic signal. The level of precision inherent in
near capacity to become oversaturated. that analysis often exceeds the accuracy of the available data. The
Level of service is a measure of the acceptability of delay levels requirement for a complete description of the signal timing plan
to motorists at a given intersection. When delays are unacceptable, is also a burden, especially when the method is being applied in
the causes of delay should be carefully examined. If an unfavorable transportation planning situations.
progression is the largest contributor to delay, changes in intersec- It is possible to obtain an approximate analysis of the level of
tion design and intersection signalization will have little impact; service at a traffic signal through the judicious use of assumed
offsets and arterial coordination should be examined for possible values for most of the data that are required. Table 9-3 contains
improvement. When progression is reasonable and unacceptable recommended default values for several data items. For planning
delays still exist, provision of greater capacity through geometric purposes, the only site-specific data that should be required are
or signal design changes should be examined. the traffic volumes and number of lanes for each movement to-
9-32 urban streets
Table 9-15. Intersection Status Criteria for be met by assumed or default values that represent reasonable or
Signalized Intersection Planning Analysis average values for operating parameters. Other data items are site-
critical v/c ratio relationship to specific and must be obtained in the field. The objective of the
(Xcm) probable capacity planning method is to minimize the need for detailed collection
Xcm ≤ 0.85 Under capacity of field data. The data requirements for this level of analysis may
0.85 < Xcm ≤ 0.95 Near capacity be met by using three worksheets that will be described in detail
0.95 < Xcm ≤ 1.00 At capacity in the next section of this chapter.
1.00 < Xcm Over capacity Much of the required data may be developed either through
judgment or by cursory observation. For each approach, it is neces-
gether with a minimal description of the signal design and other sary to answer the following questions:
operating parameters. This section sets forth a recommended tech-
nique for preparing a complete data set with minimal field data 1. Will parking be allowed?
requirements. As such, it proposes a planning level methodology 2. Will the signal be coordinated with the upstream signal on
for the analysis of signalized intersections. this approach?
3. How will left turns be accommodated?
Overview of Planning Method The treatment alternatives for left turns were described in Section
I of this chapter as ‘‘permitted,’’ ‘‘protected,’’ ‘‘protected-plus-
The concept of the planning method may be best understood by permitted,’’ and ‘‘not opposed.’’
comparison with the full operational analysis method already pre- Most of the foregoing questions may be answered easily on the
sented in this chapter. The operational analysis method is illustrated basis of existing operation. If the answers are not known, the
in Figure 9-3. The LOS estimates (A–F) are based on a detailed evalu- assumptions in the following sections are suggested.
ation of the control delay per vehicle in each lane group.
From a planning perspective, the data requirements for this pro- Parking
cedure are usually considered somewhat excessive, and the need
for an approximate analysis is apparent. The concept of the plan- If the parking restrictions have not been determined, the plan-
ning method is to apply the required approximation to the input ning method may be used as a decision tool. Both conditions (i.e.,
data and not the computational procedures. This provides a link parking and no parking) may be analyzed and compared.
between the planning and operational analyses and allows the same
basic computational methodology to serve both levels of analysis Coordination
in estimating the level of service.
Without effective coordination, signals along an arterial can
A set of worksheets to be described in the next section of this
create poor operating conditions. The closer the spacing of signal-
chapter is used to determine the critical v/c ratio, Xcm, which has
ized intersections without adequate coordination, the more delay
been described earlier as an approximate indicator of the overall
vehicles can encounter. Conversely, closely spaced signalized in-
sufficiency of the intersection geometrics. The computational
tersections with good coordination can be an enhancement to arte-
method involves the summation of conflicting critical lane volumes
rial flow. When signalized intersections are placed far enough from
for the intersection. The computations themselves depend on the
each other, their effect on slowing or enhancing arterial flow may
traffic signal phasing, which in turn depends on the type of protec-
be minimal. On the major street, coordination should be assumed
tion assigned to each left turn. The critical volume summation
if the upstream signalized intersection is less than 2,000 ft away.
divided by the computed intersection capacity represents the criti-
On the minor street, the corresponding distance is 1,200 ft. Minor
cal v/c ratio, Xcm. Although it is not possible to assign a level of
roads are usually shorter and their through traffic travels less dis-
service to the intersection based on Xcm, it is possible to evaluate
tance than on major arterial roads.
the operational status of the intersection for planning purposes.
Table 9-15 expresses the operational status as ‘‘over,’’ ‘‘at,’’ Requirement for Left-Turn Protection
‘‘near,’’ or ‘‘under’’ capacity.
One of the by-products of the critical volume summation is the For planning purposes the actual left-turn treatment should be
synthesis of a ‘‘reasonable and effective’’ signal timing plan for used. If this is unknown, the choice should be made using local
the intersection. When this timing plan is combined with assumed policies or practices. Many agencies use the product of the left-
values for other operating parameters, all of the data required to turning volume and the oncoming through traffic volume, which
apply the full operational analysis will be available. As an exten- is entered on the Lane Volume Worksheet (see Figure 9-23). Al-
sion of the planning analysis, it is therefore possible to obtain an though threshold values vary, one common practice suggests that
estimate of the level of service on each of the lane groups and left turns may require protection when this value exceeds 50,000
approaches and for the intersection as a whole. The accuracy of with one opposing lane (90,000 with two lanes, and 110,000 with
such estimates will depend heavily on the quality of the input three lanes) and the left-turn volume itself exceeds 90 vph. If the
data. If the traffic volumes are rough approximations of future left-turn volume exceeds 240 vph or if more than one turning lane
conditions, the planning analysis should not be taken beyond the is provided, protection is required regardless of the magnitude of
evaluation of intersection status. the product.
Note that these thresholds should only be applied for planning
Field Data Requirements purposes. For design and operational purposes there are many
other factors that should be considered, including accident experi-
The overall data requirements are summarized in the following ence, field observations, and conditions that may exist outside of
discussion. It should be noted that some of the requirements may the analysis period.
Protected left turns may also be allowed to proceed on a permit- fied on the Input Module Worksheet (see Figure 9-14). Values
ted phase (protected-plus-permitted phasing). This is an opera- appropriate to the local jurisdiction should be chosen. Cycle
tional detail that may not be available. The existence of a permitted lengths normally vary between 60 and 120 sec. In areas where
left-turn phase combined with a protected phase is not reflected heavy congestion is anticipated, cycle lengths as high as 150 sec
in the planning worksheets, except that two sneakers per cycle are are occasionally found. Default values of 60 sec minimum and
removed from left-turn volumes under protected-plus-permitted 120 sec maximum are recommended for planning purposes. These
phasing. values should be used in the absence of more specific information.
Unprotected left turns from exclusive lanes receive no explicit The field data requirements will be satisfied by the items de-
assignment of green time because they are assumed to be accom- scribed above. The remainder of this discussion deals with the
modated by the concurrent through movement. It is therefore possi- items that may be determined either by assuming default values
ble to produce an unreasonably optimistic assessment of the critical or by performing worksheet computations.
v/c ratio if protected phasing is not provided for heavy left-turn
volumes. The procedure to be described later includes a check for
left-turn capacity. Default Values
Split-Phase Operation To minimize the need for field data for planning analyses, a
Split-phase operation provides complete separation between number of assumptions are built into the process as default values
movements in opposing directions by allowing all movements in to represent generalized traffic characteristics and traffic signal
only one direction to proceed at the same time. This alternative operating parameters. The default values for approach grade, heavy
should only be assumed for planning purposes if vehicles, and bus blockage have already been presented in Table
9-3. Lane utilization adjustment factors of 1.0 are suggested, which
1. A pair of opposing approaches is offset; are consistent with the requirements of a planning-level analysis.
2. Protected left-turn phasing must be provided to two opposing No pedestrian conflicts are considered because data at this level
single-lane approaches; or of detail are not usually available.
3. Both opposing left turns are protected and one of the left These assumptions allow the determination of average condi-
turns is accommodated with an exclusive lane plus an optional tions for each lane group. The variables area type, saturation flow
lane for through and left-turning traffic. rate, pedestrian crossing volume, lost time, and yellow plus all red
In addition to the movement-specific data just described, there include default values that are representative of suburban intersec-
are three items that apply to the intersection as a whole: the area tion conditions. The presence of traffic signal coordination must
type, the peak-hour factor, and the cycle length requirements for be identified on each intersection approach. On those approaches
the signal operation. where coordination exists, Arrival Type 4 should be used. On
those approaches where coordination does not exist, Arrival Type
Area Type 3 should be used.
Of course, any of these default values may be overridden during
The choices offered by the operational analysis method are the analysis. Overriding a given value should produce a more
‘‘central business district’’ (CBD) or ‘‘other.’’ Some judgment is accurate assessment of the capacity and level of service at the
required here. Unless the intersection is known to be within the expense of consistency of treatment among intersections. There is
CBD, the ‘‘other’’ category should be assumed. a clear trade-off here, and the decision is up to the analyst.
Peak-Hour Factor
Synthesis of Signal Operation
The peak-hour factor (PHF) is used to focus the analysis on the
peak 15 min of the hour. It is an important feature of the operational
analysis method. However, for planning purposes, the appropriate The LOS computations for planning purposes are carried out as
value for the PHF will depend on the nature of the application. an optional step using the operational analysis method described
For near-term approximation of intersection level of service, the in this chapter. The signal design parameters to be synthesized by
use of a PHF may be desirable. If no data are available, a value the technique described here are intended as direct input data to
of 0.9 should be assumed. the operational analysis method. Worksheets are presented for the
For longer-term projections of roadway sufficiency in heavily computation of all of the design parameters. It is, however, antici-
populated areas, the balance between hourly volumes and capaci- pated that a computerized version of the technique would be em-
ties may be of more interest. If this is the case, a PHF of 1.0 may ployed in most practical applications. Implementation of the
be more appropriate. However, if 15-min peaking occurs within computations by hand would be time consuming because of the
the hour, failure to use a PHF will result in an underestimation of detailed nature of the process.
delay if the planning analysis is extended to evaluate the level of It is not essential, nor is it practical, for planning applications
service. to define a fully optimized signal timing plan for the intersection.
It is only necessary to ensure that the analysis be based on a
Cycle Length Requirements ‘‘reasonable and effective’’ timing plan. For purposes of this chap-
ter, the following attributes apply to a reasonable and effective
The design cycle length should be used if it is known. If it is not, signal timing plan:
the cycle length may be calculated by using the Signal Operations
Worksheet presented later (see Figure 9-24). The calculations are 1. The timing plan must accommodate the critical movements
subject to minimum and maximum values, which should be speci- on all lane groups at the intersection;
9-34 urban streets
2. A cycle length should be chosen that is appropriate to the leading versus lagging left-turn protection. The timing plan does
traffic volume levels; not consider user-specified minimum green times for each phase
3. The allocation of time among competing movements should nor does it consider the optimization of phase splits.
be such that the degree of saturation is equalized for the critical The limitations just mentioned pose no problems for planning
movements in each phase; applications. The technique described here will generate a com-
4. The phasing plan should accommodate unbalanced volumes plete phasing and timing plan that represents a reasonable approxi-
with a minimum of slack time through the use of overlap phasing; mation of the conditions that might be expected to occur with the
5. The phase time for a lane group with a shared left-and- given traffic volumes and intersection configuration, assuming that
through lane and permitted left turns should be designed to accom- a reasonable and effective signal timing design is employed.
modate the through movement only, not the entire lane group (the
adequacy of the timing with respect to the left turn should be
Other Analyses
checked if necessary in a subsequent step using the operational
procedure);
As noted previously, the computational procedures in this
6. The protected phase time for a protected-permitted left turn
chapter emphasize the estimation of level of service (delay)
should accommodate the entire movement; and
based on known or projected traffic demand, signalization, and
7. Protected left-turn phasing should be used for all left turns
geometric design. Other computational applications include deter-
that would not otherwise be accommodated.
mination of
The signal operation is described in terms of a phase plan indi-
1. Volume-to-capacity ratios and service flow rates associated
cating which movements are able to proceed on each phase and a
with selected levels of service given a known signalization and
timing plan indicating the cycle length and the apportionment of
geometric design;
time to each phase in the cycle. The procedure suggested here will
2. Signal timing parameters when known inputs are a selected
produce a reasonable and effective timing plan by the foregoing
level of service, demand flow rates, and geometric design; and
definition, given the information mentioned previously. It is based
3. Geometric parameters (number of lanes, lane use allocations,
on well-established principles of critical movement analysis and
etc.) given selected level of service, demand flow rates, and
on the signal timing guidelines presented in Appendix II.
signalization.
The use of this technique should be limited to planning applica-
tions. It is not intended to produce an optimized operating plan These alternative computational sequences are discussed in the
for implementation in practical situations. The phase plan is chosen next section of this chapter and illustrated with sample
from a limited set of alternatives. No consideration is given to computations.
Detailed worksheets for computations and step-by-step instruc- T Permitted left-turn operation,
tions for their use and interpretation are presented in this section. T Protected left-turn operation, or
The operational analysis will be described first, followed by the T A combination of protected and permitted left turns.
less detailed planning analysis.
Thus there are seven different possibilities, each of which must
be handled in a slightly different manner using the worksheets.
OPERATIONAL ANALYSIS
Operational analysis is divided into five modular subanalyses: Input Module

(a) Input Module, (b) Volume Adjustment Module, (c) Saturation
Flow Rate Module, (d) Capacity Analysis Module, and (e) LOS The Input Module is essentially a summary of the geometric,
Module. The computations for each of these modules are con- traffic, and signalization characteristics needed to conduct other
ducted or summarized on the appropriate worksheet, or both, as computations. When an existing case is under study, most of these
each module is presented. data will be obtained from field studies. When future conditions
In addition to the module-related worksheets, two supplemen- are under consideration, traffic data will be forecast, and geometric
tary worksheets are provided to handle computations that are more and signal designs will be based on existing conditions or will be
complex. An overview of the information flow among all work- proposed. The Input Module Worksheet is shown in Figure 9-14.
sheets is presented in Figure 9-13, which shows the proper treat- The upper half of the worksheet contains a schematic intersec-
ment of all combinations of left-turn lanes and phasing. A given tion drawing on which basic volume and geometric data are
lane group may have recorded.
T Left turns from an exclusive lane, Step 1: Record Traffic Volumes

T Left turns from a shared lane, or
T No left turns at all. For each movement, 15-min flow rates (vph) for the analysis
period or hourly volumes are entered into the appropriate boxes
When left turns are present, the signal phasing may provide shown in each corner of the intersection diagram. Left-turn,
Figure 9-13. Worksheet information flow. (RT = right turn; LT = left turn; Prot = protected; Perm = permitted.)

9-36 urban streets
Figure 9-14. Input Module Worksheet.

through, and right-turn volumes are recorded below these boxes pushbutton detector on the subject approach is indicated by a ‘‘Y’’
at the head of the appropriate directional arrow. The sum of the or ‘‘N’’ entry. The ninth column gives the minimum green time re-
left, through, and right movements on each approach should equal quired for a pedestrian to cross the street, computed from Equation
the value shown in the approach volume box. 9-8:
Step 2: Record Geometrics Gp = 7.0 + (W/4.0) − Y

The details of lane geometrics should be shown within the inter- 8. The tenth and last column is used to indicate the quality of
section diagram. Details should include signal progression. Either the value P, the proportion of vehicles
T Number of lanes, arriving on green, or the arrival type (1–6) is entered here.
T Lane widths,
When data for some of these variables are not available or
T Traffic movements using each lane (shown by arrows),
forecasts cannot be adequately established, default values may be
T Existence and location of curb parking lanes,
used as an approximation. These may be established by judgment,
T Existence and location of bus stops,
or the default values in Table 9-3 may be used when they cannot
T Existence and length of storage bays, and
be established by other means.
T Other features such as channelization, and so forth.
When geometric conditions are not known, a design should be Step 4: Enter Signal Phasing Design
proposed based on state or local practice. Appendix I may be
consulted to assist in establishing a design for analysis. When The sequence of signal phases is diagrammed in the eight boxes
separate left-turn lanes exist, the procedures assume that the stor- at the bottom of the Input Module Worksheet. Up to an eight-
age length is adequate. This should be checked against the criteria phase signal design may be shown. Each box is used to show a
in Appendix I. single phase or subphase during which the allowable movements
The middle portion of the worksheet consists of a tabulation of remain constant.
additional traffic and roadway conditions for each movement.
1. For each phase, the allowable movements are shown with
Step 3: Enter Traffic and Roadway Conditions arrows. Permitted turns are shown with a dashed arrow, and pro-
The following parameters are entered into the tabulation in the tected turns are shown with solid arrows. Conflicting pedestrian
middle of the worksheet. Separate entries are required for each flows should be shown with dashed lines.
approach: 2. For each phase, the actual green (G) time and the actual
yellow-plus-all-red (Y) time should be shown (in seconds) on the
1. Percent grade is entered in the first column; a plus sign line labeled ‘‘Timing.’’
indicates upgrade, and a minus sign indicates downgrade. 3. Each phase should be identified as pretimed (P) or actuated
2. Percent heavy vehicles is entered in the second column. Nor- (A) in the appropriate box.
mally the average for the entire approach is used. When heavy
vehicle presence varies significantly between movements, separate When signal design is not known, two decisions should be made
percentages may be used for left-turn, through, and right-turn at this point: what type of control is going to be assumed for
movements. A heavy vehicle is defined as any vehicle with more analysis, and what phase sequence will be used? These two ques-
than four tires touching the pavement. tions are important, because they will influence the determination
3. The third and fourth columns describe parking characteristics of lane groups for analysis. This portion of the signal design should
for the approach. The third column indicates the presence of an be projected on the basis of state or local practice. For additional
adjacent parking lane at the intersection; ‘‘Y’’ or ‘‘N’’ is entered suggestions on establishing the type of control and phase sequence,
as appropriate. The fourth column indicates the number of parking Appendix II may be consulted.
maneuvers per hour occurring into and out of the parking lane The timing of the signal will not be known when signal design
within 250 ft upstream of the stop line. is to be established. It may or may not be known when actuated
4. The number of local buses stopping per hour to discharge or signals are in place, depending on whether average phase durations
pick up passengers within the confines of the intersection is rec- were observed in the field. Appendix II contains recommendations
orded in the fifth column. Any bus stop within 250 ft upstream or for establishing phase times based on an assumed signal type and
downstream from the stop line is considered to be within the phase sequence and for estimating the average phase lengths of
confines of the intersection. actuated signals when observations are not available. These esti-
5. The peak-hour factor is entered in the sixth column. This mates, however, cannot be computed until the first half of the
will be used to convert hourly values to 15-min flow rates in the Capacity Analysis Module is complete. Other computations may
event that 15-min flow rates were not entered. PHF values of 1.0 proceed without this information.
should be used for 15-min flow entries. Because the establishment of signal timing will usually involve
6. The number of pedestrians per hour using the crosswalk and iterative computations, it is preferable to simply specify a complete
conflicting with right turns from the subject approach is recorded signal timing for analysis using trial-and-error computations to
in the seventh column. For the NB approach, this is the east cross- determine an appropriate final timing. As an alternative, the timing
walk; for the SB approach, the west crosswalk; for the EB ap- plan may be synthesized using the planning method described
proach, the south crosswalk; and for the WB approach, the north previously. If a fully implementable timing plan is required, a
crosswalk. variety of professionally accepted signal timing optimization mod-
7. The eighth and ninth columns describe pedestrian controls at els may be used. Some of these models apply the methodology of
the intersection. In the eighth column, the existence of a pedestrian this chapter iteratively.
9-38 urban streets
Volume Adjustment Module right-turn flow rates are obtained from Column 5 of the worksheet,
and the total lane group flow rate is given in Column 7.
The second major analysis module focuses on (a) adjustment
of hourly movement volumes to flow rates for a peak 15-min
period within the hour and (b) establishment of lane groups for Saturation Flow Rate Module
analysis. The Volume Adjustment Module results in the prepara-
tion of demand flows in a form amenable to further analysis and In the Saturation Flow Rate Module, the total saturation flow
provides values used in subsequent analysis modules. A worksheet rate that can be accommodated by the lane group under prevailing
for volume adjustment computations is shown in Figure 9-15. conditions is computed. A worksheet for this module is shown in
Figure 9-16.
Step 1: Enter Hourly Volumes
Hourly movement volumes or 15-min flow rates (vph) are en- Step 1: Enter Description of Lane Groups
tered in Column 3 of the worksheet. These are taken directly from
Column 2 of the worksheet is used to identify the lanes and
the intersection diagram on the Input Module Worksheet.
movements included in each lane group. These are the same as the
Step 2: Convert Hourly Volumes to Peak Flow Rates entries in Column 6 of the Volume Adjustment Module Worksheet,
where lane groups are established.
If hourly volumes are used, the peak-hour factor (PHF) for each
movement is entered in Column 4. If 15-min flow rates are used, Step 2: Enter Ideal Saturation Flow Rate
1.0 is entered. Hourly volumes are divided by the PHF to compute
peak flow rates: The ideal saturation flow rate per lane is entered in Column 3
of the worksheet. For most computations, this value will be taken
vp = V/PHF to be 1,900 passenger cars per hour of green time per lane
where vp is the flow rate for the peak 15-min analysis period. The (pcphgpl), unless local data indicate that another value is appro-
result is entered in Column 5 of the worksheet. priate. Appendix IV contains guidelines for conducting local stud-
ies to determine the prevailing saturation flow rate for purposes
Step 3: Establish Lane Groups for Analysis of calibrating the ideal saturation flow rate.
Lane groups for analysis should be established on the basis of Step 3: Enter Adjustment Factors
recommendations cited in Section II, Methodology. Exclusive turn
lanes are always established as separate lane groups. Where shared The ideal saturation flow rate is multiplied by the number of
left-turn and through lanes exist on an approach with additional lanes in the lane group and by nine separate adjustment factors,
lanes for through traffic, they should be checked to determine as follows:
whether they operate in a shared equilibrium mode or as de facto
left-turn lanes. This check involves determining the proportion of 1. Enter the number of lanes in the group in Column 4 of the
left turns in the shared lane. If this value equals or exceeds 1.0, worksheet.
the shared lane should be considered an exclusive left-turn lane. 2. Enter the lane width factor, fw, obtained from Table 9-5, in
The proportion of left turns in the shared lane will be determined Column 5.
later as a part of the Saturation Flow Rate Module. 3. Enter the heavy vehicle factor, fHV, obtained from Table 9-
Lane groups are shown in Column 6 of the worksheet by enter- 6, in Column 6.
ing arrows illustrating the lanes and movements included in the 4. Enter the grade factor, fg, obtained from Table 9-7, in Col-
group. Permitted turning movements are shown with dashed umn 7.
arrows, and protected turning movements are shown with solid 5. Enter the parking factor, fp, obtained from Table 9-8, in Col-
arrows. When a turn has a protected and a permitted phase, both umn 8.
types of arrows should be shown. 6. Enter the bus blockage factor, fbb, obtained from Table 9-9,
in Column 9.
Step 4: Enter Adjusted Lane Group Rate 7. Enter the area type factor, fa, obtained from Table 9-10, in
Column 10.
Once lane groups have been established, the flow rates for in- 8. Enter the lane utilization adjustment factor, fLU, computed
cluded movements must be entered in Column 7 of the worksheet using Equation 9-11, in Column 11.
as the adjusted lane group flow rate, v. 9. Enter the right-turn factor, fRT, obtained from Table 9-11, in
Column 12.
Step 5: Enter Proportion of Left or Right Turns in Lane Group
10. Enter the left-turn factor, fLT, obtained from Table 9-12
Column 8 is provided for entering the proportion of left or right or computed using the special procedure described in Section II,
turns, or both, in the lane group volume. These values may be Methodology, for permitted left turns made from exclusive or
computed as shared lanes, in Column 13.
PLT = vLT/v Factors for each lane group are determined separately from
the prevailing conditions for the lane group. Information for
PRT = vRT/v
these determinations is taken from the Input Module Worksheet.
where PLT and PRT are the proportions of left- and right-turning The proportion of left or right turns, or both, is taken from the
vehicles using the lane group, expressed as a decimal. Left- and last column of the Volume Adjustment Module Worksheet.
Figure 9-15. Volume Adjustment Module Worksheet.
Determination of right-turn factors for protected-plus-permitted permitted phasing, whether made from an exclusive or shared lane,
phasing will require an assumption of the proportion of right- for Cases 2, 3, and 5. Figure 9-17 is used in cases in which the
turning vehicles using the protected portion of the phase. This is subject approach is opposed by an approach with more than one
basically judgmental and should be guided by field observations lane. Figure 9-18 is used in cases in which the subject approach
where possible. is opposed by a single-lane approach.
The basic methodology for each worksheet assumes that the
Step 4: Special Procedure for Estimating the Left-Turn Adjustment subject approach is a multilane approach if the opposing approach
Factor for Permitted Left Turns
is a multilane approach (Figure 9-17) and that the subject approach
Figures 9-17 and 9-18 show worksheets that are used in the is a single-lane approach if the opposing approach is a single-lane
computation of the left-turn adjustment factor when permitted left approach (Figure 9-18). For cases in which the two approaches
turns are made. These worksheets are applied to the permitted are not of the same type as well as cases of protected-plus-permit-
portion of left turns, including permitted-only and protected-plus- ted phasing and a phasing in which the opposing through move-
9-40 urban streets
Figure 9-16. Saturation Flow Rate Module Worksheet.
ment has a lead phase, the worksheets may still be used, but 3. The effective green time for the permitted phase is entered.
the special instructions cited in Section II, Methodology, must be This is generally the actual green time (above) from the Input
followed carefully. Module Worksheet plus the yellow plus all-red change-and-clear-
There is a column for each approach on the worksheets, although ance interval minus the movement’s lost time. If the permitted
only those approaches with permitted left-turn conditions would be phase is part of a protected-plus-permitted phasing or the opposing
included. Since the worksheets are quite similar, they are discussed approach has a lead phase, see the special instructions in Section
together here, noting the exceptions and differences where II, Methodology.
appropriate. 4. The effective green time for the opposing approach is entered
The first set of entries consists of input variables that should be for the permitted phase. This is generally the actual green time
entered directly from values appearing on previous worksheets, as from the Input Module Worksheet plus the yellow plus all-red
follows: change-and-clearance interval minus the movement’s lost time. If
the permitted phase is part of a protected-plus-permitted phasing
1. The cycle length is entered from the Input Module or the opposing approach has a lead phase, see the special instruc-
Worksheet. tions in Section II, Methodology.
2. The actual green time for the permitted phase is entered from 5. The number of lanes in the subject lane group is entered
the Input Module Worksheet. If the permitted phase is part of a from the Input Module Worksheet. If the left turn is opposed by
protected-plus-permitted phasing or the opposing approach has a a multilane approach (Figure 9-17), the number of lanes in the
lead phase, see the special instructions in Section II, Methodology. opposing lane group is entered from the Input Module Worksheet
Figure 9-17. Supplemental Worksheet for Permitted Left Turns: Multilane Approach.

9-42 urban streets
Figure 9-18. Supplemental Worksheet for Permitted Left Turns: Single-Lane Approach.

as well. If left or right turns are made from exclusive turn lanes Figure 9-17 except when the subject lane group contains multiple
on the opposing approach, these lanes are not included in the exclusive left-turn lanes.
number of opposing lanes.
6. The adjusted left-turn flow rate is entered from the Volume Step 5: Compute Adjusted Saturation Flow Rates
Adjustment Module Worksheet.
The adjusted saturation flow rate for each lane group is com-
7. The proportion of left turns in the lane group is entered from
puted by multiplying the ideal saturation flow rate by the number
the Volume Adjustment Module Worksheet. When an exclusive
of lanes in each lane group and by each of the nine adjustment
left-turn lane group is involved, PLT = 1.0. If the left turn is opposed
factors determined in Steps 3 and 4. This is done in accordance
by a single-lane approach (Figure 9-18), the proportion of left
with Equation 9-10:
turns in the opposing flow is entered from the Volume Adjustment
Module Worksheet. s = so N fw fHV fg fp fbb fa fLU fRT fLT
8. The adjusted opposing flow rate is entered from the Volume
Adjustment Module Worksheet. If left or right turns are made
from exclusive turn lanes on the opposing approach, these adjusted Capacity Analysis Module
volumes are not included in the opposing flow rate.
9. The lost time for the left-turn lane group is entered as deter- In the Capacity Analysis Module, information and computa-
mined from the Input Module Worksheet. tional results from the first three modules are combined to compute
the capacity of each lane group and v/c ratios for each lane group
The equations used in subsequent computations are shown on and for the intersection as a whole. A worksheet for these computa-
the remaining rows of the worksheet; these equations are based tions is shown in Figure 9-19.
on the input variables that were entered above. Some of these
computations deserve some further discussion, as follows. Step 1: Enter Lane Group Description
T The opposing platoon ratio Rpo may be determined in two Column 1 of the worksheet is once again for the description of
different ways. If the arrival type of the opposing traffic appears lane groups. Lanes and movements included in each lane group
on the Input Module Worksheet, the default platoon ratio from are entered as on the Saturation Flow Rate Module Worksheet.
Table 9-2 is used. If the proportion of arrivals on green appears
on the Input Module Worksheet, Equation 9-7 based on the g/C Step 2: Enter Phase Type
ratio is used instead. Column 2, Phase Type, is included to accommodate exclusive-
T The equation shown for gf in Figure 9-17 assumes that the lane left turns that have both protected and permitted phases. In
subject approach is a multilane approach like the opposing ap- this case, the protected phase will be the primary phase and the
proach. If the subject approach is a single-lane approach, the equa- permitted phase will be the secondary phase. The primary and
tion for gf from Figure 9-18, which assumes a single-lane approach, secondary phases must be represented by separate row entries on
should be used. Conversely, the equation shown for gf in Figure this worksheet, and certain quantities, such as lane group capacity,
9-18 assumes that the subject approach is a single-lane approach must be computed as the sum of the primary and secondary phase
like the opposing approach. If the subject approach is a multilane values. Primary phase entries should be designated ‘‘P’’ in this
approach, the equation for gf from Figure 9-17, which assumes a column. Secondary phase entries should be designated ‘‘S,’’ and
multilane approach, should be used. In either case, if the subject the row containing the total values should be designated ‘‘T.’’
lane group is an exclusive left-turn lane, then gf = 0. Note that lane groups with shared left-turn lanes have only a pri-
T For multilane lane groups (Figure 9-17), PL is computed as mary phase, as do lane groups with only protected or only permit-
the proportion of left turns in the left-hand lane of the lane group. ted phasing.
If this value is determined to be 1.0 or higher, the lane groups for
the approach should be reassigned showing this left-hand lane as Step 3: Enter Adjusted Flow Rate for Each Lane Group
an exclusive left-turn lane (a de facto left-turn lane), since it is
occupied entirely by left-turning vehicles. This requires redoing The adjusted flow rate for each lane group is obtained from the
all of the computations for this approach. If a multilane lane group Volume Adjustment Module Worksheet and entered in Column 3
is opposed by a single-lane approach, Figure 9-18 should be used, of the worksheet. In the case of lane groups with both primary
but a value of PL should be estimated and substituted for PLT, as and secondary phases, the flow rate for the lane group should be
described in Section II, Methodology. In this case, the same de entered in a row identified ‘‘T’’ in Column 2. For computation of
facto left-turn check should be applied. the critical v/c ratio, Xc, it is necessary to apportion the total flow
T Figure 9-7 is used to determine the value of EL1 based on the rate between the primary and secondary phases. As indicated in
opposing flow rate and the lane utilization factor of the opposing Section II, Methodology, it is appropriate to consider whichever
phase is displayed first to be fully saturated by left-turn traffic and
flow. For the single-lane approach (Figure 9-18), EL2 is computed
to apply any residual flow to the phase that is displayed second.
by formula, not by Figure 9-7.
T The value of fm is computed as shown. The maximum value
Step 4: Enter Adjusted Saturation Flow Rate for Each Lane Group
is 1.0 and the minimum value is 2(1 + PL)/g. These limits are used
if the computed value falls outside this range. The adjusted saturation flow rate for the primary phase for each
T The left-turn adjustment factor, fLT, is computed as shown. lane group is obtained directly from the Saturation Flow Rate
For a single-lane lane group, fLT = fm. If a multilane lane group is Module Worksheet and entered in Column 4. It is not necessary
opposed by a single-lane approach, Figure 9-18 is used, but fLT is to enter a saturation flow rate value in Row T when a secondary
calculated on the basis of fm and the number of lanes as shown in phase is involved, because this value has no significance.
9-44 urban streets
Figure 9-19. Capacity Analysis Module Worksheet.

Step 5: Compute Flow Ratio for Each Lane Group The critical v/c ratio, Xc, which indicates the degree of saturation
associated with the geometrics, volumes, and signal phasing, is
The flow ratio for each lane group is computed as v/s and entered
computed as
in Column 5 of the worksheet. This should be done for rows
representing both primary and secondary phases but not for the Xc = Yc[C/(C − L)]
row that represents the total.
The results are entered in the appropriate space at the bottom of
Step 6: Enter Green Ratio for Each Lane Group the worksheet.
At the completion of this module, the capacity characteristics
The g/C ratio for each lane group, the effective green time
of the intersection have been defined. These characteristics must
divided by the cycle length, is computed and entered in Column
be evaluated in their own right as well as in conjunction with the
6 of the worksheet. The actual green times and the cycle length
delays and levels of service resulting from the next module. Al-
may be obtained from the Input Module Worksheet. Effective
though the interpretation of capacity results is discussed in Section
green times can be taken to be equal to the actual green time
II, Methodology, some key points are summarized here:
plus the change-and-clearance interval minus the lost time for the
movement as determined from the Input Module Worksheet. When 1. A critical v/c ratio of greater than 1.0 indicates that the signal
signal timing is to be determined for cases involving permitted and geometric design cannot accommodate the combination of
left turns, these computations will be iterative. critical flows at the intersection. The given demand in these move-
ments exceeds the capacity of the intersection to handle them.
Step 7: Compute Capacity of Each Lane Group The condition may be ameliorated by any or all of the following:
The capacity of each lane group is computed from Equation 9- increased cycle length, changes in the phasing plan, and basic
3 as the saturation flow rate times the green ratio: changes in geometrics.
Note, however, that computations should be conducted using
ci = si (gi /C) arrival volumes. When the v/c ratios are less than 1.0, arrival and
The result is entered in Column 7 of the worksheet. Values should departure volumes are the same. When v/c ratios are greater than
be computed for both primary and secondary phases, and the sum 1.0, either for an individual phase or for the overall intersection,
of the values for each phase should be entered in the row designated departure volumes are less than arrival volumes. Future volume
‘‘T’’ in Column 2. forecasts are also arrival volumes, by definition. When counts of
A minimum capacity value based on sneakers per cycle must actual departure volumes are used in analysis, the actual v/c ratio
be imposed as a practical matter for all permitted left-turning cannot be greater than 1.0. Observed departure volumes cannot
movements. This value may be computed as exceed capacity. In such cases, computations should be checked
for errors. If v/c ratios greater than 1.0 persist for actual departure
3,600 (1 + PL) volumes, it is an indication that the intersection operates more
C efficiently than anticipated by these computational techniques.
2. When the critical v/c ratio is acceptable but the v/c ratios for
critical lane groups vary widely, the green time allocation should
Step 8: Compute v/c Ratios for Each Lane Group
be reexamined, because disproportionate distribution of available
The v/c ratio for the lane group is the ratio of adjusted flow to green is indicated.
capacity: 3. If permitted left turns result in extreme reductions in satura-
tion flow rate for applicable lane groups, protected phasing might
Xi = vi /ci be considered.
These values are computed and entered in Column 8 of the 4. If the critical v/c ratio exceeds 1.0, it is unlikely that the
worksheet. Entries should be made for all rows, including those existing geometric and signal design can accommodate the de-
designated ‘‘P,’’ ‘‘S,’’ and ‘‘T,’’ in Column 2. mand. Changes in either or both should be considered.
5. When v/c ratios are unacceptable and signal phasing already
Step 9: Identify Critical Lane Groups includes protective phasing for significant turning movements, it
is probable that geometric changes will be required to ameliorate
At this point in the computations, critical lane groups and lost
the condition.
time per cycle may be identified according to the guidelines dis-
cussed in Section II, Methodology. A critical lane group is defined The capacity of an intersection is a complex variable depending
as the lane group with the highest flow ratio in each phase or set upon a large number of prevailing traffic, roadway, and signaliza-
of phases. When overlapping phases exist, all possible combination conditions. Suggestions on interpretation are not meant to be
tions of critical lane groups must be examined for the combination exhaustive or complete, but merely to point out some of the more
producing the highest sum of flow ratios, as discussed previously. common problems that can be identified from the Capacity Analy-
Critical lane groups are identified by a check placed in Column 9 sis Module results.
of the worksheet. The lost time per cycle is entered as the value
L in the appropriate space at the bottom of the worksheet.
LOS Module
Step 10: Compute Critical v/c Ratio
The flow ratios for critical lane groups (i.e., those checked in The LOS Module combines the results of the Volume Adjust-
Column 9) are summed. The result is entered as the value Yc in ment, Saturation Flow Rate, and Capacity Analysis modules to
the appropriate space at the bottom of the worksheet. find the average control delay per vehicle in each lane group. The
9-46 urban streets
level of service is directly related to delay and is found from Table vehicles arriving on the green) is used in lieu of the arrival type,
9-1. The worksheet for this module is shown in Figure 9-20. PF may be computed as
Delay is found from Equations 9-22, 9-23, and 9-25, presented
in Section II. These equations are restated below for convenience. (1 − P)fP
PF = (9-24)
1 − (g/C)
d = d1PF + d2 + d3 (9-22)
0.50C(1 − g/C)2 where fP = 0.93 for Arrival Type 2, 1.15 for Arrival Type 4, and
d1 = (9-23) 1.0 for all other arrival types. For this purpose, the arrival type
1 − Min(1,X)g/C
may first be determined from Table 9-2 after calculating Rp = PC/g.
3 !(X − 1) + 4
2
8kIX Because fP is greater than 1.0 for Arrival Type 4, it is possible
d2 = 900T (X − 1) + (9-25)
cT to compute a value of PF greater than 1.0 using this equation
when g/C is very low. Because Arrival Type 4 reflects ‘‘favorable
The worksheet is designed for computation of the uniform and
progression,’’ the value of PF should be reduced to 1.0 under this
incremental delay terms separately. The uniform delay is then
condition.
multiplied by the progression adjustment factor (PF) to account
Enter the value of PF in Column 5 of the worksheet.
for the impact of progression on delay.
The values of PF and k are obtained from Tables 9-13 and 9-
Step 4: Find Incremental Delay
14, respectively.
For purposes of this chapter, the upstream filtering/metering The second term of the delay equation accounts for the ‘‘incre-
adjustment factor (I) is normally set equal to 1.0 for an isolated mental delay,’’ that is, the delay over and above uniform delay
signal analysis. due to arrivals’ being random rather than uniform and due to cycles
When no unserved demand exists from a previous time period, that fail. It is based on the v/c ratio (X) and the capacity (c) for
the residual delay term, d3, is equal to zero. When an initial un- the lane group. Incremental delay is found as follows:
served queue of vehicles exists at the start of the analysis period
(observed at the beginning of red), the procedures in Appendix VI 1. Enter the lane group capacity in Column 6 of the worksheet.
are to be used to modify the calculation of d1, to calculate d3, and 2. Determine the incremental delay calibration factor (k) from
to determine delay and level of service. Table 9-14. This value is a function of the controller type and
degree of saturation. Enter the value of k in Column 7.
Step 1: Enter Lane Group Description 3. Compute the second-term delay from Equation 9-25. Enter
the result in Column 8.
As in the case of previous worksheets, Column 1 is used to
enter the description of the lanes and movements included in the
Step 5: Find Delay and Level of Service for Each Lane Group
lane group. This description will be the same as that shown on
the Volume Adjustment Module Worksheet. Delay and level of service are found by multiplying the uniform
delay by the progression factor and adding the result to the incremen-
Step 2: Find Uniform Delay tal delay, in accordance with Equation 9-22. The result is entered in
The first term of the delay equation (Equation 9-23) accounts for Column 9 of the worksheet. The level of service corresponding to
uniform delay, that is, the delay that results in a lane group if arrivals this delay, taken from Table 9-1, is entered in Column 10.
are uniformly distributed and if no cycles experience oversaturation. In the event that the analysis period starts with an initial
It is dependent upon the v/c ratio (X) for the lane group, the green queue, the procedures in Appendix VI must be used to modify
ratio (g/C) for the lane group, and the cycle length (C), which is the calculation of d1 and to calculate the additional term, d3.
entered at the top of the worksheet. It is found as follows: Furthermore, if the analysis period is oversaturated or results
in a final unmet demand at the end of the analysis period, an
1. Enter the v/c ratio for each lane group in Column 2 of the additional analysis of the subsequent analysis period should be
worksheet. These may be obtained from the Capacity Analysis made to assess its delay.
Module Worksheet.
2. In Column 3 enter the effective green ratio for each lane Step 6: Find Delay and Level of Service for Each Approach
group from the Capacity Analysis Module Worksheet.
3. Compute the first-term delay and enter the result in Col- The average delay per vehicle is found for each approach by
umn 4. adding the product of the lane group flow rate and the delay for
a. For lane groups with only primary phases indicated on the each lane group on the approach and dividing by the total approach
Capacity Analysis Module Worksheet, compute this value in ac- flow rate. The weighted-average delay is entered in Column 11 of
cordance with Equation 9-23. the worksheet for each approach. Level of service is determined
b. For the groups with both primary and secondary phases indi- from Table 9-1 and entered in Column 12.
cated on the Capacity Analysis Module Worksheet, use the supple- As in Step 5, if the analysis period starts with an initial queue,
mental worksheet for lane groups with primary and secondary the delay and level of service for each approach will be determined
phases presented in Figure 9-21 (see discussion in the next section). using the procedures in Appendix VI.
Step 3: Determine the Progression Adjustment Step 7: Find Delay and Level of Service for Intersection
The progression adjustment factor, PF, as indicated in Table 9- The average delay per vehicle for the intersection as a whole is
13, is a function of the arrival type and g/C ratio for lane groups found by adding the product of the approach flow rate and the
with coordinated control. If the value of P (i.e., the proportion of approach delay for all approaches and dividing the sum by the
Figure 9-20. LOS Module Worksheet.
total intersection flow rate. This weighted-average delay is entered each approach and for the intersection as a whole. Level of service
in the appropriate space at the bottom of the worksheet. The overall is directly related to delay values and is assigned on that basis.
intersection level of service is found from Table 9-1 and entered LOS and delay values are best analyzed in conjunction with the
in the appropriate space at the bottom of the worksheet. results of the Capacity Analysis Module. Although the discussion
As in Step 6, if the analysis period starts with an initial queue, below is clearly not exhaustive, some of the more common situa-
the delay and level of service for the intersection will be deter- tions are as follows.
mined using the procedures in Appendix VI.
The result of this module is an estimation of the average control 1. The level of service is an indication of the general acceptabil-
delay per vehicle in each lane group as well as average values for ity of delay to drivers. It should be noted that this is somewhat
9-48 urban streets
subjective: what is acceptable in a large CBD is not necessarily group, obtained from Row T, Column 8, on the Capacity Analysis
acceptable in a less-dense environment. Module Worksheet (Figure 9-19).
2. When delay levels are acceptable for the intersection as a The following signal timing intervals must also be obtained
whole but are unacceptable for certain lane groups, the phase plan, from previous computations:
allocation of green time, or both might be examined to provide
for more efficient handling of the disadvantaged movement or 1. Primary-phase effective green, g, from the Capacity Analysis
movements. Module Worksheet (Figure 9-19);
3. When delay levels are unacceptable but v/c ratios are rela- 2. Secondary-phase effective green intervals, gq and gu, from
tively low (Capacity Analysis Module), the cycle length may be too the supplemental worksheets for permitted left turns (Figure 9-17
long for prevailing conditions, the phase plan may be inefficient, or or 9-18); and
both. It should be noted, however, that when signals are part of a 3. Red time (in seconds), r, computed as C − (g + gq + gu),
coordinated system, the cycle length at individual intersections is where C is the cycle length (in seconds).
determined by system considerations, and alterations at isolated These values are entered in the appropriate rows on the worksheet.
locations may not be practical. Note that extremely heavy opposing traffic may reduce gu to zero,
4. When both delay levels and v/c ratios are unacceptable, the which means that all of the left turns on the permitted phase
situation is critical. Delay is already high, and demand is near or will be accommodated as sneakers. The effect of sneakers was
over capacity. In such situations, the delay may increase rapidly approximated on the Saturation Flow Rate Module Worksheet
with small changes in demand. The full range of potential geomet- (Figure 9-16) by imposing a lower limit on the value of fLT. Because
ric and signal design improvements should be considered in the of the lower limit on fLT, a lower limit must also be imposed on
search for improvements in such cases. the value of gu to be entered on the Supplemental Uniform Delay
Delay and level of service, like capacity, are complex variables Worksheet. The necessary time should be transferred from gq to
depending on a wide range of traffic, roadway, and signalization gu to ensure that the value of gu does not fall below 4 sec.
conditions. The operational analysis techniques presented here are The delay computations begin with determination of the arrival
useful in estimating the performance characteristics of the intersec- and departure rates in units of vehicles per second for compatibility
tion and in providing basic insights into probable causal factors. with the remaining worksheet computations. The arrival rate is
These procedures do not, however, account for all possible con- determined by dividing the left-turn volume, v, by 3,600. This
ditions. The influences of such characteristics as specific curb- value must be adjusted to ensure that for purposes of uniform
corner radii, intersection angle, combinations of grades on various delay computation, the arrivals do not exceed the capacity of the
approaches, odd geometric features (offset intersections, narrowing intersection. If the v/c ratio, X, exceeds 1.0, the arrival rate must
on the departure lanes, etc.), and other unusual site-specific condi- be divided by X, as indicated on the worksheet.
tions are not addressed in the methodology. Field studies may be Two departure rates must be determined:
conducted in such cases to determine delay directly (see Appendix 1. The primary-phase departure rate, sp = s/3,600, where s is
III) and or to calibrate the prevailing saturation flow rate (see the adjusted saturation flow rate for the primary phase, is obtained
Appendix IV). from the Capacity Analysis Module Worksheet (Figure 9-19); and
Unusual delays may result from blockages, such as illegally 2. The secondary-phase departure rate, ss, which must be com-
parked or stopped vehicles or other factors. The analyst may also puted as
gain additional insights into intersection operations by observing
them in the field in addition to making the analyses prescribed ss = s (gq + gu)/(gu × 3,600)
in this chapter. There are also a number of more complex and
microscopic modeling techniques that could provide important where s is the adjusted saturation flow rate for the secondary phase
supplementary analyses for problems that are beyond the scope of from the Capacity Analysis Module Worksheet (Figure 9-19) and
the methods described here. the other values have already been determined as described above.
When gu is very short, the secondary-phase departures will be

mostly sneakers. Since sneakers move with very low headway, it
Supplemental Uniform Delay Worksheet is possible to have extremely high values of ss. As a practical
matter, the per-lane value of ss should not exceed the ideal satura-
Left turns from exclusive lanes that are allowed to proceed on tion flow rate for the lane group divided by 3,600.
both protected and permitted phases in the signal sequence must Next, the v/c ratios for the primary and secondary phases, Xprot
be treated as a special case for purposes of computing the uniform and Xperm, must be determined from the equations given on the
delay. Such movements are analyzed for both phases on the Capac- worksheet. Note that different equations are used for leading and
ity Analysis Module Worksheet, on which the protected phase is lagging left-turn phases. Because of the adjustment of the arrival
identified as the primary phase and the permitted phase is identified rate performed in the last step, it will not be possible for both Xprot
as the secondary phase. This terminology will be continued in the and Xperm to exceed 1.0. It will, however, be possible for one or
following description of the Supplemental Uniform Delay Work- the other to exceed 1.0. It is possible to define five separate cases
sheet, which follows the procedures outlined in Section II, for delay computation, depending on which of the X values exceed
Methodology. 1.0 and on the left-turn phasing (leading or lagging). The case
The worksheet is presented in Figure 9-21. Certain input data number, 1–5, should now be determined and entered on the
must first be obtained from other worksheets and entered here, worksheet.
namely, the adjusted left-turn volume from the Volume Adjustment When the case number is known, the size of the queue at three
Module Worksheet (Figure 9-15) and the v/c ratio, X, for the lane transition points—Qa, Qu, and Qr—may be determined from the
Figure 9-21. Supplemental Uniform Delay Worksheet for Left Turns from Exclusive Lanes with Primary and Secondary Phases.

9-50 urban streets
formulas given at the bottom of the worksheet. When these values 6. Apportion the total cycle time among the conflicting phases
have been computed and entered on the worksheet in their respec- in the phase plan on the basis of the principle of equalizing the
tive rows, it is possible to determine the uniform delay, d1, using degree of saturation for the critical movements.
the formulas given at the bottom of the worksheet. Note that the
formula is different for each of the five cases. When all of these steps have been completed, the signal timing
will be specified to the level of detail required for operational
analysis using the method given previously in this chapter.
PLANNING ANALYSIS The data to be entered on the Planning Method Input Worksheet
are self-explanatory. The following discussion covers the main
The planning analysis is intended for use in sizing the overall aspects of the lane volume and signal operations worksheets, as
geometrics of the intersection or in identifying the general capacity well as the default values. Also covered are the underlying theory
sufficiency of an intersection for planning purposes. It is based and a description of the most pertinent items.
on the sum of critical lane volumes and requires minimum input
information, mainly the demand volumes and intersection geomet-
rics. Three worksheets are provided for planning analysis. Figure Lane Volume Worksheet
9-22 is the basic worksheet on which all input information is
entered, and Figure 9-23 is the lane volume worksheet used to Description
establish the individual lane volumes on each approach.
The purpose of the lane volume worksheet (Figure 9-23) is to
Figure 9-24 is the signal operations worksheet used to synthesize
establish the individual lane volumes (vehicles per hour per lane)
the signal timing plan and to determine the operational status of
on all of the approaches. This information will be used on the
the intersection for planning purposes.
signal operations worksheet to synthesize the signal timing plan.
The lane volume worksheet contains additional items such as left-
Worksheet Operations turn treatment alternatives, parking adjustments, left-turn equiva-
lence, adjustment factors for shared lanes with permitted left turns,
The relationship between the lane volume worksheet (Figure 9- and a quick method to determine the type of left-turn protection
23) and the signal operations worksheet (Figure 9-24) is shown in if unknown. Note that the items are numbered (1–20) and that a
Figure 9-25. Note that one lane volume worksheet is required for separate worksheet must be completed for each of the four ap-
each of the four approaches. This will determine the equivalent proaches. The directional designations refer to the movements as
hourly lane volume for each approach. The hourly volumes are they approach the intersection. This is consistent with the terminol-
then combined on the signal operations worksheet to determine ogy used throughout this chapter.
the critical movement sum and the intersection status. Optionally, Computational formulas are presented on the worksheet for
the cycle length and phase times may also be determined. each data item that is computed as a function of other data
items, the step number being shown in square brackets; for
example, [11]/[12] indicates that the required data item will
Computational Requirements
result from dividing the value determined in Step 11 by the
The computations must be based on the traffic volumes and lane value determined in Step 12. Note that some data fields contain
configuration of each approach to the intersection. The steps in fixed values such as 0 or 1.0. Others are shaded to indicate
performing the analysis are as follows: that a particular value does not apply to all treatment alternatives.
This allows the same basic worksheet to be used for all treatment
1. Determine the lane volumes for each movement. The detailed alternatives.
instructions for the lane volume worksheet describe this process. For purposes of this analysis, only exclusive lanes are entered
2. Determine the type of left-turn protection for each direction. for turning movements. Shared lanes are included with the through
For planning applications, the actual left-turn protection should be lanes. Right-turn volumes from shared lanes are simply added to
used if known. A left turn is considered to be protected if it is the through volumes at one point on the worksheet. Left-turn vol-
able to proceed at some point in the cycle while the oncoming umes in shared lanes are adjusted for their through-vehicle equiva-
through movement is stopped. If the actual left-turn protection is lence, and the proportion of the shared lane that they require is
unknown, a simple method will be presented later for determining removed from the through-lane capacity. Shared lanes with ‘‘not
an appropriate choice. opposed’’ left turns are treated as shared right-turn lanes.
3. From six alternative plans, select the phase plan that will Each of the three left-turn treatment alternatives identified pre-
provide the desired degree of left-turn protection and will accom- viously must be processed differently in computing the lane vol-
modate the observed left-turn volume balance. umes. Therefore, the lane volume worksheet contains three
4. Determine the sum of the critical volumes for each phase columns, each of which represents one of the alternatives. Only
and the intersection status (under, near, at, or over capacity). one of the three columns should be used for each approach.
For planning purposes the actual left-turn treatment should be
This completes the planning portion of the analysis. If an esti-
used. If this is unknown, the choice should be made using local
mate of the level of service based on control delay is desired, it
policies or practices. A quantitative method for identifying an
is necessary to establish the signal timing plan. Two additional
appropriate treatment on the basis of the product of the volumes
steps are involved:
for left turns and opposing through movements is described in
5. Determine the cycle length that will accommodate the ob- Section II, Methodology.
served volumes with a specified degree of saturation. A saturation Failure to provide protected phasing for heavy left-turn volumes
level of 90 percent is assumed. will become evident in the operational analysis in the form of very
Figure 9-22. Planning Method Input Worksheet.
high v/c ratios for these movements. These problems will not, The lane volume worksheet does not consider the case of exclu-
however, appear in the planning-level results because unprotected sive plus shared lanes for turning movements. It is possible to
left turns are not considered in the synthesis of the traffic signal have either an exclusive lane or a shared lane for either a left or
timing plan. Therefore, failure to assume protected phases for a right turn. The case of one exclusive lane plus an optional lane
heavy left-turn volumes will generally produce an unreasonably is a complicated situation that does not lend itself to the approxima-
optimistic assessment of the critical v/c ratio. Above all, the plan- tions involved in this technique.
ning analysis presented here should never be used by itself to The treatment of shared-lane permitted left turns is a very com-
determine the need for protected left-turn phasing. plex process. It is, however, possible to approximate the signal
9-52 urban streets
Figure 9-23. Planning Method Lane Volume Worksheet.
timing parameters that will handle this situation effectively. Table through movement in the same direction must move during the
9-16 sets forth the computations for the planning-method left-turn same phase as the protected left turn. This method does not deal
factor for permitted and protected-plus-permitted operation. with simultaneous opposing left turns from shared lanes. If the
The shared-lane protected treatment alternative is only valid opposing through movement exists, the protected left turn will
when one of the two opposing left turns is protected, and the be considered protected plus permitted. If the opposing through
Figure 9-24. Planning Method Signal Operations Worksheet.

9-54 urban streets
Figure 9-25. Planning method worksheet relationships.
Table 9-16. Shared-Lane Left-Turn Adjustment timing design. This is clearly not practical, and a single-pass
Computations for Planning-Level Analysis approximation technique must therefore be sought.
permitted left turn The method presented here offers a crude approximation that
is based on the through-vehicle equivalents for left turns, obtained
Lane groups with two or more lanes: from Figure 9-7. In this model, the portion of the shared lane
[17] = {[13] − 1 + e (−[13]*[1]*[16]/600)}/[13] available to through traffic decreases as a negative exponential
Subject to a minimun value that applies at very low left-turning function of the through-vehicle equivalent of the left-turn volume.
volumes when some cycles will have no left-turn arrivals: When this value is high, the shared lane will function as a de
[17] = {[13] − 1 + e (−[1]*Cmax/3600)}/[13] facto left-turn lane. Otherwise, the through traffic will be able
to occupy a portion of the shared lane. The rate at which
Lane groups with only one lane for all movements: through-vehicle capacity is lost depends on the number of lanes.
[17] = e−{0.02* ([16]+10*[15])*[1]*Cmax/3600} As the number of through lanes increases, it is natural to expect
protected-plus-permitted left turn that through vehicles will desert the shared lane more readily.
(one direction only) The case of a single shared lane (i.e., one lane that accommo-
If [2] < 1220 dates all movements in the lane group) must be treated differently.
In this situation, the through vehicles do not have the option of
[17] = 1/{1 + [(235 + 0.435∗[2])∗[15]]/(1400 − [2])}
deserting the shared lane. Therefore it is never possible to achieve
If [2] ≥ 1220 a de facto left-turn lane regardless of the left-turn volume. A
[17] = 1/(1 + 4.525∗[15]) separate equation appears in Table 9-16 to accommodate this con-
dition. The negative exponential model is retained in this case, but
different parameters are applied to reflect captivity of the through
traffic by the shared lane. The parameters given in Table 9-16 for
movement does not exist, the protected left turn will be ‘‘not both the single-lane and multiple-lane models were selected to
opposed,’’ and therefore will move on the same phase as a permit- produce close agreement with the results of the operational analysis
ted movement. The opposing through movement may be consid- obtained by the full application of the supplemental worksheets
ered not to exist in cases of one-way streets, T-intersections, and presented earlier to specific examples.
split-phase operation. In computing the left-turn factor, it must be recognized that left
The protected-plus-permitted shared lane is handled by a turns in shared lanes have no effect on through traffic during signal
simple lookup procedure described previously as Case 6 in Table cycles in which no left turns arrive. Therefore, the minimum value
9-12. The same procedure is repeated in Table 9-16 for the of the left-turn factor is 1.0 minus the probability of zero left-
planning method. The permitted case is much more difficult turn arrivals. The minimum value for the left-turn factor is also
because it is necessary to know the signal timing, which is the determined from Table 9-16. The minimum value will occasionally
final product of the computational process described here. The govern the calculations when very low left-turn volumes are op-
operational analysis method described previously involves composed by a very heavy opposing through traffic.
plex supplemental worksheets (Figures 9-17 and 9-18) for this
Instructions
purpose that would have to be applied iteratively to resolve the
mutual dependence between the left-turn factor and the signal The following instructions cover the step-by-step procedure for
completing all of the items on the lane volume worksheet. Note 12. Parking Adjustment Factor: The parking adjustment factor
that each step is numbered to correspond with each row on the should be placed in the appropriate column, as explained in Step
worksheet. 11. This factor corresponds to the assumed value of 20 parking
maneuvers per hour applied to Table 9-8. It is dependent on the
1. Left-Turn Volume: The first item is the left-turn volume (in number of through lanes available. The values are 0.800, 0.900,
vehicles per hour) on the approach. In the case of protected-plus- and 0.933 for one, two, and three lanes, respectively. If no parking
permitted phasing with an exclusive left-turn lane, two vehicles exists, the factor equals 1.0.
per cycle should be removed from the left-turn volume to account 13. Number of Through Lanes Including Shared Lanes: This
for the effect of sneakers. If the cycle length has not been estab- step is self-explanatory. Exclusive turn lane or lanes should be
lished, the maximum cycle length should be used. To prevent excluded.
unreasonably short protected left-turn phase durations, this volume
adjustment step should not reduce the left-turn volume to a value At this point it is necessary to distinguish between exclusive
below four vehicles per cycle. left-turn lanes and shared left-turn lanes. The procedure for exclu-
2. Opposing Mainline Volume: Opposing mainline volume was sive left-turn lanes will be described first. Note that Steps 15 and
defined earlier in this chapter as the total approach volume minus 17 do not apply to exclusive left-turn lanes.
the left-turn volume from exclusive lanes or from a single lane (in
vehicles per hour). The cross product ([2] × [1]) may now be 14. Total Approach Volume (([10] + [11])/[12]): The total ap-
computed by multiplying the opposing mainline volume by the proach volume is the total of the shared lane right-turn volumes
left-turn volume. This gives a value for comparison to determine plus the through volumes. Note that the through volumes are ad-
if a protected phase should be assumed. justed (increased) by the parking adjustment factor to account for
3. Number of Exclusive Left-Turn Lanes: This would be the the effect of parking on through volumes, for example, momentary
number of lanes exclusively designated to accommodate the left- lane blockage. Note also that left-turn volumes are excluded be-
turn volumes. cause they are not a part of the lane group.
4. Left-Turn Adjustment Factor: The left-turn adjustment factor 15. Not applicable to exclusive left-turn lanes.
applies only to protected left turns from exclusive left-turn lanes 16. Left-Turn Equivalence: Left-turn equivalence, determined
or to left turns that are not opposed. This factor is derived from from Figure 9-7, is not used in lane volume calculations when
Table 9-12 as 0.95 for single lanes and is further reduced by Table exclusive left-turn lanes exist. This step is, however, required for
9-4 to 0.92 for dual lanes. If the left-turn movement is not opposed permitted left turns to assess the adequacy of the left-turn treatment
because of a one-way street or T-intersection, pedestrian interfer- in Step 20.
ence must be considered. The corresponding value of 0.85 for one 17. Not applicable to exclusive left-turn lanes.
lane and 0.75 for two lanes should be used as given in Table 9- 18. Through-Lane Volume ([14]/[13]): The total approach vol-
11 and reduced in Table 9-4. ume should be divided by the number of lanes to obtain volume
5. Left-Turn Lane Volume ([1]/([3] ∗ [4])): The total left-turn per lane, which is the basis for computing critical lane volumes.
volume from Step 1 should be divided by the product of the number 19. Critical Lane Volume: Step 19 is normally the same as Step
of exclusive left-turn lanes (Step 3) and the left-turn adjustment 18 except when the right turn has an exclusive lane or the left turn
factor (Step 4). The left-turn volume should be entered directly if is not opposed and either of these movements is more critical than
there is no exclusive left-turn lane. The result is expressed in the through movement. If both conditions apply, the critical lane
vehicles per hour per lane. Zero should always be entered if the volume will be Max ([5], [9], [18]). If a shared lane exists for the
left turns are permitted. right turn, Step 9 should be eliminated. If the left turn is permitted
6. Right-Turn Volume: Right-turn volumes (in vehicles per or protected, Step 5 should be eliminated.
hour) from either a shared through and right-turn lane or from an
The case of shared left-turn lanes is more complicated and there-
exclusive turn lane or lanes should be entered. The right-turn-on-
fore requires a more detailed procedure. Steps 14 through 18 are
red volume should be subtracted in accordance with the guidelines
used to approximate the effect that left-turning vehicles have in
presented in Section II of this chapter.
reducing available lanes for through volumes. Left-turning vehicles
7. Exclusive Lanes: This is the number of lanes assigned exclu-
blocking the shared left-turn and through lane will prevent through
sively for right turns, if any.
vehicles from proceeding until the turning vehicles have been able
8. Right-Turn Adjustment Factor: The right-turn adjustment
to make the turn.
factor is derived from Table 9-11 as 0.85 for a single lane or a
shared lane and reduced by Table 9-4 to 0.75 for two lanes. 14. Total Approach Volume: The total approach volume is com-
9,10. Right-Turn Lane Volume ([6]/([7] ∗ [8])): The total right- puted in nearly the same manner as in Step 14 for exclusive left-
turn volume from Step 6 should be divided by the product of the turn lanes, that is, ([10] + [11])/[12]. The difference is that the
number of exclusive right-turn lanes (Step 7) and the right-turn volume from Step 5 must be added to the through volume in Step
adjustment factor (Step 8). If there is no exclusive right-turn lane, 11 if the left turn is not opposed.
a value of 1.0 should be used for Step 7. The result is entered as 15. Proportion of Left Turns in Lane Group: Step 15 is self-
Step 9 if one or more exclusive right-turn lanes exist or as Step explanatory. This data item is required for the follow-up
10 if right turns must share the lane. computations.
11. Through Volume: Total through volume for the approach, 16. Left-Turn Equivalence: Determined from Figure 9-7, this
excluding left and right turns, should be placed in the appropriate is one of the factors needed to compute the applicable formulas
column to correspond with the applicable treatment for left turns from Table 9-15 for shared-lane permitted left turns. It is not used
(permitted, protected, or not opposed). at all when the left turn is protected.
9-56 urban streets
17. Left-Turn Adjustment Factor for Through Traffic fDL: The ments. A phase plan deals with only one street at a time. The
appropriate formula in Table 9-15 should be used. This is a reduc- complete signal sequence will involve two phase plans: one for
tion factor applied to the through volumes to account for the effect the east-west street and one for the north-south street. The choice
of left-turn vehicles waiting for a gap in the opposing traffic to between phase plans is made by examining the left-turn protection
make the turn. Note that for lanes that are not opposed, the factor for both pairs of opposing left turns. The alternatives include the
must be 1.0 because these vehicles will have gaps in which to turn. following:
18. Through-Lane Volume: Total through volume in the ap-
proach should be divided by the number of through lanes. Note T Plan 1: No left-turn protection in either direction. In this case,
that the number of lanes is reduced by the factor obtained in Step the phase plan includes only one phase, in which all through and
17 to account for the effect of the left-turning vehicles. left-turn movements may proceed, with the left turns yielding to
19. Critical Lane Volume: The critical lane volume is the maxi- the opposing through traffic.
mum of either the value computed by Step 18 or the right-turn T Plans 2a and 2b: These two plans involve left-turn protection
volume from an exclusive right-turn lane as computed in Step 9. for only one of the two opposing left turns. Two phases will be
20. Left-Turn Check: If one or more left turns have been desig- involved in this case. In the first phase, the protected left turn will
nated as permitted (i.e., no protected phase has been assigned), proceed with the through movement in the same direction. In the
the need for a protected phase should be reexamined at this point. second phase, the two through movements will proceed. Plans 2a
If the cross product ([2] × [1]) exceeds the adopted thresholds, a and 2b differ only in terms of which of the two opposing left turns
protected left-turn phase should be assigned for planning purposes is protected.
unless existing traffic volumes have been used and it is known T Plans 3a and 3b: Both opposing left turns are protected here.
that such a phase does not exist. In the first phase, the two opposing left turns will proceed. In
When the level of opposing traffic is such that permitted left the second, the dominant left turn will continue with the through
turns have difficulty finding acceptable gaps, the permitted left- movement in the same direction. In the third, the two through
turn capacity is derived substantially from sneakers and is therefore movements will proceed. Plans 3a and 3b differ only in terms of
limited to approximately two vehicles per cycle. For planning level the dominant left turn that governs the display in the second phase.
analysis, it should be assumed that this capacity limitation will T Plan 4: This is generally known as ‘‘split-phase’’ operation.
apply whenever the left-turn equivalence exceeds a value of 3.5. Two phases are involved, with the through and left-turn move-
Therefore, if the left-turn equivalence [16] is greater than 3.5 and ments from one of the two opposing directions proceeding on each
the left-turn volume is greater than two vehicles per cycle (i.e., phase. This has the effect of full directional separation between
[1] > 7,200/Cmax), it is most likely that the subject left turn will the two approaches. From a capacity analysis point of view, it is
not have adequate capacity without a protected phase. equivalent to two one-way streets that meet at a common point.
The selection criteria are presented in a table on the signal

Signal Operations Worksheet operations worksheet. Note that the selection is made on the basis
of the user-specified left-turn protection and the dominant left-turn
Of the six steps involved in the planning method, only the first movement identified from the lane volume worksheet.
two are carried out by the lane volume worksheet. The last four 3. Critical Phase Volume, CV: When the phase plan has been
steps are included in the signal operations worksheet, which is selected, the movement codes, critical phase volumes (CVs), and
shown in Figure 9-24. To facilitate the use of the signal operations lost time per phase may be entered on the worksheet. The appro-
worksheet, the lane volumes are transferred from the lane volume priate choice for critical lane volumes is given in the phase plan
worksheet before the computations begin. Note that the through- summary shown in Table 9-17, along with a code that identifies
movement lane volume is taken as the heavier of the through or the movements that are allowed to proceed on each phase. The
right-turning movement when an exclusive right-turn lane is pres- movement codes are defined in a note to Table 9-17. For example,
ent. In other words, if the volume of a right turn from an exclusive ‘‘NST’’ indicates that the northbound and southbound through
lane is heavier than that of the through movement, the right-turn movements have the right-of-way on the specified phase. The
lane volume will be considered as the through volume for design corresponding code for the two opposing left turns moving concur-
purposes. rently is ‘‘NSL.’’ If the northbound through and left turns are
1. Transcribed Data Items: The peak-hour factor (PHF) was moving together, the code is ‘‘NTL.’’ Note that Table 9-17 also
entered on the Planning Method Input Worksheet. The appropriate indicates the lost time to be assigned to each phase.
value is discussed in connection with the description of that work- Thus, the movement codes and CVs must be determined for
sheet. The left-turn treatment is also transcribed to the signal opera- each phase from Table 9-17 and entered on the signal operations
tions worksheet from the input worksheet. Note that it is not worksheet. When all phases have been completed, the critical sum
necessary to specify whether the treatment includes a permitted (CS) of the CVs must be entered on the next line.
phase for the left turn in addition to a protected phase. The synthe- 4. Lost Time Determination: For planning purposes, it is as-
sis of the signal timing plan does not consider protected-plus- sumed that there is a lost time value of 4 sec per phase in which
permitted operation. That, of course, does not preclude specifica- any movement is both started and stopped. For one- and two-phase
tion of this type of operation in the analysis. At this time, only plans, there is a lost time associated with each phase. For three-
determination of reasonable values for the cycle length and phase phase plans (Plans 3a and 3b), the second phase requires no lost
times is of interest. time because none of the movements are both started and stopped.
2. Phase Plan Selection: The phase plan is selected from six Thus, as a simple rule, phase Plan 1 involves 4 sec of lost time
alternatives that cover the full range of left-turn protection require- per cycle, and all other plans require 8 sec.
Table 9-17. Phase Plan Summary for Planning Analysis

east-west north-south
phase plan phase no. lost time movement code critical sum movement code critical sum
1 1 4 EWT Max(ET,EL,WT,WL) NST Max(NT,NL,ST,SL)
2a 1 4 WTL WL STL SL
2 4 EWT Max(WT-WL,ET) NST Max(ST-SL,NT)
2b 1 4 ETL EL NTL NL
2 4 EWT Max(ET-EL,WT) NST Max(NT-NL,ST)
3a 1 4 EWL WL NSL SL
2 0 ETL EL-WL NTL NL-SL
3 4 EWT Max(WT,ET-(EL-WL)) NST Max(ST,NT-(NL-SL))
3b 1 4 EWL EL NSL NL
2 0 WTL WL-EL STL SL-NL
3 4 EWT Max(ET,WT-(WL-EL)) NST Max(NT,ST-(SL-NL))
4 1 4 ETL Max(ET,EL) NTL Max(NT,NL)
2 4 WTL Max(WT,WL) STL Max(ST,SL)
Note: EWT = eastbound and westbound through; ETL = eastbound through and left; WTL = westbound through and left; NST = northbound and southbound
through; STL = southbound through and left; NTL = northbound through and left; ET = eastbound through; EL = eastbound left; WT = westbound through; WL =
westbound left; NT = northbound through; NL = northbound left; ST = southbound through; SL = southbound left.
When the lost times have been determined for each phase, the cycle. The phase times for all of the phases should be equal to the
total lost time per cycle (TL) may be computed and entered on the cycle length and should be entered on the last line of the worksheet.
worksheet.
5. Critical v/c Ratio, Xcm: The planning-level critical v/c ratio,
Xcm, is the ratio of the critical sum, CS, to the sum of the critical Limitations of Planning Method
lane volumes that could be accommodated at the maximum cycle
length, computed as
The planning analysis technique described in this chapter offers
(1 − TL /Cmax) * 1,900 * CBD * PHF a method for synthesizing a reasonable and effective signal timing
The intersection status is determined directly from Xcm using the plan based on the traffic volumes and lane utilization at an intersec-
threshold values given in Table 9-15. tion. It is possible using the worksheets included here to determine
6. Timing Plan Development: The development of a timing plan the approximate status of the operation of a signalized intersection
is optional. For many planning applications, a knowledge of the with respect to its capacity.
intersection status is sufficient. The timing plan is only required It is also possible to take the analysis considerably farther and
if the planning analysis is to be extended to estimate the level of obtain the level of service for each lane group by the operational
service. analysis method. Software has already been developed that will
The cycle length may be determined from the following formula: implement the worksheets and invoke the operational analysis
method. This development introduces a very powerful capability,
TL one from which the numerical precision of the results may greatly
C=
1 − [Min(CS,RS)/RS] exceed the accuracy of the original data. In particular, great caution
where RS is the reference sum of phase volumes representing the should be employed when traffic volumes are projected to some
theoretical maximum value that the intersection could accommo- point in the future. Unless there is strong confidence in the validity
date at an infinite cycle length. of the traffic data, this method should not be taken beyond the
The recommended value for the reference sum is (1,710 * PHF). worksheet stage.
This value should be reduced by 10 percent in CBD locations. Caution must also be used in interpreting the results of the
The value of 1,710 is 90 percent of the ideal saturation flow rate operational analysis, even with reliable traffic data. In particular,
of 1,900 pcphgpl. It will attempt to produce a 90 percent v/c ratio it must be recognized that the overall intersection level of service
for all critical movements. The cycle length determined from this represents the average of all approaches to the intersection. When
equation should be checked against reasonable minimum and max- highly directional peak periods are involved, the relatively inconse-
imum values. The determination of appropriate values is discussed quential movements on the lightly traveled approaches will have
in connection with the Planning Method Input Worksheet. minimal delay. Thus, it is possible to see a favorable average delay
The lost time per cycle must be subtracted from the total cycle and level of service for the intersection even when the critical
time to determine the effective green time per cycle, which must approaches are heavily congested.
then be apportioned among all the phases. This is based on the The planning-level critical v/c ratio Xcm is used as an indicator
proportion of the critical phase volume sum for each phase deter- of the status of the intersection with respect to its capacity. This
mined in a previous step. The phase time should be entered on measure may also be used as an indicator of the additional demand
the worksheet. volume that could be accommodated. Although lower values of
As a final step, the lost time must be added to the effective Xcm indicate that larger increases in demand volumes could be
green time for each phase to determine the total phase time per absorbed, it is important to realize that the relationship is not linear.
9-58 urban streets
Therefore, linear projections of the maximum allowable increase In Figure 9-26(b), the signal timing for a given level of service
in demand volumes based solely on Xcm might not be accurate. (delay) is desired. In this case, computations through the Saturation
Flow Rate Module are performed in the normal sequence. As in
all signal timing exercises, the phase plan must be established
PROCEDURES FOR OTHER ANALYSES before computations are made. As indicated in Figure 9-26(b),
however, determination of the signal timing for a given level of
As noted in Section II, Methodology, by starting with a known service requires some iterative calculations. This is because signal
or desired level of service, it is possible to sequence the computa- timing affects both capacity and delay, whereas capacity also af-
tions of the operational analysis procedure to solve for three un- fects delay. Further, the delay equations include g/C, C, c, and X,
knowns: (a) v/c ratios, service flow rates, or both; (b) signalization; all of which are influenced by signal timing. Thus, no one variable
or (c) geometric features. In such computations, the steps of an can be directly computed without checking its effect on the others.
operational analysis are rearranged in recognition of the fact that In this approach, signal timing is estimated on the basis of the
level of service, and therefore average stopped delay per vehicle, recommendations of Appendix II or local practice, and iterations
is a known quantity. Given knowledge of any two of the other are pursued to produce the desired delay value.
three unknowns noted above, the remaining variable may then be In Figure 9-26(c), the number of lanes in a given lane group is
calculated. Solutions for any of the above may be handled through to be computed. This is also an iterative process. For any given
iterative computations using the standard sequence of calculations. signal timing, the capacity of the lane group may be estimated
Delay results are then tabulated versus various trial values of the using the delay equations (with c as the unknown). The delay
variable of interest. equations, however, also require v/c ratios that depend heavily on
It is also possible, though computationally difficult, to work capacity. Therefore, once again it is more practical to iterate the
backward through the procedure, starting with a known delay. This number of lanes, comparing the resulting delay for several trial
is complex because relationships deal primarily with individual values.
lane groups, and changes to one virtually always imply changes The relative complexity of these other approaches makes a man-
in the operation of others at the intersection. Further, geometric ual solution difficult, and therefore the operational analysis proce-
and signalization parameters must often change in relation to one dure is presented in the mode of solving for level of service. A
another, such as an exclusive left-turn phase requiring an exclusive sample calculation is included, however, illustrating how these
left-turn lane. Nevertheless, reverse computations are feasible and alternative approaches may be accomplished.
are best carried out using computer programs designed by the As with any analysis, v/c ratio and level of service must be
analyst for the specific objective. considered as two important measures of performance. Any analy-
Figure 9-26 illustrates the computational path for such alterna- sis yielding v/c ratios exceeding 1.0 should immediately trigger
tive analyses. In Figure 9-26(a), a v/c ratio or service flow rate is consideration of alternatives. High v/c ratios in the 0.95 to 1.0
calculated for a given level of service. Calculations are made in range may also cause such consideration. This is an important
the normal sequence through the computation of capacity for each point that can save a good deal of analysis effort. In many analyses
lane group. Delay equations, however, are solved for a known [Figure 9-26(b) and (c)], v/c ratios will be obtained before delays
delay commensurate with the selected level of service with the and level of service. If an intersection is operating in an unaccept-
v/c ratio (X) as the unknown. Service flow rates may be computed able v/c range, completing computations to find delay and level
as the v/c ratio times the capacity of the lane group. of service may be a fruitless exercise.

Figure 9-26. Alternative computations using operational analysis.

9-60 urban streets
This section presents six numerical examples that illustrate the There is no parking within the confines of the intersection, and pedes-
computational principles set forth in this chapter. Sample Calcula- trian flows average 100 peds/hr/crosswalk.
tions 1 through 5 demonstrate the estimation of delay with given The computations for each module of the procedure are de-
values for all the required field data items and operating parame- scribed below in detail.
ters. Sample Calculation 6 demonstrates the reverse process of
seeking the maximum traffic volume that may be accommodated Input Module Worksheet for Calculation 1
within a specified level of service.
A wide range of operational configurations, from simple to com- Most of the information on the Input Module Worksheet for
plex, is represented in the examples. In some cases, the timing Calculation 1 (Figure 9-27) is given. One item, however, must
plan as well as certain operating parameters, such as the type of be calculated—the minimum green time for pedestrians, Gp, is
left-turn protection, must be determined as a part of the exercise. computed as
When data items are not specified, the default values given in
Table 9-3 are used. Gp = 7.0 + (W/4.0) − Y
The planning method described earlier in this chapter will be where W is the width to be crossed and Y is the yellow-plus-all-
used for all computations of timing plan parameters (cycle length red interval. Common practice is to take W as the distance from
and phase times) required by the sample calculations. Minimum the curb to the midpoint of the farthest lane to be crossed. For the
cycle lengths of 60, 70, and 80 sec will be applied to two-, three-, Main Street green (crossing Third Avenue), this is about 23 ft.
and four-phase operation, respectively. Minimum phase times are For Third Avenue (crossing Main Street), this is about 39 ft. Then
generally determined by pedestrian requirements with assumed
walking speeds. Absolute minimums of 10 and 15 sec per phase Gp (Main) = 7.0 + (23/4.0) − 4 = 8.8 sec
(including change and clearance intervals) have been imposed to Gp (Third) = 7.0 + (39/4.0) − 4 = 12.8 sec
provide a consistent treatment among sample problems. The total
intergreen time (yellow plus all red) will be assumed to be 4 sec These values are entered in the appropriate boxes on the worksheet.
for each phase. RTOR volumes are assumed to be zero.
The use of a particular design configuration or parameter does Volume Adjustment Module Worksheet for Calculation 1
not imply endorsement of its suitability for field implementation
under all conditions. Many agencies have their own policies and The computations for the Volume Adjustment Module Work-
practices regarding design configurations and parameters. It is not sheet for Calculation 1 are shown in Figure 9-28. Each approach
the intent of this section to influence these policies or practices has one lane group that will be carried through the entire analysis.
nor to prescribe design procedures but simply to illustrate the The hourly volumes are divided by the PHF to provide peak flow
computational principles set forth here for evaluating delay and rates for subsequent computations.
level of service. Proportions of left- and right-turning traffic are found by divid-
The worksheets shown in Section III and Appendix V of this ing the appropriate turning flow rates by the total lane group
chapter will be used to illustrate all of the computations. Each of flow rate.
the main worksheets will be presented in graphic form the first
time it appears in this section. To conserve space and make the
Saturation Flow Rate Module Worksheet for Calculation 1
computations easier to follow, tabular equivalents will be used
when appropriate for all subsequent presentations of the same
The computations for the Saturation Flow Rate Module are
worksheet, and worksheets that are not essential to the discussion
shown in Figure 9-29. Note the following entries:
will be omitted. Because of their complex nature, supplemental
worksheets and planning method worksheets will always be pre- 1. Lane width adjustment factors are obtained from Table 9-5.
sented in tabular form. For EB and WB approaches, 11-ft lane widths result in a factor
of 0.967, and for the NB and SB approaches, with 15-ft lanes, the
factor is 1.10.
2. The heavy vehicle adjustment factors of 0.952 and 0.926 are
CALCULATION 1: OPERATIONAL ANALYSIS OF EXISTING
PRETIMED, TWO-PHASE SIGNAL
obtained from Table 9-6 and reflect 5 and 8 percent heavy vehicles
present in each lane group.
3. Grade (level), parking conditions (none), and local bus traffic
The intersection of Third Avenue and Main Street is located in the (none) are all ideal at this intersection, and therefore each has a
central business district (CBD) of a small urban area. Figure 9-27 is factor of 1.00, which can be verified by consulting Tables 9-7,
the Input Module Worksheet for the problem, which illustrates the 9-8, and 9-9, respectively.
geometry and flows that exist at the intersection. Third Avenue is a 4. The area-type adjustment factor is 0.90, reflecting the CBD
two-lane street and Main Street is a four-lane arterial. The signal location of the intersection, as given in Table 9-10.
has a simple two-phase sequence, with phase times as shown on the 5. The lane utilization adjustment factor is applied here, at least
worksheet. There are 5 percent heavy vehicles on all movements of initially, so the analysis will seek to establish the conditions in the
the eastbound (EB) and westbound (WB) approaches and 8 percent worst lane within each lane group. If this factor were not applied,
on all movements of the northbound (NB) and southbound (SB) ap- the results would reflect the average of all lanes of the defined lane
proaches. The peak-hour factor (PHF) is 0.90 for all movements. groups. The lane utilization factor is 1.0 for the single-lane ap-
Figure 9-27. Input Module Worksheet for Calculation 1.

9-62 urban streets
Figure 9-28. Volume Adjustment Module Worksheet for Calculation 1.
proaches (NB and SB) and 0.95 for the two-lane approaches (EB Computations for this example follow the worksheet exactly
and WB). with no special cases. The EB and WB approaches are multilane
6. Right-turn adjustment factors are obtained from Tables approaches opposed by multilane approaches and use the work-
9-11a and 9-11b. Both EB and WB right turns fall under Case 5, sheet in Figure 9-30a. The NB and SB approaches are single-
and both NB and SB right turns fall under Case 7. The factors lane approaches opposed by single-lane approaches and use the
are based upon the proportion of right turns being made and the worksheet shown in Figure 9-30b to determine the left-turn adjust-
pedestrian crosswalk flow with which they conflict. ment factor. It should be remembered that the value of fm must be
7. Because all left turns are permitted, the special procedure converted to fLT for multilane approaches. This conversion is done
and worksheet for such turns must be used to obtain the left-turn on the last line of the worksheet.
adjustment factor. The top of each worksheet contains the input information
transcribed from the Input Module Worksheet. The Rpo values
The supplemental worksheets for permitted left turns (Figures are determined from the arrival type using Table 9-2. It
9-17 and 9-18) must be implemented with some care, because they should be remembered that each column refers to the value of
are complex and may involve a number of special cases. The Rpo for the opposing flow. Thus, for the EB column, the WB
current problem is the most straightforward case for which this value of Rpo is used, and for the WB column, the EB value of
worksheet is used, that is, permitted turns from a shared lane in Rpo is used.
which the green indications for the subject left turn and the oppos- When all factors are entered onto the worksheet, the ideal satura-
ing traffic are displayed simultaneously. tion flow rate of 1,900 pcphgpl is multiplied by the number of
Figure 9-29. Saturation Flow Rate Module Worksheet for Calculation 1.
lanes in the lane group and by each of the nine adjustment factors is critical, as is the SB lane group, with a v/s of 0.455. The sum
shown on the worksheet. The result is the prevailing adjusted of critical v/s ratios is therefore 0.378 + 0.455 = 0.833.
saturation flow rate for each approach. From this determination and the known signal timing parame-
ters, the critical v/c ratio, Xc, can be computed. Note that for a
simple two-phase signal and an assumption of 4.0 sec of lost time
Capacity Analysis Worksheet for Calculation 1 per movement, L = 8.0 sec/cycle and the resulting Xc = 0.940.
Green ratios are entered onto the worksheet by dividing the
The capacity analysis computations for this problem are shown effective green times by the 70-sec cycle length. Lane group capac-
in Figure 9-31. Lane group volumes are entered from the Volume ities and v/c ratios may then be computed as shown on the
Adjustment Module Worksheet, and saturation flow rates are entered worksheet.
from the Saturation Flow Rate Module Worksheet. With these two The results of the Capacity Analysis Module should be studied
values, v/s ratios can be computed and entered onto the worksheet. carefully for insights into operational problems, should they exist.
Once v/s ratios are obtained, the critical lane groups must be In this case, the eastbound lane group v/c ratio is above 100 percent,
identified. Since there are no overlapping phases to consider, the indicating oversaturation. This problem will require further consid-
highest v/s ratio between EB and WB approaches defines one eration, but the remainder of the computations will be completed
critical lane group, and the highest v/s ratio between NB and SB first. Note that the v/c ratios for the two critical lane groups are
approaches defines the other. The EB lane group, with v/s of 0.378, not equal, indicating that green time is not proportionally allocated.
9-64 urban streets
Figure 9-30a. Supplemental left-turn worksheets for EB and WB approaches (multilane).

Figure 9-30b. Supplemental left-turn worksheets for NB and SB approaches (single lane).

9-66 urban streets
Figure 9-31. Capacity Analysis Module Worksheet for Calculation 1.

Figure 9-32. LOS Module Worksheet for Calculation 1.
The EB critical lane group has a v/c ratio of 1.017, whereas the delay, progression adjustment factor, and incremental delay are
SB critical lane group has a v/c ratio of 0.885. A reallocation of implemented.
green time may be considered but should not be made without The resulting lane group delays vary from LOS B to E on the
consideration of the results of the delay computations in the next basis of the criteria in Table 9-1. Because there is only one lane
module. group per approach, approach delays and levels of service are the
same as the lane group delays and levels of service.
LOS Module Worksheet for Calculation 1
The overall intersection delay is computed as 35.4 sec, resulting
Level of service and delay are determined on the worksheet in an intersection level of service of D.
shown in Figure 9-32. Key values needed for this computation are In general, the intersection operation is marginal and could be
entered from previous worksheets, and the formulas for uniform improved. Note that the EB v/c ratio of 1.017 indicates that sub-
9-68 urban streets
Figure 9-33. Saturation Flow Adjustment Module Worksheet with no lane utilization factor for Calculation 1.
Figure 9-34. LOS Module Worksheet with no lane utilization factor for Calculation 1.
stantial queueing will take place on this approach during the peak from 1.017 to 0.945 and the delay dropped to 39.7 sec/vehicle.
15 min of the peak hour. The overall allocation of green appears The overall intersection delay is the weighted average of the ap-
to result in an inequitable service to vehicles on all approaches. proach delays with the approach flow rates used as the basis for
Some further consideration of the operation at this intersection weighting and is 29.7 sec/veh.
could be considered. The net effect was an improvement in the overall intersection
It should be recalled that because the lane utilization adjustment level of service to LOS C. If it can be accepted that drivers on
factor was used, the results reflect the operation in the worst of the EB approach would change lanes to maintain equilibrium in
two lanes. Also, it was suggested earlier in this chapter that multi- the lane distribution rather than suffering considerable extra delay
ple lanes tend to be much more evenly utilized at high v/c ratios. per vehicle, it is reasonable to conclude that the revised analysis
It would therefore be quite appropriate to repeat the analysis with- is appropriate.
out the lane utilization factors, that is, to consider the operation The given signal timing of 40 sec for the NB and SB traffic
as averaged over all lanes in the lane group. This, of course, would and 30 sec for the EB and WB traffic (total phase time) has not
make no difference to the NB and SB approaches because only balanced the delay or v/c ratios among the competing movements.
single lanes are involved. However, saturation flow rates on the A more equitable design would require a small amount of time to
EB and WB approaches were decreased by 5 percent because of be taken from the N-S phase and given to the E-W phase. This can
the two-lane approaches. The effect of eliminating this adjustment only be accomplished using an iterative trial-and-error procedure.
is seen in the Saturation Flow Rate Module Worksheet with no Although this would normally be carried out with one of several
lane utilization factor (Figure 9-33). Note that the EB saturation available traffic signal timing design programs that implement the
flow, which was critical in the previous analysis, has been in- methodology of this chapter, it is possible, given enough time, to
creased from 2,118 to 2,281 vphg. arrive at a manual solution using the worksheets.
The effect of this increase is propagated through all of the work- Since the arrival types specified on the Input Module Worksheet
sheets and is evident in the LOS Module Worksheet with no lane indicate that this intersection is part of a coordinated system, it is
utilization factor (Figure 9-34). Note that the EB v/c ratio dropped logical to retain the 70-sec cycle throughout the process of reallo-
Figure 9-35. LOS Module Worksheet with timing modifications for Calculation 1.
cating green time. It can be demonstrated that an equal v/c solution analysis of this phasing will be a key part of this example. Note
would be obtained with total phase times of 39.2 and 30.8 sec to that signal timing is not specified and must be estimated as part
the N-S and E-W phases, respectively. The final LOS Module of the solution.
Worksheet shown in Figure 9-35 indicates that the v/c ratios for 4. The volumes shown represent future conditions, and the de-
the two critical movements (SB and EB) will be equalized at 0.905 tails of signal progression are not yet known for this case. As a
under these conditions. base for analysis, random arrivals will be assumed, in which case
At this point, the operation could be considered acceptable. The the appropriate arrival type is 3.
v/c ratios are balanced on the critical approaches. There is no All other conditions shown on the Input Module Worksheet are
apparent need for phasing changes or geometric improvements. straightforward. Sixth Street NB is on a 2 percent downgrade and
has parking on both sides of the street. The parking activity of 20
CALCULATION 2: OPERATIONAL ANALYSIS OF THREE- movements per hour represents the total number of movements on
PHASE, PRETIMED SIGNAL both sides of the street, which is the appropriate value to use in de-
termining the total impact of both parking lanes on saturation flow.
The intersection of Sixth Street and Western Boulevard is shown
on the worksheet in Figure 9-36. Sixth Street is a one-way local
Volume Adjustment Module Worksheet for Calculation 2
street in an outlying area, and Western Boulevard is a four-lane
arterial. Because Sixth Street is one way, signalization must ad-
dress the existence of left turns (in this case, a heavy movement) The Volume Adjustment Module Worksheet for this calculation
in one direction only on Western Boulevard by providing an exclu- is shown in Figure 9-37. All items are straightforward for these
sive left-turn lane and protected-plus-permitted phasing for the EB computations. Movement volumes are divided by the PHF to find
left turn. peak demand flow rates. Determination of lane groups is also
The intersection is to be analyzed for the impact of volumes straightforward: the EB left-turn lane must be established as a
expected as a result of new development in the vicinity. The indi- separate lane group, with remaining EB lanes forming a second
vidual computational modules for this problem are discussed in lane group; WB and NB approaches form one lane group each.
the sections that follow. There are no de facto left-turn lanes because there are no opposed
left turns from the WB or NB approaches. Left-turn and right-turn
proportions are found by dividing the appropriate turning flow rate
Input Module Worksheet for Calculation 2 by the total lane group flow rate.
Several items are worthy of note on the Input Module Worksheet

for Calculation 2 (Figure 9-36): Saturation Flow Rate Module Worksheet for Calculation 2
1. Sixth Street is a one-way street as noted. Thus, left turns Figure 9-38 shows the Saturation Flow Rate Module Worksheet
from Sixth Street do not have an opposing flow. Their principal for this calculation. All adjustment factors are obtained directly
conflict is with pedestrians, as is the case for right turns. Therefore, from the appropriate tables except for the left-turn adjustment
in the selection of a left-turn adjustment factor for Sixth Street, factor for the EB left-turn lane group, which is complex. The other
left turns will be treated using the right-turn adjustment factor factors are found as follows:
table (Tables 9-11A and 9-11B).
2. Western Boulevard has local bus traffic stopping within the 1. All lane widths are 12 ft, which is the ideal condition. There-
confines of the intersection. An appropriate adjustment factor will fore, fw = 1.00 for all lane groups.
account for this condition. 2. There are 10 percent trucks in each EB and WB lane group
3. The signal timing provides for a protected-plus-permitted and 5 percent in the NB lane group. From Table 9-6, fHV = 0.909
left-turn phase from an exclusive lane in the EB direction. The and 0.952, respectively.
9-70 urban streets

3. The grade is level for the EB lane groups and the WB lane values of fLT must be computed for the protected and permitted
group ( fg = 1.00). The NB lane group is on a 2 percent downgrade. portions of the movement. Since no signal timing was given for
From Table 9-7, fg = 1.01. this example, it will be necessary to estimate the cycle length and
4. There is no parking on the EB and WB lane groups ( fp = phase times before proceeding further. The planning methodology
1.00). On the NB lane group, there is parking with 20 movements/ will be used for this purpose. Since this is a three-phase operation,
hr. From Table 9-8, fp = 0.90. the cycle length range will be 70 to 100 sec as prescribed earlier
5. There are 20 local buses per hour stopping within the EB in this section.
through lane group and the WB lane group. From Table 9-9, fbb = The input data for the planning method normally come from
0.96 for these groups. For the other lane groups, with no bus the simplified Planning Method Input Worksheet presented in Fig-
activity, fbb = 1.00. ure 9-22. That worksheet will be omitted from this discussion
6. The area type is not a CBD. From Table 9-10, fa = 1.00 for because all of the information may be found in the Input Module
all lane groups. Worksheet for Calculation 2 (Figure 9-36).
7. There are no right turns from the EB left-turn lane group or The planning method worksheets for this calculation are shown
the EB through lane group. Thus, fRT = 1.00 for these cases. For in Figures 9-39 (lane volume worksheet) and 9-40 (signal opera-
the WB and NB lane groups, Tables 9-11a and 9-11b are used tions worksheet).
with the proportion of right turns and 50 peds/hr in conflicting For simplicity, the results of the computations for all the lane
movements to obtain the factors shown on the worksheet. volume worksheets are shown on a single table in Figure 9-39. It
8. The left-turn adjustment factor for the NB approach is ob- should be kept in mind that a manual implementation of this proce-
tained from Table 9-12, treating left turns from the one-way street dure using the worksheet originally presented in Figure 9-23 would
as right turns. For the EB through and WB lane groups, there are require a separate worksheet for each direction. Because of the
no left turns, and fLT = 1.00. one-way street, only the NB and EB left turns exist. The following
data items merit some discussion:
The left-turn adjustment factor for the EB left-turn lane group
is more complex. Since it involves permitted left turns, the special 1. One of the first items to be specified is the left-turn treatment
procedure for such cases must be applied. Moreover, since a pro- type, which includes both the signal protection (protected, permit-
tected left-turn phase is also provided for this movement, separate ted, or not opposed) and the left-turn lane assignment (shared or
9-72 urban streets
Figure 9-39. Lane Volume Worksheet for Calculation 2.
exclusive). Only the EB left turn has an exclusive lane. The NB transferred to the signal operations worksheet shown in Figure 9-
left turn operates from a shared lane, and the other two left turns 40. The process for determining the signal timing is as follows:
do not exist. The EB left turn is protected, and the NB left turn
1. The phase plane for N-S movements and for the E-W move-
is not opposed. For purposes of this worksheet, all movements
ments is determined from the type of left-turn protection for each
that do not exist should be considered permitted, because they
movement. A choice of six phase plans is available; the selection
require no special treatment in the signal phasing.
criteria are given in Table 9-16. The appropriate choices are Plan
2. There is no need for left-turn adjustment factors (Figure 9-
2b for the E-W approaches and Plan 1 for the N-S approaches.
23, line 17) here. Left-turn adjustment factors apply only to permit-
Note that all movements are assumed to exist in Table 9-16. Those
ted left turns from shared lanes in the planning analysis. Permitted
movements that do not actually take place in the field are consid-
left turns from exclusive lanes are not represented in the signal
ered to exist with zero volume. This simplifies the selection process
timing plan synthesis. Protected left turns are adjusted in line 4 of
considerably.
the worksheet in Figure 9-23.
2. The full sequence of phases is then established by specifying
3. The critical lane volumes for the lane groups with through
the movements that proceed on each phase using the codes in Table
traffic are based on the through volume in each case. If the right-
9-16. The three phases are labeled ‘‘ETL,’’ ‘‘EWT,’’ and ‘‘NST.’’
turn lane volume on any approach were heavier than the through
This is consistent with the information presented on the Input Module
volume, the right-turn volume would be considered critical.
Worksheet for Calculation 2 (Figure 9-36). Note that the ‘‘NST’’
The critical through and right-turn lane volumes of 490, 409, code used for the third phase assumes that SB traffic would be al-
and 474 vph for the EB, WB, and NB lane groups, respectively, lowed to move on that phase if it existed. This greatly simplifies the
and the lane volume for the EB left turn (126 vph) must now be analysis without affecting the generality of the results.
Figure 9-40. Signal Operations Worksheet for Calculation 2.
3. The critical phase volume (CV) for each phase is now deter- below its capacity. It would be expected that this fact would be
mined using the critical sums given in Table 9-16 for the chosen reflected in the operational analysis when it has been completed.
phase plan. Second, the timing plan synthesized by the planning method ap-
4. The critical sum (CS), determined as the sum of the critical pears to satisfy all of the minimum green requirements and should
phase volumes for all phases, is 1,009 vph. therefore be considered to be reasonable for implementation, pro-
5. The critical v/c ratio Xcm is determined as 0.64 using the viding that the assumed minimum greens are acceptable to the
formula given in Note 4 of the signal operations worksheet (Figure operating agency. Note that this is an initial signal timing estab-
9-24). lished to allow the analysis to continue to completion. When v/c
6. The intersection status, obtained from Table 9-14, is under ratios and levels of service are established, the timing may have
capacity. to be reconsidered and altered.
7. The reference sum (RS) is determined using the formula Once the signal timing has been established, the value of fLT
given in Note 6 of the signal operations worksheet as 1,624 vph. must be calculated using the special worksheet for permitted left
This is the value of the critical sum that could be accommodated turns, which is illustrated in Figure 9-41 for this calculation. Note
at an infinite cycle length with 90 percent v/c ratio. This value the following input information:
will be used later in the estimation of the cycle length.
1. The total lane group flow rate is entered as 126 vph from
8. The lost time for each phase is determined from Table 9-16
the Volume Adjustment Module Worksheet.
on the basis of the selected phase plan.
2. The proportion of left turns in the EB left-turn lane group
9. The cycle length is determined using the formula given in
flow is 1.0.
Note 7 of the signal operations worksheet. Because of the low
3. The cycle length is 70 sec, as estimated above.
v/c ratio, the specified minimum cycle length of 70 sec will govern
4. There is no lost time applied for the permitted phase here,
in this example.
because the EB left turn is already proceeding as a protected move-
10. The phase times are determined using the formula given in
ment at the beginning of the permitted phase. This is a special
Note 8 of the signal operations worksheet. The phase times are
case that applies to multiple-phase left turns. Normally a lost time
11.2 sec (ETL), 27.5 sec (EWT), and 31.3 sec (NST).
would be applied to the permitted phase.
Two important observations may be made from the planning 5. The adjusted opposing flow is 842 vph, the total flow in
worksheets. The first is that the intersection should operate well the WB lane group, taken from the Volume Adjustment Module
9-74 urban streets
Figure 9-41. Supplemental left-turn worksheet for Calculation 2.
Worksheet. The proportion of left turns in this opposing flow tered in the fourth column from the Saturation Flow Rate Module
is zero. Worksheet. These are then used to compute the v/s ratios for each
6. The opposing platoon ratio is 1.0 because of the assumption lane group.
of Arrival Type 3. This makes the opposing queue ratio equal to Critical lane groups must now be identified. The NB lane group,
the proportion of red time for the opposing movement, or 0.66. with a v/s ratio of 0.290, is clearly critical, because it is the only
7. There is one lane in the EB left-turn lane group and there one moving during Phase 3 of the sequence, which has no overlaps.
are two opposing lanes. For the purpose of critical lane group determination, the protected
8. For exclusive lanes, gf is set at 0.0, because the first vehicle and permitted phases of the EB left-turn lane group should be
in the queue is always a left-turning vehicle. considered as separate lane groups. The EB left-turn primary phase
9. The portion of green time blocked by oncoming traffic, gq, is the only lane group moving exclusively in Phase 1. The EB
is determined by the equation given on the worksheet to be 15.19 left-turn secondary lane group and the WB lane group move exclu-
sec. The remainder of the green time, or 12.31 sec, is the unsatu- sively in Phase 2. The WB v/s (0.273) is greater than that of
rated green time, gu, and is considered to be available for left turns the EB left-turn secondary, and therefore would be chosen. The
to filter through the oncoming traffic. combination of EB left-turn (primary) and WB gives a v/s sum of
10. Subsequent computations of fm and fLT are made as indicated 0.35. The EB through lane group moves in both Phases 1 and 2
on the worksheet. For a one-lane group, fm = fLT. and has a v/s ratio of 0.328. The EB left-turn (primary) lane group
and the WB lane group are critical. Thus, the sum of critical lane
The value fLT may now be entered on the Saturation Flow Rate
group v/s ratios is 0.290 + 0.077 + 0.273 = 0.640.
Module Worksheet, and the prevailing saturation flow rates for
The determination of lost time per cycle is simple for this case.
each lane group are computed as in Equation 9-10:
All the critical lane groups effectively start and stop in a single
s = so N fw fHV fg fp fbb fa fLU fRT fLT phase. Each of the critical lane groups contributes 4 sec to the
total lost time, to give a lost time per cycle of 12 sec. Had the EB
The results are shown in the last column of the Saturation Flow
left-turn secondary been critical instead of the WB, it would be
Rate Module Worksheet for Calculation 2 (Figure 9-38).
assumed that the left turn did not stop at the interface of Phases
1 and 2 and therefore would not add any additional lost time to
Capacity Analysis Module Worksheet for Calculation 2 the sum. Using the critical lane group values, the critical v/c ratio
for the intersection, Xc, is computed as 0.640 × (70)/(70 − 12) =
The capacity analysis worksheet for this calculation is shown 0.772.
in Figure 9-42. Lane group flow rates are entered in the third Green ratios g/C are now entered in the sixth column of the
column of this worksheet from the Volume Adjustment Module worksheet. The EB left-turn secondary phase is assumed to experi-
Worksheet, and prevailing adjusted saturation flow rates are en- ence no lost time. For this lane group, g/C = 0.393.
Lane group capacities are then computed as s(g/C), and lane Because of the protected-plus-permitted EB left turn, a special
group v/c ratios are computed. All of the lane group v/c ratios are procedure for computing the uniform delay must be used as shown
acceptable. The low v/c ratio for the EB left-turn lane group sug- on the Supplemental Uniform Delay Worksheet for Calculation 2
gests that protected-plus-permitted phasing may have been over- (Figure 9-44). This worksheet implements the computations re-
kill. In subsequent trials, the feasibility of a simple two-phase quired for estimating the uniform delay as described previously in
signal at this intersection might be investigated. A ‘‘protected- this chapter. The worksheet will be explained in more detail later
only’’ left turn might also be feasible within the same cycle length. in this discussion.
The critical v/c ratio is also somewhat low, suggesting that a shorter Uniform, incremental, and total delay for each lane group are
cycle length might accommodate all of the traffic; however, mini- computed as indicated on the worksheet, and appropriate levels of
mum green times must also be recognized. None of these insights service are selected from Table 9-1. The EB left-turn and EB
should be acted on before the delay results of the final analysis through lane groups are then aggregated to obtain an approach
module are considered. delay and level of service. The average delay for these two lane
groups is weighted on the basis of lane group demand flow.
LOS Module Worksheet for Calculation 2 The overall intersection delay and level of service are deter-
mined as 21.4 sec/vehicle (LOS C) by computing the weighted
The LOS worksheet for Calculation 2 is shown in Figure 9-43. average of the approach delays.
Values of cycle length, green ratio, v/c ratio, and lane group capac- The level of service at the intersection with the trial signal
ity are entered on the worksheet from previous results. Since ran- timing is generally quite good. The WB approach has the poorest
dom arrivals have been assumed as a base condition, the LOS (C) and also has the highest v/c ratio (0.814), but both
progression adjustment factor, PF, is taken to be 1.00 and may be are in the acceptable range. Minor modifications could be
entered directly on the worksheet without performing the considered, but the balance of this discussion will focus instead
computation. on the left-turn phasing alternatives. The objective will be not
9-76 urban streets
Figure 9-44. Supplemental Uniform Delay Worksheet for Calculation 2.
Figure 9-45. Capacity Analysis Module Worksheet for protected-only phasing for Calculation 2.
to improve the intersection operation, but to gain a better phase, and the secondary phase no longer exists. The capacity is
perspective on the complex nature of the process by which left thereby reduced from 272 to 169 vph.
turns must be modeled. The reduction in capacity for this movement should cause an
This provides an excellent opportunity to examine the special increase in the delay. The delay is now estimated directly from
procedure for protected-plus-permitted phasing in more detail and Equation 9-24 instead of the special procedure in Figure 9-21. The
to compare the results with the ‘‘protected-only’’ alternative. The revised LOS worksheet shown in Figure 9-46 indicates that the
Input Module Worksheet shown in Figure 9-36 indicates that the delay for the EB left turn has increased from 17.3 to 56.0 sec/
EB left turn proceeds on a green arrow in the first phase and on vehicle. Note that supplemental worksheets were not required for
a solid green in the second phase. Suppose that this movement the left-turn saturation flow adjustment fLT nor for the uniform
were only allowed to proceed on the green arrow, facing a red delay computations because no permitted movements were
indication in Phase 2 instead of a green indication. Suppose also involved.
that the length of each signal phase does not change. A better understanding of this comparison may be seen in Figure
These suppositions would make no difference in the Volume 9-47, which shows the queue accumulation polygons that deter-
Adjustment Module Worksheet, and the only difference in the mine the value of the uniform delay. The area contained under the
Saturation Flow Rate Module Worksheet would be the elimination large lightly shaded triangle represents the delay for a protected-
of the computations for fLT for the EB permitted left turn. The only left turn that would be computed using Equation 9-24. The
Capacity Analysis Module Worksheet would remain the same for area under the two smaller and heavily shaded triangles represents
all movements except the EB left turn, as indicated in Figure the delay for a protected-plus-permitted left turn that would be
9-45. Note that the EB left turn is now served only by the primary computed by the Supplemental Uniform Delay Worksheet (Figure
Figure 9-46. LOS Module Worksheet for protected-only phasing for Calculation 2.
Figure 9-47. Queue accumulation polygons for protected and protected-plus-permitted phasing for Calculation 2.
9-21). Note that the altitudes of the two heavily shaded triangles shown in Figures 9-45 and 9-46. There will, however, be a notice-
(0.53 and 1.24 vehicles, respectively) were computed previously able difference if permitted-plus-protected phasing is used.
on the Supplemental Uniform Delay Worksheet for Calculation 2 To illustrate this difference, the order of the EB and WB phases
(Figure 9-44). The effect of this difference is propagated through is reversed, keeping the same cycle length and total phase times
to the capacity analysis worksheet for protected-only phasing and as in the original example. This will have no effect on the Volume
the LOS worksheet for protected-only phasing presented in Figures Adjustment Module Worksheet for Calculation 2 (Figure 9-37);
9-45 and 9-46, respectively. the values shown there will still apply. The changes in the Satura-
Now that protected left-turn phasing has been compared with tion Flow Rate Module Worksheet will be limited to a new fLT
protected-plus-permitted (leading) phasing, it is perfectly logical that results from minor changes in the computations performed on
to ask what would happen to the delay and level of service with the Supplemental Worksheet for Permitted Left Turns (Figure
permitted-plus-protected (lagging) left-turn phasing. The Supple- 9-48). As shown in Figures 9-48 and 9-49, the fLT will increase
mental Uniform Delay Worksheet differentiates between these two from 0.15 (Figure 9-41) to 0.18 (Figure 9-48). The capacity will
phasing alternatives. Further differences will be found in the satu- increase from 272 vph (Figure 9-42) to 366 vph (Figure 9-49).
ration flow rate and capacity computations because of the relative The Supplemental Uniform Delay Worksheet for permitted-
position of the green phase for the two opposing directions. plus-protected (lagging) left-turn phasing (Figure 9-50) indicates
Reversing the order of the EB and WB phases to provide lagging an increase in the uniform delay from 11.75 sec/vehicle (Figure
left-turn protection for the EB left turn will have no effect on the 9-44) to 19.72 sec/vehicle, reflecting the difference between lead-
operation if the left turn is protected only. The results will be ing and lagging left-turn protection. The second term of the delay
identical to those presented for the leading protected-only case equation is very small in both cases, because the v/c ratios are
9-78 urban streets
Figure 9-48. Supplemental Worksheet for Permitted Left Turns: permitted-plus-protected (lagging) left-turn phasing for Calculation 2.
Figure 9-49. Capacity Analysis Module Worksheet for permitted-plus-protected (lagging) left-turn phasing for Calculation 2.
low. The overall delay and level of service are computed on the This sample calculation has exercised several features of the
LOS worksheet shown in Figure 9-51. analysis methodology presented in this chapter. All of the computa-
The comparison of the queue accumulation and discharge poly- tions indicate that traffic volumes are well below capacity and that
gons between protected-only phasing and permitted-plus-protected delays are minimal. No operational problems would be predicted
(lagging) left-turn phasing, shown in Figure 9-52, provides a under the specified conditions.
graphical insight into the modeling process. A similar comparison
for leading left-turn phasing was given in Figure 9-47. The differ-
ence between these phasing alternatives is apparent. It should be CALCULATION 3: OPERATIONAL ANALYSIS OF
remembered that the uniform delay is given by the area contained MULTIPHASE ACTUATED SIGNAL
under the queue accumulation polygon for the specified phasing
alternative. Note how this area is reduced to a greater extent in The intersection of Fifth Avenue and Twelfth Street is a major
the case of leading left-turn protection. CBD junction of two significant arterials. Both facilities have four
Figure 9-50. Supplemental Uniform Delay Worksheet for permitted-plus-protected (lagging) left-turn phasing for Calculation 2.
Figure 9-51. LOS Module Worksheet for permitted-plus-protected (lagging) left-turn phasing for Calculation 2.
lanes, with exclusive left-turn lanes provided at the intersection shown in Figure 9-53, and the unit extension for each phase is 2.5
on all four approaches. The signal is fully actuated, with standard sec. Balanced lane utilization was also observed in the field.
protected-plus-permitted phasing on Fifth Avenue and simple per-
mitted phasing on Twelfth Street. The signal is relatively isolated Volume Adjustment Module Worksheet for Calculation 3
from adjacent signals, and arrivals can be assumed to be random
for all practical purposes. The Volume Adjustment Module Worksheet for Calculation 3
is shown in Figure 9-54. Movement volumes are divided by the
appropriate PHF to obtain peak flow rates. The designation of lane
Input Module Worksheet for Calculation 3 groups is straightforward, with each approach consisting of an
exclusive left-turn lane group and the through and right-turn lane
Figure 9-53 shows the Input Module Worksheet for this sample group. The analysis will proceed with eight separate lane groups.
calculation. Most of the information is self-explanatory. There are The proportions of left and right turns are obtained by dividing
5 percent heavy vehicles on Twelfth Street and 2 percent on Fifth the individual left- and right-turn flow rates by the total flow rate
Avenue. There is parking on Twelfth Street with 5 movements/hr in each lane group.
activity, and no parking on Fifth Avenue. PHFs are different for the
two streets because of the slightly differing traffic characteristics. Saturation Flow Rate Module Worksheet for Calculation 3
There are heavy pedestrian crossing volumes on Fifth Avenue and
more moderate crossing volumes on Twelfth Street. Again, since The Saturation Flow Rate Module Worksheet for Calculation 3
arrivals are assumed to be random, Arrival Type 3 will be used for is shown in Figure 9-55. The following adjustment factors are
all approaches. The average signal timings observed in the field are taken directly from the appropriate tables:
9-80 urban streets
Figure 9-52. Queue accumulation polygons for protected and permitted-plus-protected phasing for Calculation 2.
1. The 10-ft lane widths on Twelfth Street call for an fw of applied to the EB and WB left turns because no such phase is
0.933. Fifth Avenue has standard 12-ft lanes, for which the factor present for these movements.
is 1.00. The resulting left-turn factors are very low for the NB and SB
2. Twelfth Street has 5 percent heavy vehicles and from Table movements. This is primarily because the turns are made from
9-6, fHV = 0.952. Fifth Avenue has 2 percent heavy vehicles and exclusive lanes facing very heavy oncoming traffic. It is expected
a factor of 0.980. that most of the capacity for these movements will have to come
3. All grades are level, so fg = 1.00 for all lane groups. from the protected phase and the sneakers. The fLT values are all
4. EB and WB through lane groups have adjacent parking lanes entered on the Saturation Flow Rate Module Worksheet.
with relatively low activity (5 movements/hr). From Table 9-8,
fp = 0.938 for this case. Other lane groups do not have adjacent
parking lanes and thus have a factor of 1.00.
Capacity Analysis Module Worksheet for Calculation 3
5. There are no local buses on any approach, and fbb = 1.00.
6. Because the intersection is located in a CBD, fa = 0.90.
7. Because of the observed balanced lane utilization, fLU = 1.0. The Capacity Analysis Module Worksheet for Calculation 3 is
8. Right-turn adjustment factors are found in Tables 9-11a and shown in Figure 9-57. Lane group flows are entered from the
9-11b for the appropriate proportion of right turns and conflicting Volume Adjustment Module Worksheet (Figure 9-54), and satura-
pedestrian movements. For all left-turn lane groups, the value is tion flow rates from the Saturation Flow Rate Module Worksheet
1.00. Values for through lane groups are shown on the worksheet. (Figure 9-55). Flow ratios, v/s, are then computed and reviewed
9. The left-turn adjustment factor is 1.00 for all through lane to identify critical lane groups.
groups. For left-turn lane groups, the special procedures involving In this case, there are no overlaps in the phasing, which simpli-
permitted left turns must be implemented to compute the adjust- fies the determination of critical lane groups. Because both left
ment factor. turns in the N-S direction are protected plus permitted, both left
turns are split into two lane groups. In the first phase, only the
The special worksheet for computing the left-turn adjustment fac-
left-turn primaries move. The critical lane group is the SB left-
tor for permitted left turns is shown in Figure 9-56. The EB and WB
turn primary (0.090). In the second phase, the NB through lane
left-turn lanes involve simple permitted phasing, and the worksheet
group has the highest v/s ratio (0.522). There is a single phase in
is followed easily, except that gf is set at 0.0 because the first vehicle
the E-W direction. The critical lane group is the EB left turn with
in the queue is always a left-turner. The resulting left-turn adjustment
a v/s ratio of 0.224. This accounts for all phases. The sum of the
factors are then recorded in the bottom rows of the worksheet. Note
that the proportion of vehicles arriving on the green (for the opposing v/s ratios is 0.836.
flow) is assumed to be equal to the opposing g/C ratio. In this case, The lost time is taken to be 4.0 sec/phase. Since each of the
the EB and WB g/C ratios are equal. critical lane groups defines a single phase, there is time lost in
NB and SB left-turn lane groups involve protected-plus-permit- the cycle at each phase. The total lost time is 12 sec/cycle. The
ted phasing, and one modification must be applied in following critical v/c ratio is 0.836 × [90/(90 − 12)] = 0.964.
the worksheet. The difference is that no lost time is applied to the Green ratios are entered in Column 6 of the worksheet, and lane
permitted phase because the left turns already have a green indica- group capacities are computed as s × g/C. Lane group v/c ratios
tion at the beginning of the phase. Note that a 4-sec lost time is are then computed as shown.

9-82 urban streets
LOS Module Worksheet for Calculation 3 This results in larger queues that create a greater area for the
queue accumulation polygon discussed in detail in connection with
The LOS Module Worksheet for Calculation 3 is shown in Calculation 2. The uniform delays are computed as 6.27 sec/vehi-
Figure 9-58. Relevant values of C, v/c, g/C, and c are entered from cle NB and 25.05 sec/vehicle SB.
previous results. Lane group delays are computed as shown, and LOS values range from F for the WB left-turn lane group
levels of service are obtained from Table 9-1. Delays are then to A for the NB left-turn lane group. The wide range of delay
aggregated by approach and for the intersection as a whole, and for the critical movements suggests that the signalization is far
appropriate levels of service are assigned. too favorable to some movements, reinforcing the results of the
A supplemental uniform delay worksheet is required for the NB capacity analysis.
and SB left turns. This worksheet is shown in Figure 9-59. The The disparate delay and v/c values suggest that the signal
SB left turn will experience more delay than the NB left turn timing is inefficient and inequitable. The critical v/c ratio of
because there is heavier movement, which is opposed by more 0.964 suggests that the overall intersection will operate near
oncoming traffic. The NB movement is blocked by oncoming capacity. It is likely that a redistribution of the time allocation
traffic for 15.35 sec of the permitted phase, and the SB movement among the phases will produce a workable operation in the
is blocked for 36.48 sec as indicated by their respective gq values. near-capacity v/c range.
Figure 9-56. Supplemental Worksheet for Permitted Left Turns for Calculation 3.

9-84 urban streets
In this example, a trial-and-error procedure will again be applied structed as multilane facilities, and the intersection will experience
to seek an acceptable solution using the planning method results as substantial demand. Figure 9-61, the input worksheet for the plan-
an initial solution. The EB and WB phases clearly need more green ning analysis, contains a diagram of the expected intersection lay-
time. The NB and SB left-turn movements have ample green time; out and the forecast volumes for the intersection. Note that left-
however, it is only possible to reduce this phase a small amount with- turn lanes are expected to be incorporated on each approach. Will
out violating the minimum green time for these movements. Some the capacity of the design be adequate?
time will therefore have to be transferred from the NB and SB This is an ideal application for the planning analysis. Some
through phase or the cycle length must be increased, or both. preliminary assumptions must be made about the signal phasing.
As a result of a trial-and-error process of phase time redistribu- All left turns are heavy and face heavy opposing traffic and
tion, the following timing plan was formulated. will therefore be assumed to be protected. For planning purposes
several years in the future, it is probably appropriate to assume
Lane Group Phase Time (sec) that protected-only phasing will be used. There is no guarantee
NB and SB left turn 12.4 that protected-plus-permitted operation will be practical. For
NB and SB through 49.9 example, crash rates at this or nearby intersections could dictate
EB and WB left turn 27.7 the use of simple left-turn protection, and it could be very
Total 90 optimistic to count on the extra capacity of the permitted phase.
Skipping the intermediate computations, the effect of the green Figure 9-62 shows the composite lane volume worksheet for
time redistribution is shown in the LOS worksheet presented in this intersection, and Figure 9-63 shows the signal operations
Figure 9-60. Note that a better balance of delay among the worksheet. For planning purposes, the main item of interest is the
competing movements has been achieved, with the worst delays in intersection status, which is given as near capacity, with a critical
each phase balanced at 49 sec. The overall delay for the intersection v/c ratio of 0.92. The estimated cycle length is 120 sec.
is approximately 37 sec/vehicle. This delay value reflects LOS D. The near-capacity rating could be interpreted to mean that it is
This sample problem has examined the effect of the signal tim- uncertain whether the demand would exceed the capacity. It should
ing plan on intersection performance. The initial plan was first be kept in mind that the 20-year traffic volume projections are
evaluated and then modified by trial and error to produce a more no doubt based on some coarse assumptions and approximations.
satisfactory operation. It was mentioned earlier in this chapter that Although this status might be considered acceptable with near-
a separate and more detailed procedure for estimating signal timing term volume projections, many agencies would be more comfort-
parameters at traffic-actuated intersections is presented in Appen- able with a design that was rated as under capacity to provide
dix II. The analysis of this intersection will be continued in Appen- additional assurance that the capacity will not be exceeded.
dix II to illustrate the procedure. The effect of changes in controller The suggestions in Appendix I indicate that intersection design
operating parameters will also be demonstrated. should attempt to keep the per-lane volumes to 450 vph or less.
This is not the case for the EB approach in the proposed design
of Figure 9-61. Note that the right-turn volume is extremely high
CALCULATION 4: PLANNING ANALYSIS OF INTERSECTION on this approach. If an exclusive right-turn lane were provided,
WITH MULTILANE APPROACHES lane volumes on the remainder of the approach could be brought
below the suggested 450 vph. That alternative can be examined
The intersection of Tenth Avenue and First Street is currently with the planning method.
a minor intersection of two 2-lane, lightly used streets. In 20 years, The composite lane volume worksheet with modified geometry
major development is expected to cause both streets to be recon- is shown in Figure 9-64. This worksheet is very similar to the
Figure 9-59. Supplemental Uniform Delay Worksheet for Calculation 3.
Figure 9-60. LOS Module Worksheet with revised signal timing for Calculation 3.
corresponding worksheet for the original design in Figure 9-62. analysis without violating the minimum green time for the WB
Note, however, that the right turn is now the critical movement. left turn. This problem may best be overcome by forcing the EB
The reduced lane volume resulting from the additional lane, when and WB left turns to move simultaneously by absorbing the 2.7
analyzed on the signal operations worksheet in Figure 9-65, indi- sec EB through and left-turn phase into the 9.5-sec EB and WB
cates a reduction in the v/c ratio to 0.87. left-turn phase to create a 12.2-sec phase for both the EB and WB
There is some question at this point about whether it is appro- left turns. The timing plan to be used in the operational analysis
priate to proceed with the operational analysis on the basis of will thus include five phases as follows:
volumes projected so far into the future. It is certainly mechanically Movement Phase (sec)
possible to perform the analysis; the problem lies with the degree EWL 12.2
of confidence in the results. Therefore the operational analysis will EWT 36.4
be carried out with that limitation in mind. NSL 17.7
The signal timing plan synthesized by the planning method in NTL 4.1
Figure 9-65 cannot be transferred directly into the operational NST 29.6
9-86 urban streets
Figure 9-61. Planning Method Input Worksheet for Calculation 4.
Using default values for the rest of the data items and skipping to conclude that both the planning and operational analyses indi-
the worksheets showing intermediate computations, the discussion cate that the intersection would perform acceptably under the spec-
proceeds directly to the LOS Module Worksheet for Calculation ified conditions.
4 (Figure 9-66). The v/c ratios are reasonably well balanced and
in the range indicated by the planning analysis. In general, the CALCULATION 5: PLANNING ANALYSIS OF INTERSECTION
planning analysis tends to be a better approximation of the opera- WITH SINGLE-LANE APPROACHES
tional analysis when default values are used in the operational
analysis and when permitted left turns are avoided. This is the A large area of a semirural community has been developing
case in the example that has just been examined. It is reasonable rapidly, requiring a considerable planning effort to provide addi-
tional capacity at numerous intersection of low-type, formerly rural phasing for the WB approach and a shared-lane permitted treatment
highway facilities. The intersection of Eighth Avenue and Main on all of the other approaches. To provide a conservative assessment,
Street is one such location. It is an intersection of a two-lane protected-only phasing will be used for the WB left turn.
roadway with a four-lane roadway. No turning lanes are present The composite lane volume worksheet for this initial analysis
on any approach. The intersection is shown in Figure 9-67, along is shown in Figure 9-68. The critical lane volumes appear to be
with projected traffic volumes. Is it likely that the capacity will very high, and this observation is confirmed on the signal opera-
be exceeded at this intersection, and if so, what countermeasures tions worksheet (Figure 9-69). The critical v/c ratio is computed
should be implemented? as 0.97, producing an at-capacity status. Some countermeasures
It will be difficult to evaluate phase plans with protected left should probably be considered.
turns at this intersection because none of the approaches is shown Since the EB through and right-turn lane group volume is critical,
to have an exclusive left-turn lane. The single-lane approaches for and since this lane group includes moderately heavy right-turn move-
NB and SB traffic preclude consideration of any alternative lane ment, it is logical to examine the addition of an EB right-turn lane
use. Protected left-turn phasing on these approaches would require as a countermeasure. The exclusive right-turn lane is easily incorpo-
split-phase operation. rated into this analysis. The lane volume and signal operations work-
The EB and WB approaches each have two lanes, so it is possible sheets shown in Figures 9-70 and 9-71, respectively, indicate that
to assign one lane as an exclusive left-turn lane and the other as a the critical v/c ratio would be reduced to 0.82 by this improvement,
shared lane for through and right-turn traffic. Inspection of the EB and the intersection status would be under capacity.
and WB traffic volumes suggests that this is definitely necessary This could be considered the final solution to the problem, but
and probably practical on the WB approach. On the EB approach, a the NB and SB left turns merit further consideration. It would be
single-lane group with a shared lane for left turns is likely to be more desirable to avoid the problem of significant left-turn volumes in a
appropriate. Therefore the initial solution is to use protected left-turn pair of opposing single-lane approaches, if possible. For this reason,
9-88 urban streets
split-phase operation on the NB and SB approaches will be exam- CALCULATION 6: DETERMINING v/c AND SERVICE FLOW
ined. Under split-phase operation, there would be a complete direc- RATES—AN ALTERNATIVE USE OF OPERATIONAL
tional separation between the traffic from the two opposing ANALYSIS PROCEDURE
directions because their lane groups would proceed on different
phases. The planning method worksheets provide for split-phase op- Description
eration, which is identified on the signal operations worksheet as
Phase Plan 4. A two-lane through movement at one approach to a signalized
The lane volume worksheet for this alternative is shown in intersection has a cycle length of 90 sec with a g/C ratio of 0.50. The
Figure 9-72. Note that the NB and SB per-lane volumes are reduced arrival type is currently 3 (random), but this could be improved by
because there is no interference from the left turns in the single altering the progression. What is the maximum service flow rate
lane. However, as shown in Figure 9-73, these reduced volumes that could be accommodated at LOS B (20 sec/veh delay) on this
must now be added into the critical sum because they each move approach?
on different phases. Previously, only the heavier of the two vol-
umes was reflected in the critical sum, because the two lane groups Solution
proceeded on the same phase. The net result is an increase in
the critical v/c ratio to 1.02. Therefore, it would be difficult to Delay is based on the v/c ratio, X; the green ratio, g/C; the cycle
recommend split-phase operation at this intersection as a desirable length, C; the lane group capacity, c; and the progression factor,
solution from a capacity standpoint. PF. The lane group capacity may be computed as the saturation
To provide a protected phase for the NB and SB left turns, it flow rate for the lane group times the g/C ratio, which is known.
will be necessary to widen the approaches to the intersection to Assume that a standard analysis using the Saturation Flow Rate
construct an additional lane for exclusive left turns. This alternative Module Worksheet has been conducted and that the saturation
is easy to examine with the planning method. The composite lane flow rate for the lane group has been found to be 3,200 vphg and
volume worksheet shown in Figure 9-74 indicates a substantial the capacity 3,200 × 0.50 = 1,600 vph.
reduction in the per-lane volume for the NB and SB through move- If delay is set at 20.0 sec/veh and the known values of C, g/C,
ments. The effect on the critical v/c ratio, as shown in Figure and c are inserted into Equations 9-22, 9-23, and 9-25, the follow-
9-75, is a reduction to the under-capacity level. ing relationship is established:
Figure 9-64. Lane Volume Worksheet for Calculation 4 with geometric modifications.
20.0 = d1 PF + d2 + d3 For the solution to this problem, the lower table is most useful. For
a delay of 20 sec, maximum v/c ratios are given for Arrival Types
d1 = 0.5(90)(1 − 0.50)2/(1 − 0.50X ) 1–6. Service flow rates, SF, are computed as the v/c ratio times the
lane group capacity of 1,600 vph. Thus, for LOS B, the approach can
d2 = 225[(X − 1) + √(X − 1)2 + (X/100)] carry a maximum service flow rate of 1,126 vph under the existing
arrival type (3), but could be increased to as much as 1,491 vph for
d3 = 0 Arrival Type 5 with improved progression and even up to 1,571 for
Arrival Type 6 with near-perfect progression.
Various combinations of PF and X that result in 20.0 sec of This calculation is intended to illustrate the potential for alterna-
delay may now be solved for. If the level of service (delay) were tive computational sequences using the basic operational analysis
to be allowed to vary as well, a tabular array of X versus delay format. It should be noted, however, that this calculation addresses
and arrival type could be developed for the subject approach. Table only one lane group and that computations become far more com-
9-18 is such an array. plex when multiple lane groups are to be considered simultane-
Two presentation formats are shown in Table 9-18. The upper ously. Nevertheless, the procedure is capable of determining
portion tabulates delay for various arrival types and v/c ratios (X). service flow rates, as shown here, or geometric or signal parameters
The lower portion tabulates v/c ratio versus delay and arrival type. based on a desired level of service.

9-90 urban streets
Figure 9-65. Signal Operations Worksheet for Calculation 4 with geometric modifications.

Figure 9-67. Planning Method Input Worksheet for Calculation 5.

9-92

urban streets

signalized intersections
Figure 9-71. Signal Operations Worksheet with additional EB right-turn lane for
Calculation 5.
Figure 9-70. Lane Volume Worksheet with additional EB right-turn lane for
Calculation 5.
9-93

9-94

urban streets
Figure 9-73. Signal Operations Worksheet with NB and SB split-phase operation for
Calculation 5.
Figure 9-72. Lane Volume Worksheet with NB and SB split-phase operation for
Calculation 5.
Figure 9-75. Signal Operations Worksheet with added NB and SB left-turn lanes for
Calculation 5.
Figure 9-74. Lane Volume Worksheet with added NB and SB left-turn lanes for
Calculation 5.
9-95

9-96 urban streets
Table 9-18. Service Flow Rate Solutions for Calculation 6
flow delay
X rate d1 d2 AT = 1 AT = 2 AT = 3 AT = 4 AT = 5 AT = 6
0.0 0 11.25 0.00 18.75 13.95 11.25 8.63 3.75 0.00
0.1 160 11.84 0.12 19.87 14.81 11.97 9.21 4.07 0.12
0.2 320 12.50 0.28 21.12 15.78 12.78 9.87 4.44 0.28
0.3 480 13.24 0.48 22.54 16.89 13.72 10.63 4.89 0.48
0.4 640 14.06 0.75 24.19 18.19 14.81 11.53 5.43 0.75
0.5 800 15.00 1.12 26.12 19.72 16.12 12.62 6.11 1.12
0.6 960 16.07 1.67 28.46 21.60 17.74 14.00 7.02 1.67
0.7 1,120 17.31 2.58 31.43 24.04 19.88 15.85 8.34 2.58
0.8 1,280 18.75 4.30 35.55 27.55 23.05 18.68 10.54 4.30
0.9 1,440 20.45 8.51 42.61 33.88 28.97 24.20 15.33 8.51
1.0 1,600 22.50 22.50 60.01 50.40 45.00 39.76 29.99 22.50
max AT = 1 AT = 2 AT = 3 AT = 4 AT = 5 AT = 6
los delay SFmax X SFmax X SFmax X SFmax X SFmax X SFmax X
A 5 513 0.32 1307 0.82
10 348 0.22 1241 0.78 1457 0.91
15 191 0.12 663 0.41 1046 0.65 1429 0.89 1514 0.95
B 20 177 0.11 824 0.51 1126 0.70 1318 0.82 1491 0.93 1571 0.98
25 707 0.44 1164 0.73 1333 0.83 1448 0.91 1546 0.97
30 1043 0.65 1342 0.84 1450 0.91 1500 0.94
C 35 1259 0.79 1451 0.91 1500 0.94 1551 0.97
40 1381 0.86 1499 0.94 1550 0.97
45 1462 0.91 1548 0.97 1600 1.00
50 1508 0.94 1596 1.00
D 55 1554 0.97
60 1600 1.00
65
70
75
E 80
V. REFERENCES
The methodology in this chapter is based in part on the results 5. Berry, D.S., and Gandhi, P.K., ‘‘Headway Approach to In-
of an NCHRP study conducted by JHK & Associates (2,3). Critical tersection Capacity.’’ Highway Research Record 453, Trans-
movement capacity analysis techniques have been developed in portation Research Board, Washington, D.C. (1973).
the United States (1,4,5), Australia (6), Great Britain (7), and 6. Miller, A.J., ‘‘The Capacity of Signalized Intersections in
Sweden (8). Background for delay estimation procedures was de- Australia.’’ Australian Road Research Bulletin 3, Australian
veloped in Great Britain (7), Australia (9,10), and the United States Road Research Board, Kew, Victoria, Australia (1968).
(11). Updates to the original methodology were subsequently de- 7. Webster, F.V., and Cobbe, B.M., Traffic Signals. Her Majes-
veloped (12–20). ty’s Stationary Office, London, England (1966).
1. Messer, C.J., and Fambro, D.B., ‘‘Critical Lane Analysis for 8. Petersen, B.E., and Imre, E., Swedish Capacity Manual.
Intersection Design.’’ Transportation Research Record 644, Stockholm, Sweden (Feb. 1977).
Transportation Research Board, Washington, D.C. (1977). 9. Akcelik, R. (Ed.), ‘‘Signalized Intersection Capacity and
2. ‘‘NCHRP Signalized Intersection Capacity Method.’’ Na- Timing Guide.’’ Signalized Intersection Capacity Workshop—
tional Cooperative Highway Research Program, Project 3- Papers, Australian Road Research Board, Kew, Victoria, Aus-
28(2), JHK & Associates, Tucson, Ariz. (Feb. 1983). tralia (1979).
3. ‘‘Signalized Intersection Capacity Study.’’ Final report, Na- 10. Akcelik, R., ‘‘Traffic Signals: Capacity and Timing Analy-
tional Cooperative Highway Research Program, Project 3- sis.’’ Australian Road Research Report 123, Australian Road
28(2), JHK & Associates, Tucson, Ariz. (Dec. 1982). Research Board, Kew, Victoria, Australia (1981).
4. Berry, D.S., ‘‘Other Methods for Computing Capacity of 11. Reilly, W.R., Gardner, C.C., and Kell, J.H., ‘‘A Technique
Signalized Intersections.’’ Paper presented at the 56th Annual for Measurement of Delay at Intersections.’’ FHWA Report
Meeting of the Transportation Research Board, Washington, No. RD-76-135/137, Federal Highway Administration, Wash-
D.C. (Jan. 1977). ington, D.C. (1976).
12. Fambro, D.B., Chang, E.C.P., and Messer, C.J., ‘‘Effects of 16. Hagen, L.T., and Courage, K.G., ‘‘Comparison of Macro-
the Quality of Traffic Signal Progression on Delay,’’ National scopic Models for Signalized Intersection Analysis,’’ Trans-
Cooperative Highway Research Program, Project 3-28C, portation Research Record 1225, Transportation Research
Texas Transportation Institute, Texas A&M University, Col- Board, Washington, D.C. (1989).
lege Station (Aug. 1989). 17. Strong, D.W., ‘‘Real-World Use of the 1985 HCM—Analy-
13. Roess, R.P., Papayannoulis, V.N., Ulerio, J.M., and Lev- sis of Signalized Intersections,’’ Compendium of Technical
inson, H.S., ‘‘Levels of Service in Shared-Permissive Left- Papers, Institute of Transportation Engineers, Washington,
Turn Lane Groups at Signalized Intersections,’’ Federal High- D.C. (Sept. 1989).
way Administration, Transportation Training and Research 18. Fambro, D.B., Rouphail, N., Sloup, P., Daniel, J., and Li,
Center, Polytechnic University, Brooklyn, New York (Sept. J., ‘‘Highway Capacity Manual Revisions of Chapters 9 and
1989). 11.’’ Final Report FHWA-RD-96-088 (Jan. 1996).
14. Prassas, E.S., and Roess, R.P., ‘‘The Left-Turn Adjustment 19. Courage, K., Fambro, D.B., Akcelik, R., Lin, P., Anwar,
for Permitted Turns from Shared Lane Groups: Another M., and Viloria, F., ‘‘Capacity Analysis of Traffic-Actuated
Look,’’ paper presented at the mid-year meeting of TRB Com- Intersections.’’ Final report NCHRP Project 3-48 (Dec. 1996).
mittee A3A10, Transportation Training and Research Center, 20. Powell, J.L., ‘‘Field Measurement of Signalized Intersection
Polytechnic University, Brooklyn, New York (July 1992).
Delay for the 1997 HCM.’’ Parsons Transportation Group,
15. Bonneson, J.A., and McCoy, P.T., ‘‘Operational Analysis of
Chicago, Ill., unpublished paper (July 1997).
Exclusive Left-Turn Lanes with Protected/Permitted Phas-
ing,’’ Transportation Research Record 1114, Transportation
APPENDIX I
INTERSECTION GEOMETRICS—SUGGESTIONS FOR ESTIMATING DESIGN ELEMENTS
This appendix summarizes suggestions for establishing the geo- approach. A method for estimating the required length of the stor-
metric design of an intersection when it has not been defined by age bay is summarized in Figure I.9-1 and Table I.9-1.
existing conditions or by state or local practice. These suggestions
Figure I.9-1 shows the relationship between the left-turn volume
may also be applied when analysis indicates intersection deficien-
[expressed in passenger car equivalents (PCEs)] and the length of
cies that are to be corrected by changes in geometric design. Noth-
the turn storage bay. The relationship is based on random arrivals
ing in this appendix, however, should be construed as constituting
and 5 percent probability of storage bay overflow. Left-turn PCE
strict guidelines or standards. This material should not be used in
factors, EL1, are obtained from Figure 9-7.
place of applicable state and local standards, guidelines, policies,
The values obtained from Figure I.9-1 are for a cycle length of
or practice.
75 sec and a v/c ratio of 0.80. For other values, the length obtained
The geometric design of an intersection involves several critical
from Figure I.9-1 is multiplied by a correction factor obtained
decisions about the number and use of lanes to be provided on
from Table I.9-1. The v/c ratio for left-turn lane groups is computed
each approach, as discussed in the following sections.
on the operational analysis Capacity Analysis Module Worksheet.
EXCLUSIVE LEFT-TURN LANES EXCLUSIVE RIGHT-TURN LANES
Left-turn lanes are provided to accommodate heavy left-turn Although right turns are generally made more efficiently than
movements without disruptions to through and right-turning vehi- left-turn movements, exclusive right-turn lanes are often provided,
cles. The provision of an exclusive left-turn lane (or lanes) allows for many of the same reasons that left-turn lanes are used. Right
for the use of protected left-turn phasing and provides storage for turns may face a conflicting pedestrian flow, but do not face a
queued left-turn vehicles without disruption to other flows. The conflicting vehicular flow. In general, an exclusive right-turn lane
following suggestions are made concerning the provision of exclu- should be considered when the right-turn volume exceeds 300 vph
sive left-turn lanes: and the adjacent main-line volume exceeds 300 vphpl.
1. Where fully protected left-turn phasing is to be provided, an
exclusive left-turn lane should be provided. NUMBER OF LANES
2. Where space permits use of a left-turn lane, it should be
considered when left-turn volumes exceed 100 vph. Left-turn lanes The number of lanes required on an approach depends on a
may be provided for lower volumes as well on the basis of the variety of factors, including the signal design. In general, enough
judged need and state or local practice, or both. main roadway lanes should be provided such that the total of the
3. Where left-turn volumes exceed 300 vph, provision of a dou- through plus right-turn volume (plus left-turn volume, if present)
ble left-turn lane should be considered. does not exceed 450 vphpl. This is a very broad suggestion. Higher
4. The length of the storage bay should be sufficient to handle volumes can be accommodated on major approaches when a sub-
the turning traffic without reducing the safety or capacity of the stantial portion of available green time can be allocated to the
9-98 urban streets
Table I.9-1. Left-Turn Bay Length Adjustment

Factors
cycle length, C (sec)
v/c ratio, X
60 70 80 90 100
0.50 0.70 0.76 0.84 0.89 0.94
0.55 0.71 0.77 0.85 0.90 0.95
0.60 0.73 0.79 0.87 0.92 0.97
0.65 0.75 0.81 0.89 0.94 1.00
0.70 0.77 0.84 0.92 0.98 1.03
0.75 0.82 0.88 0.98 1.03 1.09
0.80 0.88 0.95 1.05 1.11 1.17
0.85 0.99 1.06 1.18 1.24 1.31
0.90 1.17 1.26 1.40 1.48 1.56
0.95 1.61 1.74 1.92 2.03 2.14
Source: C. J. Messer, ‘‘Guidelines for Signalized Left-Turn Treatments,’’
Implementation Package FHWA-IP-81-4, Federal Highway Administration,
Washington, D.C., 1981, Table 1.
subject approach. When the number of lanes is unknown, the fore-

going value represents a reasonable starting point for analysis
computations.
OTHER FEATURES
Figure I.9-1. Left-turn bay length versus turning volume. (Source: When lane widths are unknown, the 12-ft standard lane width
C. J. Messer, ‘‘Guidelines for Signalized Left-Turn Treatments,’’
should be assumed unless known restrictions prevent such width.
Implementation Package, FHWA-IP-81-4, Federal Highway Ad-
ministration, Washington, D.C. 1981, Fig. 2.) Parking conditions should be assumed to be consistent with local
practice. When no information exists, no curb parking and no local
buses should be assumed for analysis purposes.
APPENDIX II
SUGGESTIONS FOR ESTABLISHING SIGNAL DESIGN IN ANALYSIS
The design of traffic signal operation is a complex process in- a change-and-clearance interval that are repeated in each cycle to
volving three important decisions: produce a constant cycle length.
2. Fully Actuated Controllers: The timing on all of the ap-
1. Type of signal controller to be used; proaches to an intersection is influenced by vehicle detectors. Each
2. Phase plan to be adopted, and phase is subject to a minimum and a maximum green time, and
3. Allocation of green time among the various phases. some phases may be skipped if no demand is detected. The cycle
Each of these decisions is heavily influenced by state and local length for fully actuated control will vary from cycle to cycle.
policies, guidelines, and standards, all of which may vary consider- 3. Semiactuated Controllers: Some approaches (typically on the
ably from location to location. This appendix presents the alterna- minor street) have detectors, and some do not. The earliest form
tives available to the analyst along with a general discussion of of semiactuated control was designed to confine the green indica-
the range in which they are employed. This discussion is intended tion to the major street in the absence of minor-street actuation.
only to assist the analyst in establishing an initial signal design for Once actuated, the minor-street green is displayed for a period just
study and does not represent established standards or guidelines. long enough to accommodate the traffic demand.
Although these equipment-based definitions have persisted in
traffic engineering terminology, the evolution of traffic control
TYPE OF SIGNAL technology has complicated their function from the analyst’s per-
spective. For purposes of capacity and LOS analysis, it is no longer
Traffic engineering textbooks describe three types of traffic sig- sufficient to consider the controller type as a global descriptor
nal controllers: of the intersection operation. Instead, an expanded set of these
definitions must be applied individually to each lane group.
1. Pretimed Controllers: A preset sequence of phases is dis- Each lane group may be served by a phase that is either actuated
played in repetitive order. Each phase has a fixed green time and or nonactuated. Nonactuated phases may be coordinated with
neighboring signals on the same route, or they may function in an should be used unless conditions dictate the need for additional
isolated mode without influence from other signals. Nonactuated phases. Because the change interval between phases contributes
phases generally operate with fixed minimum green times, which to lost time in the cycle, as the number of phases increases, the
may be extended by reassigning unused green time from actuated percentage of the cycle made up of lost time generally increases
phases with low demand if such phases exist. also.
Actuated phases, on the other hand, may be used at intersections Figure II.9-1 shows a number of common phase plans that may
at which other phases are coordinated, but they may not, for pur- be used with either pretimed or actuated controllers, and Figure
poses of this appendix, be coordinated themselves. Actuated phases II.9-2 illustrates an optional phasing scheme that typically can be
are subject to being shortened on cycles with low demand. On implemented only with actuated controllers. These and other phase
cycles with no demand, they may be skipped entirely or they may plans are discussed in the sections that follow.
be displayed for their minimum duration. In systems where the
nonactuated phases are coordinated, the actuated phases are also
Two-Phase Control
subject to early termination (force-off) to accommodate the pro-
gression design for the system.
Two-phase control is the most straightforward and the simplest
If all the phases at an intersection are nonactuated, the length
of the available phase plans. Each of two intersecting streets is
of each phase, and consequently the cycle length, will be fixed
given a green phase during which all movements on the street are
for purposes of analysis. This is pretimed operation, Type 1 above.
allowed to proceed. All left and right turns are made on a permitted
In current practice, one or more phases under this type of control
basis against an opposing vehicle or pedestrian flow. The two-
will usually be coordinated. In general, if the intersection is suffi-
phase plan is shown in Figure II.9-1(a). This phase plan is generally
ciently removed from its neighbors to operate in an isolated mode,
used unless turn volumes require protected phasing.
actuated operation will produce lower delays and a better level of
service. The analysis procedures prescribed earlier in this chapter
will indicate the degree to which the delays may be reduced by Multiphase Control
actuated control on any phase.
If all the phases at an intersection are actuated, the length of Multiphase control is adopted at any intersection where it is
each phase, and consequently the cycle length, will vary with each determined that one or more left or right turns require protected
cycle. This is Type 2, fully actuated operation. No coordination phasing. It is generally the left-turn movement that requires a
with neighboring signals is possible under this control mode. Fully partially or fully protected left-turn phase. Local policy and prac-
actuated signals are generally used only at intersections where tice are again critical determinants of this need. Most agencies
distances are such that coordination would not be expected to be have guidelines for left-turn volumes that require protected phas-
beneficial. The analysis procedures prescribed in this chapter will ing. These threshold volumes are generally in the range of 100 to
support an evaluation of the comparative benefits of coordinated 200 vph turning left. Protected left-turn phasing is also considered
operation versus actuated operation. when the speed of opposing traffic is greater than 40 mph.
Type 3, semiactuated control, includes all of the cases that do Multiphase control can be provided in a wide variety of ways,
not fit into either the pretimed or fully actuated category. The depending on the number of turns requiring protected phasing
majority of coordinated arterial systems must be treated as systems and the sequence and overlaps used. Figure II.9-1(b), (c), and (d)
of semiactuated controllers with coordinated nonactuated phases presents three common plans for multiphase control.
serving the arterial approaches and isolated actuated phases serving Figure II.9-1(b) shows a three-phase plan in which an exclusive
the cross-street approaches. The cross-street approaches include left-turn phase is provided for both left-turn movements on the
minor movements such as protected left turns from all approaches. major street. It is followed by a through phase for both directions
The cycle length is constant at coordinated semiactuated intersec- of the major street, during which left turns in both directions may
tions and variable at isolated semiactuated intersections. be permitted on an optional basis.
The analysis procedures presented earlier in this chapter are The use of permitted left-turn phases following protected left-
based on the assumption of a fixed sequence of phases, each of turn phases is very much a matter of local practice. Some agencies
which is displayed for a predictable time. In the case of pretimed use protected-plus-permitted phasing extensively, whereas others
control (i.e., no actuated phases), the length of each phase is as- prefer protected-only phase plans. The phasing illustrated in Figure
sumed to be fixed and constant from cycle to cycle. Actuated II.9-1(b) can be used in either mode.
phases must be approximated for analysis purposes by their aver- Exclusive left-turn phases provide for simultaneous movement
age green time, recognizing that the actual time may differ from of opposing left turns and are most efficient when the opposing
cycle to cycle. For a given timing plan (i.e., constant or average left-turn volumes are nearly equal. When volumes are unequal, or
green times), the differences between actuated and nonactuated in cases in which only one left turn requires protected phasing,
phases are recognized by the parameters used in the incremental other phase plans are more efficient.
term of the delay equation. The three-phase plan may be expanded to a four-phase sequence
if both streets require left-turn phases. Such a sequence is shown
in Figure II.9-1(d). As described previously, left turns may be
PHASE PLANS continued on a permitted basis during the through phases if desired.
It should be noted that all approaches having an exclusive left-
The most critical aspect of any signal design is the selection of turn phase should have an exclusive left-turn lane as well.
an appropriate phase plan, which involves determination of the Figure II.9-1(c) shows what is commonly referred to as ‘‘leading
number of phases to be used and the sequence in which they are and lagging green’’ phasing. The initial phase is a through-plus-
implemented. As a general guideline, simple two-phase control left-turn phase for one direction of the major street, followed by
9-100 urban streets
Figure II.9-1. Phase plans for pretimed and traffic-actuated control.
Figure II.9-2. Dual-ring concurrent phasing scheme with assigned movements.
a through phase for both directions of the major street during green time for one or both left-turn movements. Leading or lagging
which left turns in both directions may be permitted on an optional green phases, or both, can even be used where no left turn exists as
basis. (Note that many operating agencies do not, as a matter of long as turns are permitted to continue during the through phase.
policy, use the optional permitted left turn with this type of phasing The phasing in Figure II.9-1(c) may also be expanded to incorporate
because of safety considerations.) The direction of flow started in leading or lagging green phases on both streets.
the first phase is then stopped, providing the opposing direction All the phase plans discussed to this point can be implemented
with a through-plus-left-turn phase. The final phase accommodates with pretimed or actuated controllers. The only difference in opera-
all movements on the minor street. tion would be the manner in which green time is allocated to the
Such phasing is extremely flexible. When only one left turn re- various phases: for pretimed controllers, green times are preset,
quires a protected phase, a leading green can be provided without a whereas for actuated controllers, green times vary on the basis of
lagging green phase. When left-turn volumes are unequal, the lengths detector actuations.
of the leading and lagging green can be adjusted to avoid excessive At this point, it is necessary to recognize the differences in the
way that modern traffic-actuated controllers actually implement space and time, whereas an inappropriate plan can create great
the phase plan. The phase plans shown in Figure II.9-1 are single- inefficiency. The phase plans presented and discussed in this ap-
ring, sequential phase plans in which a single phase is used to pendix represent a sampling of the more common forms used.
indicate the combination of all movements that are proceeding at They may be combined in a large number of innovative ways on
a given point in time. Modern traffic-actuated controllers do not various approaches to an intersection.
use this scheme. Instead, they implement dual-ring, concurrent Again, local practice is an important determinant in the selection
phasing, in which each phase controls only one movement but two of a phase plan. Also, system considerations should be included
phases are generally being displayed concurrently. when phase plans are established. Phasing throughout an area
The dual-ring concurrent concept is illustrated in Figure II.9-2. should generally be relatively uniform; for example, it may confuse
Note that eight phases are shown, each of which accommodates drivers if protected-plus-permitted phasing is introduced at one
one of the through or left-turning movements. A barrier separates location in an area where left turns are generally handled in exclu-
the north-south phases from the east-west phases. Any phase in sive left-turn phases.
the top group (Ring 1) may be displayed with any phase in the
bottom group (Ring 2) on the same side of the barrier without
introducing any traffic conflicts. For simplicity, the right turns are ALLOCATION OF GREEN TIME
omitted and assumed to proceed with the through movements.
The definition of a phase as presented in Figure II.9-2 is not The allocation of green time is an important input to the method-
consistent with that in Figure II.9-1 nor with the definition given ology presented earlier in this chapter for the estimation of delay.
in the introduction to this chapter. It is, however, a definition It is necessary to know the average cycle length and effective
that is universally applied in the traffic control industry. It is the green time for each lane group to be analyzed. The most desirable
responsibility of the analyst to recognize which definition is appli- way to obtain these values is by field measurement; however,
cable to any given situation. For purposes of the capacity and there are many cases when field measurement is not possible. For
delay analysis procedures presented elsewhere in this chapter, each example, the comparison of hypothetical alternatives precludes
lane group is considered to be controlled separately by a phase field measurements. Even for the evaluation of existing conditions,
with specified red, green, yellow, and all-red times, so either defini- the required data collection is beyond the resources of many
tion could apply. The examples shown throughout the chapter are agencies.
based on the single-ring sequential concept. However, the dual- A procedure for estimating signal timing characteristics is there-
ring definition must be used for estimating the timing plan at fore an important traffic analysis tool. Such a procedure is also
traffic-actuated intersections with the method presented later in useful in designing timing plans that will optimize some aspect of
this appendix. the signal operation. In this respect, pretimed and actuated control
The dual-ring phases that accommodate left turns will only be must be treated differently, because the design and analysis objec-
used if the left turns are protected. Left turns with compound tives are different. For pretimed control, the objective is to design
protection will proceed with their concurrent through movements. an implementable timing plan as an end product. In traffic-actuated
For example, none of the left-turn phases would be used by a dual- control, the timing plan is generated by the controller itself on the
ring controller to implement the two-phase plan shown in Figure basis of operating parameters that are established for each phase.
II.9-1(a). All the other phase plan examples shown in Figure II.9- This creates two separate objectives for traffic-actuated control.
1 may be created by selectively omitting left-turn phases and by The first is to determine how the controller will respond to a
reversing the order in which the through and left-turn phases are specified combination of operating parameters and traffic condi-
displayed in either ring. tions. The second is to provide some indication of the optimal
The advantage of the dual-ring concept is that it is able to values for the key operating parameters.
generate the optimal phase plan for each cycle in response to the
traffic demand. Pretimed controllers, and earlier versions of traffic-
actuated controllers, are more constrained in this regard. The maxi- TIMING PLAN DESIGN FOR PRETIMED CONTROL
mum flexibility is provided by allowing the first (usually left-turn)
phases in Rings 1 and 2 to terminate independently after their The design of an implementable timing plan is a complex and
respective demands have been satisfied. iterative process that is generally carried out with the assistance
It is also possible to constrain these phases to terminate simulta- of computer software. Several popular software products are avail-
neously in emulation of the older, less efficient equipment. For able for this purpose, some of which employ the methodology of
example, simultaneous termination of the northbound and south- this chapter at least in part. There are, however, several aspects
bound left-turn phases in Figure II.9-2 would produce the phasing of signal timing design that are beyond the scope of this manual.
example shown in Figure II.9-1(b). Independent termination of the One such aspect is the choice of the timing strategy itself. Three
two left turns would introduce an overlap phase between the left- basic strategies are commonly used for designing timing plans at
turn phase and the through movement phase in Figure II.9-1(b). pretimed signals:
The overlap phase would accommodate the heavier of the two
left turns together with the concurrent through movement, thereby 1. Equalize the v/c ratios for critical lane groups: This is the
making more effective use of the green time. The degree of benefit simplest strategy and the only one that may be implemented with-
obtained from phase overlaps of this nature depends on the degree out excessive iteration. It will be described briefly in the next
of difference in the opposing left-turn volumes. section of this appendix. It is also employed in the timing plan
Establishment of a phase plan is the most creative part of signal synthesis procedures of the planning method presented earlier in
design and deserves the careful attention of the analyst. A good the chapter and in the sample calculations presented in Section
phase plan can achieve great efficiency in the use of available IV. Under this strategy, the green time is allocated among the
9-102 urban streets
various signal phases in proportion to the flow ratio of the critical The flow ratios are shown, and it is assumed that lost times equal
lane group for each phase. the change-and-clearance interval, which is 4 sec for each phase,
2. Minimize total delay to all vehicles: This strategy is generally or 8 sec/cycle. The cycle length is computed from Equation II.9-
proposed as the ‘‘optimal’’ solution to the signal timing problem, 1 for the desired v/c ratio, Xc, which must be selected by the
often in combination with other measures such as minimization of analyst. The shortest cycle length that will avoid oversaturation
stops and fuel consumption. Many signal timing models offer this may be computed using Xc = 1.00.
optimization feature. Some use a delay estimation procedure iden-
tical to that proposed in this chapter, whereas others employ minor C(minimum) = LXc /[Xc − o (v/s) ]
i
ci (II.9-3)
departures. C(minimum) = 8 (1.0)/[1.0 − (0.45 + 0.35)]
3. Balance the level of service for all critical lane groups:
This strategy promotes a level of service on all approaches that is = 8/0.2 = 40 sec
consistent with the overall intersection level of service. Both of
This cycle length has no direct value in the design of implement-
the previous strategies tend to produce a higher delay per vehicle,
able timing plans; however, it is commonly used as a departure
and therefore a less favorable level of service, for the minor move-
point for iterative procedures that seek to minimize or equalize
ments at an intersection. This causes some difficulty in represent-
delay among lane groups.
ing the overall intersection level of service because of the
If a v/c ratio of no more than 0.8 were desired, the computation
imbalance in level of service for critical lane groups. The sample
would become
calculations presented in Section IV illustrate this phenomenon.
This problem is another example of one that can only be solved C = 8 (0.80)/[0.8 − (0.45 + 0.35)] = 6.4/0 = infinity
iteratively.
This computation indicates that a critical v/c ratio of 0.8 cannot
be provided for the demand levels existing at the intersection. Any
EQUALIZING DEGREE OF SATURATION cycle length of greater than 40 sec may be selected. For purposes
of illustration, assume a cycle length of 60 sec. In all cases, the
Once a phase plan and signal type have been established, the
cycle length assumed would be rounded to the nearest 5 sec for
allocation of green time may be estimated:
values between 30 and 90 sec and to the nearest 10 sec for higher
Xi = viC/(sigi) (9-4) values.
The actual critical v/c ratio provided by a 60-sec cycle is
Xc = o (v/s) ci [C/(C − L)] (9-5)
o (v/s) C/(C − L)
i
Equation 9-5 may be manipulated to solve for the cycle length: Xc = i (II.9-4)
i
C = LXc /[Xc − o (v/s) ]

i
ci (II.9-1) Xc = (0.45 + 0.35)(60)/(60 − 8) = 0.923
Equation 9-4 may be manipulated to solve for the green time for A number of different policies may be employed in allocating the
a particular phase, gi: available green time. A common policy for two-phase signals is
to allocate the green such that the v/c ratios for critical movements
gi = viC/siXi = (v/s)i(C/Xi) (II.9-2) in each phase are equal. Thus, for the example problem, the v/c
where ratio for each phase would be 0.923, and
C = cycle length, sec; gi = (v/s)i(C/Xi) (II.9-5)

L = lost time per cycle, sec; g1 = 0.45 (60/0.923) = 29.3 sec
Xc = critical v/c ratio for the intersection; g2 = 0.35 (60/0.923) = 22.7 sec
Xi = v/c ratio for lane group i (note that the target v/c ratio is 52.0 sec
a user-specified input with respect to this procedure; the Lost time = 8.0 sec
default value suggested in the planning method described
60.0 sec
in this chapter is 0.90);
(v/s)i = flow ratio for lane group i; and Another common policy would be to allocate the minimum
gi = effective green time for lane group i, sec. required green time to the minor approach and assign all remaining
Cycle lengths and green times may be estimated by using these green to the major approach. In that case, the v/c ratio for Phase
relationships, the flow ratios computed as part of the Capacity 2 would be 1.0, and
Analysis Module, and the desired v/c ratios. g2 = 0.35(60/1.0) = 21.0 sec
For pretimed signals, fixed green times and cycle lengths may g1 = 60 − 8 − 21 = 31.0 sec
be estimated using Equations II.9-1 and II.9-2. The procedure will
be illustrated using a sample calculation. Consider the two-phase 52.0 sec
signals shown below. Lost time = 8.0 sec
60.0 sec
Note that in both cases the entire 60-sec cycle is fully allocated
among the green times and the lost time.
The procedure for timing may be summarized as follows:
1. Estimate the cycle length for full saturation using Equation
II.9-1 and Xc = 1.0.
2. Estimate the cycle length for the desired critical v/c ratio, Xc, Worksheets are presented in the NCHRP Project 3–48 Final
using Equation II.9-1. Report (1) to describe each step of the computational process. The
3. From the results of Steps 1 and 2, select an appropriate cycle primary purpose of these worksheets is to provide a clear and
length for the signal. When system constraints determine the cycle concise description of the computations for software development
length, Steps 1 and 2 may be eliminated. purposes. The process is highly iterative, and productive applica-
4. Estimate the green times using Equation II.9-2 and v/c ratios, tion of the manual worksheets is not practical. Only the input
Xi, appropriate to the proportioning policy adopted. data worksheet will be discussed in detail in this appendix. This
5. Check the timing to ensure that the sum of the green times worksheet, presented as Worksheet 1 in Figure II.9-3, gathers to-
and the lost time equals the cycle length. Include overlapping green gether all data required by the analytical model.
times only once in this summation.
Approach-Specific Data
The top portion of Worksheet 1 summarizes the approach-spe-

TIMING PLAN ESTIMATION FOR TRAFFIC-ACTUATED cific information. A separate column is used for each of the four
CONTROL
approaches. The required items are discussed in the following
sections.
A procedure for estimating the timing plan generated by a traf-
fic-actuated controller was developed under NCHRP Project 3-48. Left-Turn Treatment Codes. The logic of the proposed model
This procedure is described fully in the final report for the project requires that the left-turn treatment be identified explicitly. The
(1). In this appendix, the procedure is summarized and examples codes used here are consistent with the definitions presented else-
of its application are presented. The procedure encompasses both where in this chapter:
a traffic-actuated control model and an analytical structure for
0 = does not exist,
implementation of the model.
1 = permitted,
2 = simple protection,
Functional Requirements of Model 3 = compound protection, and
4 = not opposed.
A practical traffic-actuated control model must be functionally
The term ‘‘simple left-turn protection’’ will refer to treatment code
capable of providing reasonable estimates of the operating charac-
2, in which the left turn moves only on the protected phase. The
teristics of traffic-actuated controllers under the normal range of
term ‘‘compound left-turn protection’’ will be used to denote either
design configurations at both isolated and coordinated intersec-
protected-plus-permitted or permitted-plus-protected treatment.
tions. It must also be sensitive to common variations in design
parameters. Examples of design parameters include Position Codes. Position codes are required to distinguish be-
tween leading and lagging left-turn protection. The terms ‘‘lead-
1. Traffic-actuated controller settings (initial interval, allowable
ing’’ and ‘‘lagging’’ apply equally to the cases of simple and
gap, maximum green time),
compound left-turn protection. These terms do not apply if the left
2. Conventional actuated versus volume-density control
turn is not protected. The worksheet offers a simple choice of
strategies,
‘‘Lead’’, ‘‘Lag,’’ or ‘‘N/A.’’ The definition is simple: leading left
3. Detector configuration (length and setback),
turns precede the movement of the opposing through traffic, and
4. Pedestrian timing (Walk signal and flashing Don’t Walk
lagging left turns follow it.
signal),
5. Left-turn treatment (permitted, protected, permitted and pro- Sneakers. This term describes the number of left turns per
tected, not opposed), and cycle that may be dismissed at the end of a permitted phase. An
6. Left-turn phase position (leading or lagging). implicit default of two per cycle is built into the supplemental
permitted left-turn worksheets for purposes of determining the
minimum saturation flow rate. Since any vehicles that rest in the
Data Requirements detection zone will extend their respective phases, a more detailed
treatment of sneakers will be required for traffic-actuated control.
The information that is already required by the procedure in this Free Queue. The current pretimed model assumes that the first
chapter is used to the extent possible to avoid the need for new permitted left turn at the stop line will block a shared lane. How-
data. Most of the additional data items relate to the operation of ever, through vehicles in the shared lane are often able to squeeze
the controller itself. The model structure is based on the standard around one or more left-turning vehicles that constitute the free
eight-phase dual-ring control scheme illustrated in Figure II.9-2. queue. The lack of a free queue parameter is a definite deficiency
This scheme is more or less universally applied in the United of the pretimed model, but it is especially critical with traffic-
States. From a capacity and LOS point of view, less complex actuated control, because vehicles in the free queue do not occupy
phasing concepts (including simple two-phase operation) may gen- the detector and therefore do not extend the green phase. A permit-
erally be represented adequately as a subset of the dual-ring ted left turn stopped on the detector must be treated entirely differ-
scheme. ently in the modeling process than one that is stopped beyond the
For the purposes of this discussion, the scheme for assignment detector.
of movements to phases presented in Figure II.9-2 will be adopted.
This will greatly simplify the illustration of all modeling proce- Approach Speed. The speed of vehicles on a signalized inter-
dures without affecting the generality of the results. section approach (SP) is not important to the current methodology
9-104 urban streets
Figure II.9-3. Worksheet 1: traffic-actuated control input data.
in this chapter. It is, however, required in the analysis of traffic- for this reason. The type of operation created by independent termi-
actuated operation. This speed determines the passage time be- nation is sometimes referred to as ‘‘phase overlap.’’
tween the detector and the stop line as well as the portion of It is, however, not essential that the phases terminate indepen-
intervehicle headways during which a presence detector is occu- dently. Older single-ring controller operation may be approximated
pied. It is a common custom in modeling the operation of vehicles by requiring that the first phases of each ring (i.e., Phases 1 and 5
at a traffic signal to assume a single value for speed that applies or 3 and 7) terminate simultaneously. In some situations involving
throughout the cycle. coordination of controllers on arterial routes, it is common to force
both rings out of their first phase simultaneously. The model to
Termination of Rings 1 and 2. The nature of dual-ring control be developed must consider simultaneous or independent termina-
requires that the second phases of Rings 1 and 2 terminate simulta- tion as legitimate alternatives.
neously because they yield control to approaches with conflicting It is possible that one or more of the first phases will not be used
traffic. However, control may pass from the first phase to the because their associated left turns are not protected. In this case, the
second phase of either ring without causing conflict. Independent question of simultaneous or independent termination will not apply.
termination of the first phases improves efficiency in the allocation This is another multiple-choice entry on the worksheet. The alterna-
of time among competing movements and is generally exploited tives are ‘‘Simultaneous,’’ ‘‘Independent,’’ and ‘‘N/A.’’
Phasing and Detector Design Parameters Maximum Initial Interval. Used only when the initial interval
is extended under volume-density control, the maximum initial
The bottom portion of the worksheet includes all the data items interval (MxI) must be long enough to ensure that a queue of
that are specific to each of the eight phases represented in Figure vehicles released at the beginning of green will be in motion at
II.9-2. A separate column is provided on the worksheet for each the detector before the green terminates.
phase. The first group of data items includes the design parameters
relating to phasing and detector placement that will affect opera- Added Initial per Actuation. Used only when the initial inter-
tion. These data items are discussed in the following sections. val is extended under volume density control, this value (AI) will
depend on the number of approach lanes. It should be long enough
Phase Type. Phase type is the first of several phase-specific to ensure that each vehicle crossing the approach detector on the
inputs that are required. A phase that is not active will not be red will add an appropriate increment of time to the initial interval.
recognized in any subsequent computations. Inactive phases are
indicated by an X in the appropriate column of the worksheet. A Minimum Allowable Gap. The minimum allowable gap (MnA)
left-turn phase will be considered active only if it accommodates is a user-specified controller parameter, the effect of which will
a protected left turn. A through phase will be considered active if be illustrated later as the analytical model is exercised. It is typi-
it accommodates any traffic at all, whether through, left, or right. cally set in the range of 2 to 3 sec and establishes the threshold
Active phases will be designated as follows: for the gap length in traffic that will cause the phase to terminate.
The value of MnA is usually influenced to some extent by local
L = phase accommodates a protected left turn on a green arrow;
practice.
T = phase accommodates through and right-turning traffic only;
It is important to distinguish between the time gap and the time
in this case, all left turns are accommodated entirely on
headway between vehicles. The time headway indicates the elapsed
another phase (i.e., simple left-turn protection);
time between the successive arrival of two consecutive vehicles
G = any left turns accommodated on this phase are opposed by
at a detector. The time gap indicates the elapsed time between the
oncoming traffic (phases with permitted left turns and those
departure of the first vehicle from the detector and the arrival of
with compound left-turn protection); and
the second. The time gap is what is left of the headway after the
N = phase accommodates, in addition to other movements, left
detector occupancy time has been subtracted. A traffic-actuated
turns that are not opposed at any time in the phase sequence
controller using presence detectors views the passage of traffic in
(T intersections, one-way streets, and cases in which the
terms of gaps, not headways.
phasing completely separates all movements on opposing
approaches). Gap Reduction Rate. The gap reduction rate (GR) determines
the rate at which the allowable gap is reduced in volume-density
Note that right turns are not referenced specifically in these desig-
controllers as the green display continues. There are subtle differ-
nations. Right turns are assumed to proceed concurrently with
ences in the definition of the gap reduction rate among controllers.
through traffic.
For the purposes of this procedure, a linear reduction rate (seconds
Phase Reversal. Normally the first (odd-numbered) phase in of reduction of gap per second of elapsed green time) will be
each ring on each side of the barrier handles protected left turns assumed.
and the second (even-numbered) phase handles the remaining traf-
Pedestrian Walk plus Don’t Walk Interval. This parameter
fic. This creates a condition of leading left-turn protection. When
(WDW) is the minimum time given to each phase when pedestrian
lagging left-turn protection is desired, the movements in the first
demand is registered or pedestrian recall is active. It includes both
and second phases are interchanged. Most controllers provide an
the pedestrian Walk and flashing Don’t Walk intervals. These are
internal function to specify phase reversal. For the purposes of
actually entered into the controller as two separate parameters but
this discussion, two phases may be swapped only if both phases
will be combined for the purposes of this procedure. If the pedes-
are active.
trian timing function is not implemented in a particular phase, the
Detector Length. Detector length (DL) is the effective dis- WDW value should be entered as zero.
tance, measured parallel to the direction of travel, through which
Maximum Green. Maximum green (MxG) is a user-specified
a vehicle will occupy the detector. It is a user-specified design
parameter the effect of which will be discussed later. Local practice
parameter influenced by local practice. The detector length influ-
often plays an important part in the determination of maximum
ences the choice of other parameters, such as the allowable gap
green times.
in traffic that will terminate the phase.
Intergreen Time. Another user-specified controller parameter
Detector Setback. Detector setback (DS) is the space between
determined in accordance with local practice, the intergreen time
the downstream edge of the detector and the stop line.
(I) consists of a yellow change interval followed by an all-red
clearance interval. These two intervals are entered separately into
Controller Settings
the controller but will be combined here to simplify the analysis.
The controller itself has several operating parameters that must
Recall Mode. The recall mode determines how a phase will
be specified for each phase. Collectively, these will be referred to
be treated in the absence of demand on the previous red phase.
as the ‘‘controller settings’’ because they must be physically set
The options are as follows:
in the controller with switches, keypads, or some other electrical
means. The following settings will exert a significant influence on None = the phase will not be displayed,
the operation of the intersection and must therefore be recognized Max = the phase will be displayed to its specified maximum
by the analysis methodology. length,
9-106 urban streets
Min = the phase will always be displayed to its specified mini- earlier in this appendix. Under this approach, the green time re-
mum length but may be extended up to its specified quirement is determined by the slope of the line representing a
maximum length by vehicle actuations, target v/c of 0.9. If the phase ends when the queue has dissipated
Ped = the phase will be given the full Walk plus flashing Don’t under these conditions, the target v/c will be achieved.
Walk interval and may be extended further, up to its The second method recognizes the way in which a traffic-actu-
specified maximum, by vehicle actuations, and ated controller actually works. It does not deal explicitly with v/c
Coord = a coordinated phase on the arterial street; this phase will ratios; in fact, it has no way of determining the v/c ratio. Instead
always be displayed for its nominal design time, which it terminates each phase when a gap of a particular length is en-
may be increased by reassigning unused green time from countered at the detector. Good practice dictates that the gap
actuated phases. threshold must be longer than the gap that would be encountered
The recall mode function will have a significant effect on the while the queue is being served. Assuming that gaps large enough
to terminate the phase can only occur after the queue service
operation of the controller. For example, the maximum recall op-
interval (based on v/c = 1.0), the average green time may be
tion will have the effect of creating a nonactuated phase.
estimated as the sum of the queue service time and the green
Minimum Phase Time for Vehicles. The minimum vehicle extension time as shown in Figure II.9-4. Each of these components
phase time (MnV) is actually a traffic engineering input that speci- will be discussed separately.
fies the minimum time during which a phase must be displayed
unless it is skipped because of lack of demand. It is implemented Queue Service Time
in a conventional traffic-actuated controller as the sum of three
The queue service time, gs, can be estimated as
intervals: the initial interval; the minimum allowable gap (MnA),
and the intergreen time (I). As a matter of design, it is important q rr
gs = f q (II.9-6)
that the controller settings be compatible with the minimum phase (s − qg)
times determined by traffic engineering considerations.
This parameter did not appear earlier in the chapter because the where
procedure described does not offer the ability to compute timing qr, qg = red arrival rate and green arrival rate, respectively, veh/
plans. However, it is impossible to deal realistically with traffic- sec;
actuated control without recognition of the existence of a minimum r = effective red time, sec;
phase time. For compatibility with other signal timing programs, s = saturation flow rate, veh/sec; and
the phase times include all intervals, including green, yellow, and fq = 1.08 − 0.1 (actual green time/maximum green time)2.
all-red. (II.9-7)
For the purposes of the worksheet, the minimum phase time
The queue calibration factor, fq, was described by Akçelik (2) as
must be replaced by the maximum phase time (MxG + I) if the
a factor required to account for randomness in arrivals in determin-
‘‘recall to maximum’’ mode is in effect.
ing the average queue service time.
Green Time Estimation Model Green Extension Time
The discussion of pretimed operation presented earlier in this To estimate the extension time analytically for a particular
appendix indicated that the determination of required green time phase, it is necessary to determine the expected waiting time for
is a relatively straightforward process when the cycle length is a gap of a specific length given the average intervehicular head-
given. However, traffic-actuated controllers do not recognize spec- ways and some assumptions about the headway distribution. An
ified cycle lengths. Instead, they determine, by a mechanical anal- analytical model for this purpose was described by Akçelik (2,3).
ogy, the required green time given the length of the previous red This model made use of earlier work by Lin (4,5). The average
period and the arrival rate. This is accomplished by holding the green extension time, ge, is estimated from the following formula,
right-of-way until the accumulated queue has been served. which is based on the use of a bunched exponential distribution
The basic principle underlying all signal timing analysis is the of arrival headways:
queue accumulation polygon (QAP), which plots the number of el(e +t −D) 1
0 0
ge = − (II.9-8)
vehicles queued at the stop line during the cycle. The QAP for a wq l
simple protected movement is illustrated in Figure II.9-4. The
queue accumulation and discharge are represented in this very where
simple case as a triangle. The accumulation takes place on the left e0 = unit extension time setting, MnA on Worksheet 1;
side of the triangle (i.e., the effective red) and the discharge takes t0 = time during which the detector is occupied by a passing
place on the right side of the triangle (i.e., the effective green). vehicle;
More complex polygons are generated when permitted movements
t0 = (Ld + Lv)/v (II.9-9)
occur and when a movement proceeds on more than one phase.
An extensive discussion of this subject may be found earlier in Lv = vehicle length, assumed for the purposes of this discussion
this chapter. to be 18 ft;
Two methods of determining the required green time given the Ld = detector length, DL on Worksheet 1;
length of the previous red are illustrated in Figure II.9-4. The first v = vehicle approach speed, SP on Worksheet 1;
employs a ‘‘target v/c’’ approach. This method is the basis for D = minimum arrival (intrabunch) headway, sec;
the planning method described earlier in the chapter and for the w = proportion of free (unbunched) vehicles; and
discussion on timing plan design for pretimed control presented l = parameter calculated as
Figure II.9-4. Queue accumulation polygon illustrating two methods of green time computation.
wq phases. Thus, a circular dependency is established whose solution

l= (II.9-10)
1 − Dq requires an iterative process. With each iteration, the green time
required by each phase, given all the green times required by the
where q is the total arrival flow in vehicles per second for all lane other phases, may be determined.
groups that actuate the phase under consideration. The logical starting point for the iterative process is the mini-
The bunched exponential distribution of arrival headways was mum times specified for each phase. If these times turn out to be
originally proposed by Cowan (6). A detailed discussion of this adequate for all phases, the cycle length will simply be the sum
model and the results of its calibration using real-life data for of the minimum phase times for the critical phases. If a particular
single-lane traffic streams and simulation data for multilane phase demands more than its minimum time, more time must be
streams are given by Akçelik and Chung (7). The following rela- given to that phase. Thus, a longer red time must be imposed on
tionship can be used for estimating the proportion of free (un- all the other phases. This in turn will increase the green time
bunched) vehicles in the traffic stream (w): required for the subject phase.
w = e−bDq (II.9-11)
where b is a bunching factor. The recommended parameter values Simple Two-Phase Example
based on the calibration of the bunched exponential model using
real-life and simulation data are as follows: The circular dependency will converge quite reliably through a
series of repeated iterations. The convergence may be demon-
Single-lane case:
strated easily using a trivial example. More complex examples
D = 1.5 b = 0.6 (II.9-12a) will be introduced later to examine the effects of controller settings
Multilane case (number of lanes = 2): and traffic volumes in a practical situation.
Consider an intersection of two streets with a single lane in each
D = 0.5 b = 0.5 (II.9-12b) direction. Each approach has identical characteristics and carries
Multilane case (number of lanes > 2): 675 vehicles per hour with no left or right turns. The average
headway is 2.0 sec/veh and the lost time per phase is 3.0 sec.
D = 0.5 b = 0.8 (II.9-12c) Detectors are 30 ft long with no setback from the stop line. The
actuated controller settings are as follows:
Computational Structure for Green Time Estimation
Setting Time (sec)
This green time estimation model is not difficult to implement, Initial interval 10
but it does not lead directly to the determination of an average Unit extension 3
cycle length or green time, because the green time required for Maximum green 46
each phase is dependent on the green time required by the other Intergreen 4
9-108 urban streets
The maximum time for each phase will be (46 + 4) = 50 sec. The el(e + t −D) 1
0 0
ge = −
minimum phase time will be 10 + 3 + 4 = 17 sec, which will be wq l
the starting point for the timing computations. So the first iteration
= [e 0.222(3.0+1.09−1.5)/(0.84 × 0.1875)] − (1/0.222)
will use a 34-sec cycle with 17 sec of green time on each approach.
Allowing for lost time, the effective red time will be 20 sec, and = 6.63 sec
the effective green time will be 14 sec for each phase. 9. Compute the total phase time:
The total lost time is the sum of two components, including the
startup lost time and the clearance lost time. In this chapter’s G = l 1 + gs + ge + Y
procedure for estimation of capacity and delay, all the lost time is = 2 + 12.86 + 6.63 + 4
assumed to be concentrated at the beginning of the green. This is = 25.49
a valid approximation for delay estimation because the lost time 10. Compute the phase time deficiency as the difference be-
is used only in the computation of effective green time, and its tween the trial phase time and the computed phase time, or 25.49
position in the phase is irrelevant. − 17.0 = 8.49 sec.
However, for traffic-actuated timing estimation, the distribution
of lost time throughout the phase will have a definite influence on This computation indicates that the trial phase time was not ade-
the results. The lost time at the beginning of the phase will influ- quate to satisfy the rules under which the controller operates. It
ence the length of the phase. The lost time at the end of the phase also suggests a new trial green time of 25.49 sec and a cycle length
will influence the delay, but it will have no effect on the phase of 50.98 sec for the next iteration.
duration. It is recommended by Courage et al. (1) that for a speci- The next iteration will still produce a green time deficiency,
fied lost time of n sec, 1 sec be assigned to the end of the phase because the red time has been increased. However, this deficiency
and n − 1 sec be assigned to the beginning. Thus, for this example will be smaller. Successive iterations will produce successively
the startup lost time (l1) will be 2.0 sec. smaller green time deficiencies until eventually the solution will
The computational process may be described as follows: converge. This process is illustrated in Figure II.9-5. The solution
converged (i.e., the green time deficiency became negligible) at a
1. Compute the arrival throughout the cycle, q: phase time of 37.5 sec, producing a cycle length of 75 sec. This
q = 675/3,600 = 0.1875 veh/sec was based on a threshold of 0.1 sec difference in the computed
cycle length between iterations. In other words, the process termi-
2. Compute the net departure rate (saturation flow rate minus nated when the cycle lengths on two successive iterations fell
arrival rate): within 0.1 sec of each other.
(s − q) = 0.5 − 0.1875 = 0.3125 veh/sec As a matter of interest, consider the effect of reducing the unit
extension time, e0, from 3.0 to 2.0 sec. This would be expected to
3. Compute the queue at the end of 20 sec of effective red time: reduce the green extension time, ge for both phases and to shorten
(qrr) = 20 ∗ 0.1875 = 3.75 veh the resulting cycle length. The extent of the reduction may be
estimated by repeating all of the steps described above with the
4. Compute the queue calibration factor fq from Equation II.9-7: new value for ge. In the first iteration, the queue service time will
fq = 1.08 − 0.1 (13/46)2 = 1.07 remain the same, but the green extension time will be reduced
from the value of 6.63 sec computed above to 4.42 sec. Repeated
5. Compute the time required to serve the queue, gs: iterations with this lower unit extension time would converge to
gs = 1.07 (3.75/0.3125) = 12.86 sec a cycle length of 65.3 sec.
In this example, the green time for both phases was determined
Thus, after 12.86 sec of effective green time, the queue will have by the sum of the queue service time and the extension time. Phase
been served and gaps will start to be observed at the detector. The times will also be constrained by their specified maximum and
startup lost time (l1 = 2) must be added to the queue service time minimum times. If the maximum phase times had been set at a
for purposes of determining the total phase time requirement. The value less than the 38 sec computed above, the iterative procedure
question now is how long one would expect to wait for a gap of would have terminated before the computed times were reached.
3.0 sec.
6. Determine the parameters of Equation II.9-8 as follows.
Minimum Phase Times
D = 1.5 b = 0.6 from Equation II.9-12a
w = e−bDq The whole question of minimum phase time requires more atten-
= e−(0.6×1.5×0.1875) = 0.84 tion. The specified minimum green time constraints are valid only
wq for pretimed phases and phases that are set to recall to the minimum
l= time regardless of demand. The real significance of the minimum
1 − Dq
= (0.84 × 0.1875)/[1 − (1.5 × 0.1875)] = 0.222 phase time for an actuated phase is that the phase must be displayed
for its specified minimum time unless it is skipped because of
7. Determine the occupancy time of the detector for a vehicle lack of demand. This situation may be addressed analytically by
length of 18 ft, a detector length of 30 ft, and an approach speed determining the probability of zero arrivals on the previous cycle.
of 30 mph: Assuming a Poisson arrival distribution, this may be computed as
t0 = (30 + 18)/(1.47 × 30) = 1.09 sec P0v = e−qC (II.9-13)
8. Apply Equation II.9-8 to determine the expected green exten- where q is the vehicle arrival rate in vehicles per second and C is
sion time, ge: the cycle length for the current iteration.
Figure II.9-5. Convergence of green time computation by elimination of green time deficiency.
Thus, assuming that the phase will be displayed for the minimum turns (from both exclusive and shared lanes), and compound left-
time except when no vehicles have arrived, the adjusted minimum turn protection (both leading and lagging).
phase time becomes Two extensions to the methodology presented to this point are
required to deal with more complex situations. The first is the
AVM = MnV(1 − P0v) (II.9-14)
extension of the QAP from its simple triangular shape to a more
where AVM is the adjusted vehicle minimum time and MnV is the complex shape that represents different arrival and departure times
specified minimum green time from the worksheet in Figure II.9- at different points in the cycle. The second is a procedure to
3. This relationship also has circular dependencies because, as the synthesize a complete single-ring equivalent sequence by combin-
adjusted minimums become shorter, the probability of zero arrivals ing critical phases in the dual-ring operation. The QAP extensions
also becomes higher, which further reduces the adjusted mini- will be considered first.
mums. Fortunately, the solution fits well into the iterative scheme Figure II.9-4 presents the familiar triangular QAP for a protected
just described. movement from an exclusive lane. There are four other cases to
The use of adjusted minimum green times offers a practical be considered, including (a) permitted left turns from an exclusive
method for dealing with phases that are not displayed on each lane, (b) permitted left turns from a shared lane, (c) protected-plus-
cycle but whose minimum durations may be determined by agency permitted left turns, and (d) permitted-plus-protected left turns. The
policy. The concept applies equally to pedestrian minimum times. QAP shapes for each of these cases are illustrated in Figures II.9-
6 through II.9-9, all of which conform to a common terminology
with respect to labeling. Intervals are illustrated along the hori-
Multiphase Operation
zontal axis as follows:
Three important concepts have been introduced for estimating the r = effective red time,
timing plan at traffic-actuated signals: (a) a model for predicting gq = portion of the permitted green time blocked by a queue of
the green time for any phase given the length of the previous red opposing vehicles,
period, (b) an iterative computational structure that converges to gu = portion of the permitted green time not blocked by a queue
a stable value for the average cycle length and green time, and (c) of opposing vehicles,
a procedure to account for minimum green times with low vol- gs = portion of the protected green time required to serve the
umes. These concepts were illustrated in a trivial example, but queue of vehicles that accumulated on the previous phases,
fortunately they are robust enough to deal with the practical com- ge = extension to the protected green time that occurs while the
plexities of traffic-actuated control. These complexities include controller waits for a gap in the arriving traffic long enough
multiphase operation (both single- and dual-ring), permitted left to terminate the phase, and
9-110 urban streets
Figure II.9-6. Queue accumulation polygon for permitted left turn from exclusive lane.
Figure II.9-7. Queue accumulation polygon for permitted left turn from shared lane.
gf = portion of the green time in which a through vehicle in a Qq = queue size at the end of the interval gq;
shared lane would not be blocked by a left-turning vehicle Qp = queue size at the end of the permitted green period;
waiting for the opposing movement to clear (this condition Qp′ = queue size at the end of the permitted green period, adjusted
occurs only at the beginning of the permitted green when for sneakers;
one or more through vehicles are at the front of the queue). Qga = queue size at the beginning of the protected green (green
arrow) period; and
Note that in each case the phases are arranged so that the protected Qf = queue size at the end of the interval gf.
phase is the last to occur. The length of this phase will be deter-
mined by its detector actuations. The actual length will be the sum The shape of each of the QAPs is based on termination by a gap
of the time required to serve the queue that exists at the beginning that exceeds the unit extension, allowing the full extension time,
of the phase plus the extension time. ge, to be displayed. When a phase terminates on the maximum
Points in the cycle at which the queue size is important to the green time, the extension time may be reduced or eliminated. If a
computations are also identified as follows: permitted left-turn phase terminates before the queue has been
served, a maximum of two sneakers will be dismissed from the
Qr = queue size at the end of the effective red; queue at that point.
Figure II.9-8. Queue accumulation polygon for protected-plus-permitted left-turn phasing with exclusive left-turn lane.
Figure II.9-9. Queue accumulation polygon for permitted-plus-protected left-turn phasing with exclusive left-turn lane.
These extensions to the QAP analysis will accommodate all three columns for the north-south movements (right side of the
of the practical conditions covered by the procedures presented barrier).
earlier in this chapter. The remaining issue to be dealt with is There are three rows for each of the two rings. The first row
the synthesis of the complete cycle by combining critical phases indicates whether the phase pair is swapped. This information was
in a dual-ring operation. This procedure may be carried out using entered on Worksheet 1 in this appendix. The next two rows give
Worksheet 2 (Figure II.9-10). The structure of this worksheet is the movements and phase times for their respective phases. If the
compatible with the dual-ring concurrent phasing illustrated in phases are not swapped, the assignments will be as shown in the
Figure II.9-2. dual-ring configuration of Figure II.9-2. If they are swapped, the
The east-west movements (left side of the barrier) are shown movements and times for the phase pair will be reversed. When
in the first three columns. The first column, labeled A, represents a phase pair is reversed, the through movements will appear in
the first phase in each ring (1 or 5). The second, or B, column Column A and the left turns in Column B. Note that the movements
represents the second phase (2 or 6). The third column will in a given phase pair cannot be swapped if the left turn is not
contain the total of the phase times for the movements in the protected. The order of the phases in the pair does not affect the
first two columns. The same format is repeated in the second total phase time entered in the last column.
9-112 urban streets
Figure II.9-10. Worksheet 2: traffic-actuated timing computations.
The seventh row contains the absolute value of the phase time more green time incrementally to the coordinated phases until the
difference between the two rings. Values are entered in each of design cycle length has been reached. The result is a timing plan
the six columns. that approximates the operation of the controller in the field.
The components of the cycle time must now be determined and The procedure for timing plan estimation in coordinated systems
entered in the Cycle Time Components row. The procedure will requires that a design timing plan be established first, with phase
depend on whether the first phase termination is simultaneous or splits that add up to the design cycle length. This becomes the
independent. For simultaneous termination, the maximum value starting point for the iterative procedure that involves the follow-
of each phase is entered in the A and B columns (Ring 1 or ing steps:
2, whichever time is greater). For independent termination, the 1. Set up the controller timing parameters for the initial timing
maximum value of the total time (A + B) from Ring 1 or 2 is plan computations. The coordinated arterial phases (usually 2 and
entered. For each side of the barrier, either the A and B columns 6) should be set for recall to maximum. The maximum green times
or the Total column will have an entry, but not all three columns. for all phases should be determined by their respective splits in
This procedure should be carried out for both sides of the barrier. the pretimed timing plan. No recall modes should be specified for
It should be recalled that the termination treatment may be different any of the actuated phases.
on either side. The cycle length may now be determined as the 2. Perform the timing computations to determine the resulting
sum of all the entries in the Cycle Time Components row. cycle length. If the maximum green times have been specified
If the computed cycle length agrees with the cycle length deter- correctly in Step 1, the computed cycle length will not exceed the
mined on the previous iteration, no further action will be necessary. specified cycle length.
If it does not, this timing plan will serve as the starting point for 3. If the computed cycle length is equal to the specified cycle
the next iteration. length, no green time is available for reassignment. In this case
the procedure will be complete and the final timing plan will be
produced.
COORDINATED SEMIACTUATED OPERATION 4. If the computed cycle length is lower than the specified cycle
length, some time should be reassigned to the arterial phases. This
It has been pointed out that nonactuated phases under semiactu- is accomplished by increasing the maximum green times for the
ated control may be coordinated with those in neighboring intersec- coordinated phases. The recommended procedure is to assign one-
tions. In the most common coordination scheme, a background half of the difference between the computed cycle and the specified
cycle length is imposed. The actuated phases receive their allot- cycle to the coordinated phases. This provides a reasonable speed
ment of green time in the usual manner, except that their maximum of convergence without overshooting the specified cycle length.
green times are controlled externally to ensure conformance with 5. Repeat Steps 2 through 4 iteratively until the computed and
the specified cycle length. If the actuated phases require all of specified cycle lengths converge.
their nominal green time allotment, the intersection operates in a The results of these computations will be illustrated in the follow-
more or less pretimed manner. If not, the unused time is reassigned ing example problem.
to the coordinated phase.
The computational structure developed under NCHRP Project A MULTI-PHASE EXAMPLE
3-48 is able to approximate this operation quite effectively. The
analysis of coordinated operation requires another iterative loop, The complete timing estimation procedure described in this ap-
which executes the procedure described in this appendix, adding pendix will now be illustrated using the multiphase example pre-
sented as Sample Calculation 3 earlier in this chapter. In the operation will also be examined at the reduced volume levels to
discussion of this example presented previously, observed green observe the reassignment of unused green time from the actuated
times were first used. The green times were then modified by trial phases to the nonactuated phases to reduce the delay to arterial
and error to arrive at a final design for analysis of capacity, delay, traffic.
and level of service. The first set of conditions to be analyzed involves 100 percent
The intersection layout for this example is shown in Figure 9- volume levels, 120 sec maximum green time on each phase, and
53. Note that all left turns take place from exclusive lanes. The short unit extensions. The input data worksheet for this example
northbound and southbound left turns have protected-plus-permit- is shown in Figure II.9-11. The resulting cycle length is 238.2 sec.
ted phasing. The eastbound and westbound left turns are permitted, Because of the dual-ring operation, an overlap phase appeared for
with no protected phases. the north-south approaches in the results. The estimated average
The final timing plan design was as follows: phase times were as follows:
Phase I: NB and SB left, 12.4 sec NB and SB left, 17.2 sec
Phase 2: NB and SB green, 49.9 sec SB through and left, 15.0 sec
Phase 3: EB and WB green, 27.7 sec NB and SB green, 124.0 sec
Cycle length, 90 sec EB and WB green, 82.0 sec
Cycle length, 238.2 sec
This timing plan was shown to accommodate all movements with
no excessive delay. The average delay per vehicle for all ap- This timing plan, when analyzed by the procedure described
proaches was 37.0 sec. earlier in this chapter, produced the results shown on the LOS
The timing estimation methodology for this example will now worksheet in Figure II.9-12. Note that the average delay per vehicle
be exercised using several different values for some of the actuated was 69.7 sec. This is considerably higher than the 37.0-sec/veh
controller settings to observe their effect on the results. The detec- delay associated with the timing plan developed in Sample Calcu-
tor configuration will use loop detectors 30 ft long and positioned lation 3. The logical conclusion here is that the peak-hour volumes
at the stop line with no setback. A 30-mph design speed will be could not be handled in an optimal manner by a fully actuated
assumed for all approaches. Both isolated and coordinated opera- controller without some influence being exerted on the timing plan
tion will be explored. Different values will be used for the unit through maximum green constraints.
extension and maximum green settings. Three different sets of unit A total of 10 alternatives similar in concept to the one just
extension settings will be used: described were analyzed using combinations of these conditions.
For each analysis, the average phase times and cycle length were
T Short values of 1.5 sec and 2.0 sec for two-lane approaches
recorded, along with the average delay per vehicle and any move-
and one-lane approaches, respectively. This represents a condition
ments that were oversaturated. The results are summarized in Table
sometimes referred to as ‘‘snappy’’ operation.
II.9-1.
T Medium values of 2.5 sec and 3.5 sec for two-lane approaches
In the first three alternatives, only the unit extensions (short,
and one-lane approaches, respectively. This will be the standard
medium, and long) were changed. Note that the cycle length in-
condition for most scenarios.
creased with the unit extensions from 238.2 sec (short) to 250.9
T Long values of 3.5 sec and 4.5 sec for two-lane approaches
sec (long). In all cases, the northbound green phase reached its
and one-lane approaches, respectively. This represents a condition
maximum of 124 sec (i.e., 120 sec green plus 4 sec intergreen).
sometimes referred to as ‘‘sluggish’’ operation.
Because the northbound phase was already at its maximum length,
These values represent the actual gap between vehicles that will it lost time proportionally as the other phases increased and there-
cause a phase to terminate. With the assumed approach speed and fore became more oversaturated. It is clearly essential that some
detector configuration, each vehicle (assumed length, 18 ft) passing maximum green times be imposed to control the apportionment
over the loop will occupy the detector for an additional 1.09 sec. of time between the competing phases.
Three different values for maximum green times will also be The next two analyses used the medium extension times and
investigated. The first will use very long maximum times (120 sec evaluated both of the strategies for setting the maximum green
for each phase) to determine how the intersection would operate intervals in proportion to their design values. This reduced the
if most phases terminated on the unit extension. The second will average delay to 41.4 sec per vehicle for the 150 percent strategy
use maximum times that are proportional to the design times for and 37.1 sec per vehicle for the 100 percent strategy. The 150
each phase. Kell and Fullerton (8) proposed that the maximum percent strategy produced slightly oversaturated conditions for the
extensions be set at 125 to 150 percent of the design green times. southbound approach. Note that the results of the 100 percent
Lin (9) proposed a more complex scheme that resulted in maximum strategy were similar to the results reported in the discussion of
times in the range of 150 to 200 percent of the design times. the design timing plan for Sample Calculation 3. The interpretation
For purposes of this exercise, the maximum times will be set at here is that when traffic volumes are close to capacity, as they are
approximately 150 percent of the design times, representing a value in this example, maximum green times must be used to apportion
somewhere in the middle of the range suggested in the literature. the time among the competing phases. The gap termination tactic
The third will use the actual design green times as maximums to does not ensure a satisfactory distribution of green times.
constrain the operation to its original design. One of the major benefits of traffic-actuated control is its
In recognition of the fact that much of the benefit of traffic- ability to respond to short-term and daily fluctuations in traffic
actuated control is derived from the ability of the controller to volumes. To illustrate this principle, the volumes were reduced
respond to fluctuations in traffic volumes throughout the day, the across the board to 70 percent of their peak values. The analysis
operation will also be examined at volume levels of 70 percent of using 120-sec maximum green times on all phases was repeated
the peak-hour levels reflected in the original data. Coordinated with the three levels of unit extension. The results indicate that
9-114 urban streets
Figure II.9-11. Traffic-actuated control data for multiphase example.
Figure II.9-12. LOS results for multiphase example.
the cycle length varied from 62.3 to 76.6 sec throughout the would have been able to control both the full- and reduced-
full range of unit extension settings. The average delays were volume settings effectively provided that the maximum greens
almost identical for these three cases, varying from 16 to 18 were optimized for the high-volume conditions.
sec/veh. This indicates that when traffic volumes drop below The final two cases deal with coordinated operation. Each of
their saturation levels, it is no longer necessary to control the the intersecting routes was assumed to be coordinated in separate
distribution of green times using the maximum green settings. cases. The design timing plan, based on a 95-sec cycle, was used
An important observation here is that the same controller settings to establish nominal splits for the coordinated operation. The me-
Table II.9-1. Comparison of Traffic-Actuated Controller Settings for Multiphase Example

estimated phase times (sec) average
volume maximum green gap sbt delay
level times (sec) (sec) nsl +l nsg ewg cycle (sec/veh) comments
Short 17.2 15.0 124.0 82.0 238.2 69.7 NB-TH v/c = 1.035
Set to 120 sec for
all phases to
100 percent Med. 19.2 16.3 124.0 85.3 244.8 76.0 NB-TH v/c = 1.064
eliminate maximum
green constraints
Long 20.8 17.4 124.0 88.7 250.9 81.8 NB-TH v/c = 1.091
Set to 150 percent
of design splits
100 percent Med. 13.4 1.6 79.0 43.0 137.0 41.4 SB-LT v/c = 1.072
indicated in Chapter
9 example
Set to 100 percent Similar to Sample Calculation 3 results
of design splits presented earlier in chapter (different
100 percent Med. 12.4 — 49.9 27.7 90.0 37.1
indicated in Chapter extensions)
9 example NB-TH v/c = 1.022
Short 6.8 1.8 31.7 22.0 62.3 16.0
Set to 120 sec for
all phases to
70 percent Med. 7.5 2.1 35.4 22.0 67.0 16.5
eliminate maximum
green constraints
Long 8.4 2.4 41.7 24.1 76.6 18.0
Set to 150 percent Med. 10.3 — 60.0 25.2 95.5 16.3 North-south arterial phases coordinated
70 percent of design splits
indicated in Chapter
9 example (no
overlap phase) Med. 12.4 — 45.4 36.0 93.8 23.4 East-west arterial phases coordinated
dium unit extensions were used in combination with maximum methodology presented earlier in this chapter. One example is the
green times that reflected 150 percent of the design timing plan. case of compound left-turn protection with opposing shared lanes
Because the design timing plan included no overlap, the input data for left turns and through movements. The methodology in this
values were adjusted to require simultaneous termination of the chapter treats this as a separate case (Case 6 in Table 9-12) and
first (i.e., left-turn) phases that accommodate the northbound and applies an empirical treatment to determine the saturation flow
southbound traffic as opposed to independent termination. adjustment factor for left turns. Simulation provides the only effec-
The results indicate that the cycle length established by the tive way to estimate the timing plan parameters for this case.
iterative computations fell within 1 sec of the design cycle length. The sample problem presented in this appendix demonstrates
In the case involving east-west coordination, the east-west phases the sensitivity of the procedure to the unit extension times set in
received more than their nominal allotment of time at the expense the controller. As expected, longer unit extension times produce
of the north-south movements. The reverse was true when coordi- longer average green times except when constrained by the maxi-
nation was established for the north-south approaches. The delay mum green time settings. Shorter extension times have the opposite
was also reduced on the approaches that were coordinated, and effect. There is, however, a lower limit to the range of unit exten-
increased on those that were not. sion times that can be modeled realistically. It is well known in
practice that when the unit extension times are too short, premature
LIMITATIONS OF TRAFFIC-ACTUATED TIMING ESTIMATION terminations of a phase may result from anomalies in the departure
PROCEDURE headways created primarily by lapses in driver attention. The traf-
fic-actuated control model described in this appendix assumes a
The traffic-actuated timing estimation procedure described in this constant departure headway and does not therefore reflect this
appendix provides a reasonable approximation of the operation of phenomenon. Simulation models introduce a stochastic element
a traffic-actuated controller for nearly all the conditions encoun- into the departure headways based on a theoretical distribution.
tered in practice. As indicated by Courage et al. (1), the results They are therefore able to invoke premature phase terminations to
obtained from this method have correlated well with extensive some extent, but they do not deal with anomalous driver behavior.
simulation data and with limited field studies. However, the proce- As a practical matter, unit extensions should reflect headways
dure involves a deterministic analytical representation of an ex- at least 50 percent longer than the expected departure headways.
tremely complex stochastic process, and therefore has some For example, assuming a 2-sec average departure headway, unit
limitations. extensions should accommodate a 3-sec departure headway with-
Some of the limitations result from unique situations that cannot out terminating the phase. Assuming a detector occupancy time
be modeled analytically in a satisfactory manner, even by the of 1 sec, this implies a 2-sec gap. So the minimum practical value
9-116 urban streets
for the unit extension would be 2 sec. Smaller values may be When traffic volumes are extremely low, the timing plan be-
appropriate in multiple-lane cases in which the average departure comes somewhat of an abstraction unless the recall function is
headways are shorter. used for each phase. In the absence of any demand, the green
The analysis of permitted left turns from shared lanes always indication rests on the phase that received the last demand, and
poses special problems. The semiempirical treatment prescribed may do so for several minutes. This implies that very long red
for shared-lane permitted left turns in this chapter does not lend times will be displayed on some phases; however, no delay will
itself to the iterative timing estimation procedure described in this be associated with these red times, because no vehicles will be
appendix. An analytical approximation of the shared-lane model affected. The procedure described in this appendix will compute
was therefore substituted to ensure stable convergence of the solu- very short equivalent red times for these phases in an attempt to
tion. This produced timing plans that agreed well with simulation provide a signal timing plan that will produce realistic delays.
results; however, the analysis of delay resulting from the timing Although the procedure produces a reasonable approximation, de-
plan did not always agree with the results of the methodology sign decisions based on comparison of delays in the LOS A range
in this chapter. It appears that an iterative method of achieving should be avoided.
equilibrium between the shared lane and the adjacent through lanes
in this methodology is a prerequisite to the development of a
satisfactory timing estimation procedure.
REFERENCES
1. Courage, K.G., Fambro, D.F., Akçelik, R., Lin, P.S., and 5. Lin, F.B., ‘‘Predictive Models of Traffic-Actuated Cycle
Anwar, M., Capacity Analysis for Actuated Intersections. Splits.’’ Transportation Research, Vol. 16B, No. 5 (1982), pp.
NCHRP Project 3-48 Final Report. University of Florida, Texas 65–72.
Transportation Institute, and ARRB Transport Research, Ltd. 6. Cowan, R.J., ‘‘Useful Headway Models.’’ Transportation Re-
(1996). search, Vol. 9, No. 6 (1975), pp. 371–375.
2. Akçelik, R., Analysis of Vehicle-Actuated Signal Operations. 7. Akçelik, R., and Chung, E., ‘‘Calibration of the Bunched
Working Paper WD TE 93/007, Australian Road Research Exponential Distribution of Arrival Headways.’’ Road and
Board (1993). Transport Research, Vol. 3, No. 1 (1994) pp. 42–59.
3. Akçelik, R., ‘‘Estimation of Green Times and Cycle Time for 8. Kell, J.H., and Fullerton, I.J., Manual of Traffic Signal
Vehicle-Actuated Signals.’’ Transportation Research Record Design. Institute of Transportation Engineers, Prentice-Hall,
1457, TRB, Washington, D.C. (1994). Inc., Englewood Cliffs, N.J. (1982).
4. Lin, F.B., ‘‘Estimation of Average Phase Durations for Full- 9. Lin, F.B., ‘‘Optimal Timing Setting and Detector Lengths of
Actuated Signals.’’ Transportation Research Record 881, TRB, Presence Mode Full-Actuated Control.’’ Transportation Re-
Washington, D.C. (1982), pp. 65–72. search Record 1010, TRB, Washington, D.C. (1985), pp. 37–45.

APPENDIX III
MEASUREMENT OF INTERSECTION CONTROL DELAY IN THE FIELD
As an alternative to estimation of average control delay per may be obtained by driving through the intersection a few times
vehicle using Equation 9-22 and the progression adjustment factor, when the light is green and no queue is present and recording the
delay at existing locations may be measured directly. There are a speed at a location least affected by signal control. Typically, the
number of measurement methods, including the use of test-car recording location should be upstream and about midblock. This
observations, path tracing of individual vehicles, and recording of value and the number of lanes in the lane group should be recorded
arrival and departure volumes on a cycle-by-cycle basis. The on the worksheet.
method summarized here is based on direct observation of vehicle- The survey should begin at the start of the red phase of the
in-queue counts at the intersection and, unless the volume is very study lane group, ideally when there are no cycle failures (no
light, requires two field personnel per lane group surveyed. Also overflow queue) from the previous green period. The reason for
needed is a multifunction digital watch with a timer that counts this is the need for consistency with the analytic delay equation,
an interval of, for example, 17 sec and then repeats the countdown which is based on delay to vehicles that arrive during the study
for another 17-sec interval, plus a volume count board that has at period, not before. If the survey does start with an overflow queue,
least two tally counters. Alternatively, a laptop computer can be the overflow vehicles need to be excluded from subsequent queue
programmed to emit audio count markers at user-selected intervals, counts; that is, only vehicles arriving during the study period are
take volume counts, and execute real-time delay computations, included in the queue and arrival volume counts. It should not be
thus simplifying data reduction. difficult to do this as long as the number of overflow vehicles
In general, this method is applicable to all signalized intersec- is small.
tions operating from low, undersaturated conditions to conditions
in which the lane group demand-to-capacity ratio approaches 1.0.
PROCEDURE
For oversaturated conditions, queue buildup normally makes the
method impractical. Under those circumstances, more personnel
will be required to complete the field study, and other methods Observer 1
may be considered, such as an input-output technique or a zoned
1. Keep track of the end of standing queues for each cycle in
survey technique. In the input-output technique, different observers
the survey period, that is, the last vehicle in each lane that stops
count arrivals separately from departures and vehicles in queue
because of the signal. This includes vehicles that arrive when the
are calculated as the accumulated difference, subject to in-process
signal is actually green but stop because vehicles in front have not
checks for vehicles leaving the queue before they reach the stop
yet started moving.
line. The zoned survey technique requires subdividing the approach
For purposes of the survey, a vehicle is considered as having
into manageable segments to which the observers are assigned;
joined the queue when it approaches within one car length of a
they then count queued vehicles in their assigned zone. Both of
stopped vehicle and is itself about to stop. This definition is used
these techniques require more personnel and are more complicated
because of the difficulty of keeping precise track of the moment
in setup and execution.
when a vehicle comes to a stop. All vehicles that join a queue are
The method described here is applicable to situations in which then included in the vehicle-in-queue counts until they cross the
the average maximum queue per cycle is no more than about 20 stop line.
to 25 vehicles per lane. When queues are long or the demand-to- 2. At regular intervals of between 10 and 20 sec, record the
capacity ratio is near 1.0, care must be taken to continue the number of vehicles in the queue (e.g., using the countdown-and-
vehicle-in-queue count past the end of the arrival count period, as repeat timer on a digital watch to control the count intervals). The
detailed below. This requirement is for consistency with the ana- regular intervals should not be an integral divisor of the cycle
lytic delay equation used earlier in the chapter. length (e.g., if the cycle is 120 sec, use 14-sec or 16-sec count
The method does not directly measure delay during deceleration intervals, not 15-sec intervals). Vehicles in queue are those that
and a portion of acceleration delay, which are very difficult to (a) are included in the queue of stopping vehicles as defined in
measure without sophisticated tracking equipment. However, this Step 1 and (b) have not yet exited the intersection. A through
method has been shown to yield a reasonable estimate of control vehicle can be considered as having exited the intersection when
delay. It includes an adjustment for errors that may occur when its rear axle crosses the stop line. A turning vehicle can be consid-
this type of sampling technique is used, as well as an acceleration- ered as having exited the intersection the instant it clears the oppos-
deceleration delay correction factor, which is a function of the ing through traffic or the pedestrians to which it must yield and
typical number of vehicles in queue during each cycle and the begins accelerating back to free-flow speed. Note that the vehicle-
normal free-flow speed when vehicles are unimpeded by the signal. in-queue count often includes some vehicles that have regained
Figure III.9-1 is an intersection control delay worksheet that cruise speed but have not yet exited the intersection.
may be used for recording observations and computation of aver- 3. Enter the vehicle-in-queue counts in the appropriate box on
age time-in-queue delay. Before beginning the detailed survey, the the worksheet. Cycles in the survey period are listed in the second
observers should estimate the average free-flow speed during the column on the worksheet, next to the column for clock time, which
study period—in other words, the speed at which vehicles would is recorded every five cycles, and interval count identifiers are
pass through the intersection if the signal were green for a long given as column headings. For ease in conducting the study, the
time. Accuracy to within 3 to 5 mph is sufficient. This estimate survey period is most conveniently defined as an integer number
9-118 urban streets
Figure III.9-1. Field intersection control delay worksheet.

of cycles, though a precisely defined time length for the survey Table III.9-1. Acceleration-Deceleration Delay
period (e.g., 15 min) could be used. The key point is that the end Correction Factor
of the survey period must be clearly defined in advance, since the correction factor (sec) by avg no. of
last arriving vehicle or vehicles in the period that stop must be vehicles stopping per lane in each cycle
free-flow
identified and counted until they have exited the intersection (see speed (mph) ≤7 8–19 20–301
Step 4). This method is necessary for consistency with the delay
≤37 +5 +2 −1
definition included in the procedure in this chapter. 38–44 +7 +4 +2
4. At the end of the survey period, continue taking vehicle-in- ≥45 +9 +7 +5
queue counts for all vehicles that arrived during the survey period 1
Vehicle-in-queue counts in excess of about 25 vehicles per lane are typically
until all of them have exited the intersection. This requires noting unreliable.
the last stopping vehicle that arrived during the survey period in
each lane of the lane group and continuing the vehicle-in-queue 0.9 = empirical adjustment factor.
counts until the last stopping vehicle or vehicles, plus all vehicles
The 0.9 adjustment factor accounts for the errors that may occur
in front of the last stopping vehicles, have exited the intersection.
when this type of sampling technique is used to derive actual
Stopping vehicles that arrive after the end of the survey period are
delay values, which normally results in an overestimate of delay.
not included in the final vehicle-in-queue counts.
Research has shown the correction required to be fairly consistent
over a variety of conditions.
3. The fraction of vehicles stopping and the average number of
Observer 2
vehicles stopping per lane in each signal cycle, as indicated on
the worksheet, are computed.
During the entire survey period, maintain separate volume 4. Table III.9-1 is used to look up a correction factor appropriate
counts of (a) total vehicles arriving during the survey period and to the lane group cruise speed and the average number of vehicles
(b) total vehicles arriving during the survey period that stop one stopping per lane in each cycle. This factor adds an adjustment
or more times; that is, a vehicle stopping multiple times is counted for deceleration and acceleration delay, which cannot be measured
only once as a stopping vehicle. Enter these volumes in the appro- directly with manual techniques.
priate cells on the worksheet. 5. The correction factor is multiplied by the fraction of vehicles
stopping; then this product is added to the time-in-queue value of
Step 2 to derive the final estimate of control delay per vehicle.
DATA REDUCTION
Figure III.9-2 presents a sample application for a study site over
1. Each column of the vehicle-in-queue counts is summed; then
a 15-min period, operating with a 115-sec cycle over almost 8
the column totals for the entire survey period are summed.
cycles. Annotations have been added to clarify the procedure. The
2. A vehicle recorded as part of a vehicle-in-queue count is in
15-sec count interval is not an integral divisor of the cycle length,
queue, on average, for the time interval between counts. The aver-
to eliminate potential survey bias caused by queue buildup in a
age time in queue per vehicle arriving in the survey period is
regular, cyclic pattern.
estimated as
Figure III.9-3 shows how the field survey would have been
Time in queue per vehicle = (I × ov iq /Vtot) × 0.9 finished if a queue still remained at the end of the 15 minute
survey period. Only the vehicles that arrived during the 15-min
period would be counted, even though other vehicles continued to
where
arrive after the 15-min period.
I = interval between vehicle-in-queue counts, sec; If the study site has an actuated signal with varying cycle and
S viq = sum of vehicle-in-queue counts; phase lengths, the count interval may be chosen as the most conve-
Vtot = total number of vehicles arriving during the survey pe- nient value for conducting the field survey on the basis of volume
riod; and and vantage point considerations.

9-120 urban streets
Figure III.9-2. Sample application of intersection control delay worksheet.

Figure III.9-3. Sample application with residual queue at end.
APPENDIX IV
DIRECT MEASUREMENT OF PREVAILING SATURATION FLOW RATES
GENERAL NOTES when their rear axles cross the stop line. The measurement starts
with the passage of the fourth vehicle. Other reference points
The default ideal saturation flow rate used in the methodology may yield different saturation flow rates. In order to maintain
of this chapter is 1,900 pcphgpl. This value must be adjusted for consistency with the method described in this chapter and to allow
prevailing traffic conditions such as lane width, left turns, right for information exchange, maintaining the roadway and vehicle
turns, heavy vehicles, grades, parking, parking blockage, area type, reference points identified here is essential.
bus blockage, and left-turn blockage. These computations are made The ideal saturation flow rate is usually stable over a period of
in the Saturation Flow Rate Module. As an alternative to these time for similar traffic conditions in a given community. Values
computations, the prevailing saturation flow rate may be measured measured in the same lane during repetitive weekday traffic condi-
directly in the field.
tions (e.g., a.m. or p.m. peak patterns) normally exhibit relatively
Saturation flows have been measured and researched in recent
narrow distributions. On the other hand, saturation flow rates for
years by various groups, including the City of Edmonton in con-
different communities or different traffic conditions and composi-
junction with the University of Alberta, the University of Ken-
tions, even at the same location, may vary significantly.
tucky, Texas Transportation Institute, JHK & Associates, and the
Australian Road Research Board. Results of these studies have For practical purposes, prevailing saturation flow rates are usu-
demonstrated that saturation flow rate can be used as the starting ally expressed in vehicles per hour of green per lane. As a result,
design and analysis variable. Measured saturation flow rates are their values also depend on traffic flow composition. The default
always preferable to estimated values. value used in the methodology of this chapter is expressed in
Saturation flow rate is the maximum discharge rate during the pecphgpl (i.e., passenger cars only).
green time. It is usually achieved after the fourth to seventh vehicle Preferably, local prevailing saturation flow rates should be ob-
has entered the intersection from a standing queue. served directly. Alternatively the computation module can be used,
The ideal saturation flow rate is defined as the discharge rate with the measured regional ideal saturation flow rates as the start-
from a standing queue in a 12-ft (3.6-m) wide lane that carries ing values. The default value should be used only as an approxi-
only through passenger cars and is otherwise unaffected by condi- mate substitute. Severe weather conditions, unusual traffic mixes,
tions such as grade, parking, and so forth. Vehicles are recorded or other special local conditions can yield saturation flow rates
9-122 urban streets
that differ markedly from those estimated using the computation a. Start stop watch at beginning of green and notify the
procedures. recorder.
The procedure for measuring prevailing saturation flow rates b. Count aloud each vehicle in the queue as its rear axle
is summarized below. A sample field worksheet for recording crosses the stop line.
observations is included as Figure IV.9-1. c. Call out the time of the fourth, tenth, and last vehicle in
the queue.
d. If queued vehicles are still entering the intersection at
MEASUREMENT TECHNIQUE the end of the green, call out ‘‘saturation through the end of
green—last vehicle was number XX.’’
The following example describes a single-lane saturation flow Note any unusual events that may have influenced the saturation
survey. A two-person field crew is recommended. One person with flow rate, such as buses, stalled vehicles, and unloading trucks.
a tape recorder, pushbutton event recorder, or a notebook computer The period of saturation flow rate begins when the rear axle of
with appropriate software will suffice. The field notes and tasks the fourth vehicle in the queue crosses the stop line or reference
identified in the following section must be adjusted accordingly. point and ends when the rear axle of the last queued vehicle at
the beginning of the green time crosses the stop line.
1. General tasks
Measurements are taken cycle by cycle. To reduce the data for
a. Measure and record the area type and width and grade
each cycle, the time recorded for the fourth vehicle is subtracted
of the lane being studied.
from the time recorded for the last vehicle in the queue. This value
b. Fill out the survey identification data shown in Figure
is total headway for (n − 4) vehicles, where n is the number of
IV.9-1 completely. the last vehicle surveyed (this may not be the last vehicle in the
c. Select an observation point where the stop line for the queue). The total headway is divided by (n − 4) to obtain the
surveyed lane and the corresponding signal heads are clearly average headway per vehicle under saturation flow. The saturation
visible. flow rate is 3,600 divided by this value.
d. The reference point is normally the stop line. Vehicles For example, if the time for the fourth vehicle was observed as
should consistently stop behind this line. When a vehicle crosses 10.2 sec and the time for the 14th and last vehicle surveyed was
it unimpeded, it has entered the intersection conflict space for the 36.5 sec, the average saturation headway per vehicle would be
purpose of saturation flow measurement. Left- or right-turning
vehicles yielding to opposing through traffic or yielding to pedes- (36.5 − 10.2)/(14 − 4) = 26.3/10 = 2.63 sec/veh
trians are not recorded until they proceed through the opposing and the prevailing saturation flow rate in that cycle would be
traffic.
2. Recorder tasks 3,600/2.63 = 1,369 vphgpl
a. Note the last vehicle in the stopped queue when the In order to obtain a statistically significant value, a minimum
signal turns green. of 15 signal cycles with more than 8 vehicles in the initial queue
b. Describe the last vehicle to the timer. is usually needed. An average of the saturation flow rate values
c. Note on the worksheet which vehicles are heavy vehicles in individual cycles represents then the prevailing local saturation
and which vehicles turn left or right. flow rate for the surveyed lane. The percentage of heavy vehicles
d. Record the time called out by the timer. and turning vehicles in the sample used in the computations should
3. Timer tasks be determined and noted for reference.

Figure IV.9-1. Field Saturation Flow Rate Study Worksheet.

9-124 urban streets
APPENDIX V
WORKSHEETS FOR USE IN ANALYSIS
WORKSHEET PAGE
Input Module Worksheet ......................................................................................................................................................................... 9-125
Volume Adjustment Module Worksheet................................................................................................................................................. 9-126
Saturation Flow Rate Module Worksheet ............................................................................................................................................... 9-127
Supplemental Worksheet for Permitted Left Turns: Multilane Approach............................................................................................. 9-128
Supplemental Worksheet for Permitted Left Turns: Single-Lane Approach......................................................................................... 9-129
Capacity Analysis Module Worksheet .................................................................................................................................................... 9-130
LOS Module Worksheet .......................................................................................................................................................................... 9-131
Supplemental Uniform Delay Worksheet for Left Turns from Exclusive Lanes with Primary and Secondary Phases...................... 9-132
Planning Method Input Worksheet.......................................................................................................................................................... 9-133
Planning Method Lane Volume Worksheet ............................................................................................................................................ 9-134
Planning Method Signal Operations Worksheet ..................................................................................................................................... 9-135
Intersection Control Delay Worksheet .................................................................................................................................................... 9-136
Field Saturation Flow Rate Study Worksheet......................................................................................................................................... 9-137
Delay/LOS Worksheet with Initial Queue .............................................................................................................................................. 9-138


9-126 urban streets


9-128 urban streets


9-130 urban streets


9-132 urban streets


9-134 urban streets


9-136 urban streets


9-138 urban streets

APPENDIX VI
EXTENSION OF SIGNAL DELAY MODELS TO INCORPORATE EFFECT OF INITIAL QUEUE
INTRODUCTION
The delay model represented by Equations 9-22 through 9-25

in this chapter is based on the assumption that there are no initial
queues at the start of an analysis period of length T. In some cases,
the analysis period starts with a residual demand that was unmet
from a previous time period. This unmet demand will be designated
Qb in vehicles. Qb is observed at the start of the red period and
excludes any vehicles in the queue present because of random,
cycle-by-cycle fluctuations in demand (overflow queue due to cy-
cle failures). When Qb ≠ 0, vehicles arriving in period T will
experience an additional delay due to the presence of an initial
queue. The magnitude of this supplemental delay depends on sev-
Figure VI.9-2. Case IV: supplemental delay with initial oversatur-
eral factors, namely, the size of the initial queue, the length of the ation demand decreasing in T. [Supplemental delay per vehicle
analysis period, and the prevailing (adjusted) volume-to-capacity (d3) in seconds = 3,600Qb/c − 1,800T [1 − Min (1,X)].]
ratio in T. Hereafter, the supplemental delay term is referred to
as d3.
Five scenarios emerge when control delay is estimated. These
are labeled Cases I through V. Cases I and II occur when there is
no initial queue and the period is either undersaturated (Case I) or
oversaturated (Case II). In both these cases, d3 = 0, and the delay
model in Equation 9-22 applies. Cases III through V are shown
in Figures VI.9-1 through VI.9-3. Case III occurs when unmet
demand Qb can be fully served in time period T. For this to happen,
the sum of Qb and the total demand in T, qT, must be less than
the available capacity, cT. Case IV occurs when there is still unmet
demand at the end of T but the size of the unmet demand is
decreasing. For this to happen, the demand in T, qT, should be
less than the capacity, cT. Finally, Case V occurs when the demand
in T, qT, exceeds the capacity, cT. Here, the unmet demand will
increase at the end of period T.
The total supplemental delay due to an initial queue that is
incurred in the average cycle is depicted as the shaded area in
Figure VI.9-3. Case V: supplemental delay with initial oversatura-
Figures VI.9-1 through VI.9-3, which is labeled D. It represents
tion demand increasing in T. [Supplemental delay per vehicle (d3)
the entire delay experienced by all vehicles arriving in T, including in seconds = 3,600Qb /c.]
delay that is experienced in subsequent time periods (Figures VI.9-
2 and VI.9-3). Excluded from this delay are two components: the
delay incurred for vehicles in the initial queue (labeled Di) and
the oversaturation delay corresponding to a zero initial queue (la-
beled Dso in Figure VI.9-3). This last term is already accounted
for in the d2 term component of the delay model in Equation 9-25.
ESTIMATION OF d3
A generalized form of d3, defined as the average supplemental

control delay per vehicle (in seconds) when an initial queue of size
Qb is present at the start of the analysis period T, was developed in
a recent FHWA-sponsored project. It is an additive term in the
delay model given in Equation 9-22 and is expressed as follows:
Figure VI.9-1. Case III: supplemental delay with initial oversatur-
ation demand clearing in T. [Supplemental delay per vehicle (d3) 1,800Qb(1 + u)t
d3 = (VI.9-1)
in seconds = 1,800Qbt/cT.]. cT
9-140 urban streets
Table VI.9-1. Selection of Delay Model Variables by Case

Case X Qb d1 d2 t u d3 Tc
I ≤ 1.0 0 Eq. 9-23 Eq. 9-25 0 0 0 T
II > 1.0 0 Eq. 9-23 Eq. 9-25 0 0 0 TX
III ≤ 1.0 >0 Eq. VI.9-5 Eq. 9-25 Eq. VI.9-2 0 Eq. VI.9-1 T
IV ≤ 1.0 >0 Eq. VI.9-5 Eq. 9-25 T Eq. VI.9-3 Eq. VI.9-1 Eq. VI.9-4
V > 1.0 >0 Eq. VI.9-5 Eq. 9-25 T 1 Eq. VI.9-1 Eq. VI.9-4
where and permitted-protected), using X=1.0 in the equation to compute

Qb = initial unmet demand at the start of period T, veh; ds and using the actual X value to compute du. For compound left
c = adjusted lane group capacity, veh/hr; turn protection, the Supplemental Uniform Delay Worksheet in
T = duration of analysis period, hr; Figure 9-21 must be used as a means to approximate the ds and
t = duration of unmet demand in T, hr; and du compounds, again using X=1.0 for ds and the actual X value for
u = delay parameter. du. When using X=1.0 for the ds component in Figure 9-21, the
left turn volume used (v) must also be adjusted by the actual X
The parameters t and u are determined according to the prevailing
value (use v′=v/X) in order to meet the basic assumptions of the
case (III, IV, or V). The following two equations may be used to
Supplemental Uniform Delay Worksheet.
estimate their values:
Note that the only place where PF is used in an initial queue
5 6
Qb
t = 0 if Qb = 0, else t = Min T, (VI.9-2) analysis of Appendix VI is for the undersaturated du portion of a
c[1 − Min(1,X)]
Case III condition. This is because the existence of the initial
cT
u = 0 if t < T, else u = 1 − [1 − Min(1,X)] (VI.9-3) queue defeats the value of the progression under all other condi-
Qb tions. Analysts are advised to be wary of a similar concern in the
where X is lane group degree of saturation (v/c). use of PF in a Case II analysis (oversaturated) using Equation
In addition to the supplemental delay term, the analyst may be 9-22 because all but the first cycle will be blocked by initial queues
interested in computing the time that the last vehicle arriving in due to the oversaturated condition.
T clears the intersection (measured from the start of the time period
T), as affected by the presence of an initial queue of length Qb.
NUMERICAL EXAMPLE OF DELAYS WITH INITIAL QUEUE
This time will be referred to as the supplemental clearing time Tc.
In Cases I–III, it is evident that all vehicles will clear at the end To illustrate the application of the delay model extension, an
of period T (in addition to the normal delays d1 + d2). For Cases analysis of the EB lane group in Sample Calculation 1 in Section
IV and V, the last vehicle arriving in T will clear the intersection IV is carried out with and without an initial queue. The following
at time Tc > T (again in addition to d1 + d2). Therefore, a general input values are taken directly from Figure 9-31, the Capacity
formula for the supplemental clearing time in the case of an initial Analysis Module Worksheet:
queue, measured from the start of the analysis period T, is
Lane group capacity (c) = 787 veh/hr,
Qb Lane group v/c ratio (X) = 1.017,
Tc = Max (T, + TX) (VI.9-4)
c Analysis period length (T) = 0.25 hr, and
To summarize the procedure for estimating control delays, Table Initial queue = 20 vehicles (across the two-lane lane group).
VI.9-1 gives a comparison of the model parameters for Cases I–
V. Note that in order to decide whether Case III (t < T) or IV (t = Scenario I: No Initial Queue
T) applies, the value of t must first be computed from Equation
VI.9-2. In Scenario I, d3 = 0 as given in Table VI.9-1, Case II. The
For Cases III, IV, and V, the uniform control delay component average control delay per vehicle is d1PF + d2 + d3 = 22.0 ∗ .923
(d1) must be evaluated using X=1.0 for the period when an oversat- + 36.3 + 0.0 = 56.6 sec, as given in Figure 9-32. The corresponding
uration queue exists (t) and using the actual X value for the remain- LOS for a control delay of 56.6 sec is E. Finally, the supplemental
der of the analysis period (T-t). Therefore, in these cases, a time- clearing time Tc = 15.3 min for Case II.
weighted value of d1 is to be used, as follows:
d1 = ds∗t/ T + du∗PF∗(T-t)/T (VI.9-5) Scenario II: Initial Queue of 20 Vehicles
Since X > 1.0 and Qb = 20, Case V in Table VI.9-1 applies. Here,
where t = 0.25 and u = 1. Substituting in Equation VI.9-1 gives the
following:
ds = is the saturated delay (d1 evaluated for X=1.0), and
du = is the undersaturated delay (d1 evaluated for the actual X 1800 ∗ 20 ∗ (1 + 1) ∗ 0.25
d3 = = 91.5 sec/veh
value). 787 ∗ 0.25
In equation VI.9-5 for Cases IV and V, the du term drops out Therefore, the average control delay is computed as
because t=T and need not be calculated. Equation 9-23 can be
d = 22.0 + 36.3 + 91.5 = 149.8 sec/veh (LOS F)
used to evaluate the ds and du components in all cases except for
left turns with compound left turn protection (protected-permitted which is more than twice the delay calculated assuming a zero
time in a 100-sec cycle. Arrivals are considered to be random

(Arrival Type 3). Calculate the delay and level of service for
vehicles arriving in each 15-min time period and for the overall
analysis period of 1 hr.
Period 1
Period 1 is undersaturated, with a degree of saturation X = 800/

1,000 = 0.80. Therefore, from Equation VI.9-6, there is no unmet
demand at the start of Period 2, Qb,2 = 0, assuming no initial queue
at the start of Period 1, Qb,1 = 0. The average control delay for
vehicles arriving in Period 1 will be labeled dc,1 and is estimated
as follows:
Figure VI.9-4. Demand profile for multiple-period analysis (15-
min periods). 0.50 ∗ 100 ∗ (1 − 40/100)2
dc,1 = ∗ 1.0 + 900 ∗ 0.25 ∗
40
1− Min(1, 0.80)
initial queue. Note that PF is not applied for Case V. Thus the 100
8 ∗ 0.50 ∗ 0.80
!
impact of an initial queue can be substantial and must be accounted
for in delay and LOS estimation. 3(0.80 − 1) + (0.80 − 1)2 +
1,000 ∗ 0.25 4
= 33.2 sec
Finally, the supplemental clearing time, Tc, is estimated from
Equation VI.9-4 as
1 2
20
Tc = Max 0.25, + 0.25 ∗ 1.017 = 0.28 hr Period 2
787
or 16.8 min from the start of the peak period. Therefore, the last
vehicle entering in the peak 15-min period will experience an Period 2 is oversaturated, with a degree of saturation X = 1,200/
additional 1.8 min delay because of the presence of the initial 1,000 = 1.20. There is no unmet demand at the start of the period,
queue of 20 vehicles. so again the two-component delay formula can be used as follows:
0.50 ∗ 100 ∗ (1 − 40/100)2

dc,2 = ∗ 1.0 + 900 ∗ 0.25 ∗
EXTENSION TO MULTIPLE TIME PERIODS 40
1− Min(1, 120)
100
The procedure just described can be extended to analyze multi- 8 ∗ 0.50 ∗ 1.20
ple time periods, each of duration T and each having a fixed
demand during T. The analysis is performed sequentially, carrying
3 !
(1.20 − 1) + (1.20 − 1)2 +
1,000 ∗ 0.25 4
= 129.7 sec
over the final unmet demand Qb (if any) from one time period to
the beginning of the next. In general, for time period i, the final
unmet demand Qb,i+1 at the start of the next time period T can be
estimated as follows: Period 3
Qb,i+1 = Max[0, Qb,i + cT(Xi − 1)] for i = 1,2, . . ., n (VI.9-6)

Period 3 is fully saturated, with a degree of saturation X =
where Qb,i and Xi are the unmet demand and the degree of saturation 1,000/1,000=1.00. The unmet demand from the previous period is
for the time period i, respectively. calculated from Equation VI.9-6 as follows:
Typically, a multiple-time-period analysis would start with an
undersaturated time period, particularly one for which Qb,1 = 0. Qb,3 = Max[0, 0 + 1,000 ∗ 0.25 ∗ (1.2 − 1)] = 50 veh
Once the unmet demand is calculated, delays are estimated ac-
cording to the method described in the previous section. An impor- Here, the supplemental delay term (d3) must be added. First the
tant feature of multiple-period analysis is that the actual counts values of t and u are determined from Equations VI.9-2 and VI.9-
taken during each time period should be used in the procedure; 3, respectively:
that is, the PHF is unity. Counts are then converted into hourly
3 4
flow rates by dividing each count by T (in hours). The procedure 50
t = Min 0.25, = 0.25
is best described using a numerical example. 1,000(1 − 1)
1,000 ∗ 0.25[1 − Min(1.0, 1.0)]
u=1− = 1.0
NUMERICAL EXAMPLE FOR MULTIPLE-PERIOD ANALYSIS 50
In this example, consider a signalized lane group with no initial Substituting in Equation VI.9-1 gives
queue that has a fixed capacity of 1,000 vph. The demand profile
based on 15-min counts (factored to hourly rates) is shown in 1,800 ∗ 50 ∗ (1 + 1) ∗ 0.25
d3 = = 180 sec
Figure VI.9-4. The lane group receives 40 sec of effective green 1,000 ∗ 0.25
9-142 urban streets
Thus, the average control delay in Period 3 is be carried out over subsequent time intervals to ensure that the
most severe LOS period has been captured.
0.50 ∗ 100 ∗ (1 − 40/100) 0.25
2
dc,3 = ∗ + 900 ∗ 0.25 ∗
40 0.25
1− 1.0 PROCEDURES FOR MAKING CALCULATIONS
100
8 ∗ 0.50 ∗ 1.0
3 !
(1.0 − 1) + (1.0 − 1)2 +
1,000 ∗ 0.25 4
+ 180 = 238.5 sec Figure VI.9-6 provides a worksheet that can be used to make the
calculations of delay with initial queue. The following steps should
be used with this worksheet.
Period 4
Step 1: Enter Time Period and Duration
Period 4 is undersaturated, with a degree of saturation X = 600/
1,000 = 0.60. The unmet demand from the previous period is Each time period will need a separate worksheet to calculate the
calculated from Equation VI.9-6 as follows: delay when any of the movements at the intersection start with an
initial queue during that time period. Each time period should be
Qb,4 = Max[0, 50 + 1,000 ∗ 0.25 ∗ (1.0 − 1)] = 50 veh numbered and the period number and time entered in the upper
In essence, since the previous period was at capacity, the residual left corner of the worksheet. The duration of the time period, in
demand at the end of the period is equivalent to that at the start hours, should also be entered. Note that in a multiperiod analysis
of the period. Again, the computation of d3 requires the values of the length of all time periods should be the same.
t and u, which are calculated as follows:
Step 2: Identify Lane Groups for Analysis
3 4
50
t = Min 0.25, = 0.125
1,000(1 − 0.60)
In Column 1, label each of the lane groups to be analyzed according
Since t < 0.25, u = 0 from Equation VI.9-3. to those used in the LOS Module Worksheet (Figure 9-20). Lane
Substituting in Equation VI.9-1 gives groups that do not have an initial queue may appear on this work-
sheet so that their delay values can be averaged with those of
1,800 ∗ 50 ∗ (1 + 0.0) ∗ 0.125 oversaturated lane groups. In this case, their d1 and d2 values will
d3 = = 45 sec
1,000 ∗ 0.25 be unchanged and d3 will be zero on the worksheet.
Since t < T, the uniform delay component is calculated using
Equation VI.9-5: Step 3: Enter First- and Second-Term Delay Information
0.50 ∗ 100 ∗ (1 − 40/100) 0.125
2
d1 = ∗ In Columns 2, 3, 6, and 8, enter the v/c ratio, lane group capacity,
1 − 40/100 ∗ 1.0 0.25
unadjusted uniform delay, and incremental delay values from Col-
0.50 ∗ 100 ∗ (1 − 40/100)2 0.25 − 0.125
+ ∗ 1.0 ∗ = 26.8 sec/veh umns 2, 6, 4x5, and 8 of the LOS Module Worksheet, respectively.
1 − 40/100 ∗ Min(1,0.60) 0.25 To get unadjusted uniform delay, enter the product of the uniform
Thus, the average control delay in Period 4 is delay (LOS Module Worksheet, Column 4) and the progression
adjustment factor (column 5).
dc,4 = 26.8 + 900 ∗ 0.25 ∗
8 ∗ 0.50 ∗ 0.60
3(0.60 − 1.0) + !(0.60 − 1.0) 2
+
1,000 ∗ 0.25 4 Step 4: Enter Initial Unmet Demand
+ 45 = 74.5 sec/veh
In Column 4, enter the initial unmet demand for each lane group.
The contribution of each of the delay terms in each period is This entry may be the initial queue observed at the start of the red
illustrated in Figure VI.9-5, which displays the general delay com- period (excluding any vehicles in queue because of random, cycle-
ponent trends. The impact of the initial queue delay term is evident, by-cycle fluctuations in demand), or it may be the final unmet
particularly for Periods 3 and 4. demand for the lane group estimated in Column 10 from the previ-
Finally, the average overall control delay to all vehicles arriving ous analysis period.
in the hour is calculated as a volume-weighted delay of the individ-
ual period delays as follows:
Step 5: Calculate Duration of Unmet Demand
dc,t = (800 ∗ 33.2 + 1,200 ∗ 129.7 + 1,000 ∗ 238.5
+ 600 ∗ 74.5)/(800 + 1,200 + 1,000 + 600) = 129.3 sec/veh In Column 5, calculate the duration of unmet demand, t, using
Equation VI.9-2. If there is no initial queue (Qb = 0), then t = 0,
It is interesting to note that although the average overall signal and the value of t is limited to be no larger than the length of the
delay in the entire period is virtually identical to the delay for time period, T.
vehicles arriving in peak Period 2, it is much smaller than the
worse delay (and level of service) that is experienced in Period 3,
which immediately follows the peak. Thus, a single-period analysis Step 6: Calculate Adjusted Uniform Delay
may not be sufficient to determine the worst level of service in
an oversaturated time period. When residual queues do occur at In Column 7, calculate the adjusted uniform delay term, d1, using
the end of a peak period, it is recommended that a delay analysis Equation VI.9-5. When t = 0, the result is the same as the unad-
Figure VI.9-5. Delay model components for multiple-period analysis.
justed value of d1. Progression effects are included, as appropriate, from the existence of the initial queue. Note that this value does
in this adjusted uniform delay result. The values of g and C for not include any of the oversaturation delay accounted for in d2.
the lane group from the LOS Module Worksheet must be used to
make this calculation. Note that the unadjusted value of d2 will be
Step 10: Find Delay and Level of Service for Each Lane Group
used from the LOS Module Worksheet in the final delay calcula-
tion and that this delay value includes the oversaturation delay
Delay and level of service are found by adding the three delay
when v/c > 1.
terms in Columns 7, 8, and 11 for each lane group. Note that the
d1 value in Column 7 includes any appropriate effects of PF on
the d1 term. The result is entered in Column 12. The level of service
Step 7: Calculate Initial Queue Parameter
corresponding to this delay, taken from Table 9-1, is entered in
Column 13.
In Column 9, calculate the initial queue parameter, u, using Equa-
tion VI.9-3. When t < T, u = 0 (Cases I-III); otherwise Equation
Step 11: Find Delay and Level of Service for Each Approach
VI.9-3 is used (Case IV) or u = 1 (Case V).
The average delay per vehicle is found for each approach by adding
the product of the lane group flow rate and the delay for each lane
Step 8: Calculate Final Unmet Demand
group on the approach and dividing the sum by the total approach
flow rate. The weighted-average delay is entered in Column 14
In Column 10, calculate the final unmet demand using Equation for each approach. Level of service is determined from Table
VI.9-6. This is the estimate of the number of vehicles in queue at 9-1 and entered in Column 15.
the end of the analysis period. If its value is nonzero, this indicates
that the subsequent analysis period should be analyzed to deter-
Step 12: Find Delay and Level of Service for Intersection
mine the average delay per vehicle that results because of this
initial queue for that time period.
The average delay per vehicle for the intersection as a whole is
found by adding the product of the approach flow rate and the
approach delay for all approaches and dividing the sum by the
Step 9: Calculate Supplemental Delay
total intersection flow rate. This weighted-average delay is entered
at the bottom of the worksheet. The overall intersection level of
In Column 11, calculate the supplemental (initial queue) delay, d3, service is found from Table 9-1 and entered at the bottom of the
using Equation VI.9-1. This is the additional delay that results worksheet.

chapter 10
UNSIGNALIZED INTERSECTIONS
PREFACE
Procedures for the capacity and level-of-service (LOS) analysis of three types of unsignalized intersections are described in this
chapter: two-way stop-controlled (TWSC) intersections, all-way stop-controlled (AWSC) intersections, and roundabouts.
These procedures are not intended for use in the analysis of totally uncontrolled intersections or for yield-controlled intersections.
Research sponsored by the National Cooperative Highway Research Program (NCHRP) provided the basis for the analysis methodolo-
gies for TWSC and AWSC intersections. The methodologies are based on data from the first comprehensive data base established for
traffic flow characteristics at unsignalized intersections in the United States, including information from 79 TWSC intersections and 41
AWSC intersections. A separate study of four intersections sponsored by the Federal Highway Administration provided the data to
calibrate the procedures for roundabouts.
CONTENTS
PART A. TWO-WAY STOP-CONTROLLED INTERSECTIONS ...................................................................................................... 10-3

i. introduction .......................................................................................................................................................................... 10-3
Variables Used in Analysis of TWSC Intersections ............................................................................................................ 10-3
Overview of Procedures ........................................................................................................................................................ 10-4
ii. methodology.......................................................................................................................................................................... 10-5
Conceptual Approach ............................................................................................................................................................ 10-5
Capacity Formula .................................................................................................................................................................. 10-6
Structure of the Methodology ............................................................................................................................................... 10-6
Input Data Requirements ................................................................................................................................................. 10-7
Conflicting Traffic............................................................................................................................................................ 10-7
Critical Gap and Follow-Up Time .................................................................................................................................. 10-10
Potential Capacity for Movement.................................................................................................................................... 10-11
Impedance Effects ............................................................................................................................................................ 10-11
Shared-Lane Capacity ...................................................................................................................................................... 10-15
Upstream Signals.............................................................................................................................................................. 10-16
Two-Stage Gap Acceptance............................................................................................................................................. 10-20
Flared Minor-Street Approaches...................................................................................................................................... 10-21
Queue Lengths.................................................................................................................................................................. 10-21
Control Delay ................................................................................................................................................................... 10-22
Delay to Major-Street Through Vehicles ........................................................................................................................ 10-23
Other Relevant Delay Estimates...................................................................................................................................... 10-23
Interpretation of Results ........................................................................................................................................................ 10-25
Shared Lanes on Minor Approach .................................................................................................................................. 10-25
Consideration of Queue Lengths ..................................................................................................................................... 10-25
Determination of Intersection Control Type ........................................................................................................................ 10-27
Field Data Requirements....................................................................................................................................................... 10-29
Sequence of Computations for Capacity .............................................................................................................................. 10-29
Capacity Adjustments ...................................................................................................................................................... 10-29
Geometric Features and Movement Definitions ............................................................................................................. 10-29
Volume Adjustment ......................................................................................................................................................... 10-29
10-2 urban streets
Site Characteristics........................................................................................................................................................... 10-29

Critical Gap and Follow-Up Time .................................................................................................................................. 10-30
Effect of Upstream Signals.............................................................................................................................................. 10-30
Computations of Impedance and Movement Capacities ................................................................................................ 10-30
Two-Stage Gap Acceptance............................................................................................................................................. 10-31
Shared-Lane Capacity ...................................................................................................................................................... 10-31
Effect of Flared Minor-Street Approaches...................................................................................................................... 10-31
Average Control Delay, Queue Length, and Level of Service ...................................................................................... 10-31
Planning and Design Applications........................................................................................................................................ 10-32
Sample Calculation A1.......................................................................................................................................................... 10-46
v. references .............................................................................................................................................................................. 10-58
PART B. ALL-WAY STOP-CONTROLLED INTERSECTIONS........................................................................................................ 10-59

i. introduction .......................................................................................................................................................................... 10-59
Variables Used in Analysis of AWSC Intersections............................................................................................................ 10-59
ii. methodology.......................................................................................................................................................................... 10-61
Capacity Concept................................................................................................................................................................... 10-62
Capacity Model...................................................................................................................................................................... 10-62
Intersection of Two One-Way Streets............................................................................................................................. 10-62
Intersection of Two Two-Way Streets ............................................................................................................................ 10-63
Generalized Model for Single-Lane Sites ....................................................................................................................... 10-63
Generalized Model for Multilane Sites ........................................................................................................................... 10-64
Control Delay ........................................................................................................................................................................ 10-64
Field Data Requirements....................................................................................................................................................... 10-67
Geometric Features and Movement Definitions................................................................................................................... 10-67
Volume Adjustment and Lane Assignment.......................................................................................................................... 10-67
Saturation Headway Adjustment Factor ............................................................................................................................... 10-67
Departure Headway and Service Time................................................................................................................................. 10-68
Worksheet B4................................................................................................................................................................... 10-68
Worksheet B4 Supplemental............................................................................................................................................ 10-69
Capacity ................................................................................................................................................................................. 10-69
Delay and Level of Service .................................................................................................................................................. 10-69
Planning and Design Applications........................................................................................................................................ 10-75
Sample Calculation B1.......................................................................................................................................................... 10-75
v. references .............................................................................................................................................................................. 10-81
PART C. ROUNDABOUTS ................................................................................................................................................................... 10-81

i. introduction .......................................................................................................................................................................... 10-81
Variables Used in Analysis of Roundabouts........................................................................................................................ 10-81
Characteristics of Roundabouts............................................................................................................................................. 10-81
ii. methodology.......................................................................................................................................................................... 10-83
unsignalized intersections 10-3

Capacity ................................................................................................................................................................................. 10-84
v. references .............................................................................................................................................................................. 10-89
PART A. TWO-WAY STOP-CONTROLLED INTERSECTIONS
I. INTRODUCTION
In Part A, procedures for analysis of two-way stop-controlled t3,LT = critical gap adjustment factor for intersection geometry
(TWSC) intersections based on research sponsored by NCHRP (1) (T-intersection);
are discussed. tf,base = base follow-up time (sec); and
A variety of terminology is introduced applying to the unique tf,HV = follow-up time adjustment factor for heavy vehicles.
characteristics of TWSC intersection capacity and LOS analyses.
For ease of reference, these terms are defined in the following
section and are more fully described as they are used in the suc-
Impedance Parameters
ceeding sections.
fp = pedestrian blockage factor;
fx = capacity adjustment factor for Movement x that accounts
VARIABLES USED IN ANALYSIS OF TWSC INTERSECTIONS for impeding effects of higher-ranked movements;
p0,x = probability that conflicting Movement x will operate in
Capacity and Delay Parameters queue-free state;
p′ = adjustment to major-street left turn, minor-street through
cm,x = movement capacity of minor Movement x (veh/hr); movement impedance factor;
cp,x = potential capacity of minor Movement x (veh/hr); p″ = product of probabilities of queue-free states of Rank 1 and
cSH = capacity of shared lane (veh/hr); Rank 2 vehicles; and
cT,x = total capacity for Movement x considering two-stage gap p*0,x = factor indicating probability that there will be no queue in
acceptance process; shared lane for major-street Movements 1 and 4, where x
cI = capacity for Stage I of two-stage gap acceptance process; is the particular movement being considered; this factor is
cII = capacity for Stage II of two-stage gap acceptance used in lieu of p0,x where shared lanes for left-turn and
process; through movements exist on major street; and
d = average control delay (sec/veh); sp = pedestrian walking speed (ft/sec).
dSH = average control delay for shared-lane case;
dsep = average control delay for separate-lane case;
dRank 1 = average control delay to Rank 1 vehicles;
X = degree of saturation, or volume-to-capacity ratio; and Miscellaneous Intersection Characteristics
T = length of study period (hr).
PHV = proportion of heavy vehicles;
G = percent grade divided by 100; and
w = lane width.
Critical Gap and Follow-Up Time Parameters
tc = critical gap, or minimum time interval in major-street

traffic stream that allows intersection entry to one minor-
Movement Definitions
stream vehicle (sec);
tf = follow-up time, or headway between departure of one i = vehicle movements of Rank 1;
vehicle from minor street and departure of next vehicle, j = vehicle movements of Rank 2;
using same major-stream gap during periods of continuous k = vehicle movements of Rank 3; and
queueing on minor-street traffic stream (sec); l = vehicle movements of Rank 4.
tc,base = base critical gap;
tc,HV = critical gap adjustment factor for heavy vehicles;
tc,G = critical gap adjustment factor for grade;
Platoon Dispersion Model Parameters
tc,T = critical gap adjustment factor for each part of two-stage
gap acceptance process; g = effective green time of upstream signalized intersection;
10-4 urban streets
gq = time required for queue to clear from upstream signal at 1985 manual (2). The current procedures are based on a study of
beginning of green phase; traffic operations at 79 TWSC intersections in the United States
F = (1 + abta)−1; (1). New critical gaps and follow-up times are presented, and new
a = platoon dispersion factor; weighting factors for conflicting flows are given. Adjustments
b = dispersion factor, or (1 + a)−1; are included in the procedures to take into account the effects of
P = proportion of vehicles at upstream signalized intersection nonrandom flows (platoons from upstream signals), two-stage gap
arriving during green; acceptance, flared minor-street approaches, pedestrians, and delay
s = saturation flow rate for upstream signalized intersection; to major through movements.
C = cycle length for upstream signalized intersection; TWSC intersections, in which stop signs are used to assign the
ta = travel time from signalized intersection to subject TWSC right-of-way, are the most prevalent type of intersection within
intersection; the United States and abroad. At TWSC intersections, the stop-
vc,max = maximum platooned flow rate in conflicting stream; controlled approaches are referred to as the minor-street ap-
vc,u,x = conflicting flow for Movement x during unblocked period; proaches and can be either public streets or private driveways. The
tp,i = duration of blocked period for Movement i; intersection approaches that are not controlled by stop signs are
px = proportion of time blocked for Movement x; referred to as the major-street approaches. A three-leg intersection
cplat,x = capacity of subject Movement x accounting for effect of is considered to be a standard type of TWSC intersection as long
platooning; and as the single minor-street approach (i.e., the stem of the T-configu-
N = number of major-street through lanes per direction. ration) is controlled by a stop sign. Three-leg intersections in which
two of the three approaches are controlled by stop signs represent
a special form of unsignalized intersection control that is not ad-
Two-Stage Gap Acceptance and Flared Right-Turn Approach dressed by the procedures described in this chapter.
Parameters Some TWSC intersections have unusual operating characteris-
tics. For example, there are situations in which one or more left-
m = number of vehicles that can be stored in median of intersec- turning movements are allowed to travel unimpeded through the
tion during two-stage gap acceptance process; intersection and are given the right-of-way over opposing through
n = number of vehicles that can be stored in flared right-turn movements. Although it is recognized that such operating condi-
approach; tions may legitimately and appropriately exist under special cir-
nmax = maximum value of n, number of vehicles that can be stored cumstances, they are relatively rare in practice, and are not
in flared right-turn approach above which it will operate addressed in this chapter.
like a separate-lane condition; TWSC intersections as defined herein assign the right-of-way
a = adjustment factor for two-stage gap acceptance process; among conflicting traffic streams according to the following
QSH = average queue length for shared-lane case for flared right- hierarchy:
turn calculations; and Rank 1. All conflicting movements yield the right-of-way to
Qsep = average queue length for separate-lane case for flared any through or right-turning vehicle on the major-street ap-
right-turn calculations. proaches. These major-street through and right-turning movements
are hereafter referred to as the highest-priority movements at a
TWSC intersection.
Volume Parameters Rank 2. (a) Vehicles turning left from the major street onto the
Vx = volume for Movement x expressed in vehicles per hour minor street yield only to conflicting major-street through and
for vehicular flows and pedestrians per hour for pedes- right-turning vehicles. All other conflicting movements at a TWSC
trian flows; intersection yield to these major-street left-turning movements.
Vc,x = conflicting volume for Movement x, that is, total volume (b) Vehicles turning right from the minor street onto the major
that conflicts with Movement x expressed in vehicles per street yield only to conflicting major-street through movements.
hour or pedestrians per hour; Rank 3. Minor-street through vehicles yield to all conflicting
vL = major left-turn flow rate; major-street through, right, and left-turning movements.
vr = shared-lane right-turn flow rate; Rank 4. Minor-street left-turning vehicles yield to all conflicting
vt = shared-lane through flow rate; major-street through, right, and left-turning vehicles, and also to
vl = shared-lane left-turn flow rate; all conflicting minor-street through and right-turning vehicles.
vSH = flow rate in shared lanes; used to compute capacity for Even though the hierarchy just described suggests that the high-
flared right-turn approach; est-priority movements experience no delay as they travel through
vsep = flow rate in separate lanes on minor-street approach; used a TWSC intersection, experience shows that their right-of-way
to compute capacity for flared right-turn approach; and is sometimes preempted by other conflicting movements. Such
vprog = progressed flow rate from upstream signalized intersection preemptions most often occur during periods of congestion when
to compute effect of upstream signals. vehicles in the conflicting movements are experiencing long delays
and queues. The procedures described in this chapter include calcu-
lations to account for such preemptions.
OVERVIEW OF PROCEDURES At TWSC intersections, drivers on the controlled approaches
are required to use judgment in selecting gaps in the major-street
The procedures described in this chapter for the analysis of flow through which to execute crossing or turning maneuvers. In
capacity and level of service of TWSC intersections represent a the presence of a queue, each driver on the controlled approach
revision of the methodology published in the 1994 update of the must also use some measurable amount of time to move into the
front-of-queue position and get ready to evaluate gaps in the major- the major street, flows may not be random but will likely have
street flow. Thus, the capacity of the controlled legs is based on some platoon structure. A procedure is included in this chapter to
three factors: account for this effect.
Often the minor-street approaches to a TWSC intersection will
1. The distribution of gaps in the major-street traffic stream,
be flared or there will be storage space in the median for vehicles
2. Driver judgment in selecting gaps through which to execute
crossing or merging with the major-street traffic streams. Proce-
their desired maneuvers, and
dures to account for the increase in capacity resulting from these
3. The follow-up time required by each driver in a queue.
conditions are included.
In the basic capacity model presented here it is assumed that gaps Pedestrians may impede the flow of both major- and minor-
in the conflicting stream are randomly distributed. When traffic stream vehicles. Impedance equations are included to adjust for
signals are within 0.25 mi upstream of the subject intersection on this effect.
II. METHODOLOGY
CONCEPTUAL APPROACH that different drivers have different gap acceptance thresholds and
even that the gap acceptance threshold of an individual driver
Capacity analysis at TWSC intersections depends upon a clear changes over time (4). Therefore, it is appropriate for the purposes
description and understanding of the interaction between drivers of this methodology to consider the critical gap and follow-up
on the minor, stop-controlled approach and drivers or vehicles on times as representative of a statistical average of the driver popula-
the major street. Both gap acceptance and empirical models have tion. Under this assumption, the statistically representative driver
been developed as a means to describe this interaction. The proce- is expected to behave in the same way every time under all similar
dures described in this chapter rely upon a gap acceptance model situations.
developed and refined in Germany (3).
Gap acceptance models begin with the recognition that TWSC
intersections give no positive indication or control to the driver Priority of Traffic Streams
on the minor street as to when it is safe to leave the stop line and
enter the major traffic stream. The driver must determine both The third element is the priority of right-of-way given to each
when a gap in the major stream is large enough to permit safe traffic stream. Some streams have absolute priority, whereas others
entry and when it is his or her turn to do so on the basis of the have to give way or yield to higher-order streams. Figure 10-1
relative priority of the competing traffic streams. This decision- illustrates the relative priority of streams at both T- and four-leg
making process has been formalized into what is commonly known intersections.
as gap acceptance theory.
Gap acceptance theory includes three basic elements: the size T Movements of Rank r = 1 (the ordinal rank, denoted by the
and distribution (availability) of gaps in the major traffic stream, subscript i in the remainder of this chapter)
the usefulness of these gaps to the minor-stream drivers, and the — Through traffic on the major street (Movements 2, 5)
relative priority of the various traffic streams at the intersection. — Right-turning traffic from the major street (Movements
3, 6)
— Pedestrians crossing the minor street (Movements 15, 16)
Availability of Gaps T Movements of Rank r = 2 (subordinate to 1 and denoted by
the subscript j in the remainder of this chapter)
The first element to consider in gap acceptance theory is the — Left-turning traffic from the major street (Movements 1, 4)
proportion of gaps of a particular size in the major traffic stream — Right-turning traffic onto the major street (Movements
offered to the driver entering from the minor stream, as well as 9, 12)
the pattern of interarrival times of vehicles. The distribution of — Pedestrians crossing the major street (Movements 13, 14)
gaps between vehicles in the different streams has a major effect T Movements of Rank r = 3 (subordinate to 1 and 2 and denoted
on the performance of the intersection. Errors associated with the by the subscript k in the remainder of this chapter)
assumption of exponential interarrival times have been found to — Through traffic on the minor street (in the case of a four-
be compensated for by offsetting errors caused by the variations leg intersection) (Movements 8, 11)
in actual critical gaps and follow-up times of individual drivers (3). — Left-turning traffic from the minor street (in the case of
a T-intersection) (Movement 7)
T Movements of Rank r = 4 (subordinate to all and denoted by
Usefulness of Gaps
the subscript l in the remainder of this chapter; four-leg intersec-
tion only)
The second element is the extent to which drivers find gaps — Left-turning traffic from the minor street (Movements 7,
of a particular size useful when they are attempting to enter the 10)
intersection. It is generally assumed in gap acceptance theory that
drivers are both consistent and homogeneous. In reality, this as- For example, if a left-turning vehicle on the major street and a
sumption is not entirely correct: past studies have demonstrated through vehicle from the minor street are waiting to cross the
10-6 urban streets
Figure 10-1. Traffic streams at TWSC intersection: (a) four-leg intersection; (b) T-intersection.
major traffic stream, the first available gap (of acceptable size) CAPACITY FORMULA
would be taken by the major street’s left-turning vehicle. The
minor-street through vehicle must wait for the second available The gap acceptance method employed in these procedures was
gap. In aggregate terms, a large number of such left-turning vehi- originally developed in Germany (7,8). The method computes the
cles could use up so many of the available gaps that minor-street potential capacity of each minor traffic stream in accordance with
through vehicles would be severely impeded or unable to make the following equation:
safe crossing movements.
Right-turning vehicles from the minor street are not assumed to exp(−vc,xtc /3,600 )
cp,x = vc,x (10-1)
use up available gaps. Because such vehicles merely merge into 1 − exp(−vc,xtf /3,600)
gaps in the right-hand lane of the stream into which they turn,
they require only a gap in that lane, not in the entire major-street where
traffic flow (this may not be true for some trucks and vans with cp,x = potential capacity of minor Movement x (veh/hr);
long wheelbases, which encroach on more than one lane in making vc,x = conflicting flow rate for Movement x;
their turn). Further, a gap in the overall major-street traffic could tc = critical gap, or minimum time interval that allows intersec-
be used simultaneously by a vehicle in an adjacent lane. For this tion entry to one minor-stream vehicle (sec) for minor
reason, the methodology does not assume that right turns from the Movement x; and
minor street impede any of the other flows using major-street gaps. tf = follow-up time, or headway between departure of one vehi-
Pedestrian movements also have priorities with respect to vehic- cle from minor street and departure of next vehicle under
ular movements. Although this may be a policy issue that varies continuous-queue conditions (sec) for minor Movement x.
by jurisdiction, in both the Policy on Geometric Design of High-
ways and Streets (5) and the Manual on Uniform Traffic Control
Devices (6), it is implied that pedestrians must use acceptable gaps STRUCTURE OF THE METHODOLOGY
in major-street (Rank 1) traffic streams and that pedestrians have
priority over all minor-street traffic at a TWSC intersection. Spe- The basic structure of the methodology is as follows:
cific rankings for pedestrian movements are shown in Figure 10-
1 and are discussed later in this chapter (section on Conflicting 1. Define existing geometric and traffic conditions for the inter-
Traffic). section under study;
2. Determine the conflicting traffic through which each minor- flow rates can be entered directly with the PHF set to 1. The
street movement and the major-street left-turn movement must adjusted flow rate for Movement x is designated vx in this chapter.
cross; By convention, subscripts 1 to 6 are used to define vehicle
3. Determine the size of the gap in the conflicting traffic stream movements on the major street and subscripts 7 to 12 to define
needed by vehicles in each movement crossing or merging into a movements on the minor street. Pedestrian flows impede all minor-
conflicting traffic stream; street traffic streams. Pedestrian volumes must be specified by
4. Determine the capacity of the gaps in the major traffic stream movement, defined by subscripts 13 to 16.
or the ability of these gaps to accommodate each of the subject The presence of traffic signals upstream from the intersection
movements that attempt to utilize these gaps; on the major street will produce nonrandom flows and has an
5. Adjust the calculated capacities to account for impedance effect on the capacity of the minor-street approaches if the signal
and the use of shared lanes; is within 0.25 mi of the intersection. The basic capacity model
6. Adjust the calculated capacities to account for the effect (Equation 10-1) assumes that the headways on the major street are
of upstream signals (platooning) on the major-street traffic flow exponentially distributed. To assess the effect of signals on capac-
headway distribution; ity, a separate analysis is provided that requires the use of signal-
7. Adjust the calculated capacities to account for a two-stage ized-intersection data (cycle length, green time, saturation flow
gap acceptance process at intersections with raised or striped medi- rate, and platooned flow from the signals).
ans or two-way left-turn lanes;
8. Adjust the calculated capacities to account for flared minor-
street approaches; and Conflicting Traffic
9. Estimate the average control delay and queue length for each
of the subject movements and determine the level of service for The nature of conflicting movements for a TWSC intersection
each movement and for the intersection. is relatively complex. Each subject movement faces a different set
of conflicts directly related to the nature of the movement. These
Each of these analysis steps is discussed in detail in the sections
conflicts are shown in Figure 10-2, which illustrates the computa-
that follow.
tion of the parameter vc,x, the conflicting flow for Movement x,
that is, the total flow that conflicts with Movement x, expressed
in vehicles per hour.
Input Data Requirements
The right-turn movement from the minor street, for example, is
in conflict with the major-street through movement only in the
Data requirements for the TWSC intersection methodology are right-hand lane into which right-turners will merge. Figure 10-2
similar to those for other capacity analysis techniques. Detailed includes half of the right-turn movement from the major street,
descriptions of the geometrics, volumes, and control at the intersec- because this flow has been found to have a somewhat inhibiting
tion are needed. effect on the subject movement. This effect may be caused by
Key geometric factors include major-street right-turning vehicles approaching without using their
1. Number and use of lanes; turn indicator, which causes the driver of a waiting vehicle to
2. Channelization; believe that the turning vehicle will travel straight through the
3. Two-way left-turn lane, raised or striped median storage, intersection, or by side frictions created as they turn into a lane
or both; adjacent to waiting vehicles.
4. Approach grade; and Left turns from the major street are in absolute conflict with the
5. Flared approaches on the minor street. total opposing through and right-turn flows, because they must
cross the through flow and merge with the right-turn flow. The
Each of these factors has a substantial effect on the basic capacity method does not differentiate between crossing and merging con-
of each minor movement under any given conflicting volume level. flicts. Left turns from the major street and the opposing right turns
The number and use of lanes are critical factors. Vehicles in from the major street are considered to merge regardless of the
adjacent lanes can use the same gap in the traffic stream simultane- number of lanes provided in the exit leg of the intersection.
ously (unless impeded by a conflicting user of the gap). When Minor-street through movements have a direct crossing or merg-
movements share lanes, only one vehicle from those movements ing conflict with all movements on the major street, as indicated
may use each gap. Channelization is also important because it can in Figure 10-2, except the right turn into the subject approach.
be used to reduce impedance by separating conflicting flows from Only half of this movement is included in the computation, for
each other. A two-way left-turn center lane, raised or stripped the same reasons as those discussed above. In addition, field re-
median, or both allow a minor-stream vehicle to cross one major search (1) has shown that the effect of left-turning vehicles is twice
traffic stream at a time. The approach grade has a direct and their actual number; this is reflected in Figure 10-2.
measurable effect on the capacity of each minor movement: rela- The left turn from the minor street is the most difficult maneuver
tive to a level approach, downgrades increase capacity and up- to execute from a TWSC intersection, and it faces the most com-
grades decrease capacity. A flared approach on the minor street plex set of conflicting flows. Conflicting flows include all major-
increases capacity by allowing more vehicles to be served street flows, in addition to the opposing right-turn and through
simultaneously. movement on the minor roadway. Only half of the opposing right-
Volumes must be specified by movement. For the analysis to turn and through movement flow rate is included as conflicting
reflect conditions during the peak 15 min, the analyst must divide because both movements are stop-controlled and thus their effect
the full-hour volumes by the peak-hour factor (PHF) before begin- on the subject left-turn movement is diminished. The additional
ning computations. If the analyst has peak 15-min flow rates, these capacity impedance effects of the opposing right-turn and through
10-8 urban streets
Figure 10-2. Definition and computation of conflicting volumes.

Figure 10-2. (continued)

10-10 urban streets
movement flow rates are taken into account elsewhere within the drivers may not always act consistently and may occasionally reject
procedure. gaps of greater length than their accepted gap.
When Figure 10-2 is used to compute conflicting flow rates, The time span between the departure of one vehicle from the
the analyst should carefully consult the footnotes to the figure, minor street and the departure of the next vehicle using the same
which allow modifications to the equations shown in special cases. major-street gap, under a condition of continuous queueing on the
Note that in the equations in Figure 10-2, the conflicting traffic minor street, is called the follow-up time, tf. Put another way, tf is
flow for Movement x, which is denoted vc,x, is computed in terms the headway that defines the saturation flow rate for the approach
of an hourly flow rate of mixed vehicles per hour. if there are no conflicting vehicles with movements of higher rank.
Pedestrians may also conflict with vehicular traffic streams. Values of tc and tf for passenger cars are given in Table 10-1.
Pedestrian movements should be included as part of the conflicting These values are based on studies of 79 intersections throughout
flow rates, since they, like vehicular flows, define the beginning the United States and are believed to be representative of a broad
or end of a gap that may be used by a minor-stream vehicle. range of conditions.
Although this methodology recognizes some peculiarities associ- Adjustments are made to account for the presence of heavy
ated with pedestrian flows, it takes a uniform approach to both vehicles, approach grade, T-intersections, and two-stage gap ac-
vehicular and pedestrian movements at an intersection. ceptance. The critical gap is computed separately for each minor
Pedestrian flow rates are also defined as vx, with x denoting the movement.
leg of the intersection being crossed (Figure 10-1):
tc = tc,base + tc,HVPHV + tc,GG − tc,T − t3,LT (10-2)
T x = 15 denotes crossing of the minor-street approach serving
Movements 7, 8, and 9; where
T x = 16 denotes crossing of the opposite minor-street approach tc = critical gap
serving Movements 10, 11, and 12; tc,base = base critical gap from Table 10-1;
T x = 13 denotes crossing of the major-street approach serving tc,HV = adjustment factor for heavy vehicles: 1.0 for two-lane
Movements 1, 2, and 3; and major streets and 2.0 for four-lane major streets;
T x = 14 denotes crossing of the major-street approach serving PHV = proportion of heavy vehicles for minor movement;
Movements 4, 5, and 6. tc,G = adjustment factor for grade: 0.1 for Movements 9 and 12
and 0.2 for Movements 7, 8, 10, and 11;
Although regulations or practices may vary between jurisdic- G = percent grade divided by 100;
tions, this methodology assumes that pedestrians crossing the sub- tc,T = adjustment factor for each part of two-stage gap accept-
ject approach or opposing approaches have Rank 1 status, whereas ance process: +1.0 for first or second stage, zero when
pedestrians crossing the two conflicting approaches to the left or only one-stage gap acceptance process; and
right of the subject minor-street approach have Rank 2 status. t3,LT = adjustment factor for intersection geometry: 0.7 for minor-
The conflicting pedestrian flow rates are included in the equations street left-turn movement at three-leg intersection, 0
shown in Figure 10-2. otherwise.
Figure 10-2 also identifies the conflicting flow rates for each
stage of the two-stage gap acceptance process that takes place at The follow-up time is computed for each minor movement.
some intersections. When a two-stage gap acceptance process is Adjustments are made for the presence of heavy vehicles:
not present, the conflicting volumes shown for Stages I and II tf = tf,base + tf,HVPHV (10-3)
should be added and considered as one conflicting volume for the
movement in question. Two-stage gap acceptance is discussed later where
in this chapter. tf = follow-up time;
tf,base = base follow-up time from Table 10-1;
Critical Gap and Follow-Up Time tf,HV = adjustment factor for heavy vehicles: 0.9 for two-lane
major streets and 1.0 for four-lane major streets; and
The critical gap, tc, is defined as the minimum time interval in PHV = proportion of heavy vehicles for minor movement.
the major-street traffic stream that allows intersection entry to one Values from Table 10-1 should apply for most typical intersec-
minor-stream vehicle (9). Thus, the driver’s critical gap is the tions. It is conceivable that the actual values will differ from these
minimum gap that the driver finds acceptable. A particular driver representative figures at different sites. If smaller values for tc and
would therefore reject any gaps less than this critical gap and tf are observed, they will lead to an increase in capacity in compari-
would accept any gaps greater than or equal to this critical gap. son with those values shown in Table 10-1. If larger values for tc
This definition, although true, makes the critical gap difficult to and tf are used, the result will be a decrease in capacity. Empirical
estimate. In particular, it is impossible to measure the critical gap observations of maximum capacity (i.e., discharge rates from a
through simple field observations. However, estimates of the criti- minor approach with continuous queueing) as well as measure-
cal gap can be made on the basis of observations of the largest ments of critical gaps and follow-up times should be used in such
rejected gap and the accepted gap sizes for a given intersection. cases to calibrate the methodology to local conditions. In all cases,
In most cases, the driver’s largest rejected gap must be less than more accurate capacity estimates will be produced if field estimates
the critical gap, and the driver’s accepted gap must be greater than can be made of the critical gap and follow-up time for the intersec-
or equal to the critical gap. That is, tion under study.
Although some previous studies have shown that factors relating
Largest rejected gap < driver’s critical gap ≤ accepted gap
to major-street speed and minor-stream sight distance may affect
Even this relationship may not always be entirely true, because the critical gap, the findings by Kyte et al. (1) did not show these
Table 10-1. Critical Gaps tc and Follow-Up Times tf for Passenger Cars at TWSC Intersections
Critical Gap tc Follow-Up
Vehicle Maneuver Two-Lane Major Road Four-Lane Major Road Time tf
Left turn, major street 4.1 4.1 2.2
Right turn, minor street 6.2 6.9 3.3
Through traffic, minor street 6.5 6.5 4.0
Left turn, minor street 7.1 7.5 3.5
effects. The effect on major-street speed is primarily one of dis- intersection. Although empirical observations have shown that
tance. Drivers make their gap acceptance decisions based on dis- such delays do occasionally occur, it is important for the analyst
tance, and this results in a time gap that is independent of speed. to keep in mind that although these delays and slowing effects are
It should also be noted that the critical-gap data for multilane sites not taken into account by the basic methodology, adjustments to
account for the actual lane distribution of traffic flows measured estimate the extent of this effect are now included in the procedures
at each site. This distribution accounts for the higher value of (see the section Delay to Major-Street Through Vehicles).
critical gap for the minor-street right turn (6.9 sec) compared with Minor traffic streams of Rank 2 (including left turns from the
the value for the minor-street through movement (6.5 sec), a result major road and right turns from the minor road) must yield only
that at first glance may seem counterintuitive. to the major-road through and right-turning traffic streams of Rank
1. There are no additional impedances from other minor traffic
streams, and thus the movement capacity of each Rank 2 traffic
Potential Capacity for Movement
stream is equal to its potential capacity:
The potential capacity of a movement is denoted cp,x (for Move- cm, j = cp, j (10-4)
ment x) and is defined as the capacity under ideal conditions for
a specific subject movement, assuming the following conditions: where j is movements of Rank 2 priority.
Minor traffic streams of Rank 3 must yield not only to the major
1. Traffic from nearby intersections does not back up into the traffic streams, but also to the conflicting major-street left-turn
intersection under consideration; movements, which are of Rank 2. Thus, not all gaps of acceptable
2. A separate lane is provided for the exclusive use of each length that pass through the intersection will normally be available
minor-street movement under consideration; for use by Rank 3 traffic streams, because some of these gaps are
3. An upstream signal does not affect the arrival pattern of likely to be used by the major-street left-turning traffic instead.
the major-street traffic; that is, the headways are exponentially Rank 3 traffic streams are therefore impeded by major-street left-
distributed; and turning traffic. The magnitude of this impedance depends, of
4. No other movements of Rank 2, 3, or 4 impede the subject course, on the probability that major-street left-turning vehicles
movement. will be waiting for an acceptable gap at the same time as vehicles
Applying Equation 10-1 and using the values presented in Table of Rank 3. A higher probability that this situation will occur means
10-1, Figure 10-3 gives the potential capacity, cp,x, in vehicles per greater capacity-reducing effects of the major-street left-turning
hour of the individual minor traffic streams for a two-lane major traffic on all Rank 3 movements.
road and Figure 10-4 gives the potential capacity for a four-lane What is of interest to the analyst, therefore, is the probability
major road. It can be seen from Figures 10-3 and 10-4 that potential that the major-street left-turning traffic will operate in a queue-
capacity is a function of conflicting flow vc,x expressed as an hourly free state. This probability can be expressed as
rate, as well as a function of the particular minor-street movement vj
being analyzed. p0, j = 1 − c (10-5)
m, j
where p0, j is the probability that conflicting Rank 2 movement j

Impedance Effects
will operate in a queue-free state and j is 1, 4 (major-street left-
Vehicle Impedance turn movements of Rank 2).
The movement capacity cm,k for all Rank 3 movements depends
It has been noted that vehicles use gaps at a TWSC intersection on calculation of a capacity adjustment factor that accounts for
in a priority-ranked manner. When traffic becomes congested in the impeding effects of higher-ranked movements. This capacity
a high-priority movement, it can impede lower-priority movements adjustment factor is denoted fk for all Movements k and for all
(i.e., streams of Ranks 3 and 4) from using gaps in the traffic Rank 3 movements can be expressed as
stream and reduce the potential capacity of these movements. It
should be noted that major-street traffic is not assumed to be fk = pp
j
0, j (10-6)
impeded at any time by minor-street flows and that impedance
effects apply only to minor-street vehicles. where k is the Rank 3 movement. The movement capacity for
Major traffic streams of Rank 1 are assumed to be unimpeded Rank 3 movements can then be computed as
by any of the minor traffic stream movements; this rank also cm,k = (cp,k)fk (10-7)
implies that vehicles in the major traffic streams are not expected
to incur any delay or slowing as they travel through the TWSC Rank 4 movements (i.e., minor-street left turns at a four-leg
10-12

urban streets
Figure 10-3. Potential capacity, two-lane roadway.

Figure 10-4. Potential capacity, four-lane roadway.

10-14 urban streets
intersection) have the potential to be impeded by the queues of p″

p′ = 0.65p″ − + 0.6√p″ (10-8)
three higher-ranked traffic streams: p″ + 3
T Major-street left-turning traffic (Rank 2), where

T Minor-street through movements (Rank 3), and p′ = adjustment to major-street left, minor-street through move-
T Minor-street right-turning movements (Rank 2). ment impedance factor;
p″ = (p0,j)(p0,k);
As before, the probability that each of these higher-ranked traffic
p0,j = probability of queue-free state for conflicting major-street
streams will operate in a queue-free state is central to determining
left-turning traffic; and
their overall impeding effects on the minor-street left-turn move-
p0,k = probability of queue-free state for conflicting minor-street
ment. At the same time, it must be recognized that not all of these
crossing traffic.
probabilities are independent of each other. Specifically, queueing
in the major-street left-turning movement affects the probability Figure 10-5 is entered on the horizontal axis with the value of
of a queue-free state in the minor-street through movement. p″ as defined in Equation 10-8. A vertical line is drawn to the
Applying the simple product of these two probabilities would curve, and a horizontal line is then drawn from the intersection
likely overestimate the impeding effects of these two movements with the curve to the vertical axis, where the result, p′, is read.
on the minor-street left-turning traffic. The capacity adjustment factor for the Rank 4 minor-street left-
The curve in Figure 10-5, based on empirical and theoretical turn movements can then be computed as follows:
work conducted in Germany (3), can be used to adjust for the
overestimate caused by the statistical dependence between queues fl = ( p′)( p0,j) (10-9)
in streams of Ranks 2 and 3. The mathematical representation of
this curve, the validity of which was ascertained through simula- where l is the minor-street left-turn movement of Rank 4 [i.e.,
tion, is given by Movement 7 or 10 in Figure 10-1 (a)] and j is the conflicting Rank
Figure 10-5. Adjustment to major left-turn, minor-through impedance factor (3).

2 minor-street right-turn movement [i.e., Movement 9 or 12 in Table 10-3. Pedestrian Impedance Factors
Figure 10-1(a)]. Vehicle Stream Pedestrian Impedance Factor, pp,x
Finally, the movement capacity for Rank 4 minor-street left-
v1 pp,16
turn movements can be determined from the following equation: v4 pp,15
cm,l = cp,l fl (10-10) v7 pp,15 · pp,13
v8 pp,15 · pp,16
v9 pp,15 · pp,14
Pedestrian Impedance v10 pp,16 · pp,14
v11 pp,15 · pp,16
Minor-street traffic streams must yield to pedestrian streams. v12 pp,16 · pp,13
Table 10-2 shows the relative hiearchy between pedestrian and
vehicular streams that is assumed in this procedure.
Pedestrian groups crossing an intersection thus impede lower-
ranked minor-street vehicles, but only one lane at a time because Equation 10-9 becomes
drivers performing a given through or turning movement tend to
fl = p′p0, j pp, x (10-14)
pass in front of or behind pedestrians once a target lane is clear.
Pedestrian flows are counted somewhat differently than vehicle where pp,x takes on the value pp,13 · pp,15 for v7 and pp,14 · pp,16 for v10.
flows. If the typical pattern is for pedestrians to cross individually,
each pedestrian should be counted individually in the pedestrian
Shared-Lane Capacity
flow. However, if pedestrians tend to cross in groups, the number
of groups should be counted in the pedestrian flow. The important
Up to this point, the methodology has assumed that each minor-
factor is to determine the number of blockages that occur. In most
street movement has the exclusive use of a lane. This geometric
cases, this number will be a combination of individual pedestrians
scenario is often not the case, and frequently two or three move-
and groups of pedestrians. Thus, as defined for the purposes of
ments share a single lane on the minor approach. When this geo-
determining the pedestrian impedance, the pedestrian volume is
metric scenario occurs, vehicles from different movements do not
the sum of individual pedestrians crossing individually and groups
have simultaneous access to gaps, nor can more than one vehicle
of pedestrians crossing together during the time period of study.
from the sharing movements use the same gap.
A factor accounting for pedestrian blockage is computed on the
Occasionally, an intersection with wide corner radii or a flared
basis of the pedestrian volume, the pedestrian walking speed, and
approach will allow vehicles approaching in the same lane to stop
the lane width:
side by side. This geometric effect, which acts to reduce or elimi-
(vx)(w/sp) nate the adverse impact of the shared lane, is considered in a later
fp = (10-11)
3,600 section (see Flared Minor-Street Approaches).
where Minor-Street Approaches

fp = pedestrian blockage factor, or proportion of time that one When several movements share the same lane and cannot stop
lane on an approach is blocked during 1 hr; side by side at the stop line of the intersection, the following
vx = pedestrian flow rate, where x is 13, 14, 15, or 16; equation is used to compute the capacity of the shared lane:
w = lane width; and
sp = pedestrian walking speed, assumed to be 4 ft/sec. vl + v t + v r
cSH = (10-15)
1 2 1 2 1 2
vl vt vr
The pedestrian impedance factor for pedestrian movement x, pp,x, is cm,l + cm,t + cm,r
pp,x = 1 − fp (10-12)
where
If pedestrians are present to a significant degree, pp,x is included
cSH = capacity of shared lane (veh/hr),
as a factor in Equations 10-6 and 10-9. Equation 10-6 becomes
vl = volume or flow rate of left-turn movement in shared lane
fk = p (p
j
)pp,x
0, j (10-13) (veh/hr),
vt = volume or flow rate of through movement in shared lane
where pp,x takes on the values shown in Table 10-3. (veh/hr),
vr = volume or flow rate of right-turn movement in shared lane
(veh/hr),
cm,l = movement capacity of left-turn movement in shared lane
Table 10-2. Relative Pedestrian-Vehicle Hierarchy
(veh/hr),
Vehicle Stream Must Yield to Pedestrian Stream cm,t = movement capacity of through movement in shared lane
v1 v16 (veh/hr), and
v4 v15 cm,r = movement capacity of right-turn movement in shared lane
v7 v15, v13 (veh/hr).
v8 v15, v16
v9 v15, v14 Only those movements included in the shared lane are included
v10 v16, v14 in the equation. For example, if the shared lane includes only right-
v11 v15, v16 turn and through movements, both numerator and denominator
v12 v16, v13 terms for left-turners are deleted in the equation.
10-16 urban streets
Major-Street Approaches intersection will disperse as the vehicles travel downstream away
from the signalized intersection.
It is important to remember that the methodology implicitly Four flow regimes, and thus four headway distributions, result
assumes that an exclusive lane is provided for all left-turning traffic as the platoons arrive at the subject intersection:
from the major street. In situations where a left-turn lane is not
provided, it is possible for major-street through (and possibly right- T Regime 1: no platoons,
turning) traffic to be delayed by left-turning vehicles waiting for T Regime 2: platoon from the left only,
an acceptable gap. To account for this possibility, the factors p*0,1 T Regime 3: platoon from the right only, and
and p*0,4 may be computed as an indication of the probability that T Regime 4: platoons from both directions.
there will be no queue in the respective major-street shared lanes: During Regime 1, minor-stream vehicles enter the subject TWSC
intersection as described by the traditional gap acceptance process.
1 − p0, j When platoons are present from both directions, during Regime
p*0, j = 1 − (10-16)
1 2
v v 4, no minor-stream vehicles are able to enter the subject intersec-
1 − i1 + i2
si1 si2 tion since the mean headways of the platoon are assumed to be
less than the critical gap. Some of the minor-stream movements
where
are blocked by the platoon during Regimes 2 and 3, and are unable
p0, j = probability of queue-free state for movement j assuming to enter the subject intersection. A minor stream is considered to
exclusive left-turn lane on major street; be blocked if a conflicting platoon is traveling through the TWSC
j = 1,4 (major-street left-turning traffic streams); intersection; the stream is considered to be unblocked if no con-
i1 = 2,5 (major-street through traffic streams); flicting platoons are traveling through the TWSC intersection.
i2 = 3,6 (major-street right-turning traffic streams); If the traffic signals at the two upstream intersections are coordi-
si1 = saturation flow rate for major-street through traffic streams nated, these patterns are predictable and occur at regular intervals
(veh/hr), which is a parameter that can be measured in during the hour. On the basis of the flow pattern that exists during
the field; each regime, the capacity can be estimated. If one or both of the
si2 = saturation flow rate for major-street right-turning traffic signals are actuated, or if the signal cycle lengths are different,
(veh/hr), which can be measured in the field; the patterns are less predictable.
vi1 = major-street through flow rate in shared lane; and The analyst needs the following data for each upstream signal:
vi2 = major-street right-turning flow rate in shared lane, or 0 if
T Cycle length (sec);
exclusive right-turn lane is provided.
T Effective green time (sec) for the major-street through move-
ment and, if applicable, for the exclusive left-turn phase from the
By using p*0,1 and p*0,4 in lieu of p0,1 and p0,4 (as computed from
minor street;
Equation 10-5), the additional influence of the potential for queues
T Saturation flow rate in vehicles per hour of green (veh/hrg);
on a major street with shared left-turn lanes may be properly taken
T Distance from the signalized intersection to the subject TWSC
into account.
intersection;
T Speed of the platoon as it progresses from the signalized
Upstream Signals intersection to the TWSC intersection;
T Through flow rate arriving at the signalized intersection on
the major-street approach and, if applicable and significant, the
The existence of nearby upstream signalized intersections (i.e.,
left-turn flow rate from the side street during an exclusive left-
traffic signals on the major street within 0.25 mi of the subject
turn phase; and
intersection) usually causes vehicles to arrive at the subject inter-
T Arrival type of vehicles at the signalized intersection.
section in platoons. Major-street vehicles arriving at a TWSC inter-
section in platoons from a single direction may cause an increase The method includes five sets of computations:
in the minor-street capacity compared with the case of random
arrivals. The greater the number of vehicles traveling in platoons, 1. Time for the queue to clear at each upstream signalized
the higher the minor-street capacity for a given opposing flow intersection,
because there is a greater proportion of large gap sizes that can 2. Proportion of time that the subject TWSC intersection is
be used by more than one minor-street vehicle. When signalized blocked as a result of platoons from each upstream intersection,
intersections exist upstream of the subject TWSC intersection in 3. Duration of the defining platoon events for each of the four
both directions, the effect is much more complex. When a traffic flow regimes,
signal is more than 0.25 mi from the subject intersection, the effect 4. Conflicting flows during each unblocked period, and
on capacity is greatly diminished, since the major-street gaps are 5. Weighted capacity for each movement.
once again negative-exponentially distributed as the platoons
Computation 1: Time for Queue To Clear at Each Upstream
disperse. Signalized Intersection
The method described here considers the flow patterns that result
from traffic signals located upstream of the subject TWSC intersec- In a typical four-leg signalized intersection, three movements
tion and the headway distribution that results from the platooned combine to constitute the exit-leg flow toward the subject TWSC
flow. The method is based on a platoon dispersion algorithm (10– intersection: the through movement on the major street and the
12). Figure 10-6 shows a generalized case of a TWSC intersection right- and left-turn movements from the minor street.
located on an arterial between two signalized intersections. The In addition, the exit-leg flow consists of two components: (a) a
queues that form during the respective red phases at each signalized stable platoon discharging at the saturation flow rate when the
Figure 10-6. Platoon dispersion from upstream signalized intersections.
The proportion of vehicles arriving on the green is computed

as follows:
P = Rp(g/C) P ≤ 1.0 (10-17)
where Rp is the platoon ratio, a function of the arrival type (Table
9-2).
The time to discharge the vehicles that arrive during the red is
given by
vprogC(1 − P)
gq1 = (10-18)
s
where vprog is either vT or vL,prot.
The time to discharge the vehicles that arrive on the green and
join the back of the queue is given by
vprogCPgq1
gq2 = (10-19)
Figure 10-7. Upstream signalized intersection. sg − vprogCP
where vprog is either vT or vL,prot.
signal changes from red to green, and (b) more or less random The total time to discharge the queue is
arrivals and departures of a platoon from another upstream signal gq = gq1 + gq2 (10-20)
passing through on the green. The first component includes both
the portion of vT that arrives during the red and the portion that where gq is less than or equal to g.
arrives during the green when the standing queue is clearing. It
also includes vL for the same periods if vL has an exclusive left- Computation 2: Time Subject TWSC Intersection Blocked Because
turn lane and a protected green phase (see Figure 10-7). The second of Upstream Platoons
component includes vR and the portion of vT (and vL, if applicable)
The discharging queue from the upstream signal will disperse
that arrives after the queue has cleared.
as it travels downstream toward the subject TWSC intersection.
The time that it takes for a standing queue to clear is dependent
A platoon dispersion model is used to determine the time duration
on the pattern of vehicles arriving at the upstream signalized inter-
during which the TWSC intersection is blocked by the densest
section. The arrival pattern, designated ‘‘arrival type’’ in Chapter
part of the platoon. The platoon headways are smaller than the
9 of this manual, is determined by the proportion of vehicles
critical gap, and thus no minor movement at the TWSC intersection
arriving during the green phase. For vT, the arrival type ranges
can enter the intersection during passage of the platoon (see Figure
from 1 (very poor progression; few vehicles arrive on the green)
10-8).
to 6 (exceptional progression; most vehicles arrive on the green
The basic platoon dispersion model parameters are as follows:
in a structured platoon). For vL,prot it is assumed to be 3 (random
arrivals). a = platoon dispersion factor obtained from Table 10-4;
10-18 urban streets
Figure 10-8. Platoon dispersion model [adapted from Bonneson and Fitts (12)].
Table 10-4. Platoon Dispersion Factor, a (12, 1) The duration of the blocked period for either the through move-
No. of Through Traffic Lanes ment or the protected left-turn movement is computed using the
following equation, where ‘‘ln’’ indicates the natural logarithm:
Median Type N=1 N=2 N=3
5
311 − 21 24
Undivided 0.55 0.50 0.40 vc,min vc,max − vprog Rp f
ln
Raised curb 0.45 0.40 0.35 sf vc,min − vprog Rp f
TWLTL 0.40 0.35 0.30 tp,i = gq − vc,max > vc (10-22)
ln(1 − F)
0 vc,max ≤ vc,min
b = (1 + a) ;
−1
ta = D/Sprog, travel time from signalized intersection to subject with all variables defined earlier. Note that the variables vc,min, vc,max,
TWSC intersection, where D is distance from upstream sig- s, and vprog must be in the same units, either vehicles per hour or
nal to subject movement and Sprog is average platoon run- vehicles per hour per lane. The subscript i is set to T when the
ning speed; blocked period caused by the through-movement platoon is com-
F = (1 + abta)−1; and puted; the subscript is set to L when the blocked period caused by
f = vprog/vc ≥ 0 and ≤ 1, proportion of conflicting flow that origi- the protected left-movement platoon is computed.
nated at upstream signal, where vprog is either vT when the The proportion of time blocked is computed using the following
platoon generated by the through movement is considered equation, considering both the through movement and the protected
or vL,prot when the platoon generated by the protected left- left-turn movement platoons:
turn movement from the minor street is considered. tp,T + tp,L
p= (10-23)
The maximum platooned flow rate in the conflicting stream is C
given by
g
vc,max = sf[1 − (1 − F) q] (10-21) Computation 3: Platoon Event Periods
The minimum platooned flow rate, vc,min, is at least larger than The purpose of the third computation is to determine the propor-
3,600N/tc, where N is the number of through lanes per direction tion of the study period during which each of the four flow regimes
on the major street. By default, it is assumed to be equal to 1,000N exists. In particular, it is important to determine, for each minor
veh/hr on the basis of simulation data (12). movement, the proportion of the study period that is unblocked.
The existence of a traffic signal on both upstream approaches for each minor movement. The results for each minor movement
will result in an overlapping platoon structure at the subject TWSC for the average case are shown in Table 10-6.
intersection. Depending on the signal timing parameters, a range
of cases may present themselves, from a best case of simultaneous Computation 4: Conflicting Flows During Each Unblocked Period
platoons from both directions to a worst case of alternating pla-
The flow for the unblocked period (that is, the time periods
toons from each direction. An average case results in a partial
when no platoons are present) is determined. This flow becomes
overlap of the platoons (Figure 10-9). Figure 10-9 can also be
the conflicting flow for the subject movement and is used to com-
interpreted to represent the expected pattern averaged over the
pute the capacity for this movement.
analysis period. The method described here is based on the aver-
The conflicting flow for Movement x during the unblocked pe-
age case.
riod is given by
If p2 and p5 represent the proportion of the study period during
which Movements 2 and 5, respectively (and their corresponding vc,x − s(1 − px)
vc,u,x = (10-28)
turning movements), are blocking the TWSC intersection, the pro- px
portion of the study period during which blockages exist can be
computed. where vc,x > s (1 − px); vc,u,x is zero otherwise; and
The dominant and subordinate platoons are determined: vc,x = total conflicting flow for Movement x as determined from
pdom = max(p2, p5) (10-24) Figure 10-2;
psubo = min(p2, p5) (10-25)

Table 10-6. Proportion of Study Period Unblocked
Two conditions exist. In the unconstrained condition, there is some for Each Minor Movement for Average Case
period of time during which neither platoon is present:
Proportion Unblocked Unconstrained Constrained
pdom + (psubo /2) ≤ 1 (10-26) for Movement, px Condition Condition
The constrained condition exists when one or both platoons are p1 1 − p5 1 − p5

p4 1 − p2 1 − p2
always present: p7 1− (pdom + psubo /2) 0
pdom + (psubo /2) > 1 (10-27) p8 1− (pdom + psubo /2) 0
p9 1 − p2 1 − p2
Table 10-5shows the proportion of the study period for each of p10 1− (pdom + psubo /2) 0
the four flow regimes for the average case. It is used to determine p11 1− (pdom + psubo /2) 0
the proportion of the study period that is blocked and unblocked p12 1 − p5 1 − p5
Figure 10-9. Various platoon overlap cases: best case—platoons completely overlap so unplatooned period is maximum; worst case—
platoons alternate so unplatooned period is minimum; average case—one-half of subordinate platoon is subsumed by dominant platoon.
Table 10-5. Proportion of Study Period for Each Flow Regime for Average Case
Flow Regime Unconstrained Condition Constrained Condition
1, no platoons 1 − (pdom + psubo /2) 0
2, dominant platoon only pdom − psubo /2 1 − psubo
3, subordinate platoon only psubo /2 1 − pdom
4, both platoons psubo /2 pdom + psubo − 1

10-20 urban streets
s = total saturation flow rate for that movement, which is con-

flicting flow for Movement x during blocked period; and
px = proportion of time that subject Movement x is unblocked
by major-street platoon, determined from Table 10-6.
Computation 5: Capacity for Subject Movement During Unblocked

Period
The capacity of subject Movement x, accounting for the effect

of platooning, is given by
cplat,x = pxcr,x (10-29)
where px is the proportion of time that Movement x is unblocked

by a platoon and cr,x is the capacity of Movement x assuming
random flow during the unblocked period, using the conflicting
flow vc,u,x computed for this unblocked period and Equation 10-1.
Two-Stage Gap Acceptance
The existence of a raised or striped median or a two-way left-

turn lane (TWLTL) on the major street often causes some degree
of a special gap acceptance phenomenon known as two-stage gap
acceptance. For example, the existence of a raised or striped me-
dian allows a significant proportion of the minor-street drivers to
cross part of the major-street approach first and then to pause in
the middle of the road to wait for a gap on the other approach.
When a TWLTL exists on the major street, the minor-street left-
turning vehicle usually merges into the TWLTL first and then
seeks a usable gap on the other approach while slowly moving for
some distance along the TWLTL. Both of these behaviors can
contribute to increased capacity.
In this procedure, the intersection is assumed to consist of two
parts with the minor-street traffic crossing the major street in two
stages. Between partial intersections I and II there is storage space Figure 10-10. Intersection with two-stage gap acceptance
for m vehicles. This area has to be passed by the left-turner from process.
the major street (Movement V1 or V4) and the minor through or
left-turn traffic. It is assumed that the usual rules for TWSC inter- The capacity for the subject movement considering the two-
stage gap acceptance process is computed as follows. An adjust-
sections are applied by drivers at the intersections. Thus the major
ment factor a and an intermediate variable y are computed:
through traffic has priority over all other movements. For example,
Movement 1 (or Movement 4) vehicles must give way to priority a = 1 − 0.32 exp(−1.3√m) for m > 0 (10-30)
Movement 5 (or Movement 2), whereas Movement 7 or 8 (or cI − cm,x
Movements 10 and 11) has to give the right-of-way to all other y= (10-31)
cII − vL − cm,x
movements (see Figure 10-10).
The conflicting volumes are defined for each minor-stream where
movement that uses the two-stage gap acceptance process using m = number of storage spaces in the median;
Figure 10-2 for both the first-stage and second-stage movements. cI = capacity for Stage I process, Equation 10-1;
For the first stage, the conflicting flows consist of the major-street cII = capacity for Stage II process, Equation 10-1;
flows from the left. For the second stage, the conflicting flows vL = major left-turn flow rate, either v1 when considering Move-
consist of the major-street flows from the right. The specific ment 1, 2, or 3 or v4 when considering Movement 4, 5, or
streams that are included in each conflicting flow are shown in 6; and
Figure10-2. cm,x = capacity of subject movement considering total conflicting
First, the capacity for the subject movement is computed assum- volume for both stages of two-stage gap acceptance process,
ing a single-stage gap acceptance process through the entire inter- Equation 10-1.
section. Next the capacity for Stage I, cI, and the capacity for Stage
The total capacity cT of the intersection for the subject movement
II, cII, are computed using the appropriate values of critical gap
considering the two-stage gap acceptance process is computed.
and follow-up time for the two-stage gap acceptance process using
For y ≠ 1:
Equations 10-2 and 10-3: c1 is the capacity considering conflicting
flows vc,I, and cII is the capacity considering conflicting flows vc,II a
cT = [y(y m − 1)(cII − vL) + (y − 1)cm,x] (10-32)
(Figure 10-2). ym+1 − 1
For y = 1: two vehicles to proceed, one beside the other, to the stop line. In
this case, the storage can be defined as n = 1, since one additional
a
cT = [m(cII − vL) + cm,x] (10-33) vehicle is able to use the stop line.
m+1 The actual capacity resulting from this configuration will be
greater than that in the case in which the right-turning vehicles
Flared Minor-Street Approaches must share the lane and less than that in the case in which the
vehicles have separate lanes. The analyst must compute the average
Geometric elements near the stop line on stop-controlled ap- queue length for each movement considering the separate-lane
proaches of many intersections may result in a greater capacity case and the actual storage available in the flared-lane area for the
than may be predicted by the shared-lane capacity formula because intersection approach under study. Figure 10-11 shows how the
at such approaches two vehicles may occupy or depart from the actual capacity can be interpolated using this information.
stop line simultaneously as a result of a large curb radius, a tapered First, the average queue length for each movement sharing the
curb, or a parking prohibition. The magnitude of this effect will right lane of the approach is computed, assuming that each move-
depend in part on the turning-movement flow rates and the resul- ment operates as a separate lane. The movement with the maximum
tant probability of there being two vehicles simultaneously at the average queue length is identified:
stop line and in part on the storage length available to feed the dsepvsep
second position at the stop line. Qsep = (10-34)
3,600
If n is defined as the number of spaces for passenger cars belong-
ing to one movement that can queue at the stop line without ob- where
structing the access to the stop line for other movements, it is clear
Qsep = average queue length for movement considered as a sepa-
that with n > 0, the capacity of the minor-street approach is in-
rate lane,
creased compared with that in the shared-lane condition. With an
dsep = average control delay for movement considered as a sepa-
increase in n, the total capacity approaches the case in which each
rate lane, and
movement has its own individual lane of infinite length. Figure
vsep = flow rate for movement.
10-11 shows a situation in which the curb lane provides space for
Next, the required length of the storage area is computed such
that the approach would operate effectively as separate lanes. This
length is the maximum value of the queue lengths computed for
each separate movement plus one vehicle length:
nmax = max round(Qsep,i + 1) (10-35)
i
where
Qsep,i = average queue length for Movement i in the shared lane
considered as separate lane;
round = round-off operator, rounding quantity in parentheses to
nearest integer;
max = operator determining maximum value of various values
of Qsep,i; and
nmax = length of storage area such that approach would operate
as separate lanes.
Finally, the capacity of the approach is computed, taking the
flare into account. The capacity is interpolated, as shown in Figure
10-11. A straight line is established using values of two points:
(csep, nmax) and (cSH, 0). The interpolated value of cact is computed
using the following equation:
1o c 2
n
cact = sep − cSH n + cSH when n ≤ nmax (10-36)
i max
= oc
i
sep when n > nmax
where
cact = actual capacity of flared approach,
csep = capacity for separate-lane case, and
cSH = capacity of shared-lane case.
Queue Lengths
Figure 10-11. Capacity approximation at intersections with flared Queue length estimation is an important consideration at unsig-
minor-street approach. nalized intersections. Theoretical studies and empirical observa-
10-22 urban streets
Figure 10-12. Estimation of 95th-percentile queue length.
tions have demonstrated that the probability distribution function Control Delay
for queue lengths for any minor movement at an unsignalized The delay experienced by a motorist is made up of a number
intersection is a function of the capacity of the movement and the of factors that relate to control, geometrics, traffic, and incidents.
volume of traffic being served during the analysis period. Figure Total delay is the difference between the travel time actually expe-
10-12 can be used to estimate the 95th-percentile queue length for rienced and the reference travel time that would result during
any minor movement at an unsignalized intersection during the conditions with ideal geometrics and in the absence of incidents,
peak 15-min period on the basis of these two parameters (13). control, and traffic. Chapters 9 and 10 of this manual quantify
The mean queue length is computed as the product of the aver- only that portion of total delay attributed to traffic control mea-
age delay per vehicle and the flow rate for the movement of inter- sures, either traffic signals or stop signs. This delay is called con-
est. The expected total delay (in vehicle hours per hour) equals trol delay and its use is consistent in Chapters 9, 10, and 11.
the expected number of vehicles in the average queue; that is, the Control delay includes initial deceleration delay, queue move-up
total hourly delay and the average queue are numerically identical. time, stopped delay, and final acceleration delay.
For example, 4 vehicle-hr/hr of delay can be used interchangeably Although the methodology here results in an estimate of average
with an average queue length of 4 during the hour. control delay, it is recommended that, where possible under ex-
isting conditions, estimates of average control delay also be ob- In the simplest procedure, the proportion of major-street Rank
tained through field measurements. 1 vehicles not being blocked (i.e., in a queue-free state) is given
Average control delay for any particular minor movement is a by p*0,j in Equation 10-16 (p*0, j should be substituted for the major
function of the capacity of the approach and the degree of satura- left-turn factor p0, j in Equation 10-6 when the capacity of lower-
tion. The analytical model used to estimate control delay (Equation ranked conflicting movements is calculated). Therefore, the pro-
10-37) assumes that demand is less than capacity for the period portion of Rank 1 vehicles being blocked is 1 − p*0, j.
of analysis. In situations where the degree of saturation is greater The average delay to Rank 1 vehicles on this approach is
than about 0.9, average control delay is significantly affected by given by
the length of the analysis period. In most cases, the recommended
1N2
analysis period is 15 min, or 0.25 hr. If demand exceeds capacity vi,1
5
(1 − p*0, j) (dM,LT)
during a 15-min period, the delay results calculated by the proce-
N>1
dure may not be accurate. In this case, the period of analysis should dRank 1 = v j,1 + vi,2 (10-38)
be lengthened to equal the period of oversaturation.
(1 − p*0, j)(dM,LT) N=1
3,600
d= + 900T
cm,x where
dRank 1 = delay to Rank 1 vehicles;
3 ! 4
1 c 21c 2
3,600 vx
2 p*0, j = proportion of Rank 1 vehicles not blocked (Equation
1 2
vx vx m,x m,x
× −1+ −1 + +5 10-16);
cm,x cm,x 450T
dM,LT = delay to major-street left-turning vehicles;
(10-37) vi,1 = flow rate of major-street through vehicles in shared
where lane; and
d = average control delay (sec/veh); vi,2 = flow rate of major-street right-turning vehicles in
vx = volume for Movement x, expressed as hourly flow rate; shared lane.
cm,x = capacity of Movement x, expressed as hourly flow rate; and Note that on a multilane road, only the major-street volumes in
T = analysis time period (hr) (for 15-min analysis period, use the lane that may be blocked should be used in the calculation as
T = 0.25). vi1 and vi2. On multilane roads if it is assumed that blocked Rank
1 vehicles do not bypass the blockage by moving into other through
The final term, a constant value of 5 sec/veh, is included in Equa-
lanes (a reasonable assumption under conditions of high major-
tion 10-37 to account for the deceleration of the vehicle from
street flows), then vi1 = v1/N.
cruise speed to the speed of the vehicles in the queue and the
Because of the unique characteristics associated with each site,
acceleration of the vehicle from the stop line to cruise speed.
the decision whether to account for this effect should be left to
Equation 10-37 is depicted graphically in Figure 10-13 for a
the analyst. Geometric design features such as an adjacent exclu-
discrete range of capacities and a 15-min analysis period.
sive right-turn lane, a large curb radius, or a wide shared left and
through travel lane may enable Rank 1 vehicles to bypass the
Delay to Major-Street Through Vehicles
blockage caused by major left-turning vehicles. Also, conflicting
traffic volumes in such adjacent bypass lanes must provide suffi-
cient gaps to accept bypassing vehicles.
Traffic engineers are also interested in knowing the effect of a
shared lane on the major-street approach where left-turning vehi-
cles may block Rank 1 through or right-turning vehicles. If no Other Relevant Delay Estimates
exclusive left-turn pocket is provided on the major street, a delayed
left-turning vehicle may block the Rank 1 vehicles behind it and Frequently it is also useful for the analyst to be aware of the
cause them some delay. This effect delays not only Rank 1 vehi- average control delay per vehicle for an entire approach and for
cles, but also lower-ranked streams. When the delayed Rank 1 the entire intersection. These parameters are important to consider
vehicles are discharged from the queue formed behind a major when estimates of delay under various types of traffic control are
left-turning vehicle, they impede lower-ranked movements with compared.
which they conflict. In this section the impedance for major-street The average approach delay for all vehicles on a particular
left-turning vehicles in a shared lane is used to estimate delay to approach can be computed as the weighted average of the control
Rank 1 vehicles. delay estimates for each individual movement on the approach:
Field observations have shown that the effect of such a blockage
is usually very small because the major street provides enough dRTvRT + dTHvTH + dLTvLT
dA = (10-39)
space for the blocked Rank 1 vehicle to sneak around the left- vRT + vTH + vLT
turning vehicle. Models could be developed from a theoretical
point of view when the major-street width does not allow a through where
vehicle to bypass the left-turning vehicle. At a minimum, incorpo- dA = average approach control delay (sec/veh);
rating this effect requires the following information: dRT, dTH, dLT = computed average control delay for right-turn,
through, and left-turn movements, respectively
1. The proportion of Rank 1 vehicles being blocked, and (sec/veh); and
2. The average delay to the major-street left-turning vehicles vRT, vTH, vLT = volume or flow rate of right-turn, through, and
that are blocking through vehicles. left-turn approach traffic, respectively (veh/hr).
10-24 urban streets
Figure 10-13. Average control delay.

Similarly, the average intersection control delay can be computed: ment). This is not inconsistent in terms of the stated criteria. Left-
turn movements will generally experience longer control delays
dA,1vA,1 + dA,2vA,2 + dA,3vA,3 + dA,4vA,4
dintx = (10-40) than other movements because of the nature and priority of the
vA,1 + vA,2 + vA,3 + vA,4 movement. If left turns are placed in a shared lane, the average
where dA,x is the average approach control delay on Approach x control delay to vehicles in that lane may indeed be less than the
(sec/veh), and vA,x is the volume or flow rate on Approach x (veh/ average control delay to left turns in a separate lane. However, all
hr). If the effects of delay to major-street through vehicles, as vehicles in the shared lane experience increased control delay over
calculated using Equation 10-38, are not be included in the analy- the condition in which left turns have a separate lane. Consider
sis, the delay for all major-street movements of Rank 1 is assumed the following:
to be zero seconds per vehicle. 1. Ten left-turners will experience an average control delay of
10 sec if they have a separate lane and 15 sec if they share a lane
LEVEL-OF-SERVICE CRITERIA with a through movement.
2. Fifty through vehicles will experience an average control
The level of service for a TWSC intersection is determined by delay of 5 sec if they have a separate lane and 6 sec if they share
the computed or measured control delay and is defined for each a lane with the 10 left-turners.
minor movement. Level of service is not defined for the intersec- If the vehicles are forced to share a lane, the average control delay
tion as a whole. LOS criteria are given in Table 10-7. to a vehicle in the shared lane will be
Average control delay less than 10 sec/veh is defined as LOS
A. Follow-up times of less than 5 sec/veh have been measured [(10 × 15) + (50 × 6)] 450
= = 7.5 sec/veh (10-41)
when there is no conflicting traffic for a minor-street movement, (10 + 50) 60
so control delays of less than 10 sec/veh are appropriate for low
flow conditions. Table 10-8 illustrates this comparison. Although each vehicle ex-
The proposed LOS criteria for TWSC intersections are some- periences increased control delay when placed in a shared lane,
what different than the criteria used in Chapter 9 of this manual the average control delay in the shared lane is less than the average
for signalized intersections. The primary reason for this difference control delay to left-turners in an exclusive lane and more than
is that drivers expect different levels of performance from different the average control delay to through vehicles in an exclusive lane.
kinds of transportation facilities. The expectation is that a signal- Thus, the level of service in the exclusive left-turn lane may be
ized intersection would be designed to carry higher traffic volumes poorer than that for the shared lane. The analyst, however, may
than an unsignalized intersection. In addition, a number of driver wish to carefully consider the aggregate impact on control delay.
behavior considerations combine to make delays at signalized in- In general, expanding a one-lane stop-controlled approach to in-
tersections less onerous than delays at unsignalized intersections. clude an exclusive left-turn or right-turn lane will decrease the
For example, drivers at signalized intersections are able to relax approach control delay, regardless of LOS designations.
during the red interval, whereas drivers on the minor approaches
to unsignalized intersections must remain attentive to the task of Consideration of Queue Lengths
identifying acceptable gaps and vehicle conflicts. Also, there is
often much more variability in the amount of delay experienced LOS F exists when there are insufficient gaps of suitable size
by individual drivers at an unsignalized intersection versus that at to allow side-street demand to safely cross through a major-street
signalized intersections. For these reasons, it is considered that the traffic stream. This level of service is generally evident from ex-
control delay threshold for any given level of service would be less tremely long control delays experienced by side-street traffic and
for an unsignalized intersection than it would be for a signalized by queueing on the minor-street approaches. The method, however,
intersection. is based on a constant critical gap size; that is, the critical gap
remains constant no matter how long the side-street motorist waits.
INTERPRETATION OF RESULTS LOS F may also appear in the form of side-street vehicles selecting
smaller-than-usual gaps. In such cases, safety may be a problem,
and some disruption to the major traffic stream may result. It is
Shared Lanes on Minor Approach important to note that LOS F may not always result in long queues
but may result in adjustments to normal gap acceptance behavior,
A movement, most often a left-turn movement, can sometimes which are more difficult to observe in the field than queueing.
have a poorer level of service if it is given a separate lane than if In most cases at TWSC intersections the critical movement is
it shares a lane with another movement (usually a through move- the minor-street left-turn movement. As such, the minor-street left-
turn movement can generally be considered the primary factor
affecting overall intersection performance. The lower threshold for
Table 10-7. Level-of-Service Criteria LOS F is set at 50 sec of delay per vehicle. There are many
Level of Service Delay Range instances, particularly in urban areas, in which the delay equations
A ≤10 will predict delays of 50 sec (LOS F) or more for minor-street
B >10 and ≤15 movements under very low volume conditions on the minor street
C >15 and ≤25 (less than 25 veh/hr). Since the first term of the equation is a
D >25 and ≤35 function only of the capacity, the LOS F threshold of 50 sec/veh
E >35 and ≤50 is reached with a movement capacity of approximately 85 veh/hr
F >50 or less.
10-26 urban streets
Table 10-8. Example Left-Turn Delay Calculation

Separate-Lane Case Shared-Lane Case
Control Cumulative Control Cumulative
Volume Delay/Vehicle Control Delay Delay/Vehicle Control Delay
Movement (veh) (sec/veh) (sec) (sec/veh) (sec)
LT 10 10 100 15 150
TH 50 5 250 6
6 7.5 300
Total 350 Total 450
Figure 10-14. Queue-versus-delay relationship.
This procedure assumes random arrivals on the major street. ranging from a low of 10 veh/hr to a high of 700 veh/hr. As can
For a typical four-lane arterial with average daily traffic volumes be seen from Figure 10-14, the current LOS F threshold of 50 sec/
in the range of 15,000 to 20,000 vehicles per day (peak hour, veh based solely on average delay can be exceeded under many
1,500 to 2,000 veh/hr), the delay equation used in the TWSC low-volume, low-v/c, and low-queue conditions. Of concern is the
capacity analysis procedure will predict 50 sec of delay or more region (denoted Region 2) that includes v/c ratios less than 1.0
(LOS F) for many urban TWSC intersections that allow minor- and average delays greater than 50 sec. In this region, the average
street left-turn movements. The LOS F threshold will be reached queue length is typically fewer than one vehicle, which indicates
regardless of the volume of minor-street left-turning traffic. Not- that, although drivers would likely experience relatively long de-
withstanding this fact, most low-volume minor-street approaches lays, it is unlikely that long queues would form because of the
would not meet any of the volume or delay warrants for signaliza- low demand volumes.
tion of the Manual on Uniform Traffic Control Devices (MUTCD) There are also conditions in which average delays will be less
since the warrants define an asymptote at 100 veh/hr on the minor than 50 sec/veh but drivers will be faced with very long queues,
approach. As a result, many public agencies that use the HCM as can be seen on the far right of Region 1 in Figure 10-14. This
level of service thresholds to determine the design adequacy of would represent conditions in which there is a long queue that is
TWSC intersections may be forced to eliminate the minor-street being served relatively fast. Region 3 includes volume-to-capacity
left-turn movement, even when the movement may not present ratios greater than 1.0. In this region drivers would be faced with
any operational problem, such as the formation of long queues on extremely long queues, extremely long delays, or both.
the minor street or driveway approach. In the performance evaluation of TWSC intersections, it is im-
This point is illustrated more clearly in Figure 10-14, which portant to consider other measures of effectiveness (MOEs) in
presents plots of average delay and average queue lengths for an addition to delay, such as v/c ratios for individual movements,
individual movement with volume-to-capacity ratios varying from average queue lengths, and 95th-percentile queue lengths. By fo-
0.2 to 1.4. The points on each of the v/c lines are for volumes cusing on a single MOE for the worst movement only, such as
delay for the minor-street left turn, users may make inappropriate control type. This set of warrants represents guidance based on
traffic control decisions. The potential for making such inappropri- collective professional consensus accumulated over many dec-
ate decisions is likely to be particularly pronounced when the HCM ades. Practicing traffic engineers can refer to these warrants
level of service thresholds are adopted as legal standards, as is the whenever issues arise regarding decisions on intersection con-
case in many public agencies. trol types.
A number of important implications result from the use of aver- The TEH points out that traffic signals do not always increase
age delay as the sole basis for determining level of service, particu- safety and reduce delay. Therefore, it is not appropriate to install
larly when only the worst movement at the intersection is signals regardless of the traffic volume conditions. The TEH cites
considered. One of the primary motivations for using average delay the following warrants for all-way stop control (from the
as the primary MOE as opposed to reserve capacity (as was done MUTCD):
in the 1985 HCM) was to provide the user community with a more
1. As an interim measure that can be installed quickly while
direct and consistent way to compare unsignalized and signalized arrangements are being made for a warranted traffic signal;
intersection operations. In doing so, however, it is important for 2. When an accident problem, as indicated by five or more
users to understand that there are other operational indicators that reported accidents in a 12-month period, is of a type that can be
must be considered as well as average delay. corrected using a multiway stop and less restrictive controls have
not been successful; and
3. For the following minimum traffic volumes: (a) the total
DETERMINATION OF INTERSECTION CONTROL TYPE
vehicle volume entering the intersection from all approaches aver-
ages at least 500 veh/hr for any 8 hr of an average day, and (b)
Determination of an appropriate control for an intersection, ei- the combined vehicular and pedestrian volume from minor streets
ther signal control or some form of stop control, is now made averages at least 200 units/hr for the same 8 hr with an average
by integrating information from several sources. Traffic signal delay to minor-street traffic of at least 30 sec/veh during the maxi-
warrants, LOS analyses, accident data, and public complaints form mum hour [but when the 85th-percentile approach speed of the
the basis for a decision to signalize an intersection or to change major-street traffic exceeds 40 mph, minimum volume warrants
to stop control. Three documents, among others, are available to are 70 percent of the requirement in (a)].
assist the traffic engineer in this assessment: the MUTCD, the ITE
Traffic Engineering Handbook (TEH) (14), and the HCM. Regarding traffic signal warrants, the TEH states: ‘‘Traffic sig-
The MUTCD provides a set of warrants to help determine the nals that are appropriately justified, properly designed, and effec-
appropriate conditions for signalization, two-way stop control, or tively operated can be expected to achieve one or more of the
all-way stop control. The following 11 signal warrants are provided following:
in the MUTCD: 1. To effect orderly traffic movement through an appropriate
assignment of right-of-way,
1. Minimum vehicular volume, 2. To provide for the progressive flow of a platoon of traffic
2. Interruption of continuous traffic, along a given route,
3. Minimum pedestrian volume, 3. To interrupt heavy traffic at intervals to allow pedestrians
4. School crossings, and cross-street traffic to cross or to enter the main street flow,
5. Progressive movement, 4. To increase the traffic handling ability of an intersection, or
6. Accident experience, 5. To reduce the frequency of occurrence of certain types of
7. Systems, accidents.’’
8. Combination of warrants,
9. Four-hour volumes, The HCM provides methodologies to compute delay and level
10. Peak-hour delay, and of service based on certain traffic volume and intersection
11. Peak-hour volume. conditions given both stop control and signal control. With the
development of new models and procedures of estimating capacity
Although only one of these warrants is required to be met before and delay at stop-controlled intersections, it is now possible to
a signal is recommended, traffic engineers should ideally consider compare different intersection control types from an operational
all these aspects when making a decision regarding an intersection perspective.
The analysis of TWSC intersections is generally applied to ex- to an existing intersection or on the projected design of a new
isting locations either to evaluate existing operational conditions intersection. The procedures, however, are easily manipulated to
under present demands or to estimate the impacts of anticipated investigate the impact of key design features on probable
new demands. The methodology is specifically structured to yield operations.
a level of service and an estimate of average control delay for an Figure 10-15 shows the procedures, which are divided into three
existing or planned TWSC intersection. Thus, operational analysis separate modules. In the first module, Initial Calculations, the
is the mode in which it is used. Design applications are treated as analyst uses Worksheets A1 through A4 to record input conditions
trial-and-error computations based on anticipated improvements and to compute the critical gap and follow-up time. Worksheets
10-28 urban streets
Figure 10-15. TWSC intersection capacity and LOS computational procedures.

A5 through A9 form the second module, Capacity Calculations, destrians. The procedure includes computations that account for
to compute the capacity of each movement and make adjustments the effects of these factors.
for the effects of upstream signals, two-stage gap acceptance,
shared lanes, or flared minor-street approaches. The third module,
Delay and LOS Calculations, includes worksheets to compute the Geometric Features and Movement Definitions
delay, queue length, and level of service for the intersection.
Worksheet A1 shows the basic features of the intersection and
the movements of interest. The intersection name, analyst’s name,
FIELD DATA REQUIREMENTS
count date, and time period are entered on this form. A north
orientation arrow is also entered on this form.
As noted previously, computations require several types of data The major-street vehicle flows are denoted v1 through v6, minor-
as inputs to the methodology, including street vehicle flows are denoted v7 through v12, and pedestrian flows
1. Vehicle volumes by movement for the period of interest; are denoted v13 through v16. The two upstream signals are denoted
2. Vehicle classification for the period of interest; S2 and S5, corresponding to the through vehicle streams that arrive
3. Peak-hour factor (if peak flow rates are being used as the at the subject intersection from the direction of these two signalized
basis for analysis); intersections. Median storage may be present in the center of the
4. Number and use of lanes on the major-street approaches; intersection; flared approaches may also be present on the two
5. Number and use of lanes on the minor-street approaches; minor-stream approaches.
6. Grade of all approaches; and
7. Other geometric features of interest, such as channelization,
Volume Adjustment
two-way left-turn lane, raised or striped median, and flared
approach.
Vehicle and pedestrian flows are entered in Worksheet A2.
To determine the effects of pedestrians, two-way left-turn lanes, Measured or forecast vehicle volumes (in vehicles per hour) are
raised or striped medians, upstream signals, and flared approaches entered on Line 1 of the worksheet. The flows should be carefully
on capacity and level of service, the following data are also entered because subsequent worksheets refer to these flows by
required: their movement number to ease computations.
1. Pedestrian volumes by movement for the period of interest; The peak-hour factor is entered on Line 2 for each movement.
2. Number of vehicles that can be stored in the median; The equivalent hourly flow rate is computed for each movement
3. Number of vehicles that can be stored in the flared minor- by dividing the volume by the peak-hour factor and entering the
street approaches; result on Line 3.
4. Distance to upstream signals on the major street; and The proportion of heavy vehicles (PHV) for the minor movement
5. Other data relating to upstream signals that are within 0.25 is entered on Line 4.
mi of the subject intersection: cycle length, green time, saturation Pedestrian flows are entered on Line 5 for each movement. The
flow rate, and platooned flow. average lane width for each approach is entered on Line 6. The
average walking speed for pedestrians in the area is entered on
Line 7; a suggested default value is 4.0 ft/sec. The percent blockage
SEQUENCE OF COMPUTATIONS FOR CAPACITY for each approach is computed using Equation 10-11 and is entered
on Line 8.
Since the methodology is based on prioritized use of gaps by
vehicles at a TWSC intersection, it is important that computations
be made in a precise order. The computational sequence is the Site Characteristics
same as the priority of gap use, and movements are considered in
the following order: Lane designation data for each approach are entered into the
first section of Worksheet A3. For example, if Movements 1, 2,
1. Right turns from the minor street; and 3 all share a single lane and there are no other lanes on the
2. Left turns from the major street; approach, the string ‘‘1,2,3’’ would be entered in the ‘‘Lane 1’’
3. Through movements from the minor street; and column on Line 1. The grade for each approach is also entered
4. Left turns from the minor street. here. If the right turn is channelized, this is noted.
To assist in maintaining the proper order of computations, work- If the minor-street approaches have flared lanes that can store
sheets are provided. The use of each of these worksheets in the right-turning vehicles, this should be noted on Lines 5 and 6. The
computations is described in the following sections. number of storage spaces should also be entered.
If there is storage in the median, this should be noted on Lines
7 and 8. The number of storage spaces should also be entered.
Capacity Adjustments If there is an upstream signalized intersection within 0.25 mi of
the intersection on the major street, this should be noted on Lines
The standard gap acceptance capacity model makes several key 9 and 10. The effective green time (in seconds), the cycle length
assumptions regarding the geometry and flow characteristics of (in seconds), the progression speed, and the distance to the signal
the intersection, including a negative exponential distribution of (in feet) should be entered on Lines 9 and 10. The saturation flow
headways for the major-street traffic flow (‘‘random arrivals’’), rate (in vehicles per hour of green) and the through flow rate
single-stage gap acceptance process, and no impedance from pe- arriving at the upstream signalized intersection on the major-street
10-30 urban streets
approach and the protected left-turn movement from the side street sion parameters are entered on Lines 1 through 5. The maximum
should also be entered on Lines 9 and 10. flow during the platooned period is computed using Equation 10-
If the analyst needs to compute the delay to major-street vehicles 21 and entered on Line 6. The minimum flow during the platooned
resulting from sharing a lane with major-street left-turning vehi- period (default = 1,000N) is entered on Line 7. Note that tpi = 0 if
cles, relevant data are entered on Lines 11 through 15. vc,max is less than 1,000N. The duration of the blocked period, tp,
The length of the study period is entered on Line 16. is computed using Equation 10-22 and is entered on Line 8. The
proportion of time blocked, p, considering both the through move-
ment and the protected left-turn movement, is computed using
Critical Gap and Follow-up Time Equation 10-23 and is entered on Line 9.
Worksheet A5c is used to determine the platoon event periods
Table 10-1 and Equations 10-2 and 10-3 are used to compute and the proportion of time that is unblocked for each minor-stream
the critical gap and follow-up time. movement. The proportion of time blocked as a result of each
The base values of the critical gap are entered on Line 1 of platoon, p2 and p5, is entered on Lines 1 and 2, respectively. The
Worksheet A4 for each movement. The adjustment factors for dominant and subordinate platoons are computed using Equations
heavy vehicles (1.0 sec for two-lane major streets and 2.0 sec for 10-24 and 10-25, and are entered on Lines 3 and 4, respectively.
four-lane major streets) are entered on Line 2. The proportion of Equations 10-26 and 10-27 are used to determine if the condition
heavy vehicles for each movement (from Line 4 of Worksheet A2) is unconstrained (i.e., there is some time during which no platoons
is entered on Line 3. The adjustment factor for approach grade are present) or constrained (i.e., one or more platoons are always
(0.1 for Movements 9 and 12 and 0.2 for Movements 7, 8, 10, present). The proportion of time that is unblocked is determined
and 11) is entered on Line 4. The grade in percent/100 (from Lines for each minor movement, px, using Table 10-6 and is entered on
1 through 4, Worksheet A3) is entered on Line 5. The adjustment Lines 6 through 13 in Column 1. Note that if the subject intersec-
factor for minor-street left turns at T-intersections is entered on tion includes a two-stage gap acceptance process, the computations
Line 6. for Stages I and II are entered separately on Lines 6 through 13
For single-stage gap acceptance computations, tc,T is zero. For in Columns 2 and 3.
two-stage gap acceptance computations, tc,T is 1 sec. This is entered Worksheet A5d is used to compute the conflicting flows during
on Line 7. the unblocked period for each minor movement. The first panel
The critical gap for each movement, entered on Lines 8 or 9, in Worksheet A5d is used for a single-stage gap acceptance pro-
is computed from cess; the second panel is used for a two-stage gap acceptance
process. The conflicting flow, vc,x, is entered on Line 1. The total
Line 1 + (Line 2 × Line 3) + (Line 4 × Line 5) − Line 6
saturation flow s is entered on Line 2. The proportion of time that
− Line 7
is unblocked is determined from Lines 6 through 13 of Worksheet
The base value of the follow-up time is entered on Line 10 of A5c for each minor movement, px, and entered on Line 3. The
Worksheet A4 for each movement. The heavy-vehicle adjustment conflicting flow for Movement x during the unblocked period,
factor (0.9 sec for two-lane major streets and 1.0 sec for four-lane vc,u,x, is computed using Equation 10-28 and is entered on Line 4.
major streets) is entered on Line 11. The proportion of heavy Worksheet A5e is used to determine the capacity for the subject
vehicles for each movement is entered on Line 12. The follow-up movement during the unblocked period. The first panel in Work-
time, entered on Line 13, is computed as the sum of Line 10 and sheet A5e is used for a single-stage gap acceptance process; the
the product of Lines 11 and 12: second panel is used for a two-stage gap acceptance process. The
proportion of time that is unblocked is determined from Lines 6
(Line 10) + (Line 11 × Line 12) through 13 of Worksheet A5c for each minor movement, px, and
is entered on Line 1. The capacity for Movement x during the
unplatooned period (assuming random flow), cr,x, is computed us-
Effect of Upstream Signals ing Equation 10-1 and is entered on Line 2. The capacity of the
subject Movement x accounting for the effect of platooning is
Worksheets A5a through 5e are used to compute the flow pat- computed using Equation 10-29 and is entered on Line 3.
terns that result from upstream signalized intersections that are The capacity computed here is the potential capacity for the
within 0.25 mi of the subject TWSC intersection. movement. It is used in Worksheet A6 (and Worksheet A7, if two-
Worksheet A5a is used to determine the length of time required stage gap acceptance is present).
for the queue to clear from the upstream signalized intersection for
both the through movement and the protected left-turn movement.
Signalized intersection parameters (progressed volume, total satu- Computations of Impedance and Movement Capacities
ration flow rate, arrival type, effective green time, cycle length)
are entered on Lines 1 through 5. The platoon ratio, Rp, is deter- The capacity for each movement is computed using Worksheet
mined from Table 9-2 and entered on Line 6. The proportion of A6. Volumes are keyed to the diagram in Worksheet A1.
vehicles arriving on the green, P, is computed using Equation 10- Computations proceed in the prescribed order, considering first
17 and entered on Line 7. The length of time for the queue to clear the right turns from the minor street, followed by left turns from
is computed using Equations 10-18 through 10-20 and entered on the major street, through movements from the minor street, and
Lines 8, 9, and 10. left turns from the minor street. The user should solve parts of
Worksheet A5b is used to determine the proportion of the time movements before proceeding to the next step; that is, both right
that the subject TWSC intersection is blocked by the passing pla- turns in Step 1 should be computed before the user proceeds to
toon from the upstream signalized intersection. The platoon disper- Step 2.
For each movement, the following sequence of computations or Equation 10-33 is used to compute the movement capacity for
is used: the two-stage process. The result is entered on Line 20.
Note that if upstream signals are present, the results from Work-
1. Compute conflicting flows, vc,i, in vehicles per hour. Figure sheet A5e are used as the potential capacities for Steps 3 and 4.
10-2 may be consulted if a further explanation of this computation
is desired.
2. Use Equation 10-1 to find the potential capacity, cp,i, in vehi- Shared-Lane Capacity
cles per hour if there is no upstream signal. Use Worksheet A5e
to find the potential capacity if there is an upstream signal.
Worksheet A8 is used to compute the shared-lane capacity.
3. Compute the pedestrian impedance factor using Equation
The equation for this computation (Equation 10-15) is given on
10-12.
Worksheet A8.
4. Compute the probability of a queue-free state for the major-
street left-turn movements and minor-street through and right-turn
movements using Equation 10-5. Effect of Flared Minor-Street Approaches
5. Compute the capacity adjustment factor due to impeding
movements, fi, for the minor-street through and left-turn move-
Worksheet A9 is used to compute the effect of flared minor-
ments using Equation 10-9 or 10-14 and Figure 10-5.
street approaches. Although three columns are provided on the
6. The movement capacity, cm,i, is computed as the product of
worksheet (for either Movements 7, 8, and 9 or Movements 10,
the potential capacity, cp,i, and the capacity adjustment factor due
11, and 12), only those movements that share the right lane on the
to impeding movements, fi.
subject approach are included in the computation.
Note that if the problem involves a two-stage gap acceptance pro- The movement capacity, or the capacity for the separate-lane
cess, Worksheets A7a and A7b must be used for Steps 3 and 4 in case, csep, is entered on Line 1 for each of the movements that
place of Steps 3 and 4 from Worksheet A6. occupy the right lane. The flow rate and delay for each of these
movements are entered on Lines 2 and 3, respectively. The average
queue length for each movement is computed using Equation 10-
Two-Stage Gap Acceptance 34 and entered on Line 4. The value on Line 4 is increased by 1
and entered on Line 5. The value on Line 5 is rounded to the
Worksheets A7a and 7b are used to compute the capacity for a nearest integer and entered on Line 6. The value of nmax is computed
two-stage gap acceptance process for the through and left-turn using Equation 10-35 and entered on Line 7. The shared-lane
movements from the minor street. The procedure requires that the capacity for the movements that share the right lane is computed
analyst compute parameters both for each stage of the two-stage using Worksheet A8 and entered on Line 8. The sum of the separate
gap acceptance case and for the single-stage case (as if no median lane capacities (Line 1) is computed and entered on Line 9. The
were present). The results from both cases are combined to yield number of storage spaces in the flare is entered on Line 10. The
the increase in capacity that results from the two-stage gap accept- actual capacity of the flared approach is computed using Equation
ance process. Steps 3 and 4 from Worksheets A7a and A7b are 10-36 and entered on Line 11.
used in place of Steps 3 and 4 from Worksheet A6.
For the minor through movement (Step 3), Worksheet A7a is
used. The analyst computes the conflicting flows, the potential Average Control Delay, Queue Length, and Level of Service
capacity, the pedestrian impedance factor, the capacity adjustment
factor, the movement capacity, and the probability of a queue-free Worksheet A10 is used to compute average control delay, aver-
state and enters the results on Lines 1 through 6 for the first stage age queue length, and level of service. The flow rate and the
of the gap acceptance process and on Lines 7 through 11 for the capacity for each minor movement are entered on Lines 1 and 2.
second stage. The same computations are completed for the single- The v/c ratio is entered on Line 3. The queue length is computed
stage process (as if there were no median storage) and the results from Figure 10-12 and entered on Line 4. The control delay for
are entered on Lines 12 through 16. Parameters a and y are com- the movement is computed using Equation 10-37 and entered on
puted using Equations 10-30 and 10-31 and are entered on Lines Line 5. The level of service is determined from Table 10-7 and
17 and 18. Either Equation 10-32 or 10-33 is used to compute the entered on Line 6. The approach delay and level of service are
movement capacity for the two-stage process. The result is entered entered on Lines 7 and 8.
on Line 19. Worksheet A11 is used to compute the delay to Rank 1 vehicles.
For the minor left-turn movement (Step 4), Worksheet A7b is The probability of a queue-free state for Movement j is computed
used. The analyst computes the conflicting flows, the potential using Equation 10-5 and entered on Line 1. The major-street
capacity, the pedestrian impedance factor, the capacity adjustment through and right-turn volumes in the lane shared with the left-
factor, and the movement capacity and enters the results on Lines turn movement are entered on Lines 2 and 3, respectively. The
1 through 5 for the first stage of the gap acceptance process. For saturation flow rates for the major-street through and right-turn
the second stage, the analyst computes the conflicting flows, the movements are entered on Lines 4 and 5. The probability of a
potential capacity, the pedestrian impedance factor, the capacity queue-free state in the major-street shared lane is computed using
adjustment factor, and the movement capacity and enters the results Equation 10-16 and entered on Line 6. The computed delay for
on Lines 6 through 10. These same parameters are computed for the major-street left-turn movement is entered on Line 7. The
the single-stage process; the results are entered on Lines 11 through number of major-street through lanes is entered on Line 8. The
17. Parameters a and y are computed using Equations 10-30 and delay to Rank 1 vehicles is computed using Equation 10-38 and
10-31 and are entered on Lines 18 and 19. Either Equation 10-32 entered on Line 9.
10-32 urban streets
PLANNING AND DESIGN APPLICATIONS proach, median storage, and upstream signals are not included in
a planning analysis.
The operational analysis method described earlier in this chapter T Worksheet A4 is generally not used, since the base values
provides a detailed procedure to evaluate the performance of a from Table 10-1 are used without adjustment.
TWSC intersection. Sometimes, however, an analyst may wish to T Worksheets A5a–A5e are not used, since the effect of up-
estimate the level of service for a long-term time horizon. This stream signals is generally not included in a planning analysis.
kind of analysis is called a planning-level analysis. It is expected T Worksheet A6 is used to compute the movement capacities.
that for such an analysis only a limited amount of input data is T Worksheets A7a and A7b may be used to include the effects
available. of two-stage gap acceptance when the effects of a divided roadway
The planning analysis method described here is based on the or the addition of two-way left-turn lanes are to be considered on
operational analysis method and requires geometric data and traffic the major street.
flow data. The base values of critical gap and follow-up time from T Worksheet A8 is used to compute shared-lane capacities if
Table 10-1 are used. The effects of upstream signals, two-stage more than one movement shares the same minor-street approach.
gap acceptance, and flared right-turn approaches are normally not T Worksheet A9 is not used, since the effect of flared minor-
accounted for in a planning analysis. However, if these data are
street approaches is generally not included in a planning analysis.
available and the project under study includes these elements, their
T Worksheet A10 is used to compute control delay, queue
effects can be included.
length, and level of service.
The planning analysis uses the same worksheets as does the
T Worksheet A11 is not used, since the impedance and delay for
operational analysis, with some exceptions:
the major through movements are not accounted for in a planning
T Worksheet A1 is used to describe basic conditions. analysis.
T Worksheet A2 is used to summarize vehicle volumes. Pedes-
trian volumes are not generally used in a planning analysis. If The operational analysis procedure described earlier in this chapter
estimates of the pedestrian volumes are available, these data can is not normally used for design purposes. However, through itera-
also be used in Worksheet A2. tion, the analyst can use a given set of traffic flow data and deter-
T Worksheet A3 is used to note the lane designation for each mine the number of lanes that would be required to produce a
movement. Generally, the corrections for flared minor-street ap- given level of service.

WORKSHEET A1: BASIC INTERSECTION INFORMATION

10-34 urban streets
WORKSHEET A2: VOLUME ADJUSTMENT

WORKSHEET A3: SITE CHARACTERISTICS

10-36 urban streets
WORKSHEET A4: CRITICAL GAP AND FOLLOW-UP TIME CALCULATION

WORKSHEET A5a: EFFECT OF UPSTREAM SIGNALS (COMPUTATION 1)
WORKSHEET A5b: EFFECT OF UPSTREAM SIGNALS (COMPUTATION 2)

10-38 urban streets
WORKSHEET A5c: EFFECT OF UPSTREAM SIGNALS (COMPUTATION 3)

WORKSHEET A5d: EFFECT OF UPSTREAM SIGNALS (COMPUTATION 4)

10-40 urban streets
WORKSHEET A5e: EFFECT OF UPSTREAM SIGNALS (COMPUTATION 5)

WORKSHEET A6: IMPEDANCE AND CAPACITY CALCULATIONS

10-42 urban streets
WORKSHEET A7a: COMPUTATION OF EFFECT OF TWO-STAGE GAP ACCEPTANCE (STEP 3)

WORKSHEET A7b: COMPUTATION OF EFFECT OF TWO-STAGE GAP ACCEPTANCE (STEP 4)

10-44 urban streets
WORKSHEET A8: SHARED-LANE CALCULATIONS

WORKSHEET A9: COMPUTATION OF EFFECT OF FLARED MINOR-STREET APPROACHES

10-46 urban streets
WORKSHEET A10: DELAY, QUEUE LENGTH, AND LEVEL OF SERVICE
WORKSHEET A11: SHARED MAJOR LT IMPEDANCE AND DELAY
Six sample calculations that illustrate the application of the SAMPLE CALCULATION A1
TWSC methodology are given in this section. The first example
is a three-leg intersection with stop control on the minor leg. It Sample Calculation A1 illustrates the TWSC capacity analysis
illustrates the basic steps involved in applying the methodology. methodology for a three-leg (T-) intersection.
The second example extends this analysis to include an examina-
tion of the effects of combining the westbound through and left-
turn lanes. The third example focuses on a four-leg intersection Description
with multiple-lane approaches on the major street. This situation
is further examined in Sample Calculations A4 through A6, in Sample Calculation A1 concerns the intersection of Market and
which the effects of upstream signals, variations in geometric con- Jones streets in an urban area with a population of 100,000 (Figure
figuration, right-turn flaring on the minor approaches, and two- 10-16). Market Street is a two-lane collector with an exclusive
stage gap acceptance are explored. left-turn lane on the westbound approach to the intersection,
Figure 10-16. Traffic volumes for Sample Calculation A1.
whereas Jones Street is a two-lane, stop-controlled local street sec, respectively. Line 2 contains the adjustment factor for heavy
serving a residential development. There is no widening in the vehicles (1.0 sec), Line 3 gives the proportion of heavy vehicles
vicinity of the intersection, so adjustments for right-turn flaring for each movement (0.10 in this case), and Line 4 gives the adjust-
are not appropriate. Area residents have complained that there is ment factor for approach grade (0.1 and 0.2 for Movements 9 and
substantial total delay for drivers turning right onto Market Street 7, respectively). Lines 5 and 6 contain the percent grade and the
in the late afternoon. Residents claim that this is due to the need critical gap adjustment factor for minor-street left turns at T-inter-
for right- and left-turners to share a lane and have requested that sections, respectively. As mentioned before, the percent grade for
a right-turn-only lane be provided. The following discussion illus- this sample problem is 0, and because the intersection is a T, an
trates the capacity analysis of this example problem. adjustment factor of −0.7 is entered into the column pertaining to
Movement 7. Line 7 is left blank since there is no two-stage gap
acceptance. The critical gap for each movement, based on the first
Solution
equation at the bottom of Worksheet A4, is then calculated and
entered on Line 8. The calculation for Movement 7 is as follows:
Steps 1 and 2: Use Worksheets A1 and A2 to enter basic data
and traffic flow volumes. The diagram in Worksheet A1 contains tc = 7.1 + 1.0 ∗ 0.10 + 0.2 ∗ 0 − 0.7 = 6.5 (10-42)
the volumes and turning movements by approach. This information
for Sample Calculation A1 is presented in Figure 10-16. Line 1 Step 4b: Use Worksheet A4 to calculate follow-up time. The
on Worksheet A2 again requires the vehicle volumes by move- base follow-up time values for each movement are entered on
ment, and entries are made on Line 2 for PHF. In this example, Line 10. According to Table 10-1, they are 2.2, 3.3, and 3.5 for
PHF = 1.00, which means that the hourly flow rates on Line 3 are Movements 4, 9, and 7, respectively. The heavy-vehicle adjustment
identical to the volumes on Line 1. The proportion of heavy vehi- factor is entered on Line 11 (0.9 sec), and the proportion of heavy
cles is entered on Line 4, which in this case is 10 percent. Lines vehicles for each movement is entered on Line 12 (again, 0.10).
5–8 pertain to pedestrian volumes and adjustments; however, such The results of the follow-up time calculations, based on the second
data were not recorded in this sample calculation. equation given at the bottom of Worksheet A4, are entered on
Step 3: Use Worksheet A3 for site characteristics. Lines 1–4 of Line 13. The calculation for Movement 7 is as follows:
Worksheet A3 are used to describe the geometry of the intersec-
tion. In this example, since both the eastbound and northbound tf = 3.5 + 0.9 ∗ 0.10 = 3.6 (10-43)
approaches have only one lane each, Movements 2 and 3 and
Movements 7 and 9 appear in the column labeled ‘‘Lane 1’’ on Step 5: Use Worksheets A5a–A5e to determine effect of upstream
Lines 1 and 3, respectively. However, since the westbound ap- signals. Since no signals are present within 0.25 mi of the intersec-
proach has an exclusive left-turn lane, Movement 4 is entered tion, these worksheets are not used.
under ‘‘Lane 1’’ and Movement 5 is entered under ‘‘Lane 2’’ on Step 6: Use Worksheet A6 to calculate impedance and capacity.
Line 2. The grade is 0 percent, and there are no channelized right Worksheet A6 is divided into four different steps, each correspond-
turns. There are no flared minor-street approaches or median stor- ing to two movements. However, given the geometry of this exam-
age areas, so an N for No is entered on Lines 5 and 7 in the ‘‘Yes ple, only the calculations associated with Movements 9, 4, and 7
or No?’’ column. There are no upstream signals either, so the third in Steps 1, 2, and 4, respectively, pertain. The first line of Step 1
panel of the worksheet is left blank. The fourth panel is left empty refers to Figure 10-2 for the computation of conflicting flows; the
as well; however, 0.25 is entered on Line 16 for T. second line refers to Equation 10-1 or 10-29 for potential capacity;
Step 4a: Use Worksheet A4 to calculate critical gap. Since the the third line refers to Equation 10-12 for the pedestrian impedance
example intersection is a T, the only columns in Worksheet A4 factor, which in this case is 1.00; and the fifth line refers to Equa-
that contain entries are those that correspond to Movements 4, 7, tion 10-5 for the probability of a queue-free state. The fourth line
and 9. The base critical gap times, taken from Table 10-1, are is for the calculation of movement capacity, which in this step
entered on Line 1 of Worksheet A4. The values in this case are equals potential capacity since there are no conflicting pedestrian
those that pertain to a two-lane major road: 4.1, 6.2, and 7.1 flows. Equations 10-44 through 10-46 calculate the conflicting
10-48 urban streets
flow, potential capacity, and probability of a queue-free state, re- Table 10-9. Impedance and Capacity Calculations
spectively, for Movement 9, as follows: Step 1—Minor Right Turn v9 v12
vc,9 = 250 + 20 + 0 = 270 (10-44) Conflicting flows 270 —
Potential capacity 750 —
e−(270*6.3)/3,600 Pedestrian impedance factor 1.00 —
cp,9 = 270 ∗ = 750 (10-45) Movement capacity 750 —
1 − e−(270*3.4)/3,600
Probability of queue-free state 0.84 1.00
120 Step 2—Major Left Turn v4 v1
p0,9 = 1 − = 0.84 (10-46)
750 Conflicting flows 290 —
Potential capacity 1,227 —
In calculating the conflicting flows, the user must always be careful
Pedestrian impedance factor 1.00 —
to read the notes to Figure 10-2. In the case of Movement 9 above, Movement capacity 1,227 —
no adjustments to the formula referred to on Worksheet A6 are Probability of queue-free state 0.88 —
necessary.
Step 4—Minor Left Turn v7 v10
The first five lines in Step 2 are identical to those in Step 1.
The equation for conflicting flows is provided for convenience, Part III—Single stage (no storage)
and again no adjustments are necessary. The equations for potential Conflicting flows 870 —
capacity, movement capacity, and probability of a queue-free state Potential capacity 312 —
are the same as those referred to in Step 1. The sixth line of Step
Major left turn, minor through
2 is left blank since the example intersection has a left-turn bay impedance factor 0.88 —
on the main street. The values corresponding to the first five lines Major left turn, minor through
under Movement 4 are 290 veh/hr, 1,227 veh/hr, 1.00, 1,227 veh/ adjustment impedance 0.91 —
hr, and 0.88, respectively. Capacity adjustment factor 0.91 —
As in Steps 1 and 2, the first line of Step 4 is used to compute Movement capacity 283 —
conflicting flows, and the second is used to calculate potential
capacity. The corresponding values for Movement 7 of this exam-
ple problem are 870 veh/hr and 312 veh/hr, respectively. No adjust-
ments are made to the equation for conflicting flow. Line 3 contains
the pedestrian impedance factor, which again is 1.00. The fourth 40 + 120
cSH = = 531 (10-51)
line of Step 4 is used to compute the major-left, minor-through (40/283) + (120/750)
impedance factor. The fifth line uses the number generated by
Line 3 to calculate the major-left, minor-through adjusted imped- Step 9: Use Worksheet A9 to compute effect of flared minor-
ance factor, the sixth line is used to compute the capacity adjust- street approaches. Since the northbound approach is not flared,
ment factor due to impeding movements, and the seventh line is Worksheet A9 is not used in this calculation.
used to determine movement capacity. Equations 10-47 through Step 10: Use Worksheet A10 to calculate delay, queue length,
10-50 demonstrate the calculations for Lines 4–6, respectively, and level of service. In this instance, only the top columns for
corresponding to Movement 7. The variables p0,1, p0,11, and Movement 4 and Movements 7 and 9 are relevant. Since Move-
p0,12 and are set to be 1.00, since in this instance Movements 1, ments 7 and 9 share one lane, the third column contains one value.
11, and 12 are not defined. The combined volume of Movements 7 and 9, 160 veh/hr, is
entered in the first row, third column. The approach capacity, 531
p″7 = 1.00 ∗ 0.88 = 0.88 (10-47) veh/hr, is entered in the second row, third column. The v/c ratio,
0.30, is simply the ratio of these two numbers. The queue length,
0.88 1.4 cars, is determined by referring to Figure 10-12. The control
p′7 = 0.65 ∗ 0.88 − + 0.6√0.88 = 0.91 (10-48)
0.88 + 3 delay is computed using Equation 10-37 and is demonstrated as
follows:
f7 = 0.91 ∗ 1.00 = 0.91 (10-49)
3,600
d=5+ + 900 ∗ 0.25
cm,7 = 0.91 ∗ 312 = 283 (10-50) 531
3 ! 4
The remaining values for Worksheet A6 are shown in Table 10-9.
1 531 2 ∗ 15312
3,600 160
Step 7: Use Worksheets A7a and A7b to compute effect of two- 2
1 2
160 160
stage gap acceptance. Since there is no two-stage gap acceptance ∗ −1+ −1 + = 14.7
531 531 450 ∗ 0.25
process, Worksheets A7a and A7b are not used.
Step 8: Use Worksheet A8 to calculate shared-lane capacity. (10-52)
Lines 1–3 of Worksheet A8 are used to calculate the shared-lane
capacity of Movements 7, 8, and 9. In this case, the last three According to Table 10-7, 14.7 sec of delay corresponds to LOS
columns of Worksheet A8 are left blank, since Movements 10, 11, B. The approach delay is the same as the control delay in this
and 12 are not defined. The volumes for Movements 7 and 9 and instance, since there is only one lane. The values for Movement
their capacities from Worksheet A6 are entered in the second 4 are presented in Table 10-10.
and fourth columns of the top row, respectively. The shared-lane Step 11. Use Worksheet A11 to calculate shared major left-turn
capacity is computed using the first equation given on Work- impedance and delay. The presence of the left-turn bay on the
sheet A8: major street precludes the use of Worksheet A11.
Table 10-10. Delay, Queue Length, and Steps 4 and 5: Worksheets A4 and A5a–5e. These worksheets
Level of Service are again identical to those created for Sample Calculation A1.
Movement Step 6: Use Worksheet A6 to calculate impedance and capacity.
The calculations for Steps 1 through 4 of Worksheet A6 are the
4 7, 9
same as those in Sample Calculation A1. However, because the
v (veh/hr) 150 160 major left and through lanes are now shared on the westbound
c (veh/hr) 1,227 531
v/c 0.12 0.30
approach, p*0,4, corresponding to Line 6 of Step 2, is now calcu-
Queue length <1 — lated as follows:
Control delay 8.3 14.7
1 − 0.88
Level of service A B p*0,4 = 1 − = 0.85 (10-53)
1 2
300
1−
1,700
SAMPLE CALCULATION A2 The effect of this change on the results in Step 4 can be seen in
Table 10-11. The variable f11, used in computing the major-left,
Sample Calculation A2 is identical to the problem described in minor-through impedance factor of Movement 7 in Step 4, is calcu-
Sample Calculation A1 with the exception that the exclusive left- lated as 0.85 = 0.85 * 1.00. Since Movement 1 is undefined, p0,1
turn lane on the westbound approach is combined with the through is set at 1.00.
lane. The purpose of this calculation is to demonstrate the use of Step 7: Use Worksheets A7a and A7b to compute effect of two-
p*0,4 as defined by Equation 10-16. stage gap acceptance. No computations are performed on Work-
sheets A7 in this sample calculation.
Description Step 8: Use Worksheet A8 to calculate shared-lane capacity.
Because Movement 7 now has a capacity of 276 veh/hr, the shared
Sample Calculation A2 increases the complexity of Sample Cal- capacity for the northbound approach is 525 veh/hr.
culation A1 by changing the major-street geometry to include a Step 9: Use Worksheet A9 to compute effect of flared minor-
shared left and through lane. This change can be seen in Figure street approaches. Again, there are no flared minor-street ap-
10-17. For the purposes of this example, it is assumed that field proaches, and therefore Worksheet A9 is not needed in this sample
studies have established the saturation flow rate for the major- calculation.
street westbound through traffic as 1,700 veh/hr. The following
discussion reveals the effects of the change in configuration.
Table 10-11. Impedance and Capacity Calculations

Solution
Step 4—Minor Through v7 v10
Steps 1 and 2: Use Worksheets A1 and A2 to enter basic data Part III—Single stage (no storage)
and traffic flow volumes. These steps are identical to those for Conflicting flows 870 —
Potential capacity 312 —
Sample Calculation A1.
Step 3: Use Worksheet A3 for site characteristics. The entries Major left turn, minor through
in Worksheet A3 are the same as those in Sample Calculation A1, impedance factor 0.85 —
except that Movement 5 on Line 2 now appears under the column Major left turn, minor through
headed ‘‘Lane 1,’’ since Movements 4 and 5 share one lane. In adjustment impedance 0.89 —
addition, 300 is entered under ‘‘Movement 5’’ on Line 11, 1,700 Capacity adjustment factor 0.89 —
is entered on Line 13, and 1 is entered on Line 15. Movement capacity 276 —

10-50 urban streets
Step 10. Use Worksheet A10 to calculate delay, queue length, Solution
and level of service. Since the northbound approach now has a
shared capacity of 525 veh/hr, the values in Worksheet A10 must Steps 1 and 2: Use Worksheets A1 and A2 to enter basic data
be recalculated. At the given levels of significance, however, the and traffic flow volumes. The volumes and turning movements by
only numbers that change are the control and approach delays, approach, as indicated in Figure 10-18, are entered on Worksheet
which increase slightly from 14.7 to 14.8 sec. A1. The proportion of heavy vehicles is 10 percent. The vehicle
Step 11: Use Worksheet A11 to calculate shared major left-turn volumes are entered again on Line 1 of Worksheet A2, and a PHF
impedance and delay. Because the major-street left-turn bay is of 1.00 is entered on Line 2. The hourly flow rates on Line 3 are
now gone, Worksheet A11 must be included. Of the two Rank 1 the same as the volumes contained on Line 1. The proportion of
movements present, only Movement 5 is affected, since Movement heavy vehicles entered on Line 4 is 10 percent. Lines 5 and 8
2 does not share a lane with left-turning vehicles. The probability contain zeros, and Lines 6 and 7 are left blank, since no pedestrian
of a queue-free state, p0,j, is entered on Line 1. In this case, as can data exist. Conflicting flows for Movements 1, 4, and 7–12 are
be seen from Worksheet A6, p0,4 is 0.88. Lines 2 and 3 contain computed (see Table 10-12).
the volumes for Streams 5 and 6, respectively, which again in this Step 3: Use Worksheet A3 for site characteristics. The entries
instance were 300 and 0. The fourth and fifth lines give the satura- in the top panel of Worksheet A3 are more complex than those
tion flow rates for Streams 5 and 6, respectively. As mentioned associated with the three-leg case. For this reason, these entries
earlier, the saturation flow rate for Stream 5 was determined to be are presented in Table 10-13. The percent grade for this example
1,700 veh/hr, and the saturation flow rate for Stream 6 is undefined. is again 0, and there is no right-turn channelization. There are no
Line 6 contains the variable p*0,4, which, according to Worksheet flared minor-street approaches or median storage areas, so N for
A6, is 0.85. The seventh line of Worksheet A11 gives the delay No is entered on Lines 5–8 in the ‘‘Yes or No?’’ columns. The
for Stream 4, which can be found after completing Worksheet third panel is left blank, since there are no signals within 0.25 mi
A10. According to the results for Sample Calculation A1, this of the intersection. There are two major through lanes from both
delay is 8.3 sec. The eighth line of Worksheet A11 gives N, the the right and left, or four major through lanes.
number of major-street through lanes, which in this case is 1. Line Step 4a: Use Worksheet A4 to calculate critical gap. Unlike in
9 uses the information presented on Lines 1–8 to compute dRank1, Sample Calculations A1 and A2, all of the columns in Worksheet
or the delay for Stream 5, using Equation 10-38. This calculation A4 should contain entries. The base critical gap times entered on
is as follows: Line 1 are those that pertain to a four-lane major road, which may
be found in Table 10-1. The times are 4.1, 6.9, 6.5, and 7.5 sec,
dRank1 = (1 − 0.85) ∗ 8.3 = 1.2 (10-54) corresponding to the major-left, minor-right, minor-through, and
minor-left movements, respectively. The adjustment factor for
SAMPLE CALCULATION A3 heavy vehicles entered on Line 2 is 2.0 sec, and the proportion of
heavy vehicles on Line 3 is 0.10. Line 4 is the same as in the
Sample Calculation A3 illustrates the TWSC capacity analysis previous two sample calculations, and Lines 5–7 contain zeros.
methodology for a four-leg intersection The critical gap for each movement is calculated in the same
manner as before and the adjusted values are entered on Line 8.
Final results for this sample calculation are presented in Table
Description 10-14.
This example focuses on the intersection of Walnut and Elm

streets (Figure 10-18). Walnut is a four-lane arterial with protected Table 10-12. Conflicting Flows
left-turn lanes on both intersection approaches. Elm is a two-lane, Movement Flow (veh/hr)
stop-controlled collector. Recently, the northbound approach on 1 400
Elm was widened to add a left-turn lane, yet local residents con- 2 —
tinue to complain that total delays are excessive. The following 3 —
discussion steps through the capacity analysis of this facility, illus- 4 300
trating important points in the methodology. 5 —
6 —
7 678
8 873
9 150
10 739
11 848
12 200
Table 10-13. Lane Usage by Approach

Approach Lane Designation
Movement Lane 1 Lane 2 Lane 3
1, 2, 3 1 2 2, 3
4, 5, 6 4 5 5, 6
7, 8, 9 7, 8, 9
Figure 10-18. Traffic volumes for Sample Calculation A3. 10, 11, 12 10, 11, 12

Table 10-14. Critical Gap and Follow-Up Time by Movement

Movement
Major Left Turn Minor Right Turn Minor Through Minor Left Turn
Parameter 1 4 9 12 8 11 7 10
tc 4.3 4.3 7.1 7.1 6.7 6.7 7.7 7.7
tf 2.3 2.3 3.4 3.4 4.1 4.1 3.6 3.6
Step 4b: Use Worksheet A4 to calculate follow-up time. The Table 10-15. Impedance and Capacity Calculations
base follow-up times for each movement entered on Line 10, ac- Step 1—Minor Right Turn v9 v12
cording to Table 10-1, are 2.2, 3.3, 4.0, and 3.5 sec for the major-
left, minor-right, minor-through, and minor-left movements, re- Conflicting flows 150 200
Potential capacity 845 783
spectively. The heavy-vehicle adjustment factor on Line 11 is 1.0
Pedestrian impedance factor 1.00 1.00
sec, and, as mentioned earlier, the proportion of heavy vehicles Movement capacity 845 783
for each movement is 10 percent. The results of the follow-up Probability of queue-free state 0.93 0.96
time calculations are displayed along with those of the critical gap
Step 2 v4 v1
calculations in Table 10-14.
Step 5: Use Worksheets A5a–A5e to determine effect of upstream Conflicting flows 300 400
signals. Worksheets A5a–A5e are not used in the analysis since Potential capacity 1,202 1,100
there are no signals within 0.25 mi of the intersection. Pedestrian impedance factor 1.00 1.00
Movement capacity 1,202 1,100
Step 6: Use Worksheet A6 to calculate impedance and capacity.
Because Sample Calculation A3 is based on a four-leg intersection,
all of the steps in Worksheet A6 are completed. The calculations Step 3 v8 v11
for Steps 1 and 2 are identical to those shown in Sample Calcula- Part III—Single stage (no storage)
tion A1. Note that, as indicated in the notes to Figure 10-2, since Conflicting flows 873 848
the major street of the sample intersection is multilane, the factors Potential capacity 273 283
0.5n6 and 0.5n12 in the equation for Movement 7 and 0.5n3 and Pedestrian impedance factor 1.00 1.00
0.5n9 in the equation for Movement 10 are omitted. Capacity adjustment factor 0.92 0.92
Movement capacity 250 259
Lines 4 and 5 of Step 3 use Equations 10-6 and 10-7 to compute
the capacity adjustment factor due to impeding movements and
movement capacity, respectively. Equations 10-55 and 10-56 dem- Step 4 v7 v10
onstrate the calculation of the capacity adjustment factor and the Part III—Single stage (no storage)
movement capacity, respectively, for Movement 8: Conflicting flows 678 739
f8 = 0.95 ∗ 0.97 = 0.92 (10-55) Pedestrian impedance factor 1.00 1.00
cm,8 = 273 ∗ 0.92 = 250 (10-56) Major left turn, minor through
0.53 0.43
impedance factor
Line 6 of Step 3, the probability of a queue-free state, employs Major left turn, minor through
0.63 0.55
the same equation as in Steps 1 and 2. The results of the calcula- adjustment impedance
Capacity adjustment factor 0.61 0.51
tions for Worksheet A6 are presented in Table 10-15. Movement capacity 196 150
Step 7: Use Worksheets A7a and A7b to compute effect of two-
stage gap acceptance. Since the intersection does not include a
raised or striped median or a two-way left-turn lane (TWLTL)
on the major street, Worksheets A7 are not used in the capacity
analysis.
Step 8: Use Worksheet A8 to calculate shared-lane capacity. Step 9: Use Worksheet A9 to compute effect of flared minor-
Because this is a four-leg intersection, data are entered on Work- street approaches. Since neither the northbound nor southbound
sheet A8 for both minor-street approaches. The volumes for Move- approach is flared, Worksheet A9 is not used.
ments 7, 8, and 9, along with their capacities from Worksheet A6, Step 10: Use Worksheet A10 to calculate delay, queue length,
are entered in Rows 1 and 2 and the volumes and capacities for and level of service. As in Worksheet A8, since this is a four-leg
Movements 10, 11, and 12 are also entered in Rows 1 and 2. The intersection, all of the columns on Worksheet A10 contain entries.
shared capacity for each approach is computed using the second Since there is only one lane on both the northbound and south-
equation given on Worksheet A8. The capacity calculation for the bound approaches, the calculations in this instance are the same
northbound approach is as follows: as those for Sample Calculation A1. However, it should be noted
44 + 132 + 55 that the results for volume and capacity include the values for
cSH = = 283 (10-57) Movement 8 as well as for Movements 7 and 9. Table 10-16
(44/196) + (132/250) + (55/845)
displays the outcome for this example.
The corresponding capacity for the southbound approach is 279 Step 11: Use Worksheet A11 to calculate shared major left-
veh/hr. turn impedance and delay. Worksheet A11 is not included in this
10-52 urban streets
Table 10-16. Delay, Queue Length, and Level P = 0.333(30/80) = 0.125 (10-58)
of Service
250 ∗ 80(1 − 0.125)
Movement gq1 = = 4.9 (10-59)
3,600
1 4 7, 8, 9 10, 11, 12
v (veh/hr) 33 66 231 149
250 ∗ 80 ∗ 0.125 ∗ 4.9
gq2 = = 0.1 (10-60)
cm (veh/hr) 1,100 1,202 283 279 (3,600 ∗ 30) − (250 ∗ 80 ∗ 0.125)
v/c 0.03 0.06 0.82 0.53
Queue — — — — gq = 4.9 + 0.1 = 5.0 (10-61)
Control delay 8.4 8.2 56.5 31.8
The remaining results are shown in Table 10-18.
Level of service A A F D
Lines 1 through 5 of Worksheet A5b, Computation 2, are used to
list the dispersion parameters, as detailed in the section Upstream
Signals under Structure of the Methodology (Part II). The first
parameter, a, is obtained from Table 10-4. Equations 10-62
analysis since both the eastbound and westbound approaches pro- through 10-65 present the calculations for Lines 2–5, again for the
vide exclusive left-turn lanes. through component of Movement 2.
b = (1 + 0.500)−1 = 0.667 (10-62)
SAMPLE CALCULATION A4
ta = 450/(35 ∗ 1.47) = 8.746 (10-63)
Description F = (1 + 0.500 ∗ 0.667 ∗ 8.746)−1 = 0.255 (10-64)
f = 250/(33 + 250 + 50) = 0.751 (10-65)
The intersection and volumes used in this sample calculation
are identical to those used in Sample Calculation A3. However, Lines 6 and 7 contain the maximum and minimum flows, respec-
this example also considers the effects of upstream signals. tively, during the platooned period, Line 8 gives the duration of
the blocked period, and Line 9 contains the proportion of time
blocked. The minimum flows are set to 1,000 veh/hr. The follow-
Solution
ing three equations illustrate the calculations for Lines 6, 8, and
9, respectively, for the through component of Movement 2:
Steps 1 and 2: Use Worksheets A1 and A2 to enter basic data
and traffic flow volumes. The first two worksheets are identical to vc,max = 3,600 ∗ 0.751 ∗ [1 − (1 − 0.255)5.0] = 2,080 (10-66)
those produced for Sample Calculation A3.
3 4
2,080 − 250 ∗ 0.125 ∗ 0.751
31 2 12,000 − 250 ∗ 0.125 ∗ 0.75124
2,000
Step 3: Use Worksheet A3 for site characteristics. The first ln 1 −
2,704
three panels of Worksheet A3 are the same as those in Sample tp = 5.0 − = 0.5 (10-67)
ln(1 − 0.255)
Calculation A3. Data pertaining to the upstream signals are entered
in the third panel. The parameters used in this instance are shown
in Table 10-17.
Step 4: Use Worksheet A4 to calculate critical gap. The presence Table 10-18. Computation 1: Queue Clearance
of upstream signals does not affect the Worksheet A4 calculations. Time (A5a)
Step 5: Use Worksheets A5a–A5e to determine effect of upstream Movement 2, Movement 5,
signals. Since upstream signalized intersections are present within vT vT
0.25 mi of the subject intersection, Worksheets A5a–A5e (Compu- 1 (4) g 30 20
tations 1–5) are included in this analysis. 2 (5) c 80 70
The first five lines of Worksheet A5a, Computation 1, contain 3 (2) s 3,600 3,600
the signalized intersection parameters already specified in Work- 4 (3) AT 1 1
sheet A3. The values given in Table 9-2 are used to determine the 5 (1) vprog 250 250
6 Rp 0.333 0.333
various platoon ratios, which are entered on Line 6. The proportion
7 P 0.125 0.095
of vehicles arriving on the green is entered on Line 7, and the 8 gq1 4.9 4.4
length of time for the queues to clear is entered on Lines 8– 9 gq2 0.1 0.1
10. Equations 10-58, 10-59, 10-60, and 10-61 demonstrate the 10 gq 5.0 4.5
calculations for Lines 7, 8, 9, and 10, respectively, for the through
component of Movement 2:
Table 10-17. Upstream Signal Parameters

Saturation
Signal Mvmt vprog Flow AT Green Time Cycle Speed Distance
9 S2 1,2,3 left turn
Through 250 1,800 1 30 80 35 450
10 S5 4,5,6 left turn
Through 250 1,800 1 20 70 30 650

0.5 + 0 400 − 1,800(1 − 1.000)

p= = 0.007 (10-68) vc,u,x = = 400 (10-69)
80 1.000
The remaining results for Computation 2 are presented in Table Table 10-21displays the full results for Computation 4.
10-19. The first line of Computation 5 again contains the proportion
The first two lines of Worksheet A5c, Computation 3, give the of time that each minor movement is unblocked, from Lines 6–
proportion of time blocked as a result of each platoon, p2 and p5, 13 of Computation 3. The second line contains the capacity for
and are taken from the last line of Computation 2. Lines 3 and 4 Movement x during the unplatooned period, and the third line
contain the dominant and subordinate platoons, and Line 5 presents gives the capacity of the subject Movement x, which accounts for
the condition (constrained or unconstrained). According to Equa- the effect of platooning. The calculations for Lines 2 and 3 for
tions 10-24 and 10-25, pdom and psubo are the maximum and mini- Movement 1 are demonstrated in Equations 10-70 and 10-71,
mum of p2 and p5, respectively. Since 0.007 + (0.000/2) ≤ 1, the respectively:
condition is unconstrained. Table 10-6 is used to determine the
e−400*4.3/3,600
proportion of time that is unblocked for each minor movement. cr,x = 400 ∗ = 1,100 (10-70)
The result is entered on Lines 6 through 13 in Column 1. The 1 − e−400*2.3/3,600
second and third columns on Lines 6–13 are blank, since two- cplat,x = 1.000 ∗ 1,100 = 1,100 (10-71)
stage gap acceptance is not present in this example. The full results The complete set of results for Computation 5 can be seen in Table
of Computation 3 are shown in Table 10-20. 10-22.
The first four rows on Worksheet A5d and the first three on Step 6: Use Worksheet A6 to calculate impedance and capacity.
Worksheet A5e, the rows pertaining to single-stage gap accept- Since upstream signals are present, the potential capacities found in
ance, are used for Computations 4 and 5. Line 1 of Computation Worksheets A5 are entered in Worksheet A6. The effects these re-
4 gives the total conflicting flows, and Line 2 contains the satura- sults have on the remaining computations can be seen in Table 10-23.
tion flow, given in Worksheet A3. Lines 6 through 13 of Computa- Step 7. Use Worksheets A7a and A7b to compute effect of two-
tion 3 are entered on Line 3, and the conflicting flow for Movement stage gap acceptance. Again, Worksheets A7 remain unused.
x during the unblocked period is entered on Line 4. The following Step 8: Use Worksheet A8 to calculate shared-lane capacity.
equation demonstrates the calculation of vc,u,x for Movement 1: Worksheet A8 is completed in the same manner as in Sample
Calculation A3 on the basis of the new capacities recorded on
Worksheet A6. In this instance, the shared-lane capacities for the
northbound and southbound approaches are 290 veh/hr and 285
Table 10-19. Computation 2: Proportion of TWSC
Intersection Time Blocked (A5b) veh/hr, respectively.
Steps 9 and 10: Use Worksheets 9 and 11 to calculate effect of
Movement 2, Movement 5, flared minor-street approaches and shared major left-turn imped-
vT vT ance and delay. As in the previous example, neither of these work-
1 a 0.500 0.500 sheets is used in the current analysis.
2 b 0.667 0.667 Step 11: Use Worksheet A10 for delay, queue length, and level
3 ta 8.746 14.739 of service. The calculations used in Worksheet A10 are the same
4 F 0.255 0.169
5 f 0.751 0.536 as those used in Sample Calculation A3. The final results are
6 vc,max 2080 1093 presented in Table 10-24.
7 vc,min 2000 2000
8 tp 0.5 0.0
9 p 0.007 0.000 SAMPLE CALCULATION A5
Sample Calculation A5 is identical to Sample Calculation A3

except that the minor-street lane configuration has changed.
Table 10-20. Computation 3: Platoon Events and
Proportion Unblocked (A5c)
Platoon Event Periods Result Description
1 p2 0.007
2 p5 0.000 The intersection pertaining to Sample Calculation A5 differs
3 pdom 0.007 from that of Sample Calculation A3 in that all four approaches
4 psubo 0.000 have the same configuration: two through lanes and an exclusive
5 Constrained or unconstrained? U left-turn lane. The intersection for this problem is shown in Figure
Proportion Unblocked px 10-19. The following discussion details the differences in the meth-
odology resulting from the change in geometry.
6 p1 1.000
7 p4 0.993
8 p7 0.993
Solution
9 p8 0.993
10 p9 0.993
11 p10 0.993 Steps 1 and 2: Use Worksheets A1 and A2 to enter basic data
12 p11 0.993 and traffic flow volumes. Since only the geometry has changed,
13 p12 1.000 Steps 1 and 2 are the same as those in Sample Calculation A3.
10-54 urban streets
Table 10-21. Computation 4 (Single-Stage Process) (A5d)

Movement
1 4 7 8 9 10 11 12
1 vc,x 400 300 678 873 150 739 848 200
2 s 3,600 3,600 3,600 3,600 3,600 3,600 3,600 3,600
3 px 1.000 0.933 0.933 0.933 0.933 0.933 0.933 1.000
4 vc,u,x 400 277 658 854 126 719 829 200
Table 10-22. Computation 5 (Single-Stage Process) (A5e)

Movement
1 4 7 8 9 10 11 12
1 px 1.000 0.993 0.993 0.993 0.993 0.993 0.993 1.000
2 cr,x 1,100 1,226 334 280 876 301 290 783
3 cplat,x 1,100 1,218 332 278 870 299 288 783
Table 10-23. Impedance and Capacity Calculations Table 10-24. Delay, Queue Length, and
Level of Service
Step 1—Minor Right Turn v9 v12
Conflicting flows 150 200 Movement
Potential capacity 870 783 1 4 7, 8, 9 10, 11, 12
v (veh/hr) 33 66 231 149
cm (veh/hr) 1,100 1,218 290 285
v/c 0.03 0.05 0.80 0.52
Step 2 v4 v1 Queue
Control delay 8.4 8.1 52.6 30.6
Conflicting flows 300 400 Level of service A A F D
Potential capacity 1,218 1,100
Movement capacity 1,218 1,100
Step 3 v8 v11
Part III—Single stage (no storage)
Conflicting flows 873 848
Step 4 v7 v10
Major left turn, minor through Table 10-25. Lane Usage by Approach
0.54 0.45
impedance factor
Major left turn, minor through Approach Lane Designation
0.64 0.56
adjustment impedance Movement Lane 1 Lane 2 Lane 3
Movement capacity 203 157 1, 2, 3 1 2 2, 3
4, 5, 6 4 5 5, 6
7, 8, 9 7 8 8, 9
10, 11, 12 10 11 11, 12
Step 3: Use Worksheet A3 for site characteristics. The lane
usage corresponding to the current example is more complicated
than in the previous example and is presented in Table 10-25. All Step 8: Use Worksheet A8 to calculate shared-lane capacity.
other entries on Worksheet A3 are the same as those made in Because the northbound and southbound approaches each provide
Sample Calculation A3. an exclusive left-turn lane and two through lanes, the shared-lane
Steps 4 through 7: Use Worksheets A4 through A7. These work- capacity calculations for this example are slightly more compli-
sheets are identical to those created in Sample Calculation A3. cated than those for Sample Calculation A3. The procedure for
calculating the shared-lane capacities is the same for both direc- SAMPLE CALCULATION A6
tions. There are two entries in each cSH row, corresponding to the
exclusive left-turn lane and the remaining approach lanes, respec- The purpose of Sample Calculation A6 is to demonstrate the ef-
tively. The capacity of the exclusive left-turn lane is the same as fects of flared minor-street approaches and two-stage gap accept-
the movement capacity calculated in Worksheet A6. The shared- ance. The intersection is the same as that used in Sample Calculation
lane capacity for the remaining two lanes is computed using the A3; however, it is now assumed that both the northbound and south-
first equation given on Worksheet A8, as is demonstrated for the bound approaches are flared and that there is either a raised or striped
northbound approach in Equation 10-72: median or a two-way left-turn lane (TWLTL) on the major street.
66 + 55
cSH(8, 9) = = 368 (10-72)
(66/250) + (55/845) Description
The remaining results from Worksheet A8 are presented in Table
10-26. The intersection pertaining to Sample Calculation A6 is shown
Step 9: Use Worksheet A9 to compute effect of flared minor- in Figure 10-20. As mentioned earlier, both minor-street ap-
street approaches. Again, since neither the northbound nor the proaches are flared, and there is either a raised or striped median
southbound approach is flared, Worksheet A9 is not used. or TWLTL on the major street. The variable n, the number of
Step 10. Use Worksheet A10 for delay, queue length, and level additional vehicles able to use the stop line at any one time, is
of service. As in Worksheet A8, there are two entries in each of equal to 1, and median storage space, m, equals 2. The following
the upper two rows of Worksheet A10 corresponding to the same discussion highlights the resulting differences in methodology as
lane combinations as before. The first entry in the row labeled ‘‘v compared with Sample Calculation A3.
(veh/hr)’’ is the volume for either Movement 7 or Movement 10
(depending on which panel is being considered), and the second
Solution
entry is the added volume of the other two approach lanes. The
values in Rows 4–6 are found using the same equations and tables
as those in Sample Calculation A3. The approach delay is a Steps 1 and 2: Use Worksheets A1 and A2 to enter basic data
weighted average of the control delays, as can be seen for the and traffic flow volumes. These two steps are the same as in Sample
northbound approach in the following equation: Calculation A3.
Step 3: Use Worksheet A3 for site characteristics. Worksheet
(28.6 ∗ 44) + (24.5 ∗ 66) + (43.5 ∗ 121) A3 differs slightly from the one created for Sample Calculation
dNB = = 35.2
44 + 66 + 121 A3. All of the N’s on Lines 5–8 are changed to Y’s for Yes,
(10-73) indicating that both Movements 9 and 12 are flared and that median
storage exists for Movements 7, 8, 10, and 11. In addition, 1 is
Step 11: Use Worksheet A11 for shared major left-turn imped-
entered in the column headed ‘‘Storage Space’’ on Lines 5 and 6,
ance and delay. The major-street lane configuration is the same
and 2 is entered in the corresponding column for Lines 7 and 8.
as in Sample Calculation A3, and therefore use of Worksheet A11
Step 4: Use Worksheet A4 to calculate critical gap. The first
is again precluded.
six lines of the critical gap calculations on Worksheet A4 are
The remaining results for Worksheet A10 are presented in Table
identical to those in Sample Calculation A3; however, since this
10-27.
intersection includes two-stage gap acceptance, 1.00 is entered on
Line 7. New critical gap times are calculated and entered on Line
9. As a result of the 1.00 on Line 7, the critical gap times for the
Table 10-26. Shared-Lane Capacities minor through and minor left-turn movements on Line 9 are 1 sec
less than those on Line 8. The bottom panel of Worksheet A4
Movement
remains unchanged from that for Sample Calculation A3.
8 9 11 12 Step 5: Use Worksheets A5 to calculate effect of upstream sig-
v (veh/hr) 132 55 110 28 nals. The absence of upstream signals again obviates the need to
cm (veh/hr) 250 845 259 783 use Worksheets A5.
v/c 0.53 0.07 0.42 0.04 Step 6: Use Worksheet A6 to calculate impedance and capacity.
cSH (veh/hr) 368 335 Steps 1 and 2 on Worksheet A6 are the same as in Sample Calcula-
Table 10-27. Delay, Queue Length, and Level of Service

Movement
11,
1 4 7 8 8, 9 10 11 12
v (veh/hr) 33 66 44 66 121 11 55 83
cm (veh/hr) 1,100 1,202 196 250 368 150 259 335
v/c 0.03 0.06 0.23 0.26 0.33 0.07 0.21 0.25
Queue
Control delay 8.4 8.2 28.6 24.5 19.5 30.9 22.6 19.3
Level of service A A D C C D C C

10-56 urban streets
tion A3. To account for the effects of two-stage gap acceptance, a = 1 − 0.32 ∗ e−1.3√ 2 = 0.95 (10-77)
however, all three parts of Steps 3 and 4 are completed using
607 − 231
Worksheets A7a and A7b. y= = 2.04 (10-78)
The computations in Part I for both Steps 3 and 4 proceed as 448 − 33 − 231
follows. First the conflicting flows for the first stage are computed 0.95
on the basis of the equations given in Figure 10-2. Equation 10- cT = [2.04 ∗ (2.042 − 1) (10-79)
2.042+1 − 1
74 demonstrates the calculation of the Part I conflicting flow for
∗ (448 − 33) + (2.04 − 1) ∗ 231] = 370
Movement 7:
Tables 10-28 and 10-29 present the complete set of values for
vc,7,x = (2 ∗ 33) + 250 + (0.5 ∗ 50) + 0 = 341 (10-74) Worksheet A6.
Next, the potential capacity is computed using the critical gap
times given on Line 8 of Worksheet A4. The Part I potential Table 10-28. Two-Stage Gap Acceptance: Step 3
capacity calculation for Movement 7 is as follows:
v8 v11
e−341*6.7/3,600
cp,7,I = 341 ∗ = 625 (10-75) Part I—First Stage
1 − e−341*3.6/3,600
The pedestrian impedance factor is 1.00. Determination of the Potential capacity 618 532
capacity adjustment factor due to impeding movements and move- Pedestrian impedance factor 1.00 1.00
ment capacity is the same as in the previous examples. Movement capacity 599 503
Part II proceeds the same as Part I; however, different equations Probability of queue-free state 0.78 0.78
are used to compute the conflicting flows. The Part II conflicting
flows for Movement 7 are calculated as follows: Part II—Second Stage
300 Potential capacity 504 601
cp,7,II = (2 ∗ 66) + + (0.5 ∗ 110) + 0 = 337 (10-76)
2 Pedestrian impedance factor 1.00 1.00
Part III begins with a conflicting volume calculation as though Movement capacity 477 583
two-stage gap acceptance were not involved. This calculation is Part III—Single Stage (no storage)
followed by a computation of potential capacity, again as though
two-stage gap acceptance were not present. Consequently, the criti- Conflicting flows 873 848
cal gap is not reduced by 1.00, and is taken from Line 8 of Work- Potential capacity 273 283
sheet A4. The pedestrian impedance factor remains 1.00. Capacity adjustment factor 0.92 0.92
Computation of the capacity adjustment factor and movement ca- Movement capacity 250 259
pacity proceeds the same as in the previous sample calculations.
(Note that this value is not the final movement capacity, as will Result for Two-Stage Process
be seen below.) a 0.95 0.95
The last portion of Part III computes the variables a, y, and cT. y 1.80 0.94
These calculations are illustrated by Equations 10-77, 10-78, and cT 391 405
10-79, respectively, for Movement 7. Probability of queue-free state 0.66 0.73

Table 10-29. Two-Stage Gap Acceptance: Step 4 Worksheet A9 is now included in the intersection analysis. Since
v7 v10 there is only one lane on each of the northbound and southbound
approaches, all three columns of the top and bottom panels contain
Part I—First Stage
entries. The various movement capacities, taken from Worksheet
Conflicting flows 341 482 A6, are entered on Line 1. The movement volumes are entered on
Potential capacity 626 514 Line 2. Line 3 contains the control delay for each of the three
Pedestrian impedance factor 1.00 1.00 movements, the calculation of which for Movement 7 is as follows:
Movement capacity 607 486 3,600
d=5+ + 900 ∗ 0.25
Part II—Second Stage 370
3 ! 4
1 370 2 ∗ 13702
3,600 44
Conflicting flows 337 257 2
1 2
44 44
Potential capacity 629 703 ∗ −1+ −1 + = 16.1
Pedestrian impedance factor 1.00 1.00 370 370 450 ∗ 0.25
Capacity adjustment factor 0.71 0.71 (10-80)
The average queue length for each movement appears on Line
4. The corresponding result for Movement 7 is demonstrated as
Conflicting flows 678 739 follows:
Pedestrian impedance factor 1.00 1.00 16.1 ∗ 44
Qsep = = 0.21 (10-81)
Major left turn, minor through im- 3,600
pedance factor 0.67 0.61 The value on Line 5 is equal to the value on Line 4 plus 1, and
Major left turn, minor through ad-
the value on Line 6 is equal to the entry on Line 5 rounded to the
justment impedance 0.74 0.69
Cap adjustment factor 0.72 0.65 nearest integer. The variable nmax, or the length of the storage area
Movement capacity 231 189 such that the approach would operate as separate lanes, is com-
puted using Equation 10-35 and entered on Line 7. The result for
Result for Two-Stage Process
the northbound approach is as follows:
a 0.95 0.95
y 2.04 1.23
round(Qsep + 1) = max[1, 2, 1] = 2 (10-82)
cT 370 347 The shared-lane capacity for all three movements, taken from
Worksheet A8, is entered on Line 8. Line 9 presents the sum of
the separate lane capacities entered on Line 1, and Line 10 gives
Table 10-30. Shared-Lane Capacities the number of storage spaces in the flare, which in this case is 1
for both approaches. The actual capacity of each approach is found
Movement
using Equation 10-36 and appears on Line 11. Equation 10-83
7 8 9 10 11 12 illustrates this calculation for Movements 7, 8, and 9:
v (veh/hr) 44 132 55 11 110 28
1
cm (veh/hr) 370 391 845 347 405 783 cact = (1,605 − 443) ∗ + 443 = 1,024 (10-83)
cSH (veh/hr) 443 439 2
The remaining results for Worksheet A9 are presented in Table
10-31.
Step 7: Use Worksheets A7 to compute effect of two-stage gap Step 10: Use Worksheet A11 for shared major left-turn imped-
acceptance. Again, Worksheets A7 are not included in this sample ance and delay. Because the major street does not contain any
calculation. shared major left turns, Worksheet A11 is not used.
Step 8: Use Worksheet A8 to calculate shared-lane capacity. Step 11: Use Worksheet A10 for delay, queue length, and level
The calculations for Worksheet A8 are the same as those in Sample of service. The calculations performed in Worksheet A10 are iden-
Calculation A3, with one exception. For Movements 7, 8, 10, and tical to those made in Sample Calculation A3. The shared capacity
11, the movement capacities employed are obtained from Work- values, however, are different. The values used in Sample Calcula-
sheet A6. The results are presented in Table 10-30. tion A5 are those computed in Worksheet A9 as opposed to Work-
Step 9: Use Worksheet A9 to compute effect of flared minor- sheet A8. The new results for Worksheet A10 are shown in Table
street approaches. Because of the flared minor-street approaches, 10-32.

10-58 urban streets
Table 10-31. Flared Minor-Street Approach Calculations

Movement 7 Movement 8 Movement 9 Movement 10 Movement 11 Movement 12
csep 370 391 845 347 405 783
Volume 44 132 55 11 110 28
Delay 16.0 18.8 9.6 15.7 17.2 9.8
Qsep 0.20 0.69 0.15 0.05 0.53 0.08
Qsep + 1 1.20 1.69 1.15 1.05 1.53 1.08
round (Qsep + 1) 1 2 1 1 2 1
nmax 2 2
cSH 443 439
∑ csep 1,605 1,535
n 1 1
cact 1,024 987
Table 10-32. Delay, Queue Length, and Level of

Service
Movement
1 4 7, 8, 9 10, 11, 12
v (veh/hr) 33 66 231 149
cm (veh/hr) 1,100 1,202 1,024 987
v/c 0.03 0.05 0.23 0.15
Queue
Control delay 8.4 8.2 9.5 9.3
Level of service A A A A
V. REFERENCES
1. Kyte, M., Tian, Z., Mir, Z., Hameedmansoor, Z., Kittel- 8. Harders, J., Die Leistungsfähigkeit nicht Signalgeregelter
son, W., Vandehey, M., Robinson, B., Brilon, W., Bond- Städtischer Verkehrsknoten. Schriftenreihe Strassenbau und
zio, L., Wu, N., and Troutbeck, R., Capacity and Level of Strassenverkehrstechnik, Heft 76, Bonn, Germany (1968).
Service at Unsignalized Intersections. Final Report, Volume 1: 9. Troutbeck, R., Estimating the Critical Acceptance Gap from
Two-Way Stop-Controlled Intersections, National Cooperative Traffic Movements. Research Report 92-5, Queensland Uni-
Highway Research Program Project 3–46, TRB, National Re- versity of Technology, Brisbane, Australia (March 1992).
search Council, Washington, D.C. (April 1996). 10. Robertson, D.I., Coordinating Traffic Signals to Reduce Fuel
2. Special Report 209: Highway Capacity Manual. TRB, Na- Consumption, Proceedings of the Royal Society, Vol. 387A
tional Research Council, Washington, D.C. (1985). (9 May 1983).
3. Brilon, W., and Grossmann, M., Aktualisiertes Berechnung- 11. Yu, L., and Van Aerde, M., Implementing TRANSYT’s
sverfahren für Knotenpunkte ohne Lichtsignalanlagen. Schrif- Macroscopic Platoon Dispersion in Microscopic Traffic Simu-
tenreihe Forschung Strassenbau und Strassenverkehrstechnik, lation Models, presented at 75th Annual Meeting of the Trans-
Heft 596, Bonn, Germany (1991). portation Research Board, Washington, D.C. (1995).
4. Kittelson, W., and Vandehey M., Delay Effects on Driver 12. Bonneson, J., and Fitts, J., Effect of an Upstream Signal on
Gap Acceptance Characteristics at Two-Way Stop-Controlled Non-Priority Movement Capacity and Delay, Transportation
Intersections. Transportation Research Record 1320, TRB, Research Record 1572, TRB, National Research Council,
National Research Council, Washington, D.C. (1991). Washington, D.C. (1997).
5. A Policy on Geometric Design of Highways and Streets. Amer- 13. Wu, N., An Approximation for the Distribution of Queue
ican Association of State Highway and Transportation Offi- Lengths at Unsignalized Intersections. Proceedings of the Sec-
cials, Washington, D.C. (1994). ond International Symposium on Highway Capacity (Volume
6. Manual on Uniform Traffic Control Devices. U.S. Department 2), Australian Road Research Board Ltd., Victoria, Australia
of Transportation (1988). (August 1994).
7. Siegloch, W., Die Leistungsermittlung an Knotenpunkten 14. Pline, J., ed., Traffic Engineering Handbook, Institute of
ohne Lichtsignalsteurung. Schriftenreihe Strassenbau und Transportation Engineers, Prentice-Hall, Englewood Cliffs,
Strassenverkehrstechnik, Heft 154, Bonn, Germany (1973). N.J. (1992).

PART B. ALL-WAY STOP-CONTROLLED INTERSECTIONS
I. INTRODUCTION
Part B of Chapter 10 contains procedures for analysis of all-

way stop-controlled (AWSC) intersections. Research sponsored by
the National Cooperative Highway Research Program provided
the basis for these procedures (1).
A variety of terminology applying to the unique characteristics
of AWSC intersection capacity and level-of-service (LOS) analy-
ses is introduced in Part B. For ease of reference, these terms are
defined in the next section. The terms are also more fully described
as they are used in the succeeding sections.
VARIABLES USED IN ANALYSIS OF AWSC INTERSECTIONS
c = capacity of subject approach (veh/hr);

d = control delay (sec/veh);
hsi = saturation headway for degree-of-conflict Case i (sec);
hd = departure headway (sec); Figure 10-21. Definition of intersection approaches.
hadj = headway adjustment to account for proportion of left turns,
right turns, and heavy vehicles (sec);
k = constant to adjust degree-of-conflict case probability to
account for interdependence of headways;
m = move-up time (sec);
PLT = proportion of left turns;
PRT = proportion of right turns;
PHV = proportion of heavy vehicles;
Padj = adjusted probability of degree of conflict;
P[Ci] = probability of degree-of-conflict Case i;
s = service time, time spent by vehicle at stop line (sec);
V = volume (veh/hr);
v = flow rate (veh/hr);
x = degree of utilization, or vhd /3,600;
a = constant to adjust degree-of-conflict case probability to
account for serial correlation between forecast head-
ways; and
lx = arrival rate for approach or lane x (veh/sec).
Figure 10-22. Saturation headway conditions for Vehicle 2.
1. The degree of conflict faced by the subject driver as measured

by the presence of vehicles on the opposing and conflicting
approaches;
OVERVIEW OF PROCEDURES 2. The intersection geometry, particularly the number of lanes
on the conflicting approaches, the opposing approach, and the
subject approach;
The methodology described in Part B for the capacity and LOS
3. The directional movements of the interacting vehicles; and
analysis of AWSC intersections represents the results of recent
4. The mix of vehicle types.
research in the United States (1). The methodology analyzes each
intersection approach independently. The approach under study The departure headway for a vehicle on the subject approach is
is called the subject approach. The subject approach, opposing defined as the difference between the successive times of departure
approach, and conflicting approaches are illustrated in Figure of that vehicle and the previous departing vehicle on the subject
10-21. approach if, when a given vehicle arrives, there is already a vehicle
The basic parameter used to compute intersection capacity is ahead of it at the stop line. A departure headway is considered to
saturation headway. On the basis of data from over 22,000 vehicle be a saturation headway if the queue is continuous (see Figure
records, it was found that the saturation headway is dependent on 10-22).
10-60 urban streets
Figure 10-23. Case 1: vehicles on subject approach only. Figure 10-24. Case 2: vehicles on subject and opposing
approaches.
At AWSC intersections every vehicle is required to stop before

proceeding into the intersection. This requirement provides a
framework for studying traffic operations at AWSC intersections.
Since each driver must stop, the judgment whether to proceed into
the intersection is a function of the traffic conditions on the other
approaches. If there is no traffic present on the other approaches,
a driver can proceed immediately after the stop has been made. If
there is traffic on one or more of the other approaches, a driver
proceeds only after determining that there are no vehicles currently
in the intersection (i.e., that it is safe) and that it is his or her turn
to proceed.
Although the rules of the road have traditionally suggested that
the right-of-way belongs to the driver on the right, the actual
operation of AWSC intersections is somewhat more complex than
that. The problem becomes one of determining, under capacity
conditions for a given approach, the factors that influence the rate
at which vehicles can successively depart from the stop line. The
manner in which these factors influence the saturation headway
for a given approach is discussed below.
Field observations show that AWSC intersections operate in Figure 10-25. Case 3: vehicles on subject and conflicting
either two-phase or four-phase patterns, based primarily on the approaches.
complexity of the intersection geometry. Flows are determined by
a consensus of right-of-way that alternates between the north-south cle on the opposing approach, and thus the saturation headway
and east-west streams (for a single-lane approach) or proceeds will be greater than that for Case 1. In Case 3 (Figure 10-25),
in turn to each intersection approach (for a multilane approach vehicles on one of the conflicting approaches further restrict the
intersection). departure rate of vehicles on the subject approach, and the satura-
If traffic is present on the subject approach only, vehicles depart tion headway will be longer than that for Case 1 or Case 2. In
as rapidly as individual drivers can safely accelerate into and clear Case 4, two vehicles are waiting either on the opposing approach
the intersection (Case 1, Figure 10-23). If traffic is present on the or on the conflicting approaches, or on both (Figure 10-26). When
other approaches as well as on the subject approach, the saturation there are vehicles on all approaches, as in Case 5 (Figure 10-27),
headway on the subject approach will increase somewhat, de- saturation headways are even longer than those for the first four
pending on the degree of conflict that results between the subject cases because the potential for conflict between vehicles is greatest.
approach vehicles and the vehicles on the other approaches. In The increasing degree of potential conflict translates directly into
Case 2 (Figure 10-24), some uncertainty is introduced by the vehi- both longer driver decision times and larger saturation headways.

Figure 10-26. Case 4: vehicles on subject and two other Figure 10-27. Case 5: vehicles on all approaches.
approaches.
II. METHODOLOGY
CONCEPTUAL APPROACH
AWSC intersections require drivers on all approaches to stop

before proceeding into the intersection. Although it is a recognized
rule to give priority to the driver on the right, it is not a sufficient
descriptor of intersection operations. What in fact happens is the
development of a consensus of right-of-way that alternates between
the drivers on the intersection approaches, a consensus that is
dependent primarily on the intersection geometry and the arrival
patterns at the stop line.
For an intersection composed of two one-way streets, drivers
alternately proceed into the intersection, one vehicle from one
approach followed by one vehicle from the other approach. This
same two-phase pattern (Figure 10-28) is observed at a standard
four-leg AWSC intersection where drivers from opposing ap-
proaches enter the intersection at roughly the same time during
capacity operations. Some interruption of this pattern occurs when
there are conflicts between certain turning maneuvers (such as a
northbound left-turning vehicle and a southbound through vehicle),
but in general the north-south streams alternate right-of-way with Figure 10-28. Two-phase operation analogy.
the east-west streams. A four-phase pattern (Figure 10-29) emerges
at multilane four-leg intersections where the development of the there are vehicles waiting on the conflicting approach, a vehicle
right-of-way consensus is more difficult. Here drivers from each from the subject approach cannot enter the intersection immedi-
approach enter the intersection together as right-of-way passes ately after the previous subject vehicle but must wait for consensus
from one approach to the next as each is served in turn. with the next conflicting vehicle. The headways between consecu-
Although these patterns are useful to describe overall intersec- tively departing subject approach vehicles will be shorter for the
tion operation, the manner in which the patterns affect the capacity first case than for the second. Thus the headway for a departing
of a subject approach must be considered next. At the intersection subject approach vehicle is dependent on the degree of conflict
of two one-way streets, the headways of vehicles departing from experienced by the subject vehicle in interacting with vehicles on
the subject approach fall into one of two cases. If there are no the conflicting approach. This degree of conflict increases with
vehicles on the conflicting approach, subject approach vehicles two factors: the number of vehicles on the other approaches and
can enter the intersection immediately after stopping. However, if the complexity of the intersection geometry.
10-62 urban streets
headway for a left-turning stream at a signalized intersection de-

pending on whether the stream movement is permitted or protected
and whether it occurs from a shared lane or an exclusive lane.
The capacity of a lane at an AWSC intersection is also dependent
on the saturation headway of that lane. Since there is no traffic
signal controlling the stream movement or allocating the right-of-
way to each conflicting traffic stream, the rate of departure is
controlled instead by the interactions between the traffic streams
themselves. A degree of conflict can be observed that increases
with the number of approaches that are loaded simultaneously. To
a lesser extent, the geometry of the intersection itself controls this
rate of departure.
For an AWSC intersection prevailing conditions mean the geom-
etry of the intersection and the distribution of flow rates on each
of the intersection approaches. Because of the interaction between
the traffic streams on each approach and because it is this interac-
tion that governs the maximum flow rate on each approach, the
concept of capacity must be carefully defined. It is defined here
as the maximum throughput on an approach given the flow rates
on the other intersection approaches. Here the question is how
Figure 10-29. Four-phase operation analogy. much the flow on the subject approach can be increased if the
flows on the other approaches remain fixed.
Two other factors affect the departure headway of a subject CAPACITY MODEL
approach vehicle: vehicle type and turning movement. The head-
way for a heavy vehicle will be longer than that for a passenger The capacity model described here is an expansion of earlier
car. Further, the headway for a left-turning vehicle will be longer work (2). The model is described for four increasingly complex
than that for a through vehicle, which in turn will be longer than cases: the intersection of two one-way streets, the intersection of
that for a right-turning vehicle. two two-way streets, a generalized model for single-lane sites, and
In summary, a generalized model for multilane sites.
1. AWSC intersections operate in either two-phase or four-
phase patterns, based primarily on the complexity of the intersec- Intersection of Two One-Way Streets
tion geometry. Flows are determined by a consensus of right-of-
way that alternates between the north-south and east-west streams The first formulation of the model is based on the intersection
(for a single-lane approach) or proceeds in turn to each intersection of two one-way streets, each stop-controlled. Vehicles on either
approach (for a multilane approach). approach travel only straight through the intersection (Figure
2. The headways between consecutively departing subject ap- 10-30).
proach vehicles are dependent on the degree of conflict between The service time for a vehicle assumes one of two values: s1 is
these vehicles and the vehicles on the other intersection ap- the service time if no vehicle is waiting on the conflicting approach,
proaches. The degree of conflict is a function of the number of
vehicles faced by the subject approach vehicle (with whom the
subject approach vehicle is competing for right-of-way) and by
the number of lanes on the intersection approaches.
3. The headway of a subject approach vehicle is also dependent
on its vehicle type and its turning maneuver.
This description of intersection operations must be translated into
computational models or procedures that can be used to calculate
the service time, capacity, and delay for given conditions of traffic
flow rates and intersection geometry.
CAPACITY CONCEPT
The capacity model for signalized intersections in the 1994 revi-

sion of the HCM is based on the saturation headway estimated for
a given approach. The saturation headway is computed from an
ideal value (1.9 sec) that is modified on the basis of intersection
geometry, traffic control parameters, and traffic flow conditions.
Estimation of the saturation headway is often complex. For exam-
ple, several models have been developed to forecast the saturation Figure 10-30. Configuration for Formulation 1.
and s2 is the service time if a vehicle is waiting on the conflicting sN = s1(1 − xE)(1 − xW) + s2[1 − (1 − xE)(1 − xW)] (10-88)
approach. The mean service time for vehicles on an approach is
Unlike the formulation for the intersection of two one-way streets,
the expected value of this bivalued distribution. For the northbound
it is not possible to solve directly for the mean service time in
approach, the mean service time is
terms of a combination of arrival rates and the bivalued service
sN = s1(1 − xW) + s2xW (10-84) times. The service time on any approach is dependent upon or
directly coupled with the traffic intensity on the two conflicting
where xW is the degree of utilization for the westbound approach
approaches. This coupling prevents a direct solution. However, it
and is equal to the probability of finding at least one vehicle on
is possible to solve iteratively for the service time on each approach
that approach. Thus 1 − xW is the probability of finding no vehicle
on the basis of a system of equations for each intersection in the
on the westbound approach.
form shown in Equation 10-89.
By symmetry, the mean service time for the westbound ap-
proach is
Generalized Model for Single-Lane Sites
sW = s1(1 − xN) + s2xN (10-85)
Since the degree of utilization x is the product of the arrival rate The generalized model is based on five saturation headway val-
l and the mean service time s, the service times for each approach ues, each reflecting a different level or degree of conflict faced by
can be found directly in terms of the bivalued service times and the subject approach driver. The conditions for each case and the
the arrival rates on each approach: probability of occurrence of each are specified in Table 10-33.
The probability of occurrence is based on the degree of utilization
s1[1 − lN(s1 + s2)] on the opposing and conflicting approaches. The essence of the
sN = (10-86)
[1 − lNlW(s21 − s22)] model, and its complexity, is evident when one realizes that the
degree of utilization for one approach is computed from its service
s1[1 − lW(s1 + s2)]
sW = (10-87) time, which in turn depends on the degree of utilization on the
[1 − lWlN(s21 − s22)] other approaches. This circularity is based on the interdependence
of the traffic flow on all of the intersection approaches and shows
the need for iterative calculations to obtain stable estimates of
Intersection of Two Two-Way Streets departure headway, service time, and, finally, capacity.
From Table 10-33, the probability, P[Ci], for each degree-of-
As before, the service time for a vehicle assumes one of two conflict case can be computed. The degrees of utilization on the
values, s1 or s2. The mean service time for vehicles on an approach opposing approach, the conflicting approach from the left, and the
is again the expected value of this bivalued distribution. As ex- conflicting approach from the right are given by xO, xCL, and xCR,
pected in this case, computing the service time is more complex respectively.
than it is for two one-way streets. A northbound vehicle will have
a service time of s1 if both the eastbound and westbound ap- P[C1] = (1 − xO)(1 − xCL)(1 − xCR) (10-89)
proaches are empty simultaneously. The probability of this event P[C2] = (xO)(1 − xCL)(1 − xCR) (10-90)
is the product of the probability of an empty westbound approach
and the probability of an empty eastbound approach. The mean P[C3] = (1 − xO)(xCL)(1 − xCR)
service time for the northbound vehicle is given as follows (see + (1 − xO) (1 − xCL)(xCR) (10-91)
Figure 10-31): P[C4] = (xO)(1 − xCL)(xCR) + (xO)(xCL)(1 − xCR)
+ (1 − xO)(xCL)(xCR) (10-92)
P[C5] = (xO)(xCL)(xCR) (10-93)
The departure headway for an approach is the expected value
of the saturation headway distribution, or
5
hd = o P[C ]h
i=1
i si (10-94)
where P[Ci] is the probability of the degree-of-conflict Case Ci
and hsi is the saturation headway for that case, given the traffic
stream and geometric conditions of the intersection approach.
The service time, required for the calculation of delay, is com-
puted on the basis of the departure headway and the move-up time:
s = hd − m (10-95)
where
s = service time,
hd = departure headway, and
m = move-up time.
The capacity is computed as follows. The volume on the subject
Figure 10-31. Configuration for Formulation 2. approach is increased incrementally until the degree of utilization
10-64 urban streets
Table 10-33. Probability of Degree-of-Conflict Case

Vehicles on Approach? (Y = Yes, N = No)
Degree-of-Conflict Conflicting Conflicting
Case Subject Opposing Left Right Probability of Occurrence
1 Y N N N (1−xO)(1−xCL)(1−xCR)
2 Y Y N N (xO)(1−xCL)(1−xCR)
3 Y N Y N (1−xO)(xCL)(1−xCR)
3 Y N N Y (1−xO)(1−xCL)(xCR)
4 Y Y N Y (xO)(1−xCL)(xCR)
4 Y Y Y N (xO)(xCL)(1−xCR)
4 Y N Y Y (1−xO)(xCL)(xCR)
5 Y Y Y Y (xO)(xCL)(xCR)
on the subject approach equals or exceeds 1. This flow rate is the and 0 indicates that there is no vehicle in the lane. As before, the
maximum possible flow or throughput on the subject approach probability of there being a vehicle at the stop line in a given lane
under prevailing conditions. is x, the degree of utilization. The product of the six degrees of
saturation (encompassing each of the six lanes on the opposing or
conflicting approaches) gives the probability that any given case
Generalized Model for Multilane Sites
will occur.
The departure headway of the approach is the expected value
It is expected that saturation headways at multilane sites are
of the saturation headway distribution:
longer than those at single-lane sites, all other factors being equal.
64
These longer saturation headways are the result of two factors. A
larger intersection geometry (i.e., a larger number of lanes) requires
hd = o P[C ]h
i=1
i si (10-96)
more travel time through the intersection, thus increasing the satu- where Ci represents each combination of lane occupancy and hsi
ration headway. Additional lanes also mean more driver confusion is the saturation headway for that specific combination.
and an increasing degree of conflict with opposing and conflicting The iterative procedure to compute the departure headways and
vehicles, again increasing the saturation headway. capacities for each approach as a function of the departure head-
In contrast, some movements may not conflict as readily with ways on the other approaches is the same as that described earlier.
each other at multilane sites as they might at single-lane sites. For The number of cases clearly increases the complexity of this com-
example, a northbound vehicle turning right may be able to depart putation, however.
simultaneously with an eastbound through movement if both vehi-
cles are able to occupy separate receiving lanes. Thus, in some CONTROL DELAY
cases the saturation headway may be lower at multilane sites.
The theory described earlier proposed that saturation headway The delay experienced by a motorist is made up of a number
is a function of the directional movement of the vehicle, the vehicle of factors that relate to control, geometrics, traffic, and incidents.
type, and the degree of conflict faced by the subject vehicle. This Total delay is the difference between the travel time actually expe-
theory is extended here for multilane sites with respect to the rienced and the reference travel time that would result during
concept of degree of conflict: saturation headway is affected to a conditions with ideal geometrics and in the absence of incidents,
large extent by the number of opposing and conflicting vehicles control, and traffic. Chapters 9 and 10 of this manual quantify
faced by the subject driver. For example, in degree-of-conflict only that portion of total delay attributed to control measures,
Case 2 for a single-lane approach intersection, the subject vehicle either traffic signals or stop signs. This delay is called control
is faced only by a vehicle on the opposing approach. At a two- delay and its use is consistent in Chapters 9, 10, and 11. Control
lane approach intersection, there can be either one or two vehicles delay includes initial deceleration delay, queue move-up time,
on the opposing approach. Each degree-of-conflict case is ex- stopped delay, and final acceleration delay.
panded to consider the number of vehicles present on each of the Average control delay for any particular minor movement is a
opposing and conflicting approaches. function of the service time for the approach and the degree of
The degree-of-conflict cases for two-lane approach and three- utilization. The analytical model used to estimate control delay
lane approach intersections are defined in Tables 10-34 and 10-35. (Equation 10-97) assumes that the demand is less than capacity
For multilane sites, separate saturation headway values have for the period of analysis. In situations where the degree of satura-
been computed for the number of vehicles faced by the subject tion is greater than about 0.9, average control delay is significantly
vehicle for each of the degree-of-conflict cases. A further extension affected by the length of the analysis period. In most cases, the
of the service-time model is required to account for this increased recommended analysis period is 15 min, or 0.25 hr. If demand
number of subcases. exceeds capacity during a 15-min period, the delay results calcu-
The 27 possible combinations of the number of vehicles on each lated by the procedure may not be accurate. In this case, the period
approach for each degree-of-conflict case for intersections with of analysis should be lengthened to include the period of
two lanes on each approach are given in Table 10-36. These combi- oversaturation.
nations can be further subdivided if a vehicle can be located on d=s
3 !(x − 1) + 1450T24 + 5 (10-97)

either one of the lanes on a given approach. Table 10-37 gives the hd x
2
64 possible combinations when alternative lane occupancies are + 900T (x − 1) +
considered, where 1 indicates that there is a vehicle in the lane
Table 10-34. Degree-of-Conflict Cases for Two-Lane Approach Intersections

Approaches with Vehicles No. of Opposing
Degree-of-Conflict Conflicting Conflicting and Conflicting
Case Opposing Left Right Vehicles
1 0
2 x 1, 2
3 5 x
x 6 1, 2
5 6
x x
4 x x 2, 3, 4
x x
5 x x x 3, 4, 5, 6
Table 10-35. Degree-of-Conflict Cases for Three-Lane Approach Intersections

Approach with Vehicles No. of Opposing
Degree-of-Conflict Conflicting Conflicting and Conflicting
Case Opposing Left Right Vehicles
1 0
2 x 1, 2, 3
3 5 x
x 6 1, 2, 3
5 6
x x
4 x x 2, 3, 4, 5, 6
x x
5 x x x 3, 4, 5, 6, 7, 8, 9
Table 10-36. Number of Vehicles by Approach for Degree-of-Conflict Cases,

Multilane AWSC Intersections (Two-Lane Approach Intersections)
No. of Vehicles on Approach
DOC Subject Opposing Conflicting Conflicting
Case/Vehiclesa Approach Approach Left Approach Right Approach
1/0 1 0 0 0
2/1 1 1 0 0
2/2 1 2 0 0
3/1 1 0 1 0
1 0 0 1
3/2 1 0 2 0
1 0 0 2
4/2 1 1 0 1
1 1 1 0
1 0 1 1
4/3 1 2 1 0
1 1 2 0
1 0 1 2
1 0 2 1
1 2 0 1
1 1 0 2
4/4 1 2 2 0
1 2 0 2
1 0 2 2
5/3 1 1 1 1
5/4 1 1 2 1
1 2 1 1
1 1 1 2
5/5 1 2 2 1
1 2 1 2
1 1 2 2
5/6 1 2 2 2
a
Degree-of-conflict case and number of vehicles on the opposing and conflicting approaches.

10-66 urban streets
Table 10-37. Occupied Lane Combinations for Degree-of-Conflict Cases,

Multilane AWSC Intersections (Two-Lane Approach Intersections)
Opposing Conflicting Conflicting
DOC Approach Left Approach Right Approach
Cia Case/Vehiclesb L1 L2 L1 L2 L1 L2
1 1/0 0 0 0 0 0 0
2 2/1 1 0 0 0 0 0
3 0 1 0 0 0 0
4 2/2 1 1 0 0 0 0
5 3/1 0 0 0 1 0 0
6 0 0 1 0 0 0
7 0 0 0 0 1 0
8 0 0 0 0 0 1
9 3/2 0 0 1 1 0 0
10 0 0 0 0 1 1
11 4/2 0 0 0 1 0 1
12 0 0 1 0 0 1
13 0 0 1 0 1 0
14 0 0 0 1 1 0
15 0 1 0 1 0 0
16 1 0 1 0 0 0
17 0 1 0 0 1 0
18 1 0 0 1 0 0
19 0 1 1 0 0 0
20 0 1 0 0 0 1
21 1 0 0 0 1 0
22 1 0 0 0 0 1
23 4/3 0 0 0 1 1 1
24 0 0 1 1 0 1
25 0 0 1 1 1 0
26 1 0 1 1 0 0
27 1 1 1 0 0 0
28 1 1 0 0 1 0
29 1 1 0 0 0 1
30 0 1 1 1 0 0
31 1 0 0 0 1 1
32 0 0 1 0 1 1
33 1 1 0 1 0 0
34 0 1 0 0 1 1
35 4/4 1 1 0 0 1 1
36 0 0 1 1 1 1
37 1 1 1 1 0 0
38 5/3 0 1 0 1 0 1
39 1 0 0 1 1 0
40 0 1 1 0 1 0
41 0 1 0 1 1 0
42 0 1 1 0 0 1
43 1 0 1 0 0 1
44 1 0 0 1 0 1
45 1 0 1 0 1 0
46 5/4 1 0 0 1 1 1
47 0 1 1 1 1 0
48 0 1 1 1 0 1
49 1 0 1 0 1 1
50 1 0 1 1 1 0
51 0 1 0 1 1 1
52 1 1 1 0 0 1
53 1 0 1 1 0 1
54 0 1 1 0 1 1
55 1 1 0 1 1 0
56 1 1 0 1 0 1
57 1 1 1 0 1 0
58 5/5 1 0 1 1 1 1
59 1 1 0 1 1 1
60 1 1 1 0 1 1
61 0 1 1 1 1 1
62 1 1 1 1 1 0
63 1 1 1 1 0 1
64 5/6 1 1 1 1 1 1
a
Utilization string used in Worksheet B4 Supplemental.
b
Degree-of-conflict case and number of vehicles on the opposing and conflicting approaches.

where Table 10-38. Level-of-Service Criteria

T = length of study period or congested period, Level of Service Delay Range (sec/veh)
x = degree of utilization,
A ≤10
s = service time, and B >10 and ≤15
hd = departure headway. C >15 and ≤25
D >25 and ≤35
A constant value of 5 sec/veh is included in Equation 10-97 to E >35 and ≤50
account for the deceleration of the vehicle from cruise speed to F >50
the speed of the vehicles in the queue and the acceleration of the
vehicle from the stop line to cruise speed. the 5-sec acceleration-deceleration term in the delay equation, im-
plies that a LOS range up to 10 sec is appropriate for LOS A.
The LOS breakpoints for AWSC intersections are somewhat
LEVEL-OF-SERVICE CRITERIA different than the criteria used in Chapter 9 of this manual for
signalized intersections. The primary reason for this difference is
that drivers expect different levels of performance from different
The LOS criteria are given in Table 10-38. Average control kinds of transportation facilities. The expectation is that a signal-
delay less than 10 sec/veh is defined as LOS A. Saturation head- ized intersection is designed to carry higher traffic volumes than
ways less than 5 sec/veh have been measured when traffic is pres- an AWSC intersection. Thus a higher level of control delay is
ent only on the subject approach. This fact, in combination with acceptable at a signalized intersection for the same level of service.
Documented in this section are the steps that make up the com- 4 so that the equivalent hourly flow rate can be computed. The
putational procedures for capacity and level of service for AWSC hourly flow rates are entered on Lines 5, 6, and 7. The total flow
intersections. Five worksheets are required to complete the compu- rate for the lane (the sum of Lines 5, 6, and 7) is entered on Line
tations defined in these procedures. 8. PHV is entered for each movement on Line 9. The number of
Four primary sections make up the procedures: lanes on the subject and opposing approaches is entered on Lines
T Initial conditions and adjustments, 10 and 11. The maximum number of lanes on the conflicting
T Headway and service time, approaches, from either the left or the right, is entered on Line
T Capacity, and 12. The geometry group, from Table 10-39, is entered on Line 13.
T Delay and level of service. The length of the analysis period is entered on Line 14.
Figure 10-32 shows the steps in the computational procedures.

SATURATION HEADWAY ADJUSTMENT FACTOR
FIELD DATA REQUIREMENTS
Worksheet B3 is used to compute the saturation headway adjust-
The following field data are needed for this procedure: ment factor for each lane. The total lane flow rate, the left-turn
1. Number and configuration of lanes on each approach, flow rate, and the right-turn flow rate as computed on Worksheet
2. Volume (V) by turning movement for each approach, B2 are entered on Lines 1 through 3 of Worksheet B3. The propor-
3. Proportion of heavy vehicles (PHV) on each approach, tion of left turns, right turns, and heavy vehicles is computed and
4. Peak-hour factor (PHF), and entered on Lines 4 through 6. The geometry group, from Work-
5. Length of study period or length of oversaturated period (T) sheet B2, is entered on Line 7.
in hours. The saturation headway adjustments for turning movements and
heavy vehicles are made using Table 10-40. Equation 10-98 de-
fines this computation:
GEOMETRIC FEATURES AND MOVEMENT DEFINITIONS
hadj = hLT−adjPLT + hRT−adjPRT + hHV−adjPHV (10-98)
Worksheet B1 shows the basic features of the intersection and
where
the movements of interest. The intersection name, the analyst’s
hadj = headway adjustment,
name, the count date, and the time period are entered on this form.
hLT-adj = headway adjustment for left turns (either 0.2 or 0.5, de-
A north orientation arrow is also entered.
pending on geometry case),
hRT-adj = headway adjustment for right turns (either −0.6 or −0.7),
VOLUME ADJUSTMENT AND LANE ASSIGNMENT hHV-adj = headway adjustment for heavy vehicles (1.7),
PLT = proportion of left-turning vehicles on the approach,
The volumes for each movement are entered on Worksheet B2, PRT = proportion of right-turning vehicles on the approach, and
Lines 1, 2, and 3 for Lanes 1 and 2. The PHF is entered on Line PHV = proportion of heavy vehicles on the approach.
10-68 urban streets
Figure 10-32. Flow for AWSC procedures.
Table 10-39. Geometry Group

No. of Lanes
Geometry Intersection Subject Opposing Conflicting
Group Configuration Approach Approach Approaches
1 Four-leg or T 1 1 1
2 Four-leg or T 1 1 2
3a/4a Four-leg or T 1 2 1
3b T 1 2 2
4b Four-leg 1 2 2
5 Four-leg or T 2 1 or 2 1 or 2
5 Four-leg or T 3 1a 1a
6 Four-leg or T 3 2 or 3 2 or 3
a
If the number of lanes on the subject approach is 3 and the number of lanes on either the opposing or conflicting
approaches is 1, the geometry group is 5. Otherwise, if the number of lanes on the subject approach is 3, the
geometry group is 6.
Table 10-40. Saturation Headway Adjustment Factors by Geometry Group

Factor Group 1 Group 2 Group 3a Group 3b Group 4a Group 4b Group 5 Group 6
Left turn 0.2 0.2 0.2 0.2 0.2 0.2 0.5 0.5
Right turn −0.6 −0.6 −0.6 −0.6 −0.6 −0.6 −0.7 −0.7
HV 1.7 1.7 1.7 1.7 1.7 1.7 1.7 1.7
DEPARTURE HEADWAY AND SERVICE TIME Worksheet B4
The computation of the service time for each lane is a complex The total lane flow rate is entered on Line 1. The initial departure
process; both Worksheets B4 and Worksheet B4 Supplemental are headway of 3.2 sec is entered on Line 2. The initial degree of utiliza-
used in this process. The process is iterative and requires that tion is computed using Equation 10-99 and entered on Line 3.
Worksheet B4 Supplemental be used several times. x = vhd /3,600 (10-99)
The degree of utilization (x) is the product of the flow rate and the AdjP[C4]j = a(P[C5] − 6P[C4]) (10-103)
departure headway divided by 3,600. The final departure headway,
AdjP[C5]j = −a(10P[C5]) (10-104)
from Worksheet B4 Supplemental, Line 69, is entered on Line 4.
(The use of Worksheet B4 Supplemental is described in the next where a is a constant to adjust the degree-of-conflict probability to
section.) account for serial correlation between forecast headways (standard
The final degree of utilization for each lane is computed using value of a = 0.01).
Equation 10-99 (with input values from Lines 1 and 4); the result The computed value of AdjP[Ci] is entered in Column 8 of Work-
is entered on Line 5. The move-up time for the lane is entered on sheet B4 Supplemental. The final value of P[Ci], the sum of Col-
Line 6; it is either 2.0 sec for geometry groups 1 through 4 or 2.3 umns 7 and 8, is entered in Column 9 for each of the 64
sec for geometry groups 5 and 6. The service time is computed combinations.
using Equation 10-95 using input values from Lines 4 and 6. The The base value of the saturation headway for each of the 64
result is entered on Line 7. combinations is entered in Column 10. The value is determined
from Table 10-42 on the basis of the lane geometry group. The
saturation headway adjustment factor for each lane is entered in
Worksheet B4 Supplemental
Column 11. This value is from Worksheet B3, Line 11. The final
saturation headway, the sum of Columns 10 and 11, is computed
One copy of Worksheet B4 Supplemental is used for each lane
and entered in Column 12.
and each iteration until the values of the departure headway for
The flow rate for the lane is entered on Line 65; this value is
each lane change by less than 0.01 sec from the previous iteration.
from Worksheet B2, Line 8. The lane designation and iteration
The worksheet is used in conjunction with Table 10-37, which
number are entered on Lines 66 and 67. The departure headway
gives the 64 possible occupied lane combinations (combinations
for each lane from the previous iteration is entered on Line 68.
of vehicles present or absent on each opposing and conflicting
The new value of the departure headway is computed for each
lane) for an intersection with one or two lanes on each approach.
lane using Equation 10-96; it is the sum of the 64 lines of products
When three lanes exist on any or all of the approaches, these
of Columns 9 and 12. The new value of the degree of utilization
worksheets cannot be used.
is computed using Equation 10-99 using Lines 69 and 70 as input.
Each of the 64 combinations is represented by a utilization string
Note that if this is not the final iteration, and if the degree of
Ci. The utilization string contains six elements, a string of 0’s and
utilization exceeds 1, then it is reset to 1.
1’s corresponding to the presence or absence of a vehicle in each
If the difference between Lines 68 and 69 is less than 0.01 for
lane on the opposing and conflicting approaches. The position of
each lane, the headway for the lanes has stabilized and convergence
a 1 or 0 in the string corresponds to a specific lane, as shown in
has been reached. Once convergence has been reached, the values
Columns 1 through 6 of this worksheet. Each element is repre-
of departure headway and degree of utilization for each lane are
sented by aj. Thus the utilization string consists of the concatenated
entered on Lines 4 and 5 of Worksheet B4.
elements a1a2a3a4a5 a6. If aj equals 0, no vehicle is present in the
lane; if aj equals 1, a vehicle is present in the lane.
The probability of each value of aj depends on both the value CAPACITY
of aj (either 0 or 1) and whether the flow vj is 0 or >0. The
probability of each value of aj is given in Table 10-41. The capacity of each approach is computed with the assumption
The probability of each value of aj is entered in Columns 1 that the flows on the opposing and conflicting approaches are held
through 6 for each of the 64 combinations on the basis of Table constant. The given flow rate on the subject lane is increased and
10-41. The probability of each of the 64 combinations is computed the departure headways are computed for each approach using
by multiplying the individual probabilities given in Columns 1 Worksheet B4 and Worksheet B4 Supplemental until the degree
through 6. This value is entered in Column 7 for each of the 64 of utilization for the subject approach reaches 1. When this occurs,
combinations. the final value of the subject approach flow rate is the maximum
The adjusted probability is computed to account for the serial possible throughput or capacity of this lane. The result is entered
correlation in the previous probability computation. The degree- on Line 5 of Worksheet B5.
of-conflict cases are given in Table 10-37, and the equations for
these degree-of-conflict cases are as follows: DELAY AND LEVEL OF SERVICE
AdjP[C1]j = a(P[C2] + 2P[C3] + 3P[C4] + 4P[C5]) (10-100) Worksheet B5 is also used to determine delay and level of service.
AdjP[C2]j = a(P[C3] + 2P[C4] + 3P[C5] − P[C2]) (10-101) The total lane flow rate is entered on Line 1. The service time, from
Worksheet B4, Line 7, is entered on Line 2. The degree of utilization,
AdjP[C3]j = a(P[C4] + 2P[C5] − 3P[C3]) (10-102) from Worksheet B4, Line 5, is entered on Line 3. The final departure
headway, from Worksheet B4, Line 4, is entered on Line 4. Average
control delay per vehicle is computed for each lane and each ap-
Table 10-41. Probability of aj proach by using Equation 10-105 and entered on Line 6.
d=s
3 !(x − 1) + 1450T24 + 5(10-105)

aj vj P[aj]
2
hd x
1 0 0 + 900T (x − 1) +
0 0 1
1 1 xj
0 1 1 − xj where d is the average control delay per vehicle and the other
Note: xj is the degree of utilization, defined in Equation 10-99. terms are as previously defined. The approach delay is the

10-70 urban streets
WORKSHEET B1: BASIC INTERSECTION INFORMATION

WORKSHEET B2: VOLUME ADJUSTMENT AND SITE CHARACTERISTICS
WORKSHEET B3: SATURATION HEADWAY ADJUSTMENT FACTOR

10-72 urban streets
WORKSHEET B4: DEPARTURE HEADWAY AND SERVICE TIME

WORKSHEET B4 SUPPLEMENTAL: COMPUTATION OF PROBABILITY STATES AND SATURATION HEADWAY

(continues)

10-74 urban streets
WORKSHEET B4 SUPPLEMENTAL: COMPUTATION OF PROBABILITY STATES AND SATURATION HEADWAY
WORKSHEET B5: CAPACITY, DELAY, AND LEVEL OF SERVICE

Table 10-42. Saturation Headway Values by Case and Geometry Group

Case NumVeha Group 1 Group 2 Group 3a Group 3b Group 4a Group 4b Group 5 Group 6
1 0 3.9 3.9 4.0 4.3 4.0 4.5 4.5 4.5
2 1 4.7 4.7 4.8 5.1 4.8 5.3 5.0 6.0
2 x x x x 6.2 6.8
≥3 7.4
3 1 5.8 5.8 5.9 6.2 5.9 6.4 6.4 6.6
2 x x x 7.2 7.3
≥3 7.8
4 2 7.0 7.0 7.1 7.4 7.1 7.6 7.6 8.1
3 x x x x x 7.8 8.7
4 x x x 9.0 9.6
≥5 12.3
5 3 9.6 9.6 9.7 10.0 9.7 10.2 9.7 10.0
4 x x x x x 9.7 11.1
5 x x x 10.0 11.4
≥6 x x 11.5 13.3
a
Number of lanes on the opposing and conflicting approaches that are occupied by a vehicle.
weighted average of the delay on each lane. The intersection delay kind of analysis is called a planning-level analysis. It is expected
is the weighted average of the delay on each of the approaches. that for such an analysis only a limited amount of input data is
The level of service for each approach and for the intersection is available.
determined using Table 10-38 and the computed values of The planning analysis method described here is based on the
stopped delay. operational analysis method. The planning method requires all
geometric data and traffic flow data required for the operational
analysis method; all computations for the planning method are
PLANNING AND DESIGN APPLICATIONS identical to those for the operational analysis method.
The operational analysis procedure described earlier in this
The operational analysis method described earlier in this chapter chapter is not normally used for design purposes. However,
provides a detailed procedure to evaluate the performance of an through iteration the analyst can use a given set of traffic flow
AWSC intersection. Sometimes, however, an analyst may wish to data and determine the number of lanes that would be required to
estimate the level of service for a long-term time horizon. This produce a given level of service.
SAMPLE CALCULATION B1 in the NB L1 and WB L1 columns with volumes of 300 and 200 veh/
hr, respectively. With a PHF of 1 and 0 percent heavy trucks, this
Sample Calculation B1 illustrates the fundamental principles of means that the hourly flow rates on Line 6 are identical to the vol-
the AWSC capacity analysis methodology. umes. The geometry group is 1 and T is assumed to be 0.25 hr.
Step 3: Complete Worksheet B3 for saturation headway adjust-
ment. No adjustments are required since there are no turning move-
Description
ments or heavy vehicles. The end result is that hadj is zero for
both lanes.
The setting for the problem is an intersection between two one-
Step 4: Complete Worksheet B4 for departure headway and
lane, one-way streets as shown in Figure 10-33. The busiest hour
service time. On Lines 1 and 2 for the NB L1 and WB L1 columns,
at this intersection involves a flow of 300 veh/hr northbound and
the total lane volumes (300 and 200) and the departure headway
200 veh/hr westbound. There are no turning movements. Important
initial values, which are assumed to be 3.2 sec, are entered as the
features of the methodology are illustrated by working step by
methodology suggests.
step through the capacity analysis of this facility.
Step 4a: Use Worksheet B4 to compute initial degrees of utiliza-
tion (Line 3). The initial degrees of utilization are computed by
Solution multiplying the mean arrival rate by the departure headway. For
the northbound and westbound approaches, the initial values are
Steps 1 and 2: Use Worksheets B1 and B2 to enter basic data and computed as follows:
traffic flow volumes. As indicated above, both streets have single- vWBhd,WB (200)(3.2)
lane approaches. The flow rates are 300 veh/hr northbound and 200 xWB = = = 0.18 (10-106)
3,600 3,600
veh/hr westbound. There are no trucks and PHF is 1.0. This means
that Worksheet B1 shows a northbound flow of 300 veh/hr and a vNBhd,NB (300)(3.2)
xNB = = = 0.27 (10-107)
westbound flow of 200 veh/hr. On Worksheet B2, Line 2 has entries 3,600 3,600
10-76 urban streets
Figure 10-33. Traffic volumes for Sample Calculation B1.
Step 4b: Complete Worksheet B4 Supplemental for probability taken from Equations 10-109 and 10-111 and Equations 10-106
states and saturation headway. Since there are only two ap- and 10-107 are repeated.
proaches, only two cases pertain: Cases 1 and 3. Therefore, only
Lines 1 and 3 contain entries. The values appearing in Columns vWBsWB (200)(4.5)
xWB = = = 0.25 (10-112)
1 through 6 are taken from Table 10-41. Their products are entered 3,600 3,600
in Column 7. For simplicity, it has been assumed that no correlation v s (300)(4.3)
xNB = NB NB = = 0.36 (10-113)
among the saturation headways exists, so a has been set equal to 3,600 3,600
zero. Consequently, Column 9 contains the same entries as Column
7. The next sample calculation presents an example in which the The departure headways converge after only a few iterations: 4.6
effects of correlation are taken into account. The two base satura- sec/veh for the westbound approach and 4.4 sec/veh for the north-
tion headway values pertaining to Cases 1 and 3, 3.9 sec/veh and bound approach. It is important to note that these are not the
5.8 sec/veh, are entered in Column 10 on Lines 1 and 3, respec- saturation headways for the approaches (i.e., the headways at ca-
tively. Since there are no adjustments, Column 12 is identical to pacity); rather they are the expected values of the headways be-
Column 10. tween vehicles departing from the stop line given the flows on
Step 4c: Use Worksheet B4 Supplemental to compute the first both approaches (i.e., 200 and 300 veh/hr). (Intuitively, one would
iteration departure headways (Lines 1 through 72). With the new not expect these to be the saturation headways. At saturation there
degrees of utilization provided by Equations 10-106 and 10-107, must be a continuous queue on at least one of the two approaches,
updated departure headways can be computed for both approaches: and with flows of 200 and 300 veh/hr, that is not likely to be the
5 case.) The approach capacities have to be computed in another
hd,NB = o P[C ]h
i=1
i si = hs1(1 − xWB) + hs3(xWB) (10-108) manner, as will be shown later.
The final degrees of utilization give another insight into the
hd,NB = (3.9)(0.78) + (5.8)(0.22) = 4.3 (10-109) intersection’s operation: the proportion of the hour during which
5 each approach is busy processing vehicles. For the northbound
hd,WB = o P[C ]h
i=1
i si = hs1(1 − xNB) + hs3(xNB) (10-11 0) approach this proportion is 37 percent of the time; for the west-
hd,WB = (3.9)(0.67) + (5.8)(0.33) = 4.5 (10-111) bound approach, it is 26 percent of the time because the final
degrees of utilization are 0.37 and 0.26, respectively. In the field,
The values are entered on Line 69 of both the supplemental work- the user should see vehicle-processing activity on the two ap-
sheet for the northbound and westbound approaches, respectively. proaches for about this proportion of the time.
Finally, the degree of utilization for each approach is computed Step 4e: Use Worksheet B4 to compute the service time for each
and entered on Line 71, as shown in the next paragraph. For approach (Line 7). Once the final departure headway values have
simplicity, it has been assumed that no correlation among the been determined, the service times for each approach can be com-
saturation headways exists, so a has been set equal to zero. The puted. If the move-up time is 2.0 sec (Line 6), the service times
next sample calculation presents an example in which the effects are computed as follows:
of correlation are taken into account.
Step 4d: Use Worksheet B4 Supplemental to iterate, using Steps sW = 4.6 − 2.0 = 2.6 sec (10-114)
4a and 4b, until the departure headways for the current iteration sN = 4.4 − 2.0 = 2.4 sec (10-115)
match those for the previous one. Steps 4a and 4b must be executed
repeatedly until the departure headways from the current iteration Step 5: Complete Worksheet B5 for capacity, delay, and level
match those from the previous one within a given tolerance limit. of service. The total lane volumes and service times are obtained
To illustrate how the process starts, compute the degrees of utiliza- from Worksheet B4. The remaining lines are to be completed as
tion for the next iteration (values for Line 71). The results are described below.
Step 5a: Compute the capacity of each approach. An iterative volumes for the intersection. These are entered on both Worksheets
process is used to compute the capacity of each approach. It is B1 and B2. The PHF is 1.0, the proportion of heavy trucks is 0
necessary to ascertain the maximum volume that can be accommo- percent, and there is only one lane per approach, so the volumes
dated on the subject approach given that the volumes on the other shown in Figure 10-34 carry forward to Line 8 of Worksheet B2.
approaches remain unchanged. Put another way, it must be deter- The intersection fits geometry Group 1, and the value of T is
mined how much the flow rate on the subject approach can be assumed to be 0.25 hr.
increased, with all other flows remaining unchanged, without the Step 3: Complete Worksheet B3 for saturation headway adjust-
overall processing capacity of the intersection being exceeded. The ment: Left-turn and right-turn adjustments are applied using the
process employed is as follows. First the northbound flow rate is factors given in Table 10-40. For the southbound approach, the
increased while the westbound flow rate remains fixed until the adjustment factor computation (Line 11) is as follows:
degree of utilization on the northbound approach reaches 1.0. At
that point, the capacity limit of the northbound approach has been hadj = (0.67)(0.2) + (0.33)(−0.6) + (0)(1.7) = −0.067 (10-118)
reached. Conversely, to obtain the capacity of the westbound ap- The resulting adjusted saturation headways are 3.83, 4.63, 5.73,
proach, its flow rate is increased while the northbound flow rate and 6.93 sec/veh for Cases 1–4.
remains fixed at 300 veh/hr until the degree of utilization on the Step 4: Complete Worksheet B4 for departure headway and
westbound approach reaches 1.0. service time. On Lines 1 and 2 the total lane volumes (0, 150,
What are the quantitative results? If the flow on the northbound 350, and 400 for the northbound, southbound, eastbound, and west-
approach is increased, the intersection reaches capacity at 800 veh/ bound L1 cells, respectively) and the departure headway initial
hr. At that flow rate, the degree of utilization on the northbound values (3.2 sec) are entered. The initial x-values are 0.17, 0.39, and
approach reaches 1.0. Conversely, if the flow on the westbound 0.44 for the southbound, eastbound, and westbound approaches,
approach is increased (while the flow on the northbound approach respectively. A sequence of iterations then follows using Work-
is left at 300 veh/hr), capacity is reached at a flow of 750 veh/hr. sheet B4 Supplemental until the departure headways converge to
(The lower value results from the fact that the northbound flow is constant values—5.46, 4.79, and 4.57 sec for the southbound, east-
greater than the westbound flow, 300 veh/hr versus 200 veh/hr.) bound, and westbound approaches, respectively.
Step 5b: Compute the delay (Lines 4, 6, and 8). This next-to- In this case, since correlation among the saturation headways
last step involves applying Equation 10-105 to compute the delay is being taken into account, the value of a is 0.01, which means
per vehicle on each approach. For the northbound approach, the that the computations involved in accounting for this effect need
computation of the average delay involves the following: to be performed. Table 10-43 shows the final set of iteration
dNB = 7.4 + (900)(0.25) values.
It can be seen that for the southbound approach, computation
3(0.37 − 1) + !(0.37 − 1) + 1 (450)(0.25) 24 = 9.9
2 (4.38)(0.37)
of the probability adjustment factors requires the following
calculations:
(10-116)
Padj[C1] = 0.0171 (10-119)
The value of 7.4 sec in the above equation includes both the 2.4-
Padj[C2] = 0.010 (10-120)
sec service time and a 5-sec adjustment to account for acceleration
Padj[C3] = −0.013 (10-121)
and deceleration delay. The corresponding value for the westbound
Padj[C4] = −0.014 (10-122)
approach is 9.2 sec/veh. These values pertain to both Lines 6 and
8 since there is only one lane on each approach. For Line 10, the In addition, the adjusted departure headway (5.46 sec/veh) is com-
weighted average delay for the intersection is found as follows: puted as follows:
(300)(9.9) + (200)(9.2) hd = 3.83 ∗ (0.26 + 0.017) + 4.63 ∗ (0.000 + 0.010)
d= = 9.6 (10-117)
300 + 200 + 5.73 ∗ (0.50 − 0.0013) + 6.93 ∗ (0.24 − 0.0014) = 5.46
Step 5c: Determine level of service. Table 10-38 provides the (10-123)
correspondence between values of delay and level of service. In
this case, since all the delays are less than 10 sec, the level of The service times are computed by subtracting the move-up time
service for both approaches is LOS A. from the adjusted departure headways. Since the move-up time is
2.0 sec, this implies service times that are 2.0 sec shorter than the
departure headways, or 3.46 sec/veh in the case of the southbound
SAMPLE CALCULATION B2 approach, for example.
Step 5: Complete Worksheet B5 for capacity, delay, and level
of service. The total lane volumes and service times are from
Description
Worksheet B4. The remaining lines are completed as was set forth
for Sample Calculation B1. The results are presented in Table
This example problem involves a three-leg (T-) AWSC intersec- 10-44.
tion between two two-lane streets. Sixth Street southbound dead-
ends at Line Street as shown in Figure 10-34.
SAMPLE CALCULATION B3
Solution Description
Steps 1 and 2: Use Worksheets B1 and B2 for basic data and The third sample problem illustrates the AWSC methodology
traffic flow volumes. Figure 10-34 presents the turning movement for a four-way AWSC intersection. The intersection shown in
10-78 urban streets
Table 10-43. Departure Headways given in Table 10-42. Left-turn and right-turn adjustments are
Approach applied using the factors given in Table 10-40. The results for
each approach are given in Table 10-45.
Parameter SB EB WB
Step 4: Complete Worksheet B4 for departure headway and
x, degree of utilization 0.23 0.47 0.51 service time. The initial departure headways are set to 3.2 sec for
P[C1] 0.26 0.38 0.41 each approach. These values converge to the final values shown in
P[C2] 0.00 0.39 0.36
P[C3] 0.50 0.11 0.12 Table 10-46 in just a few interations. With a correlation adjustment
P[C4] 0.24 0.12 0.11 parameter (a) of 0.01, the departure headway values are recom-
Padj[C1] 0.017 0.010 0.009 puted at each iteration as shown in Table 10-46, which shows the
Padj[C2] 0.010 −0.001 −0.000 final set of iteration values.
Padj[C3] −0.013 −0.002 −0.003 The service times are 2.0 sec less than the departure headways
Padj[C4] −0.014 −0.007 −0.006 since the move-up time is 2.0 sec. The resulting values for service
h, departure headway 5.46 4.79 4.57
time are shown in Table 10-47.
of service. The capacity of each approach is computed by assuming
Table 10-44. Capacity, Delay, and Level of Service
that the flows on the opposing and conflicting approaches are held
Approach constant. For each approach, the flow rate is increased using Step
Parameter SB EB WB 4 until the degree of utilization for the subject approach reaches
Approach capacity (veh/h) 585 720 760 1. This produces the results given in Table 10-48.
Approach delay (sec/veh) 10.1 11.9 12.2
Approach level of service B B B
Overall delay (sec/veh) 11.7 SAMPLE CALCULATION B4
Description
Figure 10-35 is between two two-lane arterials that are controlled
by stop signs in each direction. Each approach has 10 percent
trucks. Sample Calculation B4 concerns the intersection of Eighth Ave-
nue and Sixteenth Street. The diagram in Figure 10-36 shows that
both streets are four-lane arterials, and there are stop signs on all
Solution approaches. The lane markings are left-and-through and through-
and-left on each approach. For purposes of analysis it is assumed
Steps 1 and 2: Use Worksheets B1 and B2 for basic data and that the lanes are utilized in a balanced fashion, with an equal
traffic flow volumes. For this sample calculation, all of the turning number of vehicles per hour using each lane on each approach.
movements on Worksheets B1 and B2 have nonzero values. The The proportion of trucks is 0 percent, PHF = 1, and T is assumed
turning movement volumes and percentages of heavy vehicles are to be 0.25 hr.
shown in Figure 10-35. The PHF is 1, and T is assumed to be
0.25 hr. The hourly flow rates are the same as the volumes since
PHF = 1. Since the intersection falls into Group 1 (one lane on Solution
each approach), no lane usage distribution is necessary.
Step 3: Complete Worksheet B3 for saturation headway adjust- Steps 1 and 2: Use Worksheets B1 and B2 for basic data and traffic
ment. The base saturation headways are computed using the values flow volumes. For this sample calculation, all of the turning move-
Table 10-45. Saturation Headways Table 10-48. Capacity, Delay, and Level of Service
Approach Approach
Case NB SB EB WB Parameter NB SB EB WB
1 4.0 4.0 4.0 4.0 Approach capacity (veh/hr) 445 430 475 435
2 4.8 4.8 4.8 4.8 Approach delay (sec/veh) 29.3 23.9 42.3 26.0
3 5.9 5.9 5.9 5.9 Approach level of service D C E D
4 7.1 7.1 7.1 7.1 Overall delay (sec/veh) 31.3
5 9.7 9.7 9.7 9.7 Overall level of service D
ments on Worksheets B1 and B2 have nonzero values. The intersec-

Table 10-46. Departure Headways tion falls into Geometry Group 5 (two lanes on each approach), and
Approach the distribution of flows by lane is shown in Figure 10-36. The hourly
Parameter NB SB EB WB flow rates are the same as the volumes since PHF = 1.
Step 3: Complete Worksheet B3 for saturation headway adjust-
x, degree of utilization 0.74 0.65 0.87 0.69
ment. The base saturation headways are computed using the values
P[C1] 0.01 0.01 0.03 0.01
P[C2] 0.03 0.03 0.06 0.08 given in Table 10-42. Left-turn and right-turn adjustments are
P[C3] 0.13 0.09 0.13 0.06 applied using the factors given in Table 10-40. The results for
P[C4] 0.44 0.42 0.44 0.44 each approach are given in Table 10-49.
P[C5] 0.39 0.45 0.33 0.42 Step 4: Complete Worksheet B4 for departure headway and
Padj[C1] 0.032 0.033 0.030 0.032 service time. The departure headway for each lane begins at 3.2 sec.
Padj[C2] 0.022 0.022 0.020 0.021 Iterations ensue until convergence is reached. The final degrees of
Padj[C3] +0.008 +0.010 +0.007 0.011 utilization and departure headways for each approach are given in
Padj[C4] −0.023 −0.021 −0.023 −0.022 Table 10-50.
Padj[C5] −0.039 −0.044 −0.033 −0.042 The service times are 2.3 sec less than the departure headways
h, departure headway 7.61 7.77 7.37 7.65
since the move-up time is 2.3 sec. The resulting values for service
times are shown in Table 10-51.
of service. The capacity of each approach is computed assuming
Table 10-47. Service Times
that the flows on the opposing and conflicting approaches are held
Parameter NB SB EB WB constant. The given flow rate on the subject lane is increased and
Service time 5.6 5.8 5.4 5.7 the departure headways are computed for each approach using
Step 4 until the degree of utilization for that lane reaches 1. The
results from the capacity calculations are given in Table 10-52.

10-80 urban streets
Table 10-49. Saturation Headways
No. of NB SB EB WB
Case Vehicles Lane 1 Lane 2 Lane 1 Lane 2 Lane 1 Lane 2 Lane 1 Lane 2
1 0 4.70 4.36 4.60 4.08 4.72 4.34 4.70 4.22
2 1 5.20 4.86 5.10 4.58 5.22 4.84 5.20 4.72
2 6.40 6.06 6.30 5.78 6.42 6.04 6.40 5.92
3 1 6.60 6.26 6.50 5.98 6.62 6.24 6.60 6.12
2 7.40 7.06 7.30 6.78 7.42 7.04 7.40 6.92
4 2 7.80 7.46 7.70 7.18 7.82 7.44 7.80 7.32
3 8.00 7.66 7.90 7.38 8.02 7.64 8.00 7.52
4 9.20 8.86 9.10 8.58 9.22 8.84 9.20 8.72
5 3 9.90 9.56 9.80 9.28 9.92 9.54 9.90 9.42
4 9.90 9.56 9.80 9.28 9.92 9.54 9.90 9.42
5 10.20 9.86 10.10 9.58 10.22 9.84 10.20 9.72
≥6 11.70 11.36 11.60 11.08 11.72 11.34 11.70 11.22
Table 10-50. Degrees of Utilization and Departure Headways

NB SB EB WB
Parameter Lane 1 Lane 2 Lane 1 Lane 2 Lane 1 Lane 2 Lane 1 Lane 2
x, degree of utilization 0.61 0.58 0.60 0.57 0.56 0.53 0.61 0.58
h, departure headway 8.8 8.4 8.7 8.2 8.9 8.5 8.8 8.3
Table 10-51. Service Times

NB SB EB WB
Parameter Lane 1 Lane 2 Lane 1 Lane 2 Lane 1 Lane 2 Lane 1 Lane 2
Service time 6.5 6.1 6.4 5.9 6.6 6.2 6.5 6.0
Table 10-52. Capacity, Delay, and Level of Service

NB SB EB WB
Parameter L1 L2 L1 L2 L1 L2 L1 L2
Approach capacity (veh/hr) 400 415 405 430 395 410 400 420
Lane delay (sec/veh) 24.1 22.3 23.7 21.0 22.1 20.4 24.2 21.6
Approach delay (sec/veh) 23.2 22.3 21.2 22.9
Overall delay (sec/veh) 22.4

V. REFERENCES
1. Kyte, M., Tian, Z., Mir, Z., Hameedmansoor, Z., Kittel- 2. Richardson, A., A Delay Model for Multiway Stop-Sign Inter-
son, W., Vandehey, M., Robinson, B., Brilon, W., Bond- sections, Transportation Research Record 1112, Transportation
zio, L., Wu, N., and Troutbeck, R., Capacity and Level of Research Board, Washington, D.C. (1987).
Service at Unsignalized Intersections, Final Report, Volume 2:
All Way Stop Controlled Intersections, National Cooperative
Highway Research Program 3-46 (April 1996).
PART C. ROUNDABOUTS
I. INTRODUCTION
Procedures for analysis of roundabouts are described in this alternated among vehicles. A vehicle entering as a subordinate
section. A variety of terminology is introduced applying to the vehicle immediately becomes a priority vehicle until it exits the
unique characteristics of roundabout capacity. For ease of refer- roundabout. Some traffic circles impose control measures within
ence, these terms are defined in the next section. The terms are also the circulating roadway or are designed with weaving areas to
more fully described as they are used in the succeeding sections. resolve conflicts between movements.
These procedures are to provide guidance in the estimation of 3. The speed at which a vehicle is able to negotiate the circulat-
the capacity of a roundabout. It is acknowledged that the parame- ing roadway is controlled by the location of the central island with
ters used here to estimate capacity are based on limited U.S. data; respect to the alignment of the right entry curb and the circulating
the parameters will be adjusted as more data become available. roadway cross section. It is important that the speeds of vehicles on
the roundabout be low. This feature is responsible for the improved
safety record of roundabouts. Some large traffic circles provide
VARIABLES USED IN ANALYSIS OF ROUNDABOUTS
straight paths for major movements or are designed for higher
ca = approach capacity (veh/hr); speeds within the circulating roadway. Some small traffic circles
tc = critical gap, minimum length time interval that allows inter- do not achieve adequate deflection for speed control because of
section entry to one minor-stream vehicle (sec); the small central island diameter.
tf = follow-up time, time span between departure of one vehicle 4. No parking is allowed on the circulating roadway. Parking
from minor street and departure of next vehicle, using same maneuvers prevent the roundabout from operating in a manner
major-stream gap during periods of continuous queue on consistent with its design. Some larger traffic circles permit park-
minor traffic stream (sec); ing within the circulating roadway.
va = approach flow rate (veh/hr); and 5. No pedestrian activities take place on the central island. Pe-
vc = circulating flow rate (veh/hr). destrians are not intended to cross the circulating roadway. Some
larger traffic circles provide for pedestrian crossing to, and activi-
ties on, the central island.
CHARACTERISTICS OF ROUNDABOUTS
6. All vehicles circulate counterclockwise, passing to the right
of the central island. In some small traffic circles (sometimes called
Three main features of a roundabout are illustrated in Figure mini-traffic circles) left-turning vehicles are allowed to pass to the
10-37—the central island, the circulating roadway, and the splitter left of the central island.
island (1). A roundabout is distinguished from a traffic circle in 7. Roundabouts are designed to properly accommodate speci-
general by a set of common characteristics. Traffic circles that do fied design vehicles. Some smaller traffic circles are often unable
not exhibit these characteristics are not considered to be round- to accommodate large vehicles, usually because of right-of-way
abouts. For comparison purposes the nonconforming features
constraints.
found at some traffic circles are indicated in italics in the following
8. Roundabouts have raised splitter islands on all approaches.
list of common characteristics that define roundabouts:
Splitter islands are an essential safety feature, required to separate
1. Vehicles entering a roundabout on all approaches are re- traffic moving in opposite directions and to provide refuge for
quired to yield to vehicles within the circulating roadway. Traffic pedestrians. Some smaller traffic circles do not provide raised
circles sometimes employ stop or signal control or give priority splitter islands.
to entering vehicles. 9. When pedestrian crossings are provided for the approach
2. Circulating vehicles are not subjected to any other right-of- roads, they are placed approximately one car length back from the
way conflicts, and weaving is kept to a minimum. These character- entry point. Some traffic circles accommodate pedestrians in other
istics provide the means by which the priority is distributed and places, such as the yield point.
10-82 urban streets
Figure 10-37. Basic roundabout.
10. Entry deflection is the result of the physical features of a 5. Adequate lighting must be provided for safe operation at
roundabout. Some traffic circles rely on pavement markings to night.
promote deflection.
Although not explicit roundabout characteristics, the following
features are necessary for a roundabout to perform safely and OVERVIEW OF PROCEDURES
efficiently:
1. It must be easily identified in the road system, Roundabouts have been used successfully in many cities
2. The layout must be clearly visible and marked appropriately, throughout the world and are gaining popularity in the United
3. The layout must encourage drivers to enter the intersection States. Although extensive literature on roundabout modeling has
slowly, evolved worldwide, there is limited experience with their applica-
4. Adequate sight distance must be provided at all entry points tion in this country. It is therefore not possible to offer a compre-
to enable the driver to enter the intersection and to observe the hensive analysis methodology for all situations. The procedures
movements of pedestrians and bicycles, and described in this section make the best use of the limited field data
collected at U.S. roundabouts to modify the operating parameters as capacity and delay. Analytical models are based on the concept
of established performance analysis techniques. Although the pro- of gap acceptance theory. The choice of an analysis approach
cedures should be used with care until additional research is con- depends on the calibration data available. Empirical models are
ducted, they do provide the U.S. practitioner with basic guidelines generally better but require a number of congested roundabouts
concerning the capacity of roundabouts. for calibration. Gap acceptance models, however, can be developed
Existing intersection analysis models fall into two general cate- for uncongested roundabouts (2). Given the limited gap acceptance
gories. Empirical models rely on field data to develop relationships calibration data and the lack of empirical data at the time this
between geometric design features and performance measures such section was written, a gap acceptance approach was used.
II. METHODOLOGY
CONCEPTUAL APPROACH Drivers have been found to use longer gaps when the flow is
low. Drivers are prepared to wait for a longer gap that may come
The capacity of a roundabout can be estimated using gap accept- after a shorter gap so that driving may be more leisurely. In other
ance techniques with basic parameters of critical gap and follow- circumstances, drivers at roundabouts in other countries have been
up time. found to accept quite small gaps. This behavior has now been
It has generally been assumed that the performance of each leg found to cause the following circulating drivers to slow and the
of a roundabout can be analyzed independently from the other following headways to be reduced. This affects the predicted ca-
legs, and consequently most techniques tend to use information pacity only if circulating headways are used. Good estimates have
about only one leg in the analysis (3, 4) (see Figure 10-38). This been found for single-lane roundabouts if the circulating flows are
assumption is reasonable, up to a point. It has also been shown assumed to be random. This is the same assumption that has been
(5) that origin-destination paths at roundabouts affect capacity. used in the analysis of TWSC intersections.
This assumption is reasonable because more drivers who use a Because drivers make a right turn onto the roundabout, the gap
smaller radius when making a left turn will travel farther around the acceptance characteristics of drivers are expected to be the same
roundabout, will travel slower, and may have a longer intraplatoon as or similar to those of drivers making right turns at TWSC
headway (or lower saturation flow). This longer intraplatoon head- intersections. The concepts described in Part A of this chapter
way will reduce the opportunities for drivers to enter the round- dealing with TWSC intersections are generally applicable to round-
about, and capacity will be reduced. abouts. These concepts are only suitable, however, for single-
Figure 10-38. Analysis of one roundabout leg.

10-84 urban streets
Table 10-53. Critical Gap and Follow-Up Time

Critical Gap Follow-Up Time
(sec) (sec)
Upper-bound solution 4.1 2.6
Lower-bound solution 4.6 3.1
lane roundabouts. There are more traffic interactions at multilane

roundabouts that influence driver behavior and cause this technique
to be inappropriate. More details of the U.S. experience are re-
quired before a more elaborate procedure can be recommended.
CAPACITY
Figure 10-39. Roundabout approach capacity.
The equation to forecast the capacity of an entry to a roundabout
with one approach lane is as follows:
vc e−vc tc /3 ,600 10-40. For example, circulating traffic for entry by Streams 7, 8,
ca = (10-124)
1 − e−vc t f /3,600 and 9 is Streams 1, 2, and 10. Consequently, vc would be equal
to v1 + v2 + v10. Roundabouts can often be used to facilitate U-
where turns, and the flow of U-turns should also be included.
The foregoing methodology applies to single-lane roundabouts.
ca = approach capacity;
Experience with multiple-lane roundabouts in the United States is
vc = conflicting circulating traffic flow rate;
not sufficient to support an analysis procedure. Experience in other
tc = critical gap; and
countries indicates that capacity may be increased by increasing
tf = follow-up time.
the number of lanes on the approaches and on the circulating
Limited studies of U.S. roundabouts (6), as well as comparisons roadway, but the effect is less than that of an additional full lane.
with existing roundabout operations in countries with extensive In other words, doubling the number of entry or circulating lanes
experience in the design and operation of roundabouts (7), indicate does not double the capacity. In addition, the performance of multi-
that a range of values of critical gap and follow-up time should ple-lane roundabouts is affected to a greater extent by site geomet-
provide the U.S. analyst with a reasonable estimate of the likely rics and by driver characteristics. It is widely recognized that each
capacity of a planned roundabout. The recommended value ranges of the approach lanes is likely to have substantially different gap
are given in Table 10-53. The relationship between approach ca- acceptance characteristics.
pacity and circulating flow for these upper- and lower-bound val- When capacity values are required for multiple-lane round-
ues of critical gap and follow-up time is shown in Figure 10-39. abouts, a comprehensive roundabout analysis model should be
Until more definitive U.S. data become available, it is recom- used in lieu of the procedures presented here. Some caution is
mended that the roundabout’s present performance be based on necessary in the interpretation of the results produced by these
the lower-bound solution and that the upper-bound solution be models because their internal assumptions and parameters have
used to estimate the expected performance as drivers become more not generally been well validated in the United States.
experienced with roundabouts. A choice of roundabout analysis software is available; this soft-
The conflicting flows are calculated by evaluating the 15-min ware is based on established models developed by researchers in
volumes of vehicles passing in front of the entering vehicles. In other countries. For the most part, the models are supported by
other countries, the effect of vehicles exiting into the road where much more extensive field data than the method described here.
drivers are entering has been found to be a second-order effect. At Although the applicability of these data to U.S. conditions is un-
most well-designed roundabouts the exiting traffic can be ignored. known, the models themselves have demonstrated credibility in a
In practice, it is necessary to convert intersection turning move- wide range of analysis tasks and therefore should be recognized
ments into circulating flows; flows v1 to v12 are shown in Figure as useful tools for evaluating capacity and delay at roundabouts.
The basic structure of the methodology is described in this sec- 1 through 3 of the worksheet. The approach volumes are entered
tion. Worksheet C1 is provided to assist the analyst in completing on Line 1 for each turning movement. The PHF is entered on Line
the computations. This worksheet is only applicable to single-lane 2. The approach flow rates for each turning movement are entered
roundabouts with circulating flows less than 1,200 veh/hr. on Line 3. The total approach flow is computed and entered on
1. Define the existing geometry and traffic conditions for the Lines 4 through 7.
intersection under study. For each leg, the approach traffic needs 2. Determine the conflicting (circulating) traffic at each leg of
to be defined as shown earlier. The input data are entered on Lines the roundabout. For each leg, the approach and the circulating
Figure 10-40. Flow stream definitions.
traffic need to be defined as shown above. The circulating flows 4. The analyst can make an assessment of the sufficiency of
are entered on Lines 8 through 11. If the circulating flow exceeds the capacity of the roundabout based on the v/c ratio. It is
1,200 veh/hr, this procedure should not be used unless field data important to remember that this assessment is of the adequacy
have been collected for the critical gap and follow-up time. of the geometry of the roundabout design and not of the level
3. Determine the capacity of the entry lanes using Equation 10- of service provided to the driver. The result is entered on Lines
124. The analyst must select the appropriate values of the critical 14 and 15.
gap and follow-up time from Table 10-53. The results are entered
on Lines 12 and 13.

10-86 urban streets
WORKSHEET C1: CAPACITY CALCULATIONS FOR ROUNDABOUTS

SAMPLE CALCULATION C1 the eastbound approach is composed of the westbound left-turning,

the southbound left-turning, and the southbound through volumes
Sample Calculation C1 illustrates the use of the roundabout (103 + 254 + 94 = 451 veh/hr). It is important to note that each
capacity analysis procedure. of the four circulating flow volumes (451, 597, 809, and 639) is
less than 1,200 veh/hr, part of the basis upon which it was deter-
mined that this methodology could be applied.
Description
For each approach, Lines 12 and 13 present the capacity calcula-
tions based on Equation 10-124. For the eastbound approach, the
The juncture of Buena Vista and El Moro is a four-leg round-
numerical values involved are as follows:
about similar to the one shown in Figure 10-41. The methodology
described here can be applied since the roundabout is one lane 451e−(451)(4.1)/3,600
wide and its circulating flows are all less than 1,200 veh/hr. cupper = = 971
1 − e−(451)(2.6)/3,600
Figure 10-41 shows the turning-movement volumes. The east-
bound approach has the highest volume at 660 veh/hr, whereas 451e−(451)(4.6)/3,600
clower = = 788
the northbound approach has the smallest, 427 veh/hr. Heavy left- 1 − e−(451)(3.1)/3,600
turn volumes can be found on both the eastbound and southbound
Further, the v/c ratio for this approach ranges from 0.68 to 0.84.
approaches at 247 and 254 veh/hr, respectively (see Figure 10-42,
From an interpretive standpoint, Figure 10-43 portrays the rela-
for example).
tionship between the circulating volumes and the approach capaci-
ties (from Equation 10-124). The approach volume values are also
Solution plotted so that a graphical representation of the v/c ratios can be
displayed. For example, in the case of the westbound approach, it
Figure 10-42 shows the worksheet for the capacity calculations. was indicated earlier that the circulating flow was 597 veh/hr. This
Columns v1 through v12 contain the turning-movement volumes, is shown at W on the Circulating Volume axis. The corresponding
and Lines 4 through 7 summarize these values by approach. Lines capacity is 770 veh/hr, as can be seen at W on the Capacity or
8 through 11 contain the circulating flow calculations and show Volume axis. The actual approach flow is 619 veh/hr, as indicated
how these flows are derived from the turning movements. To by the data point in the middle of the graph, the vertical location
illustrate, the circulating volume of 451 veh/hr conflicting with of which relative to the capacity indicates the v/c ratio. One can
immediately see that this v/c ratio is the highest (the data point is
closest to the capacity line). The next highest v/c ratio pertains
to the eastbound approach (E), followed by the southbound and
northbound approaches.
Parametric Analysis
As a concluding comment, it is useful to show how variations

in the critical gap and follow-up time values can affect the analysis
results. The results obtained above can be used as a point of
reference. For the capacities and v/c ratios presented in Figure 10-
42, tc = 4.35 and tf = 2.85, inasmuch as these are midway between
the upper-bound and lower-bound values presented in Table 10-
54. To see what happens when alternative values are used, the
results obtained from these values are compared with those ob-
tained by using the upper-bound and lower-bound combinations.
Table 10-54 shows that shifting to the upper-bound solution pro-
duces about an 11 percent increase in the v/c ratio, whereas shifting
to the lower-bound values produces about a 10 percent decrease.
For the westbound approach, the capacity increases to 864 veh/hr
for the upper-bound values and decreases to 693 veh/hr with the
lower-bound values. The implication is that variations of about
610 percent can be obtained by deviating from the midpoints of
Figure 10-41. Traffic volumes for Sample Calculation C1. the value ranges presented in Table 10-54.

10-88 urban streets
Figure 10-42. Worksheet for Sample Calculation C1.

Figure 10-43. Sample Calculation C1 capacity and volume analysis.
Table 10-54. Effects of Changes in Critical Gap and

Move-Up Time
Scenario
A B C
tf 3.10 2.85 2.60
tc 4.60 4.35 4.10
Capacity (Eq. 10-124)
EB 788 871 971
WB 693 770 864
NB 573 644 728
SB 667 744 835
v/c Ratio
EB 0.84 0.76 0.68
WB 0.89 0.80 0.72
NB 0.74 0.66 0.59
SB 0.75 0.67 0.60
V. REFERENCES
1. Florida Roundabout Guide, Florida Department of Transporta- 4. Troutbeck, R.J., The Capacity and Design of Traffic Circles
tion (March 1996). in Australia, Transportation Research Record 1398, Transpor-
2. Kyte, M., Tian, Z., Mir, Z., Hameedmansoor, Z., Kittel- tation Research Board, Washington, D.C. (1994).
son, W., Vandehey, M., Robinson, B., Brilon, W., Bond- 5. Akcelik, R., and M. Besley, SIDRA 5 User’s Guide, ARRB
zio, L., Wu, N., and Troutbeck, R., Capacity and Level of Transport Research Limited (1996).
Service at Unsignalized Intersections, Final Report, Volume 1, 6. Flannery, A., and Datta, T., Operational Performance Mea-
Two-Way Stop-Controlled Intersections, National Cooperative sures of American Roundabouts, Transportation Research Rec-
Highway Research Program Project 3-46, TRB, National Re- ord 1572, Transportation Research Board, Washington, D.C.
search Council, Washington, D.C. (April 1996). (1997).
3. Brilon, W., and Stuwe, B., Capacity and Design of Round- 7. Troutbeck, R., The Analysis of the Performance of Round-
abouts in Germany, Transportation Research Record 1398, abouts, unpublished Technical Memorandum to Committee on
Transportation Research Board, Washington, D.C. (1994). Highway Capacity and Quality of Service (1996).

chapter 11
ARTERIAL STREETS
CONTENTS
i. introduction .......................................................................................................................................................................... 11-2

Applications ........................................................................................................................................................................... 11-2
Characteristics of Arterial Flow............................................................................................................................................ 11-2
Arterial Level of Service....................................................................................................................................................... 11-4
ii. methodology and procedures for application .............................................................................................................. 11-4
Step 1—Establish Arterial To Be Considered...................................................................................................................... 11-5
Step 2—Determine Arterial Class and Free-Flow Speed .................................................................................................... 11-6
Step 3—Divide Arterial into Sections .................................................................................................................................. 11-6
Step 4—Compute Arterial Running Time............................................................................................................................ 11-6
Step 5—Tabulate Intersection Information and Compute Delay ........................................................................................ 11-9
Step 6—Compute Average Travel Speed............................................................................................................................. 11-12
Step 7—Assess Level of Service.......................................................................................................................................... 11-13
iii. planning applications.......................................................................................................................................................... 11-15
Objectives .............................................................................................................................................................................. 11-15
Data Requirements ................................................................................................................................................................ 11-15
Computational Steps.............................................................................................................................................................. 11-16
Interpretation of Results ........................................................................................................................................................ 11-17
Calculation 1—Arterial Classification .................................................................................................................................. 11-17
Calculation 2—Computation of Arterial Level of Service .................................................................................................. 11-17
Step 1. Establish Arterial To Be Considered.................................................................................................................. 11-18
Step 2. Determine Arterial Classification ....................................................................................................................... 11-18
Step 3. Define Arterial Sections...................................................................................................................................... 11-18
Step 4. Compute Running Time...................................................................................................................................... 11-20
Step 5. Compute Intersection Delay................................................................................................................................ 11-20
Step 6. Compute Average Travel Speed ......................................................................................................................... 11-20
Step 7. Assess Level of Service ...................................................................................................................................... 11-20
Calculation 3—Computation of Arterial Level of Service .................................................................................................. 11-20
Calculation 4—Effect of Traffic Flow Rate on Arterial Level of Service ......................................................................... 11-21
Calculation 5—Effect of Traffic Flow Rate and Length on Arterial Level of Service ..................................................... 11-23
Calculation 6—Evaluation Based on Field Data.................................................................................................................. 11-23
Calculation 7—Arterial with Large Signal Spacings........................................................................................................... 11-25
Calculation 8—Planning Application: Determining Level of Service ................................................................................ 11-29
Calculation 9—Planning Application: Determining Volumes............................................................................................. 11-31
Calculation 10—Stop Control on Arterial............................................................................................................................ 11-32
Calculation 11—Two-Lane Arterial ..................................................................................................................................... 11-35
appendix i. Test-Car Method for Existing Arterials.............................................................................................................. 11-40
appendix ii. Worksheets for Use in Analysis ........................................................................................................................ 11-41

11-2 urban streets
I. INTRODUCTION
Urban and suburban arterials are signalized streets that primarily arterial capacity analysis by analyzing the capacity of the signal-
serve through traffic; secondarily, they provide access to abutting ized intersections and other such points. It is important to note
properties. For purposes of this manual, they are defined generally that capacity analysis of signalized intersections is necessary be-
as facilities with lengths of at least 1 mi in downtown areas and cause when demand exceeds capacity at any point along the arte-
at least 2 mi in other areas, with a signalized intersection spacing rial, the arterial evaluation methodology based on average travel
ranging from as little as 200 ft in downtown areas and 400 ft for speed becomes inappropriate.
interchanges and elsewhere to as long as 2 mi, and with turning The methodology of this chapter is oriented toward the evalua-
movements at intersections that usually do not exceed 20 percent tion of an existing operations situation or a specific design proposal
of total traffic volume. Roadside development along arterials can by a level-of-service (LOS) determination. The person doing
be intense, producing friction for through traffic that generally such design or operations work will be able to investigate the
limits a driver’s ability to travel at the desired speed. effect of signal spacing, arterial classification (as defined here),
In the hierarchy of urban highway transportation facilities, arte- and traffic flow on the arterial level of service. The methodology
rial streets are ranked between collector and downtown streets on uses the signalized intersection procedure in Chapter 9 for the
one level and multilane suburban highways and rural roads on lane group containing the through traffic. By redefining lane
another. The difference in ranking is mainly determined by func- arrangement (e.g., presence or absence of left-turn lanes, number
tion and by the character and intensity of roadside development. of lanes), the analyst may influence the projected traffic flow
Collector streets provide both land access and traffic circulation in the through-traffic lane group and the capacity of the lane
service within residential, commercial, and industrial areas. Their group. This redefinition, in turn, influences the arterial LOS
access function is more important than that of arterials, and, unlike determination by changing the intersection evaluation and possi-
arterials, their operation is not always dominated by traffic signals. bly the arterial classification.
Downtown streets are signalized facilities that often resemble Those interested in planning applications may use the entire
arterials. They not only move through traffic but also provide arterial methodology in a straightforward but somewhat simplified
access to local business by passenger cars, transit buses, and trucks. way by computing control delay using certain default values as
Turning movements at downtown intersections are often greater outlined in Chapter 9. Knowledge of the intended signal timing
than 20 percent of total traffic because downtown flow involves and quality of progression, however, is vital. If it is lacking or
a substantial amount of circulatory traffic. cannot be estimated, no meaningful estimation of arterial level of
Typical of downtown streets are numerous pedestrian conflicts service is possible, even on a planning level.
and lane obstructions caused by stopping or standing taxicabs, buses, LOS criteria can be applied when travel time and delay runs
trucks, and parking vehicles that cause turbulence in the traffic flow. are used to assess the impact of optimizing signal timing or other
Downtown street function may change with the time of the day, and improvements to the arterial and periodically to evaluate the entire
for this reason certain strategically located downtown streets are con- arterial system in an urban area. Arterial level of service also can
verted to arterial-type operation during peak traffic hours. be estimated by arterial traffic models, provided that
Multilane suburban highways and rural roads differ from arterials
1. Input parameters such as running speeds and saturation flow
in the following features: (a) roadside development is not as intense,
rates are determined in a manner consistent with the procedures
(b) density of traffic access points is not as high, and (c) signalized
in this manual,
intersections are more than 2 mi apart. These conditions result in a
2. The delay calculated or estimated by the model is defined
smaller number of traffic conflicts, a smoother flow, and a dissipation
of the platoon structure associated with arterial traffic. consistent with the definitions in this manual, and
3. The delay outputs from the model are based on the delay
Urban and suburban arterials include multilane divided arterials;
equations in this manual or have been validated with field data.
multilane undivided arterials; two-lane, two-way arterials (one
travel lane in each direction); and one-way arterials. Federal High- These applications of the methodology always require determi-
way Administration (FHWA) statistics from the early 1980s indi- nation of the level of service and associated measures of effective-
cate the following distribution of urban and suburban arterial miles ness (i.e., travel time, delay, speed). In certain cases determination
in urbanized areas of more than 100,000 people: multilane divided of LOS values is the final objective; in other cases LOS values
arterials constitute approximately 37 percent; multilane undivided associated with different alternatives are computed, and decisions
arterials total 27 percent; two-lane, two-way arterials make up 33 are made using these values.
percent; and one-way arterials represent the remaining 3 percent.
CHARACTERISTICS OF ARTERIAL FLOW

APPLICATIONS
The operation of vehicles on arterial streets is influenced by
The methodology contained in this chapter can be used by those three main factors: arterial environment, interaction among vehi-
concerned with the planning, design, and operation of arterials to cles, and effect of traffic signals. These factors contribute to the
evaluate the level of service of an existing or proposed facility. The capacity of an arterial street and the quality of service offered to
methodology does not address arterial capacity, which is generally its users. They constitute the basic elements of the methodology
determined by the capacity of an arterial’s signalized intersections, discussed in Section II of this chapter.
addressed in Chapter 9. In some cases, special midblock restric- Arterial environment includes the geometric characteristics of
tions also limit capacity. In general, the user can best conduct an the facility and adjacent land uses. Number of lanes and lane
arterial streets 11-3
Figure 11-1. Typical time-space trajectories of vehicles on one-lane arterial segment.
width, type of median, driveway-access-point density, and spacing The average delay per vehicle depends mainly on the proportion
between signalized intersections are among the environmental fac- of red time displayed to the arterial segment, the proportion of
tors, as are the existence of parking, level of pedestrian activity, vehicles arriving on green (or the quality of traffic signal
speed limit, and population of the city. progression), and the traffic volume. The travel speed over an
The arterial environment affects a driver’s notion of safe speed. arterial segment (which includes time lost due to intersection
Even if the effect of the other factors is negligible, the environment effects, including stops and all associated control delay for the
restricts a driver’s desired speed, that is, the maximum speed at through movements) is generally lower than the corresponding
which a driver would like to travel under a given set of environ- running speed. Similarly, the average travel speed of all vehicles
mental conditions. The average desired speed of all drivers on an on the segment is lower than their average running speed unless
arterial segment or section is termed free-flow speed in this chapter. no vehicles stop.
Interaction among vehicles is determined by traffic density, the Figure 11-1 shows simplified time-space trajectories of represen-
proportion of trucks and buses, and turning movements. This inter- tative vehicles along one lane of an arterial. Vehicles 1 and 2 turned
action affects the operation of vehicles at intersections and, to a onto the arterial from side streets, and the rest were discharged from
lesser extent, between signals. the upstream signal. Vehicles 1, 2, and 3 arrived at the downstream
Seldom can a driver travel at the desired speed. Most of the signal approach during the red interval and had to stop. Vehicle 4
time, the presence of other vehicles restricts the speed of a vehicle could have arrived at the stop line on green but had to stop because
in motion because desired speeds differ among drivers or because it was blocked by Vehicle 3, which was not yet in motion. Vehicles
downstream vehicles are accelerating from a stop and have not 5, 6, and 7 did not stop but had to reduce their speeds because they
yet reached their drivers’ desired speeds. Therefore, the average were still affected by the stoppages caused by the signal. Vehicle 8
speed of a vehicle in motion over a certain length of road, or was delayed because its driver’s desired speed was higher than
running speed, is usually lower than the desired speed of its driver that of Vehicle 7’s driver. Vehicles 9 and 10 traveled at their
because of the effect of vehicle interactions. Likewise, the average drivers’ desired speeds. The travel speeds of Vehicles 1, 2, 3, and
running speed of all vehicles on an arterial segment is usually 4 were lower than their respective running speeds, which in turn
lower than the free-flow speed of the segment. were lower than the desired speeds of their drivers. The travel
Traffic signals force vehicles to stop and to remain stopped for speeds of Vehicles 5, 6, 7, and 8 were equal to their corresponding
a certain time, and then release them in platoons. The delays and running speeds, but lower than their drivers’ desired speeds. Finally,
speed changes caused by traffic signal operation considerably re- for Vehicles 9 and 10, whose drivers were traveling at their desired
duce the quality of traffic flow on arterial streets. speeds, the three types of speeds have the same value.
11-4 urban streets
ARTERIAL LEVEL OF SERVICE the traffic stream is only slightly restricted and delays are not
bothersome.
Arterial level of service is based on average through-vehicle 3. LOS C represents stable operations; however, ability to ma-
travel speed for the segment, section, or entire arterial under con- neuver and change lanes in midblock locations may be more re-
sideration. This parameter is the basic measure of effectiveness stricted than in LOS B, and longer queues, adverse signal
for Chapter 11. The average travel speed is computed from the coordination, or both may contribute to lower average travel speeds
running time on the arterial segment or segments and the control of about 50 percent of the average free-flow speed for the arterial
delay for through movements at all intersections. To ensure that classification.
the arterial is of sufficient length so that average travel speed is a 4. LOS D borders on a range in which small increases in flow
reasonable measure of effectiveness, the arterial’s length generally may cause substantial increases in approach delay and hence de-
should be at least 1 mi in downtown areas and at least 2 mi in creases in arterial speed. LOS D may be due to adverse signal
other areas. progression, inappropriate signal timing, high volumes, or some
Arterial level of service is defined in terms of average travel combination of these. Average travel speeds are about 40 percent
speed of all through vehicles on the arterial. It is strongly influ- of free-flow speed.
enced by the number of signals per mile and the average intersec- 5. LOS E is characterized by significant delays and average
tion control delay. On a given facility, such factors as inappropriate travel speeds of one-third the free-flow speed or less. Such opera-
signal timing, poor progression, and increasing traffic flow can tions are caused by some combination of adverse progression, high
substantially degrade arterial level of service. Arterials with me- signal density, high volumes, extensive delays at critical intersec-
dium to high signal densities (more than two signalized intersections, and inappropriate signal timing.
tions per mile) are even more susceptible to these factors, and poor 6. LOS F characterizes arterial flow at extremely low speeds,
arterial level of service will probably be observed even before from less than one-third to one-quarter of the free-flow speed.
substantial intersection problems occur. Intersection congestion is likely at critical signalized locations,
The following general statements may be made regarding arte- with long delays and extensive queuing.
rial level of service:
Table 11-1 contains the arterial LOS definitions, which are based
1. LOS A describes primarily free-flow operations at average on average travel speed over the arterial segment being considered
travel speeds, usually about 90 percent of the free-flow speed for (up to and including the entire facility). It should be noted that if
the arterial classification. Vehicles are seldom impeded in their demand volume exceeds capacity at any point on the facility,
ability to maneuver in the traffic stream. Delay at signalized inter- average travel speed may not be a good measure of the arterial
sections is minimal. level of service. Thus, intersection demand-to-capacity ratios
2. LOS B represents reasonably unimpeded operations at aver- greater than 1.0 will probably result in an unacceptable level of
age travel speeds, usually about 70 percent of the free-flow service on the arterial. The arterial classification concept in Table
speed for the arterial classification. The ability to maneuver in 11-1 is defined as part of the methodology to follow.
II. METHODOLOGY AND PROCEDURES FOR APPLICATION
This methodology provides the framework for arterial evalua- for arriving at accurate arterial evaluations. If field data are un-
tion. If field data are available, this framework can be used to available, arterial traffic models are an alternative that can be
determine the level of service of a given arterial without reference used provided certain conditions are met. Input parameters such
to running time and intersection delay estimates. Instead of treating as running speeds and saturation flow rates must be determined
field evaluation as a less desirable method than estimation, the in a manner consistent with the procedures in this manual, the
transportation analyst should consider field data a better alternative delay calculated or estimated by the model must be defined consis-
Table 11-1. Arterial Levels of Service

arterial classification
i ii iii iv
Range of free-flow speeds 45 to 55 35 to 45 30 to 35 25 to 35
Typical free-flow speeds 50 40 33 30
level of service average travel speed
A ≥42 ≥35 ≥30 ≥25
B ≥34 ≥28 ≥24 ≥19
C ≥27 ≥22 ≥18 ≥13
D ≥21 ≥17 ≥14 ≥9
E ≥16 ≥13 ≥10 ≥7
F <16 <13 <10 <7
NOTE: Units are miles per hour.

tent with the definitions in this manual, and the delay outputs from green ratio; X, the v/c ratio; and c, the capacity of the through lane
the model must be based on the delay equations in this manual or group) and the quality of the signal progression;
must have been validated with field data. 6. Compute average travel speed (a) by section to prepare a
Note that field data on free-flow speed will help in determining speed profile and (b) over the entire facility; and
the arterial classification. In cases where the specific arterial does 7. Assess the level of service by referring to Table 11-1.
not yet exist, data on free-flow speed at comparable facilities are
The methodology should be applied twice on two-way arterials
recommended as an estimate.
if the level of service is to be assessed in each direction.
The procedure to determine arterial level of service involves
Steps 4 through 6 can be superseded by field data measurements
seven steps, as shown in Figure 11-2:
of the average travel speed by doing travel time and delay studies
1. Establish the location and length of arterial to be considered; along the arterial. Appendix I presents the field data collection
2. Determine the arterial classification using the classification procedures needed to provide the necessary data. Steps 4 through
scheme presented here in conjunction with the measurement of 6 can also be superseded by arterial traffic model estimates of
free-flow speed; average travel speeds and control delays for the arterial through
3. Divide the arterial for the purpose of the evaluation into movement provided the estimates are either calculated on the basis
sections that each contain one or more arterial segments; of procedures in this manual or validated with field data.
4. Compute the arterial running time for each segment, and if Each of the steps is addressed in the remainder of this section.
any sections are larger than the individual segments, aggregate for
the sections; STEP 1—ESTABLISH ARTERIAL TO BE CONSIDERED
5. Tabulate the necessary information on each intersection, and
compute the control delay for the arterial through movements at At the start of the analysis, it is useful to define the location
each intersection, taking into account intersection parameters for and length of the arterial to be considered and identify all relevant
the through movement (C, the cycle length; g/C, the effective physical, signal, and traffic data.
Figure 11-2. Arterial LOS method.

11-6 urban streets
The arterial being analyzed should be at least 1 mi long in undivided one way, or two lane. It may have some separate or
downtown areas and at least 2 mi long in other areas. If it is not, continuous left-turn lanes and some portions with parking permit-
the analyst should consider whether to add more sections. ted. It has a higher density of roadside development than the typical
suburban design, and it usually has 4 to 10 signals per mile. Speed
STEP 2—DETERMINE ARTERIAL CLASS AND FREE-FLOW limits are normally 30 to 40 mph.
SPEED Typical urban design represents an arterial with a high drive-
way-access-point density. It frequently is an undivided one-way
Four arterial classifications are defined in this chapter on the or two-way facility with two or more lanes. Parking is usually
basis of arterial function and design. Each classification includes permitted. Generally, there are few separate left-turn lanes, and
a range of free-flow speeds. In some cases, measurement of free- some pedestrian interference is present. The arterial commonly
flow speed is a valuable aid in determining proper arterial classifi- has 6 to 12 signals per mile. Roadside development is densely
cation because of ambiguities in the classification categories. commercial. Speed limits range from 25 to 35 mph.
Both free-flow speed and actual average travel speed can be In addition to these definitions, Table 11-2 should be used as
obtained by arterial travel time studies. Thus, the application of this an aid in the determination of functional and design categories.
chapter can be based entirely on field measurements. Appendix I Once the functional and design categories have been established,
presents the necessary field procedures. the arterial classification may be established by referring to Table
Free-flow speed is the average speed of drivers over the portions 11-3.
of arterial segments that are not close to signalized intersections, As a practical matter, there are sometimes ambiguities in de-
as observed during very low traffic volume conditions while driv- termining the proper categories. Measurement or estimation of
ers are not constrained by other vehicles or by traffic signals. free-flow speed is a great aid in this determination because each
Average free-flow speed should approximate drivers’ desired arterial classification has a characteristic range of free-flow speeds,
speeds for the facility and its use. Free-flow speeds may be mea- as shown in Table 11-1. Free-flow speed alone cannot be used to
sured by test cars or by spot speed observations away from determine arterial classification, but it can be used as an effective
intersections. check in the arterial classification scheme. Information on arterial
In all cases, the arterial should be classified first by functional classification is used in Steps 4 and 7 of the methodology.
category and then by design category. The functional category is
either principal or minor arterial.
STEP 3—DIVIDE ARTERIAL INTO SECTIONS
A principal arterial serves major through movements between
important centers of activities in a metropolitan area and a substan-
The basic unit of the arterial is the segment, which is the one-
tial portion of trips entering and leaving the area. It also connects
directional distance from one signalized intersection to the next.
freeways with major traffic generators. In small cities (less than
Figure 11-4 illustrates the segment concept on one- and two-way
50,000), its importance is derived from the service provided to
arterials.
traffic passing through the urban area. Service to abutting land is
If two or more consecutive segments are comparable in arterial
subordinate to the function of moving through traffic.
classification, segment length, speed limit, and general land use
A minor arterial is a facility that connects and augments the
and activity, the analyst may wish to aggregate these into a section.
principal arterial system. Although its main function is still traffic
If the segments are aggregated into a section, all results would
mobility, it performs this function at a somewhat lower level and
then focus on the section rather than on the smaller component.
places more emphasis on land access than does a principal arterial.
When a section is defined, the average segment length may be
A system of minor arterials serves trips of moderate length and
used in finding the running time per mile in the next step.
distributes travel to geographical areas smaller than those served
by a principal arterial.
Within the functional classification, the arterial is further classi- STEP 4—COMPUTE ARTERIAL RUNNING TIME
fied by its design category. Figure 11-3 shows some typical exam-
ples of the four design categories. Two principal components make up the total time that a vehicle
Typical high speed design represents an arterial with a low spends in a section and on the arterial: arterial running time and
driveway-access-point density, separate left-turn lanes, and no control delay for the through movement. This step is focused on
parking. It may be a multilane divided or undivided arterial or a computing the first of these terms so that it may be used in the
two-lane facility with shoulders. Signals are infrequent and spaced denominator of the following equation:
at long distances (no more than two signals per mile). Roadside 3,600 * (length)
development is low density, and the speed limits are typically 45 to ART SPD =
[(running time/mile) * (length) + (∑ inters. control delay)]
55 mph. This design category includes many arterials in nonurban
(11-1)
settings.
Typical suburban design represents an arterial with a low drive- where
way-access-point density, separate left-turn lanes, and no parking. ART SPD = arterial or section average travel speed
It may be a multilane divided or undivided arterial or a two-lane (mph),
facility with shoulders. Signals are spaced for good progressive length = arterial or section length (mi),
movement (one to five signals per mile or signals spaced at even running time/mile = total of the running time per mile on all
greater distances). Roadside development is of low to medium segments in arterial or section (sec), and
density, and the speed limits are usually 40 to 45 mph. inters. control delay = summation of control delays for through
Typical intermediate design represents an arterial with a moder- movements at all signalized intersections
ate driveway-access-point density. It may be multilane divided, in arterial or section (sec).
Figure 11-3. Design categories: top left, typical high speed design; top right, typical suburban design; bottom left, typical intermediate
design; bottom right, typical urban design.
The 3,600 sec/hr is a conversion factor to compute ART SPD in the effect of length directly. In addition, running time per mile
miles per hour. may be influenced by such factors as the presence of parking,
In special cases, unusual midblock delays may be caused by opportunities for side friction, and local development and street
regular vehicle stops at pedestrian crosswalks. Other such delays use. In this chapter, these factors are assumed to influence the
may be caused by bus stops or driveway interference. Such delays free-flow speed, so observation of free-flow speed includes the
may be added to the intersection control delay in the denominator effect of these factors. Once free-flow speed is estimated, the
of Equation 11-1. running speed used also reflects the effect of these factors; Table
To compute the running time in a segment, the analyst must 11-4 contains higher running times for the lower free-flow speeds
know within each classification.
If it is not possible to observe the free-flow speed on the actual
T Arterial classification, facility or on comparable existing facilities, a note to Table 11-4
T Segment or section length in miles, and gives default values to use; however, a local history of free-flow
T Free-flow speed in miles per hour. speeds on different arterial types should be available.
The segment running time may then be found by using Table 11-4
(based on research conducted by FHWA and others). Example: What is the running time on a segment that is 0.20
If a section has been defined that encompasses several segments, mi long and has a free-flow speed of 40 mph? The arterial is a
the average segment length should be used in finding the running principal arterial, suburban design.
time per mile from Table 11-4. Running time per mile is then Solution: Note that on the basis of Tables 11-2 and 11-3, the
multiplied by the section length. arterial falls in Classification II. Table 11-4 estimates the running
In each arterial classification, a number of factors can influence time per mile at 115 sec, so that the segment running time is 115
actual free-flow speed and running time per mile. Table 11-4 shows × 0.20 = 23 sec.
11-8 urban streets
Table 11-2. Aid in Establishing Arterial Classification

functional category
criterion principal arterials minor arterials
Mobility function Very important Important
Access function Generally minor Substantial
Points connected Freeways, important activity centers, major traffic Principal arterials
generators
Predominant trips Relatively long trips between points connected, Trips of moderate lengths within relatively small
served through trips entering, leaving, going through city geographical areas
design category
suburban intermediate urban
criterion high speed design design design design
Driveway access density Low density Low density Moderate density High density
Cross section Multilane divided or Multilane divided; Multilane divided; Undivided one way;
undivided multilane undivided; multilane undivided; two way, two or
two lane with one way; two lane more lanes
shoulders
Parking No No Some Usually
Separate left-turn lanes Yes Yes Usually Some
Signal per mile 1 to 2 1 to 5 4 to 10 6 to 12
Speed limits 45 to 55 mph 40 to 45 mph 30 to 40 mph 25 to 35 mph
Pedestrian interference None Little Some Usually
Roadside development Low density Low to medium Medium/moderate High density
density density
Table 11-3. Arterial Classification According to 20 percent of the average segment (see Step 3 of the methodology).
Functional and Design Categories From Table 11-4, the running time per mile for a Classification
functional category IV arterial with 30 mph free-flow speed is 150 sec for a 0.15 mi
segment (the average of the three segment lengths in this section).
principal minor
The actual running time is computed as follows:
design category arterial arterial
High speed design
(150) × (0.15 + 0.17 + 0.13) = 67.5 sec
and control I Not applicable
Typical suburban
design and Example: What is the southbound running time for the same
control II II section?
Intermediate Solution: The southbound running time is found in the same
design II III or IV way, and the answer is therefore the same. This example is a useful
Typical urban reminder that frequently two-way arterials should be evaluated in
design III or IV IV each direction; generally the answers will be different because of
the influence of intersection delay (the effect of different signal
progression quality in the two directions will contribute to this
difference).
Example: Consider the foregoing case, but with an average 30 As noted in Table 11-4, it is logical that segment running time
sec midblock delay due to a pedestrian crosswalk. How should should depend on traffic flow rate; however, arterial research con-
the analysis be done? ducted for FHWA in the early 1980s did not establish a quantitative
Solution: The analysis should be done as above, but the 30 sec relation for such a dependence. It logically exists, but is not strong,
should be added to the third term in the denominator of Equation certainly not as strong as the effect of segment length on segment
11-1 when the computations are done. running time. Nor is it as strong as the substantial variation of
Example: Three consecutive segments on a north-south two- intersection control delay with traffic flow rate.
lane two-way facility (i.e., one lane in each direction) are 0.15, As a practical matter, computation of arterial travel speed for
0.17, and 0.13 mi long, respectively, all with a free-flow speed of different traffic flow rates is dominated by changes in control delay
30 mph. The arterial is a Classification IV principal arterial. What for the arterial through movements, whether or not the segment
is the northbound running time on the section? running time volume dependence is clearly identified. Thus, the
Solution: Note that it is reasonable to define a single section if absence of such an explicit factor does not affect the practical
all necessary conditions are met, including all lengths being within result, namely, the computation of arterial travel speed.
STEP 5—TABULATE INTERSECTION INFORMATION AND The correct delay to use in the arterial evaluation is the intersec-
COMPUTE DELAY tion control delay for the through movement. In general, the analyst
has the necessary information because the intersections are evalu-
To compute arterial or section speed, the analyst needs to deter- ated individually as part of the overall analysis. Geometric and
mine individual intersection delays. Because the arterial function traffic delay have already been taken into account in the segment
is to serve through traffic, the lane group that includes the through running times in Table 11-4.
traffic is used to characterize the arterial. The equations for computing average control delay per vehi-
cle are
d = d1 × PF + d2 + d3 (11-2)
0.5C [1 − (g/C)]2
d1 = (11-3)
1 − (g/C) [min(X,1.0)]
d2 = 900T [(X − 1) + √(X − 1)2 + 8kIX/Tc] (11-4)
where
d = control delay (sec/veh),
d1 = uniform delay (sec/veh),
d2 = incremental delay (sec/veh),
d3 = residual demand delay (sec/veh) (see Appendix 9-VI),
PF = uniform delay adjustment for quality of progression,
c = capacity of lane group (veh/hr),
X = v/c ratio for lane group with v representing demand flow
rate,
C = cycle length (sec),
g = effective green time for lane group (sec),
T = duration of the analysis period (hr),
k = incremental delay adjustment for actuated control, and
I = incremental delay adjustment for filtering and metering by
Figure 11-4. Types of segments. upstream signals.
Table 11-4. Segment Running Time Per Mile

arterial
classification i ii iii iv
free-flow
speed (mph) 55 50 45 45 40 35 35 30 35 30 25
segment running time per mile
length (mi) (sec/mi)
0.05 227 265
0.10 145 155 165 180 220
0.15 135 141 140 150 180
0.20 109 115 125 128 134 130 140 165
0.25 97 100 104 104 110 119 120 127 122 132 153
0.30 92 95 99 99 102 110
0.40 82 86 94 94 96 105
0.50 73 78 88 88 93 103
1.00 65 72 80 80 90 103
NOTES:
1. It is best to have an estimate of free-flow speed based on field observations of the facility or comparable facilities. If an estimate is lacking, however, the ana-
lyst can use the table by assuming the following default values:
Classification Free-Flow Speed (mph)
I 50
II 40
III 33
IV 30
2. For very long segment lengths on Classification I or II arterials (1 mi or longer), free-flow speeds may be used to compute running time per mile. These times
are shown in the entries for a 1.0-mile segment length.
3. If a Classification I or II arterial has a segment length less than 0.20 mi, the user should (a) reevaluate the classification and (b) if reevaluation confirms the
classification, use the values for 0.20 mi.
4. Likewise, Classification III or IV arterials with segment lengths longer than 0.25 mi should first be reevaluated (i.e., the classification should be confirmed). If
necessary, values for a segment of this length can be extrapolated.
5. Although this table does not show segment running time dependent on traffic flow rate, such a dependence is logical; however, the dependence of intersection
delay on traffic flow rate is much stronger and thus dominates in the computation of arterial travel speed.

11-10 urban streets
The components of these equations are discussed in the sections type is representative of operations characterized by highly dis-
that follow. persed platoons at isolated and noninterconnected signalized inter-
sections. It may also be used to represent coordinated operation
in which the benefits of progression are minimal.
Uniform Delay (d1)
T Arrival Type 4: Moderately dense platoon arriving in the
Equation 11-3 estimates control delay assuming perfectly uni- middle of the green phase or dispersed platoon containing 40 to
form arrivals and stable flow. It is based on the first term of 80 percent of the lane group volume arriving throughout the green
Webster’s delay formulation and is widely accepted as an accurate phase. This arrival type is representative of favorable progression
depiction of delay for the idealized case of uniform arrivals. Note quality on a two-way arterial.
that values of X higher than 1.0 are not used in the computation T Arrival Type 5: Dense to moderately dense platoon containing
of d1. more than 80 percent of the approach volume and arriving at the
start of the green phase. This arrival type is representative of highly
favorable progression quality, which may occur on routes that have
Incremental Delay (d2) a low to moderate number of side street entries and receive high
priority in the signal timing plan design.
Equation 11-4 estimates the incremental delay due to nonuniform T Arrival Type 6: This arrival type is reserved for exceptional
arrivals and individual cycle failures (random delay) as well as progression quality on routes with nearly ideal progression charac-
that caused by sustained periods of oversaturation (oversaturation teristics. This arrival type is representative of very dense platoons
delay). It is sensitive to the degree of saturation of the lane group progressing over a number of closely spaced intersections with
(X), the duration of the analysis period of interest (T ), the capacity minimal or negligible side street entries.
of the lane group (c), the type of signal control, as reflected by the
control type parameter (k), and the upstream filtering/metering pa- Arrival type is best observed in the field, but can be approxi-
rameter (I). The incremental delay term is valid for all degrees of mated by examining time-space diagrams for the arterial or street
saturation (X), including highly oversaturated lane groups. The equa- in question. The arrival type should be determined as accurately
tion assumes that no unmet demand causes residual queues at the as possible because it has a significant impact on delay estimates
start of the analysis period (T ). and LOS determination. Although no definitive parameters pre-
cisely quantify arrival type, the following ratio is a useful value:
1g2
Residual Demand Delay (d3) C
Rp = P × (11-5)
When demand from a previous time period causes a residual
queue to occur at the start of the analysis period (T ), additional where
delay is experienced by the vehicles arriving in the analysis period Rp = platoon ratio,
because the residual queues must first clear the intersection. A P = proportion of all vehicles in movement arriving during
procedure for determining residual demand delay is described in green phase,
detail in Appendix 9-VI. This procedure is also used to analyze C = the cycle length (sec), and
delay over multiple time periods, each having a duration (T ), in g = effective green time for movement (sec).
which a residual demand may be carried from one time period to
P may be estimated or observed in the field, while C and g are
the next.
computed from the signal timing. When P is estimated, note that
its value may not exceed 1.0. As shown in Table 11-5, the approxi-
Progression Adjustment Factor (PF) mate ranges of Rp are related to arrival type, and default values
are suggested for use in subsequent computations.
One of the most critical traffic characteristics that must be quan- Good signal progression results in the arrival of a high propor-
tified to complete an operational analysis of an arterial or a signal- tion of vehicles on the green. Poor signal progression results in
ized intersection is the quality of the progression. The parameter the arrival of a low percentage of vehicles on the green. The
that best describes this characteristic is the arrival type for each progression adjustment factor, PF, applies to all coordinated lane
lane group. This parameter is a general categorization that repre- groups, including both pretimed control and nonactuated lane
sents the quality of progression in an approximate manner. Six groups in semiactuated arterial control systems. Progression pri-
arrival types are defined for the dominant arrival flow as follows: marily affects uniform delay, and for this reason, the adjustment
is applied only to d1. The value of PF may be determined by
T Arrival Type 1: Dense platoon containing more than 80 per-
cent of the lane group volume and arriving at the start of the red (1 − P) f p
phase. This arrival type is representative of arterials that experience PF = (11-6)
(1 − g/C)
very poor progression quality as a result of conditions such as lack
of overall network signal optimization. where g/C = effective green time ratio, and fp = supplemental
T Arrival Type 2: Moderately dense platoon arriving in the adjustment factor for platoon arriving during the green.
middle of the red phase or dispersed platoon containing 40 to 80 The default values for fp are 0.93 for Arrival Type 2, 1.15 for
percent of the lane group volume arriving throughout the red phase. Arrival Type 4, and 1.0 for all other arrival types.
This arrival type is representative of unfavorable progression qual- As mentioned previously, the value of P may be measured in
ity on a two-way arterial. the field or estimated from the arrival type. If field measurements
T Arrival Type 3: Random arrivals in which the main platoon are carried out, P should be determined as the proportion of vehi-
contains less than 40 percent of the lane group volume. This arrival cles in the cycle that arrives at the stop line or joins the queue
Table 11-5. Relationship Between Arrival Type and Platoon Ratio (Rp)
arrival range of platoon default progression
type ratio (Rp) value (Rp) quality
1 ≤0.50 0.333 Very poor
2 >0.50 and ≤0.85 0.667 Unfavorable
3 >0.85 and ≤1.15 1.000 Random arrivals
4 >1.15 and ≤1.50 1.333 Favorable
5 >1.50 and ≤2.00 1.667 Highly favorable
6 >2.00 2.000 Exceptional
NOTE: Rp = P × (C/g).
Table 11-6. Uniform Delay (d1) Progression Adjustment Factor (PF)

green ratio arrival type (at)
(g/C) AT 1 AT 2 AT 3 AT 4 AT 5 AT 6
0.20 1.167 1.007 1.000 1.000 0.833 0.750
0.30 1.286 1.063 1.000 0.986 0.714 0.571
0.40 1.445 1.136 1.000 0.895 0.555 0.333
0.50 1.667 1.240 1.000 0.767 0.333 0.000
0.60 2.001 1.395 1.000 0.576 0.000 0.000
0.70 2.556 1.653 1.000 0.256 0.000 0.000
Default, fp 1.00 0.93 1.00 1.15 1.00 1.00
Default, Rp 0.333 0.667 1.000 1.333 1.667 2.000
NOTES: 1. PF = (1 − P)fp /(1 − g/C).
2. Tabulation is based on default values of fp and Rp.
3. P = Rp ∗ g/C (may not exceed 1.0).
4. PF may not exceed 1.0 for AT 3 through AT 6.
(stationary or moving) while the green phase is displayed. The a k-value of 0.50 is used. This value is based on a queuing process
value of PF may be computed from measured values of P using with random arrivals and uniform service equivalent to the lane
the default values for fp. Alternately, Table 11-6 may be used to group capacity. Actuated controllers, on the other hand, have the
determine PF as a function of the arrival type based on the default ability to adjust the green time to the cyclic demand, thus reducing
values for P (i.e., Rp × g/C) and fp associated with each arrival the overall incremental delay component. The delay reduction de-
type. If PF is estimated by Equation 11-6, its calculated value may pends in part on the controller’s unit extension and the prevailing
exceed 1.0 for Arrival Type 4 with extremely low values of g/C. degree of saturation. Recent research indicates that lower unit
As a practical matter, PF should be assigned a maximum value extensions result in lower values of k and d2; however, when the
of 1.0 for Arrival Type 4. This constraint has already been taken degree of saturation approaches 1.0, an actuated controller behaves
into consideration in the values shown in Table 11-6. similarly to a pretimed controller, resulting in k-values of 0.50 at
Application of the progression adjustment factor requires de- X ≥ 1.0. Table 11-7 illustrates recommended k-values for pretimed
tailed knowledge of offsets, travel speeds, and intersection signal- and actuated lane groups with different unit extensions and degrees
ization. When delay is estimated for future situations involving of saturation.
coordination, particularly when alternatives are analyzed, it is ad-
visable to assume Arrival Type 4 as a base condition for coordi- Upstream Filtering/Metering Adjustment Factor (I )
nated lane groups, in which case P may be estimated from Table
11-5 and Equation 11-5 as Rp × g/C. Arrival Type 3 should be The incremental delay adjustment term I in Equation 11-4 incor-
assumed for all uncoordinated lane groups. porates the effects of metering arrivals from upstream signals. For
Movements made from exclusive left-turn lanes on protected isolated signals, an I-value of 1.0 is used. This value is based on
phases are not usually provided with good progression. Thus, Ar- a queuing process with random arrivals such that the ratio of the
rival Type 3 is usually assumed for coordinated left turns. When variance to mean arrivals per cycle is equal to 1.0. Upstream
the actual arrival type is known, it should be used. When the signals decrease the variance by metering arrivals at the down-
coordinated left turn is part of a protected-permitted phasing, only stream intersection, thus reducing the ratio of the variance to mean
the effective green for the protected phase should be used to deter- arrivals per cycle. The I-value and the resultant delay reduction
mine the PF since the protected phase is normally associated with depend on the through movement’s degree of saturation at the
platooned coordination. When a lane group contains movements upstream intersection and the amount of entering and exiting traffic
that have different levels of coordination, a flow-weighted average between the two intersections. Table 11-8 illustrates recommended
of P should be used in determining the PF. values of I for different upstream degrees of saturation at the
upstream intersection.
Actuated Control Adjustment Factor (k) Example
The incremental delay adjustment term k in Equation 11-4 incor- Delay is a complicated variable that is sensitive to a variety of
porates the effect of controller type on delay. For pretimed signals, local and environmental conditions. The procedures provided here
11-12 urban streets
Table 11-7. Recommended k-Values for Lane Groups Under Actuated and Pretimed Control
unit degree of saturation (X)
extension
(UE) (sec) ≤0.50 0.60 0.70 0.80 0.90 ≥1.0
≤2.0 0.04 0.13 0.22 0.32 0.41 0.50
2.5 0.08 0.16 0.25 0.33 0.42 0.50
3.0 0.11 0.19 0.27 0.34 0.42 0.50
3.5 0.13 0.20 0.28 0.35 0.43 0.50
4.0 0.15 0.22 0.29 0.36 0.43 0.50
4.5 0.19 0.25 0.31 0.38 0.44 0.50
5.0 0.23 0.28 0.34 0.39 0.45 0.50
NOTES: 1. k = 0.50 for nonactuated lane groups.
2. For a given UE and its kmin value at X p 0.5, k = (1 − 2kmin) (X − 0.5) + kmin.
3. For UE > 5.0, extrapolate to find k, keeping k ≤ 0.5.
Table 11-8. Recommended I-Values For Lane Groups With Upstream Signals
degree of saturation at upstream intersection (Xu)
0.40 0.50 0.60 0.70 0.80 0.90 ≥1.0
I 0.922 0.858 0.769 0.650 0.500 0.314 0.090
NOTE: I = 1.0 − 0.91Xu2.68 and Xu ≤1.0.
present reasonable estimates for delays expected for average condi- is a summary worksheet for intersection delay computations. An
tions. They are most useful when used to compare operational additional blank worksheet may be found in Appendix II to this
conditions for various geometric or signalization designs. When chapter.
existing conditions are evaluated, it is advisable to measure delay
in the field. Appendix III of Chapter 9 contains guidelines for
intersection delay measurements using lane occupancy and volume STEP 6—COMPUTE AVERAGE TRAVEL SPEED
counts.
Description: Consider an arterial segment with a through lane
group with two lanes, a demand volume of 1,500 veh/hr, and The average speed is to be computed by segment and over the
peak hour factor of 0.91. Further, a pretimed signal has a cycle entire arterial. It is recommended that the user also prepare a speed
length of 90 sec, the g/C ratio is 0.60, and X or the v/c ratio profile of the facility and supplement the LOS assessment with
is 0.90. Vehicles arrive as a dense platoon at the beginning of insights gained from the speed profile and the levels of service of
the green. What is the estimated control delay for the through the individual intersections.
lane group? Figure 11-6 shows a worksheet, with some illustrative data filled
Solution: To use Equations 11-3 and 11-4 to compute control in, which is provided to ease the task of assembling the
delay, it is necessary to know C, g/C, X, and c. The last term must information.
be computed. Equation 11-1 is used to compute the arterial speed for each
The adjusted demand flow rate is segment and for the overall facility. Performing these computations
results in the speed profile shown in Figure 11-7. For segments 1
v = (1,500/0.91) = 1,648 veh/hr
and 9, the running time per mile for a segment 0.10 mi long is
Once it is known that X or the v/c ratio is 0.90, c can be calcu- used, but is multiplied by the actual segment lengths.
lated as Sample Computation. Fourth Avenue is a principal arterial of
intermediate design with a 35-mph free-flow speed (Figure 11-6).
c = v/(v/c) From Table 11-3, it is arterial Classification III. In Section 2 of
c = 1,648/(0.90) = 1,831 veh/hr the arterial, the average segment length is 0.20 mi. From Table
11-4, the running time per mile is 128 sec.
The intersection control delay is computed as d = 17.8 + 3.5 = The total running time in the section is given by
21.3 sec/veh.
From the description of the arriving vehicles, the arrival type is 128 × (0.20 + 0.20 + 0.20) = 76.8 sec
5. On the basis of a pretimed signal, a g/C ratio of 0.60, and
Arrival Type 5, Tables 11-6, 11-7, and 11-8 are consulted to find The control delay for the arterial through movements at the
PF = 0.0, k = 0.5, and I = 1.0, respectively. Thus, the estimated three intersections in Section 2 is given in Figure 11-6 as 5.0 +
control delay = 3.5 sec/veh. 7.0 + 10.0 = 22.0 sec, so the total travel time is 76.8 + 22.0 =
The computations must be done for each signalized intersection 98.8 sec.
or obtained from the results of Chapter 9 evaluations. Figure 11-5 The arterial speed in the section is 3,600 x 0.60/98.8 = 21.9 mph.
Figure 11-5. Worksheet for summary of arterial intersection delay estimates.
STEP 7—ASSESS LEVEL OF SERVICE tations while driving on that facility and the lower the speed associ-
ated with a given level of service. Thus, a Classification III arterial
A distinct set of arterial LOS criteria has been established for provides LOS B at a lower speed than does a Classification I
each arterial classification. These sets of criteria are based on the arterial.
differing expectations drivers are judged to have for the different The analyst should be aware of this relationship in explaining
classes of arterials. before-and-after assessments of arterials when upgrading is in-
In the arterial LOS definitions, both the free-flow speed of the volved. If reconstruction results in upgrading a facility from Classi-
arterial classification and the intersection LOS definitions are taken fication II to Classification I, it is possible that the level of service
into account. In general, the arterial LOS criteria are based on the will not change (or may even technically degrade), despite higher
smooth and efficient movement of through traffic along an entire average speed and other improvements, because expectations are
arterial. Therefore, it is necessary to expect less delay per segment higher.
than the corresponding intersection level of service. Note that the concept of an overall arterial level of service is
Table 11-1 gives the arterial LOS criteria for each of the four generally meaningful only when all segments on the arterial are
arterial classifications. The lower the arterial classification (i.e., of the same classification. If different arterial classifications are
the larger the classification number), the lower the driver’s expec- represented, the LOS criteria are different.
11-14 urban streets
Figure 11-6. Worksheet for computation of arterial level of service.

Figure 11-7. Speed profile by arterial section.
III. PLANNING APPLICATIONS
OBJECTIVES this assumption, planning application results should not be used

for intersection design or traffic operations analyses.
The objective of an arterial LOS analysis at a planning level is
to approximate the operating conditions of the facility. A major DATA REQUIREMENTS
use for this type of analysis is related to growth management
issues. The accuracy of the planning LOS analysis is largely depen- To conduct a planning analysis, traffic, roadway, and signal
dent upon the degree of generalization of input data and should not input values or assumed defaults are needed for the following
be used for design or operational analyses. The planning method is characteristics:
most applicable when Traffic characteristics:
Annual average daily traffic (AADT),
1. LOS estimates are desired,
Planning analysis hour factor (K),
2. Field data are lacking,
Directional distribution factor (D),
3. Relatively long planning horizons are used, and
Peak hour factor (PHF),
4. Individuals with limited transportation planning experience
Adjusted saturation flow rate,
are involved.
Percentage of turns from exclusive lanes;
A major difference between the planning analysis of signalized Roadway characteristics:
intersections and that of arterials is the treatment of turning vehi- Number of through lanes (N),
cles. Whereas the purpose of a signalized intersection is to move Free-flow speed,
vehicles (including turning vehicles) past a point, the purpose of Arterial classification,
an arterial is to move (through) vehicles over a reasonable length Medians,
of roadway. Because the emphasis of an arterial is on through Left-turn bays or exclusive left-turn lanes;
Signal characteristics:
movement, the major simplifying assumption in this planning ap-
Arrival type,
plication is that left turns are accommodated by providing left-turn
Signal type,
bays at major intersections and controlling the left-turn movement
Cycle length (C),
with a separate phase that is properly timed. With this simplifying
Effective green ratio (g/C).
assumption, many of the inputs and complexities of intersection
analyses can be handled abstractly as default values, allowing a Some of these characteristics are discussed in the remainder of
relatively easy-to-use planning application; however, as a result of this section.
11-16 urban streets
Planning Analysis Hour Factor (K ) saturation flow rate may be reduced 5 percent for roadways that
do not have medians.
The planning analysis hour factor represents the percentage of
AADT occurring in the peak hour. For planning purposes many
possible peak hours may be appropriate. K30 (the 30 highest hour Left-Turn Bays or Exclusive Left-Turn Lanes
volumes of the year) is widely accepted as the design hour in
nonurban areas. K100 approximates the typical weekday peak hour Left-turn bays or lanes are storage areas at signalized intersec-
during the peak season in developed areas and is frequently used tions to accommodate left-turn movements. These bays or lanes
in long-range urban transportation models. K200 to 400 is a better must be long enough to accommodate left turns without impeding
representation of a typical peak hour of the year. In many urban the through movement. For planning purposes, the saturation flow
areas, general ranges for K30, K100, and K200 to 400 are 8.5 to rate should be reduced 20 percent for roadways that do not have
11.0 percent, 8.0 to 10.0 percent, and 7.0 to 9.0 percent, respec- left-turn bays at major intersections. (This value is a 15 percent
tively. The analyst needs to determine the appropriate peak hour. additional reduction for a roadway that does not have a median.)
Adjusted Saturation Flow Rate Effective Green Ratio (g /C)
Numerous factors affect the saturation flow rate per lane (see The parameter g/C is the ratio of the time allocated for the
Chapter 9). For a planning analysis, these adjustments may reason- through traffic movement (red clearance minus the startup lost
ably be combined and multiplied by the ideal saturation flow rate time minus effective green time) to the cycle length (C). An arteri-
to determine an adjusted saturation flow rate. On the basis of an al’s through g/C for each intersection is desirable; however, for
ideal saturation flow rate of 1,900 passenger cars per hour of green broad planning purposes a weighted g/C may be appropriate. The
time per lane (pcphgpl), a reasonable range for urban arterials weighted g/C of an arterial is the average of the critical-intersection
during the peak hour is 1,750–1,850 pcphgpl. through g/C and the average-intersection through g/C. For exam-
ple, if an arterial section has three signalized intersections with
Percentage Turns from Exclusive Lanes effective green ratios of 0.4, 0.7, and 0.7, the critical intersection
has a g/C of 0.4 (the lowest g/C); the average intersection has a
Turns from exclusive lanes represent the percentage of vehicles g/C of 0.6 [(0.4 + 0.7 + 0.7)/3], and the weighted g/C is 0.5 [the
performing left- or right-turn movements at signalized intersections average of the critical g/C and the average g/C, (0.6 + 0.4)/2].
from lanes dedicated solely to turning movements. The planning Thus, the weighted g/C takes into account the adverse impact of
methodology assumes that left turns are accommodated by separate the critical intersection and the overall quality of flow for the
lanes and phases so that they have minimal effect on through arterial length. Average weighted effective green ratios for arterials
vehicles. Where a separate right-turn lane exists, it is reasonable vary by road purposes and by areas.
to add the percentage of right turns to the percentage of left turns
(assuming a left-turn bay or lane) to determine the percentage of
COMPUTATIONAL STEPS
turns from exclusive lanes.
The calculation process for determining arterial level of service

Number of Through Lanes is illustrated in Figure 11-8 and consists of the following steps:
Because significant delays seldom occur in midblock locations, 1. Convert daily volumes to the planning analysis hour by an
a parameter of importance is the number of through and shared appropriate planning analysis hour factor (K).
right-turn lanes at signalized intersections; however, when signifi- 2. Multiply K by the directional distribution factor (D) to obtain
cant midblock delays occur or reasonable lane continuity between hourly directional volumes.
intersections is not maintained, caution should be used in strictly 3. Adjust the hourly directional volumes based on PHF and
applying the concept of the number of such lanes. turns from exclusive lanes to yield estimated through volumes for
15-min service flow rates.
4. Calculate the running time on the basis of arterial classifica-
Free-Flow Speed tion, intersection spacing, and free-flow speed.
5. Calculate the intersection control delay on the basis of ad-
For planning purposes, an arterial’s free-flow speed should be justed saturation flow rates, number of lanes (N), arrival type,
based on actual studies of the road or on studies of similar roads signal type, cycle length (C), and g/C for each intersection using
and should be consistent with arterial classifications. The actual Equations 11-3 and 11-4.
or probable posted speed limit may be used as a surrogate for free- 6. Calculate the average travel speed using running time and
flow speed if comparable roadway free-flow studies do not exist. intersection control delay.
7. Obtain arterial level of service on the basis of the average
travel speed.
Medians
Calculation 8 in Section IV of this chapter illustrates the computa-
Medians are painted, raised, or grassed areas that separate optional steps in a planning analysis.
posing midblock traffic lanes and that are wide enough to serve Frequently in a planning analysis, however, the level of service
as bays for turning vehicles. For planning purposes, the adjusted may be given and the desired outcome is a volume—hourly direc-
tional, hourly nondirectional, or daily. For such applications, the

calculation process is essentially reversed, as follows:
1. Select the level of service and the corresponding average
travel speed (range or minimum) based on arterial type and free-
flow speed as selected from Table 11-1.
2. Compute the total section running time for the given arterial
type, number of intersections, free-flow speed, and section length.
3. Calculate the control delay at all intersections (d) using Equa-
tion 11-1 and Steps 1 and 2.
4. Compute v by using the value of average control delay, given
number of lanes, adjusted saturated flow rate, arrival type, C, and
weighted g/C in Equations 11-2 to 11-4.
5. Determine the hourly directional volume for the design hour
using the percentage of turns from exclusive lanes, basic through
15-min volumes, and the PHF.
6. Calculate the two-way hourly directional volume for the de-
sign hour using the hourly directional volume and the directional
distribution factor.
7. Determine AADT using the two-way hourly directional vol-
ume and the applicable K factor.
INTERPRETATION OF RESULTS
Planning analysis results range from a rough estimate of level

of service to an operational analysis, depending primarily on the
degree to which default values are used as input. For example,
using statewide defaults for appropriate traffic, roadway, and signal
characteristics results in rough LOS estimates. Using area- or road-
way-specific data but treating all signal characteristics the same
(e.g., using a weighted g/C approach) should provide more accurate
LOS estimates. Using specific traffic, roadway, and signal data
for each road segment and traffic signal should provide an even
Figure 11-8. Arterial LOS calculation process. more accurate LOS estimate. The next level of precision is a
detailed treatment of turning movements and signal timing, which
is essentially an operational analysis except that projected rather
than actual traffic volumes are used.
CALCULATION 1—ARTERIAL CLASSIFICATION nitions to be used in evaluating the arterial. Further, lacking more
specific information, one can expect a free-flow speed on the order
1. Description: An arterial with three lanes in each direction of 33 mph (refer to the top of Table 11-1), with a range of 30 to
and signal spacing of 0.15 mi passes through an area with moderate 35 mph.
roadside development. Virtually all of the traffic passes through
the area; there is very little pedestrian activity. Identify the arterial
classification.
CALCULATION 2—COMPUTATION OF ARTERIAL LEVEL OF
2. Solution: To determine the arterial classification, it is neces-
SERVICE
sary to decide the design and functional categories of the arterial
and then to use Table 11-3 to specify the arterial classification.
The statement that ‘‘virtually all of the traffic passes through the 1. Description: A multilane divided facility functions as a princi-
area’’ defines the functional category: the roadway is a principal pal arterial. There is significant access control, no parking, and a
arterial. Table 11-2 can be used to assist in determining the design signal spacing of approximately 0.30 mi between pretimed signals.
category. Note that the arterial is a multilane undivided facility The arterial has little roadside development, two lanes in each
with approximately seven signals per mile (based on 0.15-mi spac- direction, and a measured free-flow speed of 39 mph.
ing), moderate roadside development, and very little pedestrian Detailed information on the intersection parameters and the arte-
activity. The design category is therefore intermediate. rial segments for the southbound flow is contained in Figures
Referring to Table 11-3, one concludes that the arterial is Classi- 11-9 and 11-10. The progression is excellent in the southbound
fication III. This information is used in determining the LOS defi- direction.
11-18 urban streets
Figure 11-9. Calculation 2, description: using worksheet for summary of arterial intersection delay estimates.
Determine the arterial level of service by segment and for the T Multilane divided,
entire facility. Do not aggregate the segments. T Significant access control,
2. Solution: This solution proceeds according to the steps out- T No parking,
lined in Figure 11-2. In some applications, it may not be necessary T Little roadside development,
to perform all steps, or it may be easier to do certain steps before T Seven signals in 2.1 mi (three signals per mile).
others. For instance, if the intersection evaluations have been done
previously (or if the summary information is available), that infor- The facility clearly belongs to the suburban design category.
mation may be entered on the appropriate worksheet (Figure On the basis of a functional category of principal arterial and a
11-5) and the control delay computed before the arterial running design category of suburban, the facility is found to be a Classifica-
times are computed. tion II arterial by referring to Table 11-3.
Step 1. Establish Arterial To Be Considered

Step 3. Define Arterial Sections
This step has been performed in the preceding statement.
Step 2. Determine Arterial Classification Step 3 may be skipped because the instructions in the description
were not to aggregate the segments. Nonetheless, note that some
The functional category, principal arterial, is given. The design sections could be aggregated on the basis of average segment
category may be established by referring to Table 11-2 and noting lengths and volume pattern. For instance, the following aggrega-
the following characteristics: tions could be made:
Figure 11-10. Calculation 2, description: using worksheet for computation of arterial level of service.

11-20 urban streets
Segment Section Figure 11-13 shows the speed profile for the arterial. This draw-
1 1 ing is a valuable depiction of the operation and should be con-
2 1 structed as part of each evaluation.
3 2
4 2
Step 7. Assess Level of Service
5 2
6 3
With all of the preliminary work done, the final determination
7 3
of LOS values is straightforward. The speeds computed in the
summary arterial LOS worksheet can be compared with the defini-
If the volume differences make the user uncomfortable with this
tions for the appropriate arterial classification (in this case, Classi-
aggregation, it could be checked after the intersection delay has
fication II, as established in Step 2) given in Table 11-1. These
been estimated.
are entered on the arterial LOS worksheet in Figure 11-12 and,
together with the intersection levels of service determined pre-
Step 4. Compute Running Time viously, on the speed profile in Figure 11-13.
As stated in Section I, intersection LOS values are generally
The arterial is Classification II with a free-flow speed of 39 better than the arterial LOS values. This difference is logical,
mph, which establishes the relationship to be used for the running because an intersection with less than 5 sec of delay per vehicle
time computation (see Table 11-4). is certainly LOS A, whereas an arterial on which one could travel
Consider Segment 1. For a Classification II arterial and a seg- at a speed of 39 mph but instead has to travel at 30 to 35 mph is
ment length of 0.20 mi, Table 11-4 indicates a running time per somewhat less than LOS A.
mile of 115 sec for a free-flow speed of 40 mph and 125 sec for
a free-flow speed of 35 mph. It may be interpolated that for 39 mph CALCULATION 3—COMPUTATION OF ARTERIAL LEVEL OF
the running time would be 117 × 0.20 = 23.4 sec. This information is SERVICE
entered on the computation of arterial LOS worksheet.
1. Description: The northbound side of the arterial described in
Step 5. Compute Intersection Delay Calculation 2 has intersection traffic as shown in Figure 11-14 and
very poor progression, with virtually the entire northbound platoon
Figure 11-11 shows the arterial intersection delay estimates for arriving in the middle of the red at each intersection.
Calculation 2. Note that because this computation is an arterial Determine the arterial level of service by segment and for the
evaluation, the information must be for the lane group containing entire facility. Do not aggregate the segments.
the principal part of the through movement. This information is 2. Solution: The calculations for this solution are identical in
generally available for the desired lane group from evaluations form and sequence to those of Calculation 2 and will not be re-
of individual intersections based on procedures described in peated; however, certain key points must be highlighted:
Chapter 9. T The evaluation of an arterial is by direction, and a two-way
Equations 11-3 and 11-4 are used to compute the uniform delay arterial usually requires two evaluations, one for each direction,
(d1) and the incremental delay (d2), which can then be entered on just as was required in Calculation 2.
the summary worksheet. T The arrival types in the two directions are generally different
The selection of the arrival type for the approaching vehicles is because the progression of the signal timing is often set to favor
a special consideration. In this case, it is straightforward because one direction over the other. This difference has a major impact
of the information given in the description that progression is on the intersection delay estimates.
excellent in the southbound direction. When this information is T It is useful to include the segment numbers in the speed profile
matched with the arrival type definitions, Arrival Type 5 is selected (as shown in Figure 11-13), to make the final presentation clear.
because it is defined as a dense platoon arriving at the beginning It is also useful to mark the direction of travel clearly.
of the green phase with a highly favorable progression quality. T The intersections analyzed are those at the input and output
Table 11-6 shows the progression factors (PFs) for the pretimed ends of each segment.
signals and arrival types given. As shown in Figure 11-11, all the
intersections have g/C ratios of 0.60, so a PF of 0.00 is used for The results of the computations are shown in Figures 11-15 and
all of them. 11-16, and the speed profile is shown in Figure 11-17, which for
The results of the intersection computations are shown trans- comparative purposes also shows the southbound speed profile
ferred to the arterial LOS worksheet in Figure 11-12. as well as the intersection and arterial levels of service for both
directions.
One additional point stands out: the determination of arrival
Step 6. Compute Average Travel Speed type so that correct PFs may be selected. The description states
that there is ‘‘very poor progression, with virtually the entire north-
With the running time from Step 4 and the delay time from bound platoon arriving in the middle of the red at each intersec-
Step 5, the computations may be summarized using the arterial tion.’’ It is important to note that this situation is not the worst
LOS worksheet. The completed worksheet is shown in Figure 11- condition: a careful reading of the arrival type descriptions makes
12, with the calculation for each section (in this case, each seg- it clear that Type 2 covers the present case, whereas the worst
ment) identical in form to that shown on the bottom of the work- case—Type 1—is reserved for a dense platoon arriving at the
sheet for the entire arterial. beginning of the red phase.
Figure 11-11. Calculation 2, solution: using worksheet for summary of arterial intersection delay estimates.
CALCULATION 4—EFFECT OF TRAFFIC FLOW RATE ON found from Table 11-4 to be 128 sec for a segment length of 0.20
ARTERIAL LEVEL OF SERVICE mi. The running time in the segment is therefore 128 × 0.20 =
25.6 sec.
1. Description: An arterial with two lanes in each direction and The intersection control delay is based on Equations 11-3 and
a 35-mph free-flow speed has been found to be a Classification 11-4 and the application of the PF. Two parameters are given (C =
III arterial. Ten signals are spaced 0.20 mi apart. The intersections 60 sec and g/C = 0.50). The other two, arterial lane group capacity
all have pretimed signals with a 60-sec cycle length and g/C of (c) and v/c ratio (X), are not directly given.
0.50. The progression is excellent. Without specific information on the lane group capacity, it is
For a range of adjusted traffic demand from a flow rate of 600 both possible and necessary to compute c = 1,600 × 2 × 0.50 =
to 1,600 veh/hr, plot the arterial segment speed and find the arterial 1,600 veh/hr, for all segments. If the g/C differed from segment
level of service, as well as the intersection levels of service. to segment, the computed value would also differ. When this rela-
2. Solution: The relationships shown in this chapter for arterial tionship is used for a specific site, the evaluation becomes highly
running time do not depend explicitly on arterial volume or flow approximate; however, this sample calculation is for a typical or
rate (see Note 5, Table 11-4). The arterial speed is sensitive to
representative arterial.
traffic volume because the intersection delay is dependent on that
In the information given, the adjusted demand flow rate varies
volume. Recall that the basic relation is Equation 11-1, which is
from v = 600 veh/hr to v = 1,600 veh/hr. For each value of v, the
repeated here for convenience:
corresponding value of X = v/1,600, because c = 1,600 veh/hr was
3,600 * (length) just computed above.
ART SPD = (11-1)
[(running time/mile) * (length) + (∑ inters. control delay)] The arrival type is 5 because the progression is excellent. The
PF is selected from Table 11-6 for Arrival Type 5 and a g/C of
For the stated situation, the segment running time per mile is 0.50. The results of the computations are given in Table 11-9. The
11-22 urban streets
Figure 11-12. Calculation 2, solution: using worksheet for computation of arterial level of service.
estimated control delay is the uniform delay multiplied by the PF on control delay and is shown for the lane group containing
plus the incremental delay. the through traffic.
The levels of service are identified by referring to Table 11- Figure 11-18 is a plot of arterial segment speed as a function of
1 for a Classification III arterial and to Table 9-1 for the arterial volume for the stated condition of a 0.20-mi segment length.
intersections. Note that the intersection level-of-service is based Note that the intersection approach delay ranges from 13 to 32 per-
Figure 11-13. Speed profile for Calculation 2, southbound traffic.
cent of the total time spent on the segment, depending on the traffic The levels of service are again identified by referring to Table
flow rate. Figure 11-19 is a plot of arterial segment average travel 11-1 for a Classification III arterial and to Table 9-1 for the
speed as a function of arterial flow rate for a 0.10-mi segment length. intersections.
For comparative purposes, the plot for a 0.20-mi segment length is Table 11-10 illustrates the following point: because of the
also shown. The facts that the speeds are much lower and that the close signal spacing and the control delay per unit length, it is
arterial level of service is now significantly lower than the intersec- possible for arterial level of service to be two or even three
tion level of service deserve attention. levels worse than that of a typical intersection. (As will be
First, it is necessary to observe that with the 0.10-mi segment, shown in Calculation 7, it is also possible for the arterial level
the intersection delay per mile has increased (relative to that in of service to be better than the intersection level of service
Calculation 4) because there are now more intersections per mile: when the segment is very long.)
5 intersections per mile for the 0.20-mile segment have become Note that in this calculation, the intersection delay ranges from 20
10 intersections per mile for the 0.10-mi segments. Thus, a delay to 45 percent of the total time spent on the segment, depending on
of 8.0 sec/veh per intersection now contributes 10 × 8.0 = 80 sec/ the traffic flow rate. In Calculation 4, the range was 13 to 32 percent.
mi to the arterial travel time, whereas it was 5 × 8.0 = 40 sec/mi
in the previous computation. Thus, two radically different arterials
are being compared. CALCULATION 6—EVALUATION BASED ON FIELD DATA
The driver’s expectation is more demanding for an arterial than
for an individual intersection. With 10 signals per mile, very little 1. Description: On a given multilane two-way divided arterial
delay per intersection is required to degrade the quality of flow with left-turn bays and good access control, the free-flow speed
for through traffic; however, any intersection with less than 5.0 is measured along its length as 45 mph. The following data are
sec of stopped delay is operating rather well (i.e., LOS A is a collected along its eight eastbound segments, using the field data
realistic evaluation of such an intersection). procedures of Appendix I:
Average Average
CALCULATION 5—EFFECT OF TRAFFIC FLOW RATE AND Travel Time Control Delay
LENGTH ON ARTERIAL LEVEL OF SERVICE Segment Length (mi) (sec) (sec/veh)
1 0.20 28.3 3.4
1. Description: Reevaluate Calculation 4 using a signal spacing 2 0.15 19.2 1.7
of 0.10 mi. All other information is the same as that given in 3 0.15 21.8 3.6
4 0.20 29.4 5.3
Calculation 4, including the arterial classification. 5 0.25 49.7 17.6
2. Solution: Numerically, the computations are the same as in 6 0.25 40.6 0.5
Calculation 4, and all the introductory remarks are the same. The 7 0.25 35.2 6.2
results of the computations are given in Table 11-10. 8 0.20 28.1 3.2

11-24 urban streets
Figure 11-14. Calculation 3, description: using worksheet for summary of arterial intersection delay estimates.
These data are based on an appropriate number of travel time II. The ranges of free-flow speed shown in Table 11-4 indicate
runs that include both the running time and the intersection con- that a measured free-flow speed of 45 mph is consistent with
trol delay. arterial Classification II.
Find the arterial level of service, by segment and for the entire The field data can also be used to compute the arterial speed
facility, as well as the intersection levels of service. by segment and for the entire facility without any need to use
2. Solution: To determine the arterial classification, consult Table 11-4. The computations of the arterial speed are shown in
Tables 11-2 and 11-3 and note that the summary of calculations on the completed arterial LOS work-
sheet in Figure 11-20. The speed calculations are straightforward;
T The facility is multilane divided,
for instance, for Segment 1, ART SPD = 3,600 × 0.20/28.3 =
T Access control is good,
25.4 mph.
T There are eight signals in 1.65 mi, or about five signals per
The LOS determination is made by referring to Table 11-1 for
mile.
arterial Classification II and applying the definitions; for instance,
It is likely that the design category would be suburban on the basis Segment 1 with a computed speed of 25.4 mph is LOS C.
of Table 11-2. Because the facility is a principal arterial, Table Figure 11-21 shows the speed profile for the arterial and graphi-
11-3 leads one to determine that it belongs in arterial Classification cally demonstrates where the problem occurs. Note that the overall
level of service does not highlight the problem as well as the speed CALCULATION 7—ARTERIAL WITH LARGE SIGNAL
profile or the set of segment levels of service does. SPACINGS
The field data also allow a direct determination of intersection
levels of service, based on measured control delay. With the LOS 1. Description: Route 25 is a suburban arterial with a free-flow
definitions of Table 9-1, the determination is straightforward: speed of 51 mph measured in field studies. It is an undivided
facility, with two lanes in each direction and left-turn bays, and
Measured
Intersection Control Delay is dominated by its signals. A pretimed set of signals is used on
Segment LOS (sec/veh) the portion of the facility of interest. The following information
1 A 3.4 is available for westbound traffic for the period being studied:
2 A 1.7
3 A 3.6
4 A 5.3 Length C c
5 B 17.6 Segment (mile) (sec) g/C X (veh/hr)
6 B 10.5 1 0.7 70 0.60 0.89 1,800
7 A 6.2 2 0.6 70 0.57 0.97 1,710
8 A 3.2 3 0.7 70 0.60 0.94 1,800
4 0.7 70 0.60 0.94 1,800
These levels of service are also shown in Figure 11-21. 5 0.7 70 0.60 0.94 1,800

11-26 urban streets

Figure 11-17. Speed profile for Calculation 3, northbound traffic.
Table 11-9. Computations for Sample Calculation 4

dist = 0.20 miles, cycle length = 60 seconds, g/C = 0.50
intersection segment average
v/c uniform incremental control approach run- sum travel arterial
flow capacity ratio delay delay level of control ning time speed level of
(vph) (vph) X d1 PF I d2 service delay time (sec) (mph) service
600 1,600 0.38 9.2 0.33 0.93 0.6 A 3.7 25.6 29.3 24.6 B
700 1,600 0.44 9.6 0.33 0.90 0.8 A 4.0 25.6 29.6 24.3 B
800 1,600 0.50 10.0 0.33 0.86 1.0 A 4.3 25.6 29.9 24.1 B
900 1,600 0.56 10.4 0.33 0.81 1.2 A 4.6 25.6 30.2 23.8 C
1,000 1,600 0.63 10.9 0.33 0.74 1.4 A 5.0 25.6 30.6 23.5 C
1,100 1,600 0.69 11.4 0.33 0.67 1.6 A 5.4 25.6 31.0 23.2 C
1,200 1,600 0.75 12.0 0.33 0.58 1.9 A 5.9 25.6 31.5 22.8 C
1,300 1,600 0.81 12.6 0.33 0.48 2.3 A 6.5 25.6 32.1 22.4 C
1,400 1,600 0.88 13.3 0.33 0.36 2.7 A 7.2 25.6 32.8 22.0 C
1,500 1,600 0.94 14.1 0.33 0.23 3.5 A 8.2 25.6 33.8 21.3 C
1,600 1,600 1.00 15.0 0.33 0.09 6.8 B 11.8 25.6 37.4 19.3 C
The signal progression is good, with less than 10 percent of the Table 11-6 indicates the following PF values for pretimed con-
through traffic stopping. trol and Arrival Type 5:
Determine the arterial level of service by segment and for the Progression
entire facility. g/C Ratio Factor, PF
2. Solution: On the basis of the free-flow speed, the facility is 0.50 0.333
arterial Classification I. Refer to Table 11-1 or Table 11-4. 0.60 0.000
The intersection delay may be computed using Equations 11-3
and 11-4, with the computations summarized on the worksheet for From these values, the PF for a g/C ratio of 0.57 may be interpo-
summary of arterial intersection delay estimates (Figure 11-22). lated as 0.099.
On the basis of PF descriptions in this chapter, the arrival type Given the free-flow speed of 51 mph and the relatively long
is 5—a dense platoon arriving at the beginning of the green phase, signal spacing, the free-flow speed can be used as the arterial
a highly favorable progression. This judgment is based on the speed in computing the running time:
given condition that the signal progression is good, with less than
10 percent of the through traffic stopping. Segment running time = 3,600 × (segment length)/(ART SPD)
11-28 urban streets
Figure 11-18. Sample calculation speed as a function of arterial flow rate.
Figure 11-19. Calculation 5 speed as a function of arterial flow rate on two different segment lengths.
For instance, in Segment 1, segment length of 0.50 mi and a free-flow speed of 50 mph, the
segment running time is (78/72) or 1.08 times the value of a 1.0-
Segment running time = 3,600 × (0.70)/(51 mph) = 49.4 sec
mi segment. Thus, more precise estimates for such segment lengths
To this computed running time is added the intersection delay time as 0.60, 0.80, and 0.90 mi could be generated for a free-flow speed
in the usual way, as shown in Figure 11-23. of 51 mph by similar logic. However, the better and more accurate
If Table 11-4 is inspected carefully, a more precise estimate of approach would be to rely on field data for such an arterial.
the computed running time can be generated. For instance, for a Figure 11-23 also indicates the level of service for each arterial
Table 11-10. Computations for Sample Calculation 5

dist = 0.10 miles, cycle length = 60 seconds, g/C = 0.50
intersection segment average
v/c uniform incremental control approach run- sum travel arterial
flow capacity ratio delay delay level of control ning time speed level of
(vph) (vph) X d1 PF I d2 service delay time (sec) (mph) service
600 1,600 0.38 9.2 0.33 0.93 0.6 A 3.7 14.5 18.2 19.8 C
700 1,600 0.44 9.6 0.33 0.90 0.8 A 4.0 14.5 18.5 19.5 C
800 1,600 0.50 10.0 0.33 0.86 1.0 A 4.3 14.5 18.8 19.2 C
900 1,600 0.56 10.4 0.33 0.81 1.2 A 4.6 14.5 19.1 18.8 C
1,000 1,600 0.63 10.9 0.33 0.74 1.4 A 5.0 14.5 19.5 18.4 C
1,100 1,600 0.69 11.4 0.33 0.67 1.6 A 5.4 14.5 19.9 18.1 C
1,200 1,600 0.75 12.0 0.33 0.58 1.9 A 5.9 14.5 20.4 17.6 D
1,300 1,600 0.81 12.6 0.33 0.48 2.3 A 6.5 14.5 21.0 17.2 D
1,400 1,600 0.88 13.3 0.33 0.36 2.7 A 7.2 14.5 21.7 16.6 D
1,500 1,600 0.94 14.1 0.33 0.23 3.5 A 8.2 14.5 22.7 15.8 D
1,600 1,600 1.00 15.0 0.33 0.09 6.8 B 11.8 14.5 26.3 13.7 E
segment, based on the fact that the arterial is Classification I, and 2. Hourly directional volume based on the predominant direc-
referring to the LOS boundaries in Table 11-1. Figure 11-24 de- tional flow,
picts the speed profile for the arterial and also indicates the arterial 3. Basic through-volume 15-min flow rate,
and intersection levels of service based on the average travel speed 4. Running time,
and control delay values, respectively. 5. Control delay,
Note that with the large signal spacings on such an arterial, one 6. Average travel speed, and
can expect the intersections to provide the driver with poorer levels 7. Level of service for the arterial section.
of service than the arterial, simply on the basis of LOS criteria for
2. Solution: The solution is reached with the following steps:
arterials and signalized intersections. Even on a Classification I
arterial, LOS A can be achieved with a speed of 42 mph or greater; Step 1. The two-way hourly volume for the planning analysis
however, more than 5.0 sec of stopped delay per vehicle removes hour is 2,730 (AADT × K = 30,000 × 0.091).
an intersection from LOS A (refer to Table 9-1). Step 2. The hourly directional volume for the planning analysis
hour is 1,550 (two-way hourly volume × D = 2,730 × 0.568).
Step 3. The basic through-volume 15-min flow rate is 1,475 or
CALCULATION 8—PLANNING APPLICATION: DETERMINING the hourly directional volume divided by the product of the PHF
LEVEL OF SERVICE and the quantity 1 minus the percentage of turns from exclusive
lanes [1,550/0.925 × (1 − 0.12)].
1. Description: The following information has been determined Step 4. The running time of 88 sec/mi is obtained directly from
about an arterial section for the predominant directional flow: Table 11-4 with arterial Classification II, a segment length of 0.5
Traffic characteristics: mi, and a free-flow speed of 45 mph as entries.
AADT = 30,000, Step 5. The control delay (d) for all the intersections of 140.0
K100 = 0.091, sec is obtained using Equations 11- 2, 11-3, and 11-4, the number
D = 0.568, of signalized intersections, and the following inputs: adjusted satu-
PHF = 0.925, ration flow rate, number of through and through/right lanes, arrival
Adjusted saturation flow = 1,850 pcphgpl, type, signal type, C, g/C, progression adjustment factor (PF), and
Percentage of turns from exclusive lanes = 12; incremental delay adjustment factors (k,I). The 140.0 sec is calcu-
Roadway characteristics: lated from Equations 11-2, 11-3, and 11-4 with the inputs that
Through lanes = 4 (2 through lanes in each direction), follow.
Arterial classification = II,
Free-flow speed = 45 mph, d = d1 × PF + d2 + d3
Section length = 2 mi, 0.5C f1 − sg/Cdg2
Median = yes, d1 =
1 − sg/Cd [min (X,1.0)]
Left-turn bays = yes;
Signal characteristics:
d2 = 900T [(X − 1) + √(X − 1)2 + 8kIX/Tc ]
Signalized intersections = 4 (thus, average segment length =
0.5 mi), where
Arrival type = 3, d1 = 33.6 sec,
Signal type = actuated, d2 = 3.9 sec,
C = 120 sec, d3 = 0 sec,
Weighted g/C = 0.42. d = 33.6 + 3.9 + 0 = 37.5 sec,
Find the following: and Sd = 37.5 × 4 = 140.0 sec, c = 1,850 × 2 × 0.42 = 1,554; the
1. Two-way hourly volume for the planning analysis hour, v/c ratio (X) = 1,475/(1,850 × 2 × 0.42) = 0.949; PF = 1.00 (Table
11-30 urban streets

Figure 11-21. Speed profile for Calculation 6.
11-6); T = 0.25 hr; k = 0.5 (for planning purposes); I = 0.209 Left-turn bays = yes;
(Table 11-8); and the number of signalized intersections = 4. Signal characteristics:
Step 6. The average travel speed is 22.8 mph, which is calculated Signalized intersections = 6 (thus, segment length = 0.33
by applying Equation 11-1 with running time and control delay mi),
for all signalized intersections as inputs. Arrival type = 5,
Signal type = semiactuated,
ART SPD = (3,600 × 2)/[(88 × 2) + (D)]
C = 120 sec,
= 7,200/(176 + 140.0)
Weighted g/C = 0.42.
= 7,200/316.0
= 22.8 mph Find the following:
Step 7. On the basis of an average travel speed of 22.8 mph 1. The lowest acceptable average travel speed for LOS C,
and the criteria in Table 11-1 for a Classification II arterial, the 2. The maximum acceptable hourly directional volume based
arterial’s level of service is C. on the predominant directional flow,
3. The maximum acceptable two-way hourly directional vol-
ume, and
CALCULATION 9—PLANNING APPLICATION: DETERMINING
VOLUMES
4. The maximum acceptable AADT.
2. Solution: The solution is found as follows:
1. Description: In preliminary design it is desired to know the
maximum volume of vehicles that a six-lane facility could handle Step 1. From Table 11-1, the lowest acceptable average travel
at LOS C given the following traffic, roadway, and signal speed for arterial Classification II and LOS C is 22 mph.
characteristics: Step 2. A running time of 100 sec/mi is obtained by interpolation
from Table 11-4 with arterial Classification II, a segment length
Traffic characteristics: of 0.33 mi, and a free-flow speed of 40 mph as entries.
K30 = 0.095, Step 3. The control delay for all the intersections of 127.3 sec
D = 0.55, is calculated by applying Equation 11-1 with average travel speed
PHF = 0.95, and running time as inputs:
Adjusted saturation flow = 1,750 pcphgpl,
22 = (3,600 × 2)/[(100 × 2) + d]
Percentage of turns from exclusive lanes = 12;
Roadway characteristics: Solving for d: d = 127.3 sec/veh control delay.
Through lanes = 6 (3 through lanes in each direction), Step 4. The average control delay per vehicle per intersection
Arterial classification = II, of 21.2 sec is calculated as follows:
Free-flow speed = 40 mph,
d = 127.3/6
Section length = 2 mi,
Median = yes, Solving for d: d = 21.2 sec/veh.
11-32 urban streets
Step 5. The basic 15-min flow rate of 2,156 is obtained by using It should be noted that this example applies to a preliminary
Equations 11-2, 11-3, and 11-4; the average control delay of 21.2 design problem, and the planning application and results obtained
sec/veh; PFs from Table 11-6; and the following inputs: adjusted should not be used beyond preliminary design. A more detailed
saturation flow rate, number of through and through/right lanes, planning analysis could have used signal-specific effective green
arrival type, signal type, C, g/C, k, and T. ratios, variable turning movements, variable lengths between sig-
The capacity of the lane group (c) = 1,750 × 3 × 0.42 = 2,205; nalized interesections, and other traffic, roadway, and signal char-
PF = 0.508 (Table 11-6); T = 0.25 hr (15-min period); k = 0.5 (for acteristics. The use of 12 percent for turns from exclusive lanes
planning purposes). and the application of a weighted effective green ratio of 0.42
Solving for v/c ratio: X = 0.978. to all signalized intersections in this problem are broad planning
Solving for the flow rate: v = 2,156. assumptions and are inappropriate for design and operational
Step 6. The hourly directional volume for the design hour is analyses.
2,328, the product of the basic 15-min through volume and the
PHF divided by the quantity 1 minus the percentage of turns from
exclusive lanes [2,156 × 0.95/(1 − 0.12)]. CALCULATION 10—STOP CONTROL ON ARTERIAL
Step 7. The two-way hourly directional volume for the design
hour is 4,232, the hourly directional volume divided by the direc- 1. Description: Humboldt Boulevard is a minor arterial of inter-
tional distribution factor (2,328/0.55). mediate design with a free-flow speed of 30 mph. It is an undivided
Step 8. The AADT based on the design hour and LOS C is facility with two lanes per direction and the layout and intersection
44,545, the two-way hourly directional volume divided by design spacing given in Figure 11-25. The intersections of Keefe Avenue,
hour factor (4,232/0.095). Locust Street, Center Street, and North Avenue are signal con-

11-34 urban streets
Figure 11-24. Speed profile for Calculation 7.
trolled, while Wright Street is all-way-stop controlled (AWSC).

The signals are pretimed and coordinated with a 90-sec cycle
length. The coordinated through movements have an arrival type
of 4. Table 11-11 provides further information.
Determine the southbound arterial level of service by segment
and for the entire facility. Do not aggregate the segments.
2. Solution: This is a slightly different problem because one of
the intersections on the arterial is stop controlled. Neverthless, the
methodology in Chapter 11 can be used, provided that the follow-
ing conditions are met:
T The control delay at the stop-controlled intersection is calcu-
lated with the procedures presented in Chapter 10.
T Arrival Type 3 is used for the intersection directly downstream
from the stop-controlled intersection. This should be done because
stop control breaks up platoons and results in random arrivals
downstream from the stop-controlled intersection.
T A stop-controlled intersection has a filtering effect similar to
that of a traffic signal, and therefore the value of the filtering and
metering factor (I) downstream from a stop-controlled intersection
should be calculated as for a signal-controlled intersection.
Table 11-11. Input Data for Sample Calculation 10

green degree of capacity
ratio saturation (vph)
intersection segment g/C X c
Locust Street 1 0.411 0.403 1,628
Center Street 2 0.706 0.284 2,302
Wright Street 3 —a 0.880 749
North Avenue 4 0.467 0.396 1,655
a
All-way-stop control. Figure 11-25. Arterial geometry for Calculation 10.
A summary of the arterial delay estimates is given in Figure CALCULATION 11—TWO-LANE ARTERIAL
11-26. The value of I is selected from Table 11-8. The intersection
control delay is calculated using Equation 11-2 for the signal- 1. Description: Park Avenue is an undivided two-lane arterial
controlled intersection and Equation 10-1 for the stop-controlled with heavy traffic in both directions during the peak hour and
intersection. relatively heavy turning volumes. The intersection spacing is given
Based on the functional and design categories, free-flow speed, in Figure 11-28. Most of the arterial intersection approaches have
and intersection spacing, the arterial can be identified as belonging exclusive left-turn lanes with permitted left-turn phasing. Several
to Classification III. The segment running times are calculated as movements fail during the afternoon peak hour, when the signals
before from Table 11-4. Because the 0.50-mi length of Segment 1 operate on a 90-sec cycle. Table 11-12 provides information ob-
exceeds the values given in Table 11-4, the free-flow speed is used tained from a capacity analysis of the individual intersections.
to calculate the running time on the segment. The results of the Because of heavy side volumes, the through movements have an
LOS computations are shown in Figure 11-27. arrival type of 3. The free-flow speed on the facility is 30 mph.
3. Discussion: The intersection levels of service for Segments Determine the arterial level of service in the eastbound direction
1 to 4 are C, A, D, and B, respectively. Note that the same delay at by segment and for the entire facility. Do not aggregate the
signalized and stop-contolled intersections may represent different segments.
levels of service, because of the different LOS thresholds for sig- 2. Solution: The facility can be classified as a minor arterial
nalized and stop-contolled intersections. From Figure 11-27 it can because it has a substantial access function. In terms of design,
be seen that the stop-controlled intersection LOS is slightly worse the arterial falls in the intermediate category because no parking
than the arterial LOS. This difference is not surprising because all is allowed along the facility, separate left-turn lanes are provided
vehicles should stop at the stop-controlled intersection, whereas at most intersections, and four signals are found in approximately
only a portion of all vehicles stop at signal-controlled intersections. a mile. Minor arterials in the intermediate design category can be
11-36 urban streets

identified as Classification III or IV. In this case Classification III lated using Equation 11-2. The segment running times are calculated
was selected because of the relatively high density of access points. from Table 11-4. For segment lengths outside the values given in
A summary of the arterial delay estimates is given in Figure 11- Table 11-4, running times per mile are extrapolated. The results of
29. The value of I is selected from Table 11-8. Park Avenue is a the LOS computations are shown in Figure 11-30.
westbound one-way facility west of Atherton Street; therefore inflow 3. Discussion: As shown in Figure 11-30 and in the speed profile
to Segment 1 takes place only from the cross streets. Consequently, in Figure 11-31, the average speed during the peak hour along the
an I of 1.0 is assumed for Segment 1. The intersection delay is calcu- arterial is 14 mph in the eastbound direction. The corresponding
level of service is C, which is somewhat better than the level field
observations indicate. A possible reason for this discrepancy may
be that queue interaction occurs because of the high degrees of
saturation. Queue interaction occurs whenever a downstream
queue reduces saturation flow at an upstream intersection, thereby
reducing capacity and increasing delay. In this case, the queue
from the oversaturated eastbound approach at University Drive
may extend back far enough to affect capacity at Bigler Street and
intersections even further upstream. Queue interaction may also
occur between left-turn movements and the adjacent through
movements. Because high left-turn volumes are serviced by per-
mitted phasing only, left-turn queues may spill out of the left-turn
bays and block the through movements, reducing saturation and
capacity and increasing delay.
This example illustrates the care required when analysing over-
Figure 11-28. Arterial geometry for Calculation 11. saturated conditions, because the methodology presented herein
does not take into account queue interaction.
Table 11-12. Input Data For Sample Calculation 11

effective green degree of capacity
time ratio saturation (vph)
intersection Segment g/C X c
EB WB EB WB EB WB EB WB
Atherton — 1 — 0.289 — 0.022 — 538
Allen 1 2 0.556 0.556 0.667 0.951 1,004 1,007
Shortlidge 2 3 0.633 0.467 0.906 0.977 1,037 869
Bigler 3 4 0.556 0.556 0.589 1.105 1,033 1,018
University 4 — 0.489 0.600 1.075 0.456 911 1,115

11-38 urban streets


11-40 urban streets
Figure 11-31. Speed profile for Calculation 11, eastbound traffic.
APPENDIX I
TEST-CAR METHOD FOR EXISTING ARTERIALS d. Some midblock speedometer readings should also be recorded
to check on unimpeded travel speeds and to see how they relate
The following steps are used when applying the test-car method to free-flow speed.
for determining levels of service on existing urban and suburban e. Data should be summarized for each segment and time period
arterials. to provide the average travel time and the average delay time
for a signal and for other delays and events (four-way stops,
1. Identify and inventory the geometry and the access control parking disruptions, etc.).
of each arterial segment, the segment lengths, existing signal tim- f. The minimum number of test-car runs depends on the variance
ing, and the 15-min flow rates for selected times of the day (such in the data and the occurrence desired. Six to 12 runs for each
as the peak a.m. period, the peak p.m. period, and a representative traffic-volume condition may be adequate. (See HRB Proc.,
off-peak period, by direction of flow). 1952, pp. 864–866.)
2. Determine the appropriate free-flow speed for the arterial g. An instrumented test car should be used if available to reduce
section being evaluated. For existing arterials, this speed may be labor requirements and to facilitate recording and analysis. Com-
determined by making runs with a test car equipped with a cali- puter-produced summaries of test-car runs, with all data re-
brated speedometer at times of low volumes. An observer should corded and analyzed by the computer, are now common.
read the speedometer at midblock locations where the vehicle is 5. For each segment and time period, the average travel speed
not impeded by other vehicles, and readings should be recorded should be determined by using travel times and segment lengths.
for each segment of an arterial. These observations may be supple-
Average travel speed for each arterial section should also be
mented by spot speed studies made at typical midblock locations
determined.
during low-volume conditions. Other data, such as design type,
6. Table 11-1 should be used to obtain a LOS value for each
access points, roadside development, and speed limit, may be con-
arterial segment and for the overall arterial, for each time period
sidered also.
and each direction of flow. This determination is made by compar-
3. Use Tables 11-2 and 11-3 along with the physical information
ing the average travel speed obtained in Step 5 with the speed
and free-flow speed already cited to determine the arterial
values given in Table 11-1 for the appropriate arterial
classification.
4. Make test-car travel time runs over the arterial section during classification.
the selected times of the day. 7. The test-car data can be modified to permit evaluation of
a. The observer should use appropriate measurement equipment different signal timing plans. As shown in Table 11-6, the adjust-
to obtain the information specified in the travel time field work- ment factors can be applied to control delays to evaluate effects
sheet in Appendix II. The equipment may be a computer-based on control delay of changes in offsets. It is then possible to evaluate
collection system or a pair of stopwatches. effects of these changes on average travel speeds and levels of
b. Travel times between centers of signalized intersections service.
should be recorded, along with the location, cause, and duration
of each delay.
c. Test-car runs should begin at different times in the signal
cycle to avoid making all trips the first in a platoon.

APPENDIX II
WORKSHEETS FOR USE IN ANALYSIS
WORKSHEETS Page
Summary of Arterial Intersection Delay Estimates................................................................................................................................. 11-42
Computation of Arterial Level of Service ............................................................................................................................................... 11-43
Travel Time (TT) Field Worksheet.......................................................................................................................................................... 11-44

11-42 urban streets


11-44 urban streets

chapter 12
TRANSIT CAPACITY
CONTENTS
i. introduction .......................................................................................................................................................................... 12-2

Context................................................................................................................................................................................... 12-2
Concepts................................................................................................................................................................................. 12-4
Person Movement............................................................................................................................................................. 12-4
Person-Capacity................................................................................................................................................................ 12-4
Basic Factors and Equations............................................................................................................................................ 12-4
Level of Service ............................................................................................................................................................... 12-7
Vehicle Capacities and Loading Criteria ........................................................................................................................ 12-7
ii. transit capacity experience ............................................................................................................................................... 12-9
Bus Capacity Experience ...................................................................................................................................................... 12-9
Bus Flow and Equivalency Studies................................................................................................................................. 12-10
Effects of Buses on Vehicular Capacity ......................................................................................................................... 12-10
Observed Bus Flows—Streets and Highways................................................................................................................. 12-10
Observed Bus Flows—Terminals .................................................................................................................................... 12-11
Passenger Service Times and Bus Headways................................................................................................................. 12-11
General Capacity Ranges................................................................................................................................................. 12-12
Rail Transit Capacity Experience ......................................................................................................................................... 12-14
Observed Train and Passenger Flows ............................................................................................................................. 12-15
General Capacity Ranges................................................................................................................................................. 12-16
iii. methods and procedures—on-street transit ................................................................................................................. 12-18
Bus Berth and System Capacity ........................................................................................................................................... 12-18
General Considerations .................................................................................................................................................... 12-18
Equations and Guidelines ................................................................................................................................................ 12-19
Applications...................................................................................................................................................................... 12-26
Bus Priority Treatments ........................................................................................................................................................ 12-30
Operational Overview ...................................................................................................................................................... 12-30
Planning Considerations................................................................................................................................................... 12-30
Guidelines for Specific Treatments ................................................................................................................................. 12-33
iv. applications and sample problems .................................................................................................................................... 12-34
General Approach.................................................................................................................................................................. 12-34
Types of Problems................................................................................................................................................................. 12-39
Sample Calculations .............................................................................................................................................................. 12-40
Calculation 1—Person-Flow ............................................................................................................................................ 12-40
Calculation 2—Person-Capacity ...................................................................................................................................... 12-40
Calculation 3—Effect of Buses on Freeway Capacity ................................................................................................... 12-40
Calculation 4—Effect of Buses on Arterials .................................................................................................................. 12-41
Calculation 5—Passenger Service Times (Bus Stop)..................................................................................................... 12-41
Calculation 6—Passenger Service Times (Bus Routes) ................................................................................................. 12-41
Calculation 7—Planning Applications, Downtown Street, Level of Service ................................................................ 12-42
Calculation 8—Bus Terminal (Transit Center)............................................................................................................... 12-42
Calculation 9—Berth Capacity for Loading ................................................................................................................... 12-42
Calculation 10—Bus Berth Unloading............................................................................................................................ 12-44
Calculation 11—Berth Capacity for Loading at Major Stops........................................................................................ 12-45
12-1
12-2 urban streets
Calculation 12—Arterial Street Capacity........................................................................................................................ 12-45

Calculation 13—CBD Busway ........................................................................................................................................ 12-46
Calculation 14—Arterial Bus Turnout ............................................................................................................................ 12-46
Calculation 15—Rail Rapid Transit ................................................................................................................................ 12-47
Calculation 16—Light Rail Transit on City Street......................................................................................................... 12-47
v. references .............................................................................................................................................................................. 12-47
appendix i. Bus Capacity Experience..................................................................................................................................... 12-49
appendix ii. Rail Capacity Experience ................................................................................................................................... 12-55
appendix iii. Examples of Boarding and Alighting Time ..................................................................................................... 12-59
I. INTRODUCTION
This chapter contains guidelines and procedures for estimating They also analyzed bus berth capacity. The findings on bus transit
bus and rail transit capacities. It defines basic capacity concepts capacity were summarized in ‘‘Bus Capacity Analysis,’’ Transpor-
and principles; summarizes previous studies and current experi- tation Research Record 546 (5).
ence; develops analytical relationships; presents capacity guide- Rail transit capacity has a long history of actual operating
lines; and sets forth illustrative applications. experience and analysis. The Board of Supervising Engineers
The guidelines and procedures may be used to estimate: for Chicago Traction, for example, analyzed street railway capacity
in 1912, and passenger dwell times by door width in 1916. Lang
1. The effects of bus flows on freeway and signalized intersec- and Soberman derived formulas for rapid transit track capacity in
tion capacity. 1964 (40). More recent studies by Homburger, Pushkarev, and
2. Total passenger or person flow based on roadway operating
Vuchic further addressed rail transit capacity theory and practice
conditions, and the prevailing mix of cars and buses (or rail
(7,8,9). Relevant materials from the more recent references are
vehicles).
incorporated into this chapter.
3. Generalized ranges of bus capacities for arterial streets,
downtown streets, and busways.
4. Bus berth (stop) requirements at terminals and along down-
town busways, bus-only streets, and other city streets. CONTEXT
5. Passenger flows on rail transit lines for varying car sizes,
train lengths, service frequencies, and loading conditions—for both
light rail and rapid transit lines. Transport system management solutions to urban transport
problems have increased the interest in the person-capacity char-
The chapter also provides ways to address various questions acteristics of transportation facilities in addition to their vehicle-
normally encountered in transit service planning and operations. capacity characteristics. The underlying rationale is that al-
For instance: though buses and rail transit cars require more street space per
vehicle than private automobiles, they carry many more passengers
1. How many passengers can be carried per unit of time? per vehicle than automobiles, especially during peak hours. Thus,
2. How many transit vehicles are needed to provide a specific
public transportation emerges as an important way to increase the
rate of passenger flow?
number of people carried by urban transportation systems.
3. How many passengers can be carried with a given vehicle
Transit vehicles carry a substantial number and proportion of
fleet?
peak-hour person trips to and from the downtown areas, and along
It emphasizes bus transit capacities because buses operate over the many urban freeways, arterials, and downtown streets.
street and highway systems in most urban areas. However, it also Table 12-1 indicates the peak period use of public transport,
presents salient rail transit characteristics and capacities. It builds bus and rail combined, by persons entering the central business
upon and extends previous transit capacity analysis. districts of selected cities in Canada and the United States (1).
The initial research on transit capacity relative to streets and Transit carries more than two-thirds of all peak-hour travelers
highways was developed in 1961 by the Transit Subcommittee of to or from the New York, Chicago, Philadelphia, and Toronto
the Committee on Highway Capacity and Quality of Service. It downtown areas, and more than a third of all peak-hour travelers
summarized operating experience in the United States and con- entering or leaving most other CBD’s. The variations in transit
tained broad guidelines for passenger dwell times and vehicle use reflect differences in population, central business district
occupancies. employment, extent of bus and rail transit services, and geo-
More detailed analyses of bus capacity were contained in graphic characteristics.
NCHRP Reports 143 and 155 on Bus Use of Highways (2,4). Buses carry over 85 percent of all peak-hour person-trips
These reports summarized, synthesized, and interpreted avail- through the Lincoln Tunnel in the City of New York, account
able information on bus flows, passengers, and service times. for about half of all peak-hour travelers on the Shirley Highway
transit capacity 12-3
(I-95), Virginia, and the Long Island and Gowanus Expressways Table 12-1. Peak-hour Use of Public Transport by
(New York City), and for more than a quarter of all passengers Persons Entering or Leaving the Central Business
on radial freeways approaching or leaving other large-city District
CBD’s. percent by public
Buses carry an even higher proportion of peak-hour travelers transport in peak
on many city streets. More than 80 percent of all peak-hour urban area year direction
passengers on Hillside Avenue and Madison Avenue in New New York, New York 1982 89a
York City, Market Street in Philadelphia, and Main Street in Chicago, Illinois 1974 82a
Dallas are carried by buses. Buses accommodate more than half Toronto, Ontario 1970 68a
Boston, Massachusetts 1974 49a
of all peak-hour person-trips on downtown streets in many other
Cleveland, Ohio 1970 44a
cities (2). Ottawa, Ontario 1974 40
These observations do not necessarily represent maximum possi- Vancouver, British Columbia 1970 40
ble bus volumes or total traffic volumes. They do, however, clearly Los Angeles, California 1974 37
indicate that while buses account for a relatively small proportion Washington, D.C. 1979 36b
of the vehicles in a traffic stream, they carry a sizable part of the Detroit, Michigan 1974 35
Baltimore, Maryland 1982 33
total person flow.
Denver, Colorado 1977 30
Rail transit, operating mainly off-street, becomes important Dallas, Texas 1971 28
in serving large, intensively developed city centers where it ac- Milwaukee, Wisconsin 1974 25
counts for more than half of all people entering or leaving in Providence, Rhode Island 1977 21
the peak hour. New Haven, Connecticut 1982 20
Minneapolis, Minnesota 1965 20
Houston, Texas 1971 14
a
With rail transit.
b
CONCEPTS Includes Pentagon area; data for 6:30–9:30 AM
SOURCE: Cordon Counts for each city, mainly compiled in Ref. 1.
Transit capacity is more complex and less precise than high-

way capacity: it deals with the movement of both people and service frequencies and allowable passenger loadings. Accordingly
vehicles, depends on the size of the transit vehicles and how the traditional concepts applied to highway capacity must be
often they operate, and reflects the interaction between passenger adapted and broadened.
traffic concentrations and vehicle flow. It depends on the op- Table 12-2 defines the important terms that relate to transit
erating policy of the transit agency, which normally specifies capacity.
Table 12-2. Important Terms in Transit Capacity

T Clearance Time—All time losses at a stop other than passenger dwell times, in seconds. It can be
viewed as the minimum time, in seconds, between one transit vehicle leaving a stop and the following
vehicle entering, i.e., the clearance time between successive buses should not be less than 15 sec.
T Crush Capacity—The maximum number of passengers that can be physically accommodated on a
transit vehicle. It is also defined as level-of-service F. It can be viewed as an ‘‘offered’’ capacity, since
it cannot be achieved on all vehicles for any sustained period of time.
T Dwell Time—The time, in seconds, that a transit vehicle is stopped for the purpose of serving passengers.
It includes the total passenger service time plus the time needed to open and close doors.
T Interrupted Flow—Transit vehicles moving along a roadway or track and having to make service stops
at regular intervals.
T Maximum Load Point—The point, actually section, along a transit route at which the greatest number
of passengers is being carried.
T Passenger Service Time—The time, in seconds, that is required for a passenger to board or alight from
a transit vehicle.
T Person-Capacity—The maximum number of persons that can be carried past a given location during a
given time period under specified operating conditions without unreasonable delay, hazard, or restriction.
Usually measured in terms of persons per hour.
T Person Level of Service—The quality of service offered the passenger within a transit vehicle, as
determined by the available space per passenger.
T Productive Capacity—A measure of efficiency or performance. The product of passenger capacity
along a transit line and speed.
T Seat Capacity—The number of passenger seats on a transit vehicle.
T Standees—The number of standing passengers on a transit vehicle. The ratio of total passengers carried
to the number of seats during a specified time period is called the load factor. The percent standees
represents the number of standing passengers expressed as a percentage of the number of seats. A
transit vehicle with 40 seats and 60 passengers has a load factor of 1.5 and 50 percent standees.
T Uninterrupted Flow—Transit vehicles moving along a roadway or track without stopping. This term
is most applicable to transit service on freeways or on its own right-of-way.
12-4 urban streets
Person Movement capacity drops from 1,800 to 1,620. (Note that this figure only
refers to capacity, not demand or actual use.) If each car carried
Each roadway or transit facility should be analyzed in terms five passengers, then with 1,320 cars and 300 buses, the total
of the number of people it carries in a specific time period. person-capacity would be 21,600.
This calls for knowing both the number and occupancies of A slightly different approach should be used for downtown
each type of vehicle. streets. The person-capacity of the bus or street car lane (as-
For example, an urban freeway lane carrying 1,800 passenger suming only transit use) can be estimated by the procedures
cars per lane per hour with an average occupancy of 1.5 persons outlined in this chapter (Sec. II and III). The vehicular capacities
would have a person movement of 2,700 people per hour. of the general traffic lanes can be estimated based on the pro-
Likewise, an arterial street carrying 600 automobiles per hour cedures outlined in Chapter 9 and weighted by their passenger
and 50 buses per hour, with occupancies of 1.5 and 40, respec- occupancies. The total person-capacity equals the sum, and may
tively, would have a total person movement of 2,900 persons be higher than figures based entirely on actual usage. Note that
per hour of which approximately 70 percent would be carried this approach is different from that for freeways, where it is
by public transport. usually unrealistic to preempt an entire lane for buses; however
it could be applied where dedicated freeway bus lanes are
considered, by taking into account the limits on bus capacity
Person-Capacity
resulting from approaches to and from the freeway, as well as
stops along it.
The person-capacity or passenger-carrying capability for any

given transport route can be defined as ‘‘the maximum number
of people that can be carried past a given location during a Basic Factors and Equations
given time period under specified operating conditions without
unreasonable delay, hazard, or restriction, and with reasonable
certainty.’’ The passenger capacity of a transit line is the product of the
This definition is less absolute than definitions for vehicle- number of vehicles per hour (usually past the busiest stop) and
capacity, because it recognizes that when dealing with transit, the number of passengers that each vehicle can carry. Four basic
additional considerations enter the picture. More specifically, per- factors determine the maximum passenger capacity:
son-capacity depends on the mix in the traffic stream, including 1. The maximum number of vehicles per transit unit (bus,
the number and occupancy of each type of vehicle that can reason- car, train).
ably be expected to pass a point on a roadway. It is a function of 2. The passenger capacity of the individual transit vehicles.
vehicle size, type, occupancy, and headway. 3. The minimum possible headway or time spacing between
The number of transit vehicles should be based on a specified individual vehicles or trains.
flow. The number of cars should reflect the auto capacity of the 4. The number of movement channels or loading positions.
facility after deducting the passenger car equivalents of the buses.
The total person-capacity then represents the number of people The many variables that influence these factors and transit ca-
that can be carried by the specified number of buses and the pacities are given in Table 12-3. Some affect the number of passen-
remaining passenger car capacity. gers per unit, while others affect the number of units that can pass
The person-capacity of a freeway lane with bus and car traffic a given location within a specified time period.
under prevailing conditions of flow can be estimated as follows: The capacity of a transit line varies along the route. Limi-
tations may occur (1) between stops (i.e., way capacity) (2) at
cp = f ′ O1 + [(1,800 − 1.5 f ′) O2 ] (12-1) stops or stations (i.e., station capacity), (3) at major intersections
with cross traffic, or (4) at terminals (station capacity). In most
where:
cases station capacity rather than way capacity is the critical
f′ = number of buses per hour; constraint.
O1 = bus occupancy; Capacities are generally governed by the critical stops where
O2 = car occupancy; and major passenger boarding or alighting takes place, or where
cp = person-capacity, people per hour. vehicles terminate or turn around. This is similar to estimating
arterial street system capacity based on critical intersections
The number of persons that can be carried in buses depends along a route. Sometimes, however, outlying rail transit ter-
on the number of buses scheduled. This may be below the minals limit system capacity due to heavy passenger boardings,
maximum capacity of a street to accommodate buses. It is and track configurations or operating practices that limit train
certainly the case for most urban freeways, as illustrated by the turnarounds.
following example. The actual mix of automobiles and transit vehicles in a traffic
Figure 12-1 shows the person-capacities for an urban freeway stream results from the choice of travel mode by the traveler and
lane, with various numbers of buses in the traffic stream. This from the number of transit vehicles scheduled over the facility.
example assumes a maximum freeway capacity of 1,800 vph The number of persons that can be carried by a given bus or rail
without buses, a bus-passenger car equivalency of 1.5, and oc- line, therefore, reflects the operating policy of the transit property
cupancies of 1.5 and 50 for cars and buses respectively. As the with respect to minimum service frequency and passenger loading
number of buses on the freeway increases to 300, the total person- conditions (i.e., number of standees).
capacity increases from 2,700 to nearly 17,000, while the vehicle- The following considerations are important:
Figure 12-1. Example of freeway person-capacity.
1. A transit line with a relatively uniform distribution of 3. The maximum rate of passenger flow is usually constrained
boarding passengers among stops will usually have a higher by such factors as acceptable levels of passenger comfort, the
capacity than one where passenger boarding is concentrated at presence of other traffic sharing the same right-of-way, and safety
a single stop. considerations. Therefore, transit operators generally are more con-
2. Short-term fluctuations in ridership demand must be concerned with the realistic rates of flow that can be achieved by
sidered to avoid unacceptable passenger queuing or overcrowd- different modes, rather than with physical capacity in the theoreti-
ing. Variations in arrival patterns and dwell times at stops will cal engineering sense.
tend to reduce capacity. 4. Operations at ‘‘capacity’’ tend to strain transit systems,
12-6 urban streets
Table 12-3. Factors That Influence Transit Capacity

1. Vehicle Characteristics
T Allowable number of vehicles per transit unit (i.e., single unit bus, or several unit-cars per train)
T Vehicle dimensions
T Seating configuration and capacity
T Number, location, width of doors
T Number and height of steps
T Maximum speed
T Acceleration and deceleration rates
T Type of door actuation control
2. Right-of-Way Characteristics
T Cross-section design (i.e., number of lanes or tracks)
T Degree of separation from other traffic
T Intersection design (at grade or grade separated, type of traffic controls)
T Horizontal and vertical alignment
3. Stop Characteristics
T Spacing (frequency) and duration
T Design (on-line or off-line)
T Platform height (high level or low level loading)
T Number and length of loading positions
T Method of fare collection (prepayments, pay when entering vehicle; pay when leaving vehicle)
T Type of fare (single-coin, penny, exact)
T Common or separate areas for passenger boarding and alighting
T Passenger accessibility to stops
4. Operating Characteristics
T Intercity versus suburban operations at terminals
T Layover and schedule adjustment practices
T Time losses to obtain clock headways or provide driver relief
T Regularity of arrivals at a given stop
5. Passenger Traffic Characteristics
T Passenger concentrations and distribution at major stops
T Peaking of ridership (i.e., peak-hour factor)
6. Street Traffic Characteristics
T Volume and nature of other traffic (on shared right-of-way)
T Cross traffic at intersections if at grade
7. Method of Headway Control
T Automatic or by driver/trainman
T Policy spacing between vehicles
SOURCE: Adapted from Canadian Transit Handbook (Ref. 12).
and do not represent desirable operating conditions. Moreover, D = dwell time at major stop under consideration, in sec;
most U.S. transit systems operate at capacity for a relatively short S = passengers per vehicle;
period of time, if at all. n = vehicles per unit (n = 1 for buses; n = 1 to n = 11 for rail
5. Capacity relates closely to system performance and service vehicles); and
quality in terms of speed, comfort, and service reliability. A single R = reductive factor to compensate for dwell time and arrival
fixed number often can be misleading. The concept of ‘‘productive variations.
capacity,’’ the product of passenger flow and speed, provides an The factor R reduces the capacity to account for variations in
important index of system efficiency (9). bus arrival patterns and in dwell times at stop. It may approach
6. Capacities obtained by analytical methods must be cross- 1.00 for a rail transit system on private right-of-way with wayside
checked against actual operating experience for reasonableness. cab signal control or with automatic train operations. For bus
The capacity of a transit line can be estimated from the following operations, especially on city streets, it is always less than 1—a
equations: value of 0.833 is suggested for maximum capacity. Using this
3,600 R 3,600 R factor, the term (3,600 R) in Eqs. 12-2a and 12-2b becomes 3,000
cv = = (12-2a) for maximum capacity. In effect, it increases headways by 20
h D + tc
percent.
3,600 nSR These equations, with further adjustments for the reductive ef-
cp = nScv = (12-2b)
D + tc fects of traffic signals, form the basis for all transit capacity compu-
tations. In this case the basic equation becomes:
where:
(g/C) 3,600 nSR
cv = vehicles per hour per channel or berth (maximum); cp = (12-2c)
cp = people per hour per channel or berth (maximum); (g/C) D + tc
h = headway between successive units, in sec; where: g = green time, in seconds, and C =cycle length in
tc = clearance between successive vehicles, in sec; seconds.
Equations 12-2b and 12-2c may be used to estimate passenger capacity’’ will be influenced by the type of technology (rail
capacity when the number of effective loading positions is taken versus bus), the method of operation (private right-of-way versus
into account. They provide a realistic estimate where loading shared), and the spacing of stops (9).
patterns and/or door configurations enable vehicles to fill up as Two aspects of level of service are important from a capacity
they reach the maximum load point. Where this condition is perspective: the number of passengers per vehicle, and the number
not likely, as along many bus routes, more detailed analyses of vehicles per hour. Capacity-related level-of-service criteria
are needed. In such cases, the detailed methods described in should reflect both. Figure 12-2 illustrates this two-dimensional
Section III should be applied. nature of urban transit capacity.
It can be seen that it is possible to operate many transit vehicles,
each carrying few passengers. From a roadway capacity perspec-
Level of Service tive, the number of vehicles could be at or near capacity, even if
they run nearly empty.
The concept of level of service (LOS) for transit is far more A few vehicles could operate, each overcrowded. This repre-
complex than for highways. It includes such factors as coverage sents a poor level of service from a passenger comfort (user)
of major residential and activity areas, comfort, speed, and perspective. Long waiting times would also detract from user
reliability (i.e., on-time performance). Convenient schedules, convenience.
comfortable vehicles, and frequent, fast, and reliable service Finally the domain of peak-period operation commonly involves
contribute to LOS. Speed is influenced not only by the number a large number of vehicles, each heavily loaded.
of riders using a transit line, but, to an even greater extent, by
stop frequency and dwell times, traffic interferences, and right-
of-way design. Vehicle Capacities and Loading Criteria
Productive capacity, the product of passenger capacity and
speed, is an important measure of transport system efficiency. Typical transit vehicle types, dimensions, and passenger capaci-
It is important in that it distinguishes between equal passenger ties are given in Table 12-4. The total passengers carried varies
throughputs achieved at different speeds. Thus, express bus depending on bus or rail car capacity and the tradeoff between
service normally has a higher productive capacity than local bus seated capacity and standees. The largest number of seats and
service; similarly, commuter rail line operating at 40 mph is lowest number of standees should occur on longer suburban bus
twice as ‘‘productive’’ as an urban rail transit line carrying the routes or on commuter rail routes where higher levels of comfort
same number of people at 20 mph. In general, ‘‘productive are essential.
Figure 12-2. The two-dimensional nature of transit level of service as related to transit capacity.
12-8 urban streets
Table 12-4. Characteristics of Typical Transit Vehicles—United States and Canada

type of vehicle or length width typical capacitya
train (ft) (ft) seats standeesb total remarks
Minibus-short haul 18–25 6.5–8.0 15–25 0–15 15–40
Transit bus 30.0 8.0 36 19 55 Example: General Motors,
35.0 8.0 45 25 80 RTS II, 1978
40.0 8.5 53 32 85
Articulated transit bus 55.0 8.5 66 34 100 Chicago-am General-MAN
59.7 8.5 73 37 110 am General-MAN
Street car 46.7 9.0 59 40–80 99–139 P.C.C.c
Light rail car train 151.2 8.7 128 248–272 376–400 San Diego—6-axle car,
2-car train (Duewag)
142.0 8.8 104 250–356 354–460 Boston—6-axle car, 2-car
train (Boeing Vertol)
Rail rapid transit train 605.0 10.0 500 1,300–1,700 1,800–2,200 10-car train, IND
New York
600.0 10.0 576 1,224–1,664 1,800–2,240 8-car train, R-46 cars,
New York
448.6 10.3 504 876–1,356 1,380–1,860 8-car train, Toronto
Commuter rail train 85.0 10.5 1,100 200–1,200 1,300–2,300 Regular car, 10-car train
a
In any transit vehicle the total passenger capacity can be increased by removing seats and by making more standing room available, and vice-versa.
b
Higher figures denote crush capacity; lower figures, schedule-design capacity.
c
Presidents’ Conference Committee Cars.
SOURCE: Refs. 8 and 34.
Table 12-5. Passenger Loading Standards and Levels of Service for Bus Transit
Vehicles (50-Seat, 340-Sq Ft Bus)
peak-hour approx. pass./seat
level of service passengers sq ft/pass. (approx.)
A 0 to 26 13.1 or more 0.00 to 0.50
B 27 to 40 13.0 to 8.5 0.51 to 0.75
C 41 to 53 8.4 to 6.4 0.76 to 1.00
D 54 to 66 6.3 to 5.2 1.01 to 1.25
E (Max. scheduled load) 67 to 80 5.1 to 4.3 1.26 to 1.50
F (Crush load) 81 to 85 <4.3 1.51 to 1.60
SOURCE: Ref. 34.
A typical 40-ft urban transit bus can normally seat 53 passengers a reasonable balance between operating economy and passenger
and can carry up to 32 additional standees. Similarly, a 60-ft articu- comfort. It is consistent with the use of 5.4 sq ft per passenger
lated bus can carry 69 passengers and 41 standees. suggested by Pushkarev and Zupan as a realistic passenger capacity
An 8-car train of 75-ft rail transit cars normally can seat about for rapid transit lines (7).
500 and carry a ‘‘crush load’’ of over 2,000 people. Level-of-service E is synonymous with ‘‘capacity’’ assuming a
Doorways on buses range from 22 to 30 in. each, while doors reasonable number of standees. It represents the upper limit for
on rail vehicles typically average 50 in. each. scheduling purposes. These maximum scheduled loads are nor-
Table 12-5 gives suggested ‘‘passenger’’ levels-of-service for mally 65 to 75 percent of the crush loads.
a conventional 40-ft bus, based on 53 passengers per bus and 340 Level-of-service F defines ‘‘crush load’’ conditions in which
gross square feet per vehicle. These approximate comfort-related standees and other passengers are subject to unreasonable discom-
levels of service are from the perspective of passengers on the fort. Such loads are unacceptable to passengers. Although LOS
vehicle rather than the number of vehicles in a given channel. F represents the theoretically offered capacity it cannot be
They are based on local bus operations where short trips at rela- sustained on every vehicle for any given period, and it exceeds
tively slow speed allow standees. Express bus service on express- the maximum utilized capacity. Moreover, it is not reasonable
ways and busways should not allow standees; hence, their to assume that passengers will be equally distributed among all
scheduling should be guided by level-of-service C. cars of all trains. Therefore, level F should not be used for
Suggested passenger levels of service for urban rail transit transit capacity calculations. Note, however, that when the
vehicles are given in Table 12-6. LOS D, which allows up to 2 maximum schedule loads are used, some transit units will operate
persons per seat and a minimum 5.0 sq ft per person, provides at LOS F.
Table 12-6. Passenger Loading Standards and Levels Table 12-7. Typical Space Requirements for Seated
of Service for Urban Rail Transit Vehicles and Standing Passengers
approx. sq approx. sq ft per pass. (net)a
peak-hour level of service ft/pass. pass./seat Seated Passenger
A 15.4 or more 0.00 to 0.65 Typical commuter rail 4 to 6
B 15.2 to 10.0 0.66 to 1.00 Typical urban rail transit 3 to 5
C 9.9 to 7.5 1.01 to 1.50 Typical urban bus transit 3 to 4
D 6.6 to 5.0 1.51 to 2.00 Standing Passenger
E-1 4.9 to 4.0 2.01 to 2.50 Spacing of persons in unconstrained 4 to 9
E-2 (Maximum scheduled load) 3.9 to 3.3 2.51 to 3.00 condition
F (Crush load) 3.2 to 2.6a 3.01 to 3.80 Minimum space requirement to avoid 2.4 to 2.8
a
The maximum crush load can be realized in a single car, but not in every car contact (maximum schedule load
on the train. LOS E)
NOTE: Fifty percent standees reflects a load factor of 1.5 passengers per seat. Duewag Standard—commonly used 2.7
SOURCES: H.S. Levinson and W.R. Reilly as reported in Ref. 34. in German LRT systems
NYCTA—maximum ‘‘practical’’ 1.8
capacity (crush loads)
a
Excludes nonusable space. For seated passengers includes space consumed
by seat plus space between seats for legs. For standing passengers, based on
clear floor area per standee.
SOURCE: Ref 37.
The gross passenger loading criteria provide a reasonable ap- An 8.5 ft by 40 ft 53-passenger bus would have the following
proximation of passengers’ levels of service. However, because capacities under maximum load schedule (LOS E) conditions:
such loading criteria do not reflect specific space criteria for seated
and standing passengers, more refined computations sometimes 340 Gross sq ft
may be desirable. Table 12-7 gives suggested net space require- 245 Net sq ft
ments for various types of transit that can be applied to specific 53 Seats at 3.3 sq ft/seat = 175 sq ft
vehicle sizes and seating considerations. The standing passenger Net area for standees 245 − 175 = 70 sq ft
criteria reflect LOS E, schedule capacity. Standees at 2.6 sq ft/person = 27 standees
The precise passenger capacity of a transit vehicle can be esti- Total capacity 80
mated by the following relationship: Maximum capacity Table 12-5
based on gross floor area: 80
An
Si = s n + (12-2d)
Li
where:
Car dimensions, seats, and schedule and crush capacities for
sn = seats per vehicle;
specific U.S. and Canadian rail transit lines are contained in
An = net area for standees;
Appendix II. More detailed information on specific transit vehicle
Li = net sq ft/standee for service level i; and
characteristics and capacities is contained in Ref. 39.
Si = passengers/vehicle or passenger spaces/vehicle, for service
The data in Tables 12-4 through 12-7 may be used to estimate
level i.
transit vehicle requirements for specified passenger demands at
Li should equal 2.6 for maximum schedule loads (level-of-ser- the maximum load points. They also can be used to assess the
vice E) and 2.0 for crush load conditions. ‘‘level of service’’ from the passengers’ standpoint.
II. TRANSIT CAPACITY EXPERIENCE
This section presents bus and rail transit operating experiences. BUS CAPACITY EXPERIENCE
It identifies service frequencies, passengers carried, and passenger
car equivalents; and it indicates the ranges in capacity based on The number of buses that can operate past any point in a given
this experience. period of time varies according to specific roadway conditions and
12-10 urban streets
operating practices. Results of both theoretical studies and actual per hour one-way. Bus dwell times averaged 19 sec (ranging from
operating experience are summarized as follows. 9 to 32 at individual stops), and bus speeds averaged 2 to 3 mph.
However, about 200 other vehicles also used the 36-ft wide, signal-
controlled street each way (32).
Bus Flow and Equivalency Studies
Effects of Buses on Vehicular Capacity

Several studies have analyzed the effects of buses on the capac-
ity of mixed-traffic roadways and have estimated the capacity of
a bus lane. The reductive effect of buses on vehicular capacity varies ac-
cording to the method of operation. The time available for other
1. Theoretical capacities—Simulation analysis and field obser- vehicles generally will be reduced by the time preempted by buses.
vations of passenger car equivalents have shown that capacities of This time loss depends on the number of buses in the traffic flow
1,400 or more buses per lane per hour can be achieved on exclusive and their service time requirements at stops.
bus roadways with uninterrupted flow and no stops for passengers. Consequently, for uninterrupted flow, buses are the equivalent
They compare with some 700 to 750 buses per hour moving of 1.5 passenger car units in the lane where they operate. At
through the Lincoln Tunnel—the highest bus flows found in the bus stops buses have a greater reductive effect because of the
United States (4, 10, 11). time involved in discharging and receiving passengers. The
Theoretical simulation studies based on buses with 30-sec equivalency factors for these conditions depend on the specific
dwell times that operate in platoons of six between stations 0.3 duration of the bus stop and its reductive effect on arterial street
mi apart result in capacities ranging from 350 to 400 buses per green time.
hour on an exclusive grade-separated busway (14). These results The reductive effects of local transit buses on other vehicles in
have not been verified, because reported bus volumes of this an arterial street lane can be estimated as follows:
magnitude occur only under express operations without stops.
1. Where the buses stop in a lane that is not used by moving
Maximum hourly bus flows in a single lane on city streets in
traffic (for example in a curb parking lane), the time loss to other
the United States rarely exceed 100.
vehicles is approximately 3 to 4 sec per bus. For this case, buses
2. Bus headways and passenger car equivalents—freeways—
would either accelerate or decelerate across the intersection,
Field studies of bus-car equivalency factors conducted by the
thereby reducing the impeditive effects to other traffic.
Port of New York Authority in the Lincoln Tunnel found an
2. Where buses stop in a normal traffic lane, the time loss
equivalent of 1.5 cars per bus (15). A nationwide study of
involves the dwell time for buses plus a time loss for stopping
mixed traffic flows on seven expressways conducted by the
and starting, and the associated queuing effects on other traffic.
Bureau of Public Roads found an equivalency factor of 1.6 (13).
The time loss can be estimated from the following equation for
The effects of grades on bus flows are summarized in other
the lane in which the buses operate.
chapters of the Manual.
The similarity of these findings indicates that when buses are TL = (g/C) × N × (D + L) (12-3)
in motion either in exclusively bus traffic or in mixed traffic,
under uninterrupted flow conditions over a broad range of levels where:
of service, their equivalency factor will be approximately 1.5 pas- TL = time loss, in sec per hr;
senger cars. g/C = green time/cycle time ratio;
3. Capacity of freeway bus lane (no stops)—The capacity or N = buses per hour that stop;
service volume of an exclusive bus lane with uninterrupted flow D = average dwell time, in sec; and
can be computed by applying the 1.5 car equivalency factor to the L = additional time loss due to stopping, starting, and
computed capacity or corresponding service volume in passenger queuing, in sec (L = 6 to 8 sec, assuming average
cars per hour. For example, a roadway lane having a capacity of conditions).
1,500 passenger cars per hour would have an equivalency of 1,000
buses per hour. Corresponding uninterrupted bus-flow capacities Equivalent passenger car units derived from this equation for
for various freeway levels of service are as follows, assuming 70- various rates of vehicle flow, dwell times, g/C ratios, and bus
mi per hour design speeds: volumes are given in Table 12-8. Alternatively, the (effective)
green for the lane in which the buses operate can be obtained
by deducing the time loss. The data are precise for near side
Passenger Cars/
bus stops and a reasonable approximation for far side stops.
LOS Lane/Hour Buses/Lane/Hour
A 700 467
B 1,100 733 Observed Bus Flows—Streets and Highways
C 1,550 1,033
D 1,850 1,233 Observed bus volumes on urban freeways, city streets, and
E 2,000 1,333 bus-only streets clearly show the reductive effects of bus stops
These uninterrupted bus flow volumes require that bus stops be on bus capacity. The highest bus volumes, 735 buses per hour
located off of the travel lane and that adequate acceleration and through the Lincoln Tunnel and on the Port Authority Midtown
deceleration lanes be provided. Bus Terminal access ramps, are achieved on an exclusive right-
4. Arterial streets—A bus capacity demonstration study on Ho- of-way where buses make no stops. Where bus stops or layovers
tel Street in Honolulu found a capacity of 95 to 100 buses are involved, reported bus volumes are much less.
Stopping a bus to receive or discharge passengers limits the Table 12-8. Passenger Car Equivalency of Urban
capacity of a bus lane. Time must be allowed for acceleration, Buses at Signalized Intersections (Applies Where
deceleration, and stop clearance, as well as for the time when the Buses Block Cars)
doors are open. percent green time on street with buses
duration of
When intermediate stops are made bus volumes rarely exceed stop (sec) 30% 40% 50% 60%
120 buses per hour. However, volumes of 180 to 200 buses per
5 2 2 3 3
hour are formed where buses may use two or more lanes to allow
10 2 3 4 5
bus passing, especially where stops are short. An example is Hill- 15 3 4 5 6
side Avenue, New York City. Two parallel bus lanes in the same 20 4 5 7 8
direction, as found along Madison Avenue, New York, and Port- 25 5 6 8 9
land’s Fifth and Sixth Street Transit Mall also achieve this flow 30 5 7 9 11
rate. Chicago’s State Street Mall moves up to 45 buses one-way 45 8 10 13 15
in a single lane in 15 min (180/hour); however, this is achieved 60 10 13 19 20
by advance marshalling of buses into 3-bus platoons, and by auxil- NOTE: Computations are based on the following relationship:
iary rear-door fare collection during the evening peak hours to g (D + 6)
Pass. car equivalent per bus = ×
expedite passenger loading. C h
Several downtown streets carry bus volumes of 80 to 100 buses where:
per hour, where there are two or three boarding positions per stop, h = 2 sec per car;
and where passenger boarding is not concentrated at a single stop. g/C = green time/cycle ratio;
(This frequency corresponds to about 5,000 to 7,500 passengers 6 = additional time loss due to starting, stopping, and queuing, sec; and
D = dwell time per bus, sec.
per hour, depending on load factors.)
These bus volumes provide initial capacity ranges that are suit- SOURCE: Computed.
able for general planning purposes. They compare with maximum
streetcar volumes on city streets some 50 years ago approaching
150 cars per track per hour, under conditions of extensive queuing
and platoon loading at heavy stops (16). However, the street cars Dwell times may be governed by boarding demand (e.g., in the
had two-person operations, and large rear platforms where board- PM peak when relatively empty buses arrive at a heavily used stop),
ing passengers could assemble. alighting demand (e.g., in the AM peak at the same location), or
total interchanging passenger demand (e.g., at a major transfer
point on the system). In all cases, dwell times are proportional
Observed Bus Flows—Terminals to boarding and/or alighting volumes times the service time per
passenger.
Peak-hour bus flows at 13 major bus terminals in the United
1. Clearance times—Field observations of bus clearance times
States and Canada range from 2.5 buses per berth at the George
are limited. A British study (17) reported ‘‘dead time’’ (the time
Washington Bridge Terminal to 19 at the Eglinton Station,
spent standing at a stop with the doors closed) of 2 to 5 sec. On-
Toronto. The mean is 8.1; median, 8.0; and standard devia-
bus analysis of time spent at stops in New Haven and Boston
tion, 4.2.
suggests a dead time of about 4 to 6 sec, for opening and closing
The high berth productivity in Toronto reflects the special design
doors (18).
of the terminal (with multiple positions in each berthing area), the
Scheel and Foote (10) found that bus start-up times also range
wide doors on the trolley buses using the terminal, and the free
from 2 to 5 sec. The time for a bus to travel its own length after
transfer between bus and subway. The relatively low productivity
starting ranges from 5 to 10 sec, depending on acceleration and
at the New York terminals reflects the substantial number of inter-
traffic conditions. Accordingly, the following ranges are reason-
city buses that use the terminals and the single-entrance doors
able for normal operations:
available on many suburban buses.
This current experience suggests about 8 to 10 buses per berth T Start-up time: 2 to 5 sec
for commuter operations. Intercity berths can accommodate 1 to T Clearance: 5 to 10 sec
2 buses per hour. T Lag time (before passengers board): 2 to 5 sec
Thus, a reasonable range for clearance time appears to be be-
Passenger Service Times and Bus Headways tween 9 and 20 sec for conventional buses, depending on traffic
conditions and driver behavior.
2. Passenger service times—The amount of time required by
The passenger service times and dwell times at bus stops are
each boarding or alighting passenger depends on many factors (19,
necessary for estimating bus and passenger capacities, and the
20). These include:
capacity increases that would result from changes in equipment
or operating practices. More specifically, they provide the key T Number and widths of doors used.
parameters for capacity calculations. T Number and height of steps.
The minimum headway of buses at a stop consists of (1) actual T Type of door actuation control.
dwell time when the bus doors are open for boarding and alighting, T Fare collection system.
plus (2) clearance times between buses. The time lost in opening T Amount of baggage or parcels carried by passengers.
and closing doors may be added to the dwell times, or incorporated T Procedures and time required to serve wheelchair passen-
in the clearance intervals. gers.
12-12 urban streets
Table 12-9. Passenger Boarding and Alighting Times Observed ranges in passenger service times for various bus op-
Related to Service Conditions erating and fare collection procedures in U.S. cities are summa-
time (seconds rized in Table 12-9. Boarding times are greater than alighting
per times. American experience with single-door buses shows passen-
conditions passenger) ger boarding times ranging from 2 sec (single-coin) to over 8 sec
Unloading (Alighting) for multiple-zone fares. Alighting times vary from about 11⁄2 to
21⁄2 sec for typical urban conditions to 6 sec or more where baggage
Very little hand baggage and parcels; few transfers 1.5 to 2.5
is involved.
Moderate amount hand baggage or many transfers 2.5 to 4.0
Considerable baggage from racks (intercity runs) 4.0 to 6.0 The importance of fare collection procedures to bus berth capac-
ities is apparent. A simplified method can substantially reduce
Loadinga (Boarding) service times per boarding passenger. Zone fare collection
Prepayment before entering bus or pay when leaving 1.5 to 2.5 schemes, which require monitoring of access and egress points,
bus are the most time consuming.
Single coin or token with fare box 2.0 to 3.0 Ranges in bus service time in relation to door width, methods
Odd-penny multiple-coin cash fares, paid on vehicle 3.0 to 4.0 of operation, and fare collection practices are given in Table 12-
Zone fares prepaid and registered on bus 4.0 to 6.0
10. These suggested operating service times (seconds per pas-
Multiple zone fares; cash; including registration on 6.0 to 8.0
bus senger) based on current experience provide a basis for esti-
a
mating bus dwell times at stop and, in turn, bus and person
Add 1 sec where fare receipts are involved.
NOTE: Assumes single channel loading. capacity.
SOURCE: Adapted from Ref. 4. Suggested service times for typical operating conditions—single
door loading, pay on bus—are:
T Seating configuration.
T Aisle width. Boarding
T The mix of alighting vs. boarding. 2.6 sec single coin
T The condition and configuration of the pavement, curb, and 3.0 sec exact fare
stop area. 3.5 sec exact fare—standees on bus
Time at stops is reduced and there are less boarding accidents Alighting
when the vehicle floor is flush with the station platform. This is 1.7 to 2.0 sec
commonly achieved in rapid transit, but is not currently used in
bus operation for safety reasons. Passenger service times decrease as the number of door chan-
Research on passenger service times found the following (19, nels available to passengers increases. The time values in Table
20): 12-10 reflect inefficiencies in using additional doorway capacity.
For example, one passenger may occupy a double door; moreover,
T There is no difference between front and rear door alighting passengers do not distribute themselves uniformly among doorway
times. openings. The values do not, however, reflect doorway and aisle
T Using both doors to alight requires more than one-half the turbulence at points of heavy simultaneous boarding and alighting
time than it does to alight from one door. Time reductions of 27 (see Refs. 19 and 20 for more details).
to 80 percent were observed. The values assume that prepayment would reduce passenger
T For alighting passengers, double stream doors require 27 to service time, a reasonable assumption for downtown busways
46 percent less time than single stream doors. and bus terminals. However, many of the values for multi-
T Rear door boarding times for double stream doors were ob- channel doors and multidoor loading are derived, because rel-
served to be 0.4 sec per passenger faster than for double stream atively little operating experience is available in the United States.
front doors, a reduction of 30 percent. 3. Dwell times at bus stops—The amount of time that buses
T The use of boarding through both doors requires less time spend at specific stops reflects the time of day, location of stop,
than for one door, but the time requirements for two doors is more surrounding land uses, and the number of interchanging transit
than half that required for one door. lines. Stops during the PM rush hour generally average less than
T Reducing double seats on each side of the vehicle to a single 15 to 20 sec; however buses may spend 30 to 60 sec at major
seat on one side of the vehicle may reduce passenger service time transfer points, terminals, or rail-bus interchange locations. Within
during periods of peak flow. the central business district, dwell times average 50 to 60 sec at
T Boarding service time requirements exceed those for busy locations, although individual stops may be as long as 2
alighting. min (21).
T Alighting times are greater when boarding passengers are 4. Queuing at stops—Studies of bus flow found that queues of
present. 2 to 4 buses develop approximately 20 percent of the time when
T Fewer delays to alighting and boarding passengers occur when bus volumes exceed 100 per hour at various locations along Michi-
boarding queues are organized and orderly. gan Avenue, Chicago (22).
T The presence of standees increases passenger service time.
Observations of bus operations on Bloomfield Avenue in Newark,
New Jersey, indicated an increase of 20 percent in boarding and
General Capacity Ranges
alighting times when standees were present. It was observed that
standees did not always interfere with the boarding and alighting
of passengers. The observed peak-hour bus movements along freeways, city
Table 12-10. Typical Bus Passenger Boarding and Alighting Service Times for Selected Bus Types and Door
Configurations (Seconds Per Passenger)
available doors or channels typical boarding service timesa typical alighting
bus type number location prepaymentb single coin farec service times
Conventional 1 Front 2.0 sec 2.6 to 3.0 sec 1.7 to 2.0 sec
1 Rear 2.0 sec N.A. 1.7 to 2.0 sec
2 Front 1.2 sec 1.8 to 2.0 sec 1.0 to 1.2 sec
2 Rear 1.2 sec N.A. 1.0 to 1.2 sec
2 Front, Reard 1.2 sec N.A. 0.9 sec
4 Front, Rearf 0.7 sec N.A. 0.6 sec
Articulated 3 Front, Rear, Center 0.9 secf N.A. 0.8 sec
2 Rear 1.2 secg N.A. —
2 Front, Centerd — — 0.6 sec
6 Front, Rear, Centere 0.5 sec N.A. 0.4 sec
Special Single Unit 6 3 Double Doorsh 0.5 sec N.A. 0.4 sec
a
Typical interval in seconds between successive boarding and alighting passengers. Does not allow for clearance times between successive buses or dead time at
stop.
b
Also applies to pay-on-leave or free transfer situations.
c
Not applicable with rear-door boarding. Higher end of range is for exact fare.
d
One each.
e
Two double doors each position.
f
Less use of separated doors for simultaneous loading and unloading.
g
Double door rear loading with single exits, typical European design. Provides one-way flow within vehicle, reducing internal congestion. Desirable for line-
haul, especially if 2-person operation is feasible. May not be best configuration for busway operation.
h
Examples: Neoplan TR-40 Mobile Lounge designed by Trepal Systems, Inc., for airport apron use.
SOURCE: Refs. 4, 17, 31, 36.
streets, and to or from bus terminals provide guidelines for 1. Bus capacity—Suggested arterial street bus capacity ranges
estimating the capacity of similar facilities. They also provide based on actual operating experience are given in Table 12-11.
means of checking or verifying more detailed capacity calculations. This table gives representative service volumes for downtown
General guidelines for planning purposes follow: streets and arterial streets leading to the city center for each
Table 12-11. Suggested Bus Flow Service Volumes for Planning

Purposes (Flow Rates for Exclusive or Near-Exclusive Lane)
level of service description buses/lane/hour midvalue
Arterial Streets
A Free Flow 25 or less 15
B Stable Flow, 26 to 45 35
Unconstrained
C Stable Flow, 46 to 75 60
Interference
D Stable Flow, 76 to 105 90
Some Platooning
E Unstable Flow, 106 to 135 120
Queuing
F Forced Flow, over 135a 150a
Poor Operation
Main CBD Street
A Free Flow 20 or less 15
B Stable Flow, 21 to 40 30
Unconstrained
C Stable Flow, 41 to 60 50
Interference
D Stable Flow, 61 to 80 70
Some Platooning
E Unstable Flow, 81 to 100 90
Queuing
F Forced Flow, over 100a 110a
Poor Operation
a
Results in more than one-lane operation.
SOURCE. Adapted from Refs. 5 and 34.
12-14 urban streets
level of service. Where stops are not heavily patronized, as along Wilshire Boulevard. The Transportation and Traffic Engineering
outlying arterial streets, volumes could be increased by about 25 Handbook suggests peak-hour factors of 0.70 to 0.95. A peak-
percent. hour factor close to 1.0 may well indicate system overloading
These service volumes may be used for planning purposes. More (underservicing) and reveal the potential for more service.
precise values for operations and design purposes should be com- A peak-hour factor of 0.80 would result in a maximum one-
puted from the capacity relationships and procedures set forth in way hourly passenger volume of about 6,000 persons on downtown
the following sections. streets and 8,000 persons on arterial streets.
THE VALUES FOR LOS F, FORCED FLOW CONDITIONS,
SHOULD NOT BE USED FOR PLANNING OR DESIGN. They RAIL TRANSIT CAPACITY EXPERIENCE
are merely given for comparative purposes.
2. Person capacity—The people per hour that can be served by This section briefly overviews peak-hour rail transit ridership
varying bus flow rates and passenger load factors are given in in the United States and Canada, and its passenger capacity impli-
Table 12-12. This table provides a broad person-capacity planning cations. More detailed information on rail transit ridership and
guide assuming that key boarding points are sufficiently dispersed capacity is set forth in a variety of references (3,7,8,9,12,24,25,26).
to achieve these bus loads. It suggests maximum person-flow rates The rail transit capacities have been included to provide a com-
of about 7,500 people per hour per lane on downtown streets and plete picture of urban transit capacities and to indicate the passen-
10,000 people per hour per lane on arterial streets. Corresponding ger volume ranges for which rail transit may be appropriate. Thus,
maximum values for seated passenger flow rates are 5,000 and they provide an important input for modal planning decisions. In
6,750 people respectively. Exclusive use of articulated buses would addition, light rail transit operates on city streets and affects street
increase these values 15 to 20 percent. operations.
3. Peak-hour factor—These person-flow rates indicate the num- Rail transit encompasses a variety of modes—each with distinc-
ber of people that can be carried, assuming uniform flow during tive service and performance characteristics. It includes commuter
the peak-hour. Accordingly, appropriate peak-hour factors should rail lines (both electric and diesel); urban rapid transit (both city
be used in discounting these values to reflect flow-variations within and suburban systems), street car and light-rail transit with both
the 15-min peak hour. Preferably, these flows should be compared on- and off-street running. All U.S. and Canadian systems, except
directly with the observed 15-min volume multiplied by four. for Montreal’s rubber-tire Metro, operate on conventional steel
The peak-hour factor (PHF) is defined as the hourly volume rails.
divided by four times the highest 15-min volume occurring within All types of rail transit except street car and light rail lines
the hour. The actual hourly volume can be calculated by: operate totally off-street. Light rail transit vehicles (LRV’s) operate
singly or in trains (1) on streets in mixed traffic or within reserved
HV = (Peak 15-min volume) (4) (PHF) (12-4) areas or (2) off-street in exclusive rights-of-way.
Differences among rail transit modes also exist in station spacing
Typical peak-hour factors range from 0.60 to 0.95 for transit and design, fare structure and collection methods, train length and
lines. The Los Angeles SCRTD reports peak-hour factors of propulsion, degree of access control and markets served. Some-
0.66 for commuter buses and 0.74 for local buses on Route 83, times, however, such differences may be difficult to discern.
Table 12-12. Suggested Bus Passenger Service Volumes for Planning Purposes (Hourly Flow Rates Based on
50 Seats Per Bus)
level of level of service (passengers)

service buses/ A B C D E
(street) passengers/seat 0.00–0.50 0.51–0.75 0.76–1.00 1.01–1.25 1.26–1.50
Arterial Streets
A 25 or less 625 940 1,250 1,560 1,875
B 26 to 45 1,125 1,690 2,250 2,810 3,375
C 46 to 80 2,000 3,000 4,000 5,000 6,000
D 81 to 105 2,625 3,940 5,250 6,560 7,875
E 106 to 135 3,375 5,060 6,750 8,440 10,125
CBD Streets
A 20 or less 500 750 1,000 1,250 1,500
B 21 to 40 1,000 1,500 2,000 2,500 3,000
C 41 to 60 1,500 2,250 3,000 3,750 4,500
D 61 to 80 2,000 3,000 4,000 5,000 6,000
E 81 to 100 2,500 3,750 5,000 6,250 7,500
NOTE: Ratio shown for level of service (passengers) is ‘‘passengers per seat’’ on average bus. Thus 1.00 means 50 passengers for the assumed 50 seats.
Values would be 6 percent higher for a 53-seat bus.
Values for articulated buses would be 15 to 20 percent greater.
SOURCE: Computed.
Table 12-13. Observed Peak-Hour Passenger Volumes on U.S. and Canadian Rapid Transit Systems
(in Peak Directions)
approx. persons/
trains cars car hour in peak passengers
per per headway length ft direction (max. per train
city and year line/location hour hour seconds (rounded) load section) (rounded)
New York City 1982 IND E, F, 53rd St. Tunnel 26 208 128 75 54,500 2,100
IND A, D, 8th Ave. Express 21 210 159 60, 75 43,500 2,070
IRT 4, 5, Lexington Ave. Exp. 25 250 157 50 38,100 1,520
PATH-World Trade Centera 38 266 98 50 25,500 670
1960 IND E, F, 53rd St. Tunnel 32 320 112 60 61,400 1,920
IND A, D, 8th Ave. Express 30 300 120 60 62,000 2,070
IND 4, 5, Lexington Ave. Exp. 31 310 116 50 44,500 1,430
IND 2, 3, 7th Ave. Express 24 240 150 50 36,800 1,530
Toronto 1978 Yonge St. 30 210 120 75 32,000 1,060
1974 Yonge St. 28 168 129 75 36,000 1,290
1960 Yonge St. 28 224 129 57 32,200 1,260
Montreal 1976 N Line 23 207 157 56 28,200 940
Chicago 1984 Milwaukee 17 136 212 50 12,400 730
Lake-Ryan 19 152 189 50 12,300 647
North-South 15 120 240 50 11,400 760
1978 Lake-Ryan 21 168 111 50 16,500 790
North-South 20 160 180 50 14,000 700
Philadelphia 1976 North Broad (2 tracks) 23 126 157 67 10,600 460
Boston 1977–78 Red Line 17 68 212 70 13,000 460
Orange Line 13 52 277 55 8,400 650
San Francisco 1977 BART-Transbay 11 98 327 75 8,000 730
BART-Mission 10 85 360 75 6,500 650
Washington 1980 Blue-Orange 20 120 180 75 13,000 650
Atlanta 1980 East Line 6 36 600 75 4,250 710
Cleveland 1976 West Side 14 52 258 50, 70 5,400 390
1960 West Side 20 80 180 50 6,200 360
a
Multiple track terminal.
SOURCE: Refs. 1, 7, 8, 9. New York Metropolitan Transportation Council, Chicago Transit Authority.
For rail systems other than street cars, travel times between 1. Rapid transit—There is a wide range of peak-hour passen-
stations are relatively unaffected by increased passenger volumes gers carried by the various rapid transit lines. This range reflects
or service frequency. factors such as the number, length, and frequency of the trains
operated. Especially important are the peak-period trains assigned
for scheduling purposes, the demands in the specific corridors, and
Observed Train and Passenger Flows the configuration or constraints of principal stations and switching
points.
The operating experience for typical rail rapid transit and light- There are generally less than 30 trains per track during the
rail transit lines is given in Tables 12-13 and 12-14, respectively. peak hour in the United States and Canada, although during
These tables give typical peak-hour peak-direction passenger vol- portions of this period slightly shorter headways are operated.
umes, service frequencies, and train lengths for principal U.S. and In general, the 90-sec headway that is possible with modern
Canadian rail transit lines. signaling systems is not realized on an hourly basis in the
These figures mainly reflect current (1976–1984) experience. United States and Canada. The single exception is the PATH
However, since many of the lines carried substantially higher pas- system, which operates 38 trains per hour on a single track
senger flows in peak years, 1946–1960 data are shown for compar- under the Hudson River from a multitrack World Trade Center
ative purposes. Thus, the observed number of peak-hour Terminal in New York City; in this case signals and interlocking
passengers at the maximum load point usually reflects demand points limit capacities.
rather than capacity. Peak 15- to 20-min volumes expressed as Before the State Street subway was opened, Chicago’s elevated
hourly flow rates are about 15 percent higher. Loop carried 78 trains/track with 438 cars in a single hour.
12-16 urban streets
Table 12-14. Observed Peak-Hour Passenger Volumes on Street Car and Light Rail Systems in United States
and Canada (Peak Direction)
trains cars length passengers/hour
per per headway of car in peak passengers/car
city location year hour hour seconds or train direction or train equipment
ON STREET
Pittsburgh Smithfield St. Bridge 1949 120 120 30 46.5 9,000a 75a PCC
Pittsburgh Smithfield St. 1976 51 51 71 46.5 3,800 74 PCC
San Francisco Market Street 1977 68 68 53 46.0 4,900 72 PCC
(before subway)
Toronto Queen St. East 1978 66 66 55 46.5 4,200 64 PCC
IN TUNNEL OR OFF STREET

Philadelphia Market St. 1956 133 133 27 46.0 9,000 67 PCC
Boston Green Line 1976 36 88 100 46.5 6,900 192 PCC
(Boylston St.)
Philadelphia Market Street 1978 73 73 180 46.0 3,700 151 PCC
San Francisco Market Street 1983 NA 62 NA 70.0 6,340 109 Boeing LRV
Cleveland Shaker Hts. 1976 30a 60a 120a 50.0 4,400 143 PCC
Boston Green Line 1978 16 48 225 46.5 1,500 94 PCC
(Lechmere)
Newark City Subway 1978 30 30 120 46.5 1,500 50 PCC
Edmonton LRT Line 1978 12 24 300 77.0 2,100 87 Duewag
San Diego LRT 1981 3 6 1,200 151 600 200 Duewag
a
Estimated.
SOURCE: Refs. 7, 8, 9.
This high movement of trains resulted from manual train control on the Newark City subway and Boston’s Lechmere line to over
and platoon loading of trains at stations. The Loop under cab 6,000 persons per hour on Boston’s Boylston Street subway and
signal control has carried up to 35 trains per track in peak hours. San Francisco’s Market Street subway.
Train lengths of 4 to 10 cars are commonly operated. Maximum The observed volumes generally reflect passenger demands and
train lengths range up to 8 cars in Chicago, San Francisco and scheduling policies rather than maximum possible flows. Flows as
Toronto, and 10 cars in New York City. The IRT Flushing Line high as 15,000 persons per hour have been observed in the past.
in New York City is the only line that operates 11 cars per train. Moreover several European systems report peak flows ranging up
Rail car lengths range from about 50 ft in Chicago and New to 18,000 persons per hour. Philadelphia’s Market Street subway has
York City (IRT, PATH) to 75 ft in Washington, San Francisco, carried 140 cars per hour with a minimum headway of 23 sec. It
and New York (new cars). Maximum train lengths are 600 ft. has carried as many as 12,000 people per hour. Boston’s Tremont-
Peak-hour passengers carried per track past the maximum load Boylston LRT subways traditionally scheduled 60 trains and 250
point range upward from 5,400 in Cleveland to 36,000 in Toronto cars per hour. The MBTA estimated the capacity of 15,000 persons
and over 50,000 in New York City, as of 1974–1983. The highest per hour in 1971 when inbound peak flows approximated 12,000
volumes carried are found on the Queens-Manhattan trains passing persons (27).
through the 53rd Street Tunnel: 53,000 persons per hour per track
in one direction. This line carried more than 60,000 passengers General Capacity Ranges
per hour, one-way, in previous years.
The capacity of a rail line is determined by station capacity or
Lines in cities such as Moscow, Tokyo, and São Paulo carry
way capacity, whichever is smaller. In most cases, station (or stop)
peak-hour flows of 50,000 to 60,000 persons per track per hour.
capacity governs.
2. Light rail transit—Operating characteristics of U.S. and Ca-
Capacity depends on: (1) car size and train-station length, (2)
nadian LRT and streetcar lines are given in Table 12-14.
allowable standees as determined by scheduling policy, and (3)
Post-World War II streetcars operated at 30-sec headways both the minimum spacing (headway) between trains. This minimum
on-street (Pittsburgh) and in tunnel (Philadelphia). Peak-hour pas- headway is a function not only of dwell times at major stations,
senger flows approximated 9,000 persons per hour. but also train length, acceleration and deceleration rates (including
Current operating experience shows up to 75 single cars per deceleration), and train control systems.
track carrying 5,000 passenger per hour. San Francisco’s Market Time-space diagrams can be used to estimate the ‘‘safe separa-
street surface routes carried 4,900 peak-hour one-way passengers tion’’ or minimum headway between trains. Theoretical ap-
per hour before they were placed underground. Approximately proaches to estimating the minimum spacing are sometimes used;
4,000 passengers per hour are carried by Toronto’s Queen Street examples are given in Appendix II. A more common practice is
line, and Pittsburgh’s Smithfield Street Bridge lines. to obtain the minimum spacing between trains based on actual
Both 50-ft and 70-ft to 75-ft cars operate in two and three car experience, station dwell times, and signal control systems.
trains. Up to 40 trains per track with 90 cars are operated. Passenger capacity in the peak direction during the peak hour
Off-street passenger volumes range from 1,500 people/hour can be estimated from the following equations:
Trains Cars Seats Passengers T The ‘‘crush load’’ capacities are shown for comparative pur-
Passengers/hour = × × ×
Hour Train Car Seat poses. They should not be used in determining capacity.
(12-5a) T LOS D (5.0 sq ft per passenger) represents a realistic value for
use in transit operations and planning. It results in capacities
or ranging from 18,000 to 30,000 persons per hour for train
lengths up to 600 ft. These figures compare with 20,000 to
Cars Seats Passengers 34,000 persons per hour per track derived by Pushkarev as a
Passengers/hour = × × (12-5b) comfortable peak-hour capacity (7).
Hour Car Seat
Where signal controls, station dwell times, and operating policies
An alternative formulation, based on allowable levels of pedes- allow closer than 2-min headways, capacities can be increased
trian space, is as follows: accordingly.
In estimating rail transit capacities and levels of service for over-
Ft2 Ft2
@Passenger
Trains Cars crowding, it is essential to analyze the peak 15-min period. For exam-
Passengers/hour = × ×
Hour Train Car ple, a ‘‘scheduled load’’ of 200 percent standees (3.3 sq ft per
(12-6) passenger) would relate to the peak 15-min period. Similarly, if an
hourly capacity of 27,000 people is provided by 6-car trains with
This latter formulation derives a passenger capacity that is inde- 200 percent standees, this implies that the peak 15 min would carry
pendent of the seating configuration and that directly relates to 27,000/4 = 6,750 people. If half of the peak-hour passengers moved
the area of each car. Cars that maximize total passenger capacity in 15 min, the effective hourly capacity would be 13,500. In this
generally minimize the number of seats. case, the peak-hour factor would be 0.5; therefore, the hourly service
The precise values for these equations will vary among individ- volume would be 0.5 × 27,000 or 13,500. These peaking characteris-
ual transit properties depending on the type of equipment used and tics further explain the differences between observed passengers and
operating policy. theoretical capacities (i.e., utilized and offered capacities).
2. Light rail transit—The passenger carrying capacity of light
1. Rapid transit—Typical ranges in rail rapid transit capacities rail transit (LRT) depends on vehicle size, train length, and
are summarized in Table 12-15 for U.S. and Canadian operating headway. However, the realizable LRT capacities also depend
experiences, based on 30 trains per track per hour. Ranges reflect on design and policy considerations that reflect specific local
varying car lengths (50 ft and 75 ft) and train sizes (6, 8, or 10 constraints of station design, at-grade operations, and type of
cars) and passenger load factors. These capacities can be adjusted right-of-way.
upward or downward based on specific operating policies. Detailed Where trains operate on-street, capacity estimates can be derived
car dimensions, seated passengers, schedule-loads, and crush loads by adapting the equations for bus transit (Section III) to account for
are contained in Appendix II. differing vehicle sizes, train lengths, and clearance requirements.
Levels of service are also shown in Table 12-15 for various Capacity estimates for off-street operations may be derived from
load factors, i.e., percent standees. These are keyed to the levels the approaches set forth for rail transit.
of service given in Table 12-6. LRT trains usually are limited to a maximum of three cars,
Table 12-15. Typical Rail Transit Capacities—30 Trains Per Track Per Hour, 2-Min Headway (Flow Rate)
passengers per hour
a
0% 50% 100% 150% 200% 250%
standees standees standees standees standees standees
car/length approx. seat load =
cars/train cars/hour (ft) seats/train (1.00)b (1.50)b (2.00)b (2.50)a (3.00)b (3.50)b
6 180 50 300 9,000 13,500 18,000 22,500 27,000 40,500
75 450 13,500 20,250 27,000 33,750 40,500 60,750
8 240 50 400 12,000 18,000 24,000 30,000 36,000 54,000

75 600 18,000 27,000 36,000 45,000 54,000 81,000
10 300 50 500 15,000 22,500 30,000 37,500 45,000 67,500

75c 750 22,500 33,750 45,000 56,250 67,500 101,250
FT2/Passenger: 10.0 6.7 5.0 4.0 3.3 2.6
Passenger level of service
B C D E-1 E-2 F
(U.S. & Canada Conditions)
Maximum Not attainable
Comments: schedule on a train
loads basis
a
Approximate.
b
Passengers per seat.
c
This condition does not exist in the United States.
SOURCE: Ref. 34.
12-18 urban streets
where on-street operation is involved: (1) longer trains could not ridership demands, route convergence limitations, and terminal
operate on city streets without simultaneously occupying more constraints.
than the space between adjacent cross streets when traversing short Typical ranges in capacities are as follows:
blocks; (2) long trains cannot clear at-grade intersections rapidly;
and (3) they need long platform lengths at stations. Pass. Level of Service
Minimum headways for light-rail systems will depend on train
length, platform design (high versus low), fare collection methods D E
(prepayment versus pay on train), and headway controls (manual Units 5.0 Sq Ft Max. Schedule
versus block signals). Under manual operations, 80 to 100 single- Per Per Loads 3.3 sq ft
unit cars per track per hour could be accommodated. When trains run Hour Person Per Person
under block signal controls, as is common with rapid transit systems, Street cars
120-sec headways are common, although shorter headways could (single 46–50 ft 90 7,500 12,000
unit on street)
be realized.
LRT—Off street 30 11,000 17,500
At 120-sec headways, a high-speed LRT system operating on (three 75-ft car 35 13,000 20,000
mainly reserved right-of-way with three-unit Boeing vehicle units)
trains would have a line capacity slightly in excess of 6,000
seated and 19,000 total passengers per hour (thirty 3-car units
at 211 persons/car). Under single-vehicle manual operation at
lower speeds, closer headways are feasible. At 60-sec headways, Current operating experience in the United States and Canada
single Boeing LRT units have a capacity of 4,000 seated and suggests maximum realizable capacities of 12,000 to 15,000 per-
13,000 total passengers per hour (schedule load) (29). However, sons per track per hour. However, the European experience
in practice these capacities are not realized because of limited shows up to 20,000 persons per hour (Appendix II).
III. METHODS AND PROCEDURES—ON-STREET TRANSIT
This section sets forth detailed methods and procedures for esti- The capacities of bus routes, terminals and busways, in persons
mating on-street transit capacity. It contains approaches for: carried, are generally limited by the ability of stops or loading
areas to pick up and discharge passengers. Just as the critical
1. Estimating bus berth and system capacity. signalized intersection usually determines arterial street capacity,
2. Planning bus priority treatments. bus route capacity is determined by the passenger service times at
major loading and unloading points.
Although the methods are keyed to bus transit, many can be One of the basic considerations in analyzing bus capacity,
applied to on-street light rail transit operations. Illustrative applica- therefore, is the bus berth or bus stop, and its ability to process
tions show how the methods can be used. buses and passengers. Capacity is influenced by many factors,
including the type and number of berths, number of boarding
and alighting passengers at major stops, design of the vehicle,
BUS BERTH AND SYSTEM CAPACITY
method of fare collection, location of the berth, bus layover
The section describes detailed methods for estimating the capac- practices at terminals, other operating policies, and traffic signal
ity of a bus berth, bus stop, or bus route. These approaches should controls.
be used for operations and design purposes. Capacity values as- Each bus requires a certain amount of service time at stops
sume that the bus lane or stop area would be exclusively for bus that varies with the number of boarding and alighting passengers,
use. Where other traffic shares a lane with transit vehicles, the door configuration of buses, and methods of fare collection. The
time needed for this traffic should be deducted. The net time should minimum safe spacing between buses in motion and the number
then be used for transit capacity analysis. of loading positions available at any stop also influence the total
number of buses and people that a given stream can carry. Bus
volume may be increased where vehicles can overtake or pass
General Considerations each other in entering or leaving loading positions.
The most common form of berth is the linear bus stop at a
The capacity of a bus system is determined by the capacity street curb. The length of such a berth may be adjusted to
of the most heavily used stop or the capacity of the line, whichever simultaneously handle multiple buses within reason, and buses not
is less. Consequently, theoretical capacities for uninterrupted stopping may pass stopped buses in other lanes where street
flow have little practical application for other than express width permits. The same type of stop may be provided under
nonstop runs. two conditions on busways:
1. In the travel lane (on-line), in which case following buses passengers board a bus at 3-sec headways, the total dwell time
may not pass the stopped bus. could be 35 sec. As formulated in the various capacity analyses
2. Out of the travel lane (off-line), in which case following herein, this door opening and closing time is included in the
buses can pass stopped vehicles. clearance time between buses. Whichever approach is used, the
door opening and closing time must be considered.
Berths in bus terminals may be linear, or they may take various The reductive factor R is 0.833 for maximum bus capacity. This
other forms. Angle berths are limited to one bus per berth, and occurs when one-third of all buses are waiting in approach queues,
they require buses to back out. Drive-thru angle berths are also reducing the capacity of the berth to about three-quarters of the
feasible, and may accommodate multiple vehicles. Shallow ‘‘saw- ideal value. Thus, 3,000 normally replaces 3,600R in the equation.
tooth’’ berths are popular in urban bus-rail interchange terminals, The minimum headway, h′, can be obtained as follows:
and are designed to permit independent movements into and out
of each berth. Boarding only; one way flow
h′ = bB + tc (12-8a)
Equations and Guidelines Alighting only; one way flow
The following equations show how the capacity of a busway, h′ = aA + tc (12-8b)

bus terminal, or city street can be estimated in terms of (1) buses
Two-way flow through door
per hour, and (2) people per hour. They establish ranges in typical
time requirements for each of the operations at a bus berth, and they
h′ = [aA + bB + tc] (12-8c)
identify relationships between bus passenger line-haul capacity,
boarding and alighting volumes, and types of bus equipment. They where:
should be applied to the peak 15 min in each rush hour since
this period is when the maximum boarding and alighting volumes A = alighting passengers per bus, in peak 15 min;
normally occur. a = alighting service time, in seconds per passenger;
The number of buses that can be handled at stops without devel- B = boarding passengers per bus, in peak 15 min; and
oping unacceptably long queues (and associated waiting lines) b = boarding service time, in seconds per passenger.
varies principally with the service time per bus and, to a lesser
degree, with the number of loading positions. Additional loading Where passengers enter via the front door, and exit via the rear
spaces (or additional length of bus zones) increase the capacity, door, the greater result from Eqs. 12-8a and 12-8b determines the
but at a decreasing rate as the number of spaces increases. minimum headways. For heavy two-way flow through a single
door, the headways in Eq. 12-8c could be increased by 20 percent.
1. Capacity of a bus berth (vehicles)—The number of buses Substituting the appropriate values of h′ in Eq. 12-7 produces
that can use a bus berth (or stop) varies inversely with passenger the following equations for the maximum number of buses per
service (dwell) times, D, and bus clearance times, tc. The passenger berth per hour (R is assumed as 0.833):
service times depend on the number of boarding, alighting, or
interchanging passengers, fare collection practices, and door con- Boarding only; one way flow
figurations. The clearance time should include door opening and
closing times, when they are not incorporated into the dwell times 3,600 R 3,600 (0.833) 3,000
f′ = = = (12-9a)
directly. It includes all time losses associated with a bus entering h′ bB + tc bB + tc
and leaving a stop, other than passenger loading.
Alighting only; one-way flow
a. Uninterrupted flow, no delays caused by traffic signals—The
number of buses per berth per hour can be estimated from the 3,600 R 3,000
following set of equations. f′ = = (12-9b)
aA + tc aA + tc
3,600 R 3,600 R
f′ = = (12-7) Two-way flow through door
h′ D + tc
3,600 R 3,000
where: f′ = = (12-9c)
aA + bB + tc aA + bB + tc
h′ = minimum headway at the bus berth or stop;
D = dwell times at bus berth or stop; Equations 12-7 and 12-8 are precise where there are no delays
tc = clearance time between successive buses; due to traffic signals, as along a busway or at a terminal. For
R = reductive factor to compensate for dwell time and arrival city street operations at signalized intersections, they provide an
variations; and upper limit of berth (stop) capacity. Several downward adjust-
f ′ = maximum buses per berth per hour. ments are necessary, especially for dwell times less than 60 sec
per stop.
In estimating the total time that buses spend at a stop, the b. Bus flow interrupted by traffic signals—The number of
time spent opening and closing doors should be taken into buses that can stop for passengers and then pass through a
account. This normally approximates 4 to 5 sec. Thus, if 10 signalized intersection can be estimated from Eq. 12-10a and
12-20 urban streets
Table 12-16. Estimated Maximum Capacity of Bus A 60-sec dwell time per bus with a 15-sec clearance time would
Stops—Buses Per Hour (Flow Rate) result in a capacity of 33 buses per berth per hour for a g/C of
clearance time (sec) tc 0.5 and 40 buses per hour for a g/C of 1.0. These values are based
on 3,000 ‘‘effective’’ sec per hour. Corresponding values assuming
10 15
perfect schedule reliability and a uniform distribution of dwell
dwell time g/C g/C g/C g/C
(sec) 0.5 1.0 0.5 1.0 times during the peak 15 min would be 40 and 48 buses per hour
(i.e., R = 1.00).
15 86 120 67 100
30 60 75 50 67
The number of buses that are stopped at a traffic signal must
45 46 54 40 50 fit within the available block length. For short blocks, the block
60a 38 42 33 40 spacing may limit capacity. (This is even more true where LRT
75 32 35 28 33 trains run on-street.)
90 28 30 25 28 Where passenger boardings are dispensed along the transit line,
105 24 26 22 25 the passenger capacity can be estimated by multiplying the number
120 22 23 20 22
150 18 18 17 18 of buses per berth obtained from Table 12-16, or from Eq. 12-10b
180b 15 16 14 16 by (1) the number of effective loading positions and (2) the specific
a
Typical CBD Stop-PM peak.
loading standard, i.e., persons per vehicle.
b
Maximum CBD Stop-PM peak. d. Levels of service—Suggested levels of service for the num-
SOURCE: Computed from Eq. 12-10. Assumes R = 0.833. ber of buses per berth (i.e., per stop) are given in Table 12-17.
The levels of service are keyed to the approximate probability or
likelihood of queues forming behind the bus stop.
The number of buses per berth that can be accommodated at
any level of service can be estimated as follows:
12-10b. These equations assume that the time spent loading 3,000 (LOS Factor)i
cv(i) = (g/C) (12-11)
and/or discharging passengers on the green, g, and red, r, phases tc + D (g/C)
is proportional to the g/C and r/C ratios, respectively. The yellow
time is assumed as part of the green time. The equations are where: cv(i) is the capacity at service level i and (LOS Factor)i are
precise for near side stops and provide a reasonable approximation the index values given in column 4 of Table 12-17.
for far side stops. For example, assuming a g/C of 0.5, a dwell time, D, of 60 sec,
and a clearance of 15 sec, the number of buses per berth at service
g C would be
f ′c = (12-10a)
tc + D (g/C)
0.5(3,000) (0.80) 1,500 (0.80)
cv(C) = = = 26.7, say 27
3,600 R 3,000 15 + 60 (0.50) 45
f ′ = (g/C) = g/C (12-10b)
tc + D (g/C) tc + D (g/C)
Finally, if the peak-hour factor, PHF, were 0.67, the service
where: volume at LOS C would be 0.67 (27) or 18 buses per berth.
Typical (rounded) values for a 60-sec dwell time, 15-sec clear-
g = green (plus yellow) time per cycle, in sec; ance, and g/C ratios of 0.5 and 1.0 are given in Table 12-18.
C = cycle length, in sec; e. Berth use efficiency—Each loading position at a multiple-
f′c = buses per cycle; berth stop does not have the same capacity as a single berth stop.
D = dwell time per bus resulting from loading and/or unloading This is because it is not likely that the loading positions at a
passengers (i.e., bB, aA, [aA + bB], or if inflow and outflow multiberth stop will be equally used, or that passengers will distrib-
both heavy, use 1.2 [aA + bB]); ute equally among loading positions. Moreover, where stops are
tc = clearance time per bus; and designated for specific routes, bus schedules may not permit an
f ′ = buses per hour. even distribution of buses among loading positions. Buses also
may be delayed in entering or leaving a berth by buses in adjacent
Note that as g/C approaches 1, Eqs. 12-7 and 12-10b become loading positions.
identical. Both equations assume that there is no other traffic in The actual efficiency of a system of loading positions will also
the bus lane and that buses do not pass and overtake each other. vary with the type of design. Consequently, the design of the bus
c. Berth capacity values—The number of buses per hour is loading areas influences capacities.
given in Table 12-16 for g/C ratios of 0.5 and 1.0. Values are Suggested ‘‘berth efficiency factors’’ are given in Table 12-19
tabulated for clearance times of 10 and 15 sec, and dwell times for ‘‘on-line’’ and ‘‘off-line’’ stops. These factors are based on
ranging from 15 to 180 sec. This table can be used to estimate the experience at the Port Authority of New York and New Jersey’s
number of buses per hour that can be served by a single berth. Midtown Bus Terminal. The table suggests that four or five ‘‘on-
Values for g/C times between 0.5 and 1.0 can be interpolated; line’’ positions could have a maximum efficiency of 2.5 berths.
values for g/C times less than 0.5 and for other dwell times can Five ‘‘off-line’’ positions would have an efficiency of about 3.75
be computed directly from Eq. 12-10b. berths.
The 10-sec clearance time represents the absolute minimum Note that to provide for two ‘‘effective’’ berths, three physical
time spacing possible at a stop for conventional buses. However, berths would have to be provided (partial berths are never built).
for most situations the 15-sec clearance values should be used. Thus, Nb is not the number of berths which must necessarily
Table 12-17. Levels of Service for Bus Stops

1 2 3 4 5
effective index approx. probability
level of R sec/hour los e = of queues forming
service (los) value (3,600 R) 1.00 behind bus stop
A 0.400 1,200 0.40 1
B 0.500 1,800 0.60 2.5
C 0.667 2,400 0.80 10
D 0.750 2,700 0.90 20
E (Capacity) 0.833 3,000 1.00 30
Capacity
E- Perfect 1.000 3,600 50
Conditions
NOTE: For use in this equation:
3,600 R (g/C) 3,600 R
Cv = or
D + tc tc + (g/C) D
be built. Table 12-19 may be entered with knowledge of Nb to Table 12-18. Typical Service Levels, Single Stop, No
find the number that must be provided. Passing (15-Sec Bus Clearance Time, 60-Sec Stop)
Also note that Table 12-19 applies only to linear berths. All buses per hour (flow rate)
level of
other types of multiple berths are fully effective.
service g/C = 0.5 g/C = 1.0
f. Guidelines—Estimated capacities of on-line bus berths are
given in Table 12-20. This table shows the number of buses per A 13 16
B 20 24
hour for varying clearance times, dwell times, g/C ratios, and C 26 32
loading positions. The maximum capacities attainable are 2.5 times D 30 36
those for a single berth. E 33 40
Thus, for a 60-sec dwell time per bus, and a 15-sec clearance SOURCE: Computed from Tables 12-16 and 12-17.
time, a 5-berth stop would result in capacities of 82 and 100 buses
per hour at g/C ratios of 0.5 and 1.0 respectively.
Figure 12-3 provides a further planning guide for estimating
bus berth capacity. It shows the number of buses per hour for
selected dwell times and g/C ratios based on a 15-sec clearance passenger service times. Where such data are lacking or cannot
time. Increasing the number of loading positions has a much be obtained easily, the following representative values can be
smaller effect on changes in capacity than reducing dwell times. used: 60 sec per CBD stop, 30 sec per major outlying stop, and
Note that for dwell times more than 60 sec, the differences between 15 sec per typical outlying stop.
a g/C of 0.5 and 1.0 are small. Table 12-20 and Figure 12-3 provide a means of estimating
The application of Table 12-20 and Figure 12-3 calls for the number of buses per hour that can pass through the busiest
estimates of the approximate dwell times at the major stops. stop. The number of people that these buses can carry depends
These can be obtained from field observations or from counts on seated and standing passengers per bus—assuming that these
of the number of people boarding each bus and their associated ‘‘places’’ are filled as the bus reaches its maximum load point.
Table 12-19. Efficiency of Multiple Linear Bus Berths

on-line stations off-line stations
berth efficiency no. of cumulative efficiency no. of cumulative
no. % effective berths % effective berths
1 100 1.00 100 1.00
2 75 1.75 85 1.85
3 50 2.25 75 2.60
4 20 2.45 65 3.25
5 5 2.50 50 3.75
NOTE: On-line station figures assume that buses do not overtake each other. In Ref. 3, efficiency values were (1) 100
percent, (2) 73 percent, (3) 41 percent, (4) 27 percent, and (5) 18 percent. The resulting capacity factors (cumulative)
were 1.00, 1.73, 2.14, 2.41, and 2.54.
12-22 urban streets
Table 12-20. Estimated Capacity of On-Line Bus Stops by Number of Berths (Buses Per Hour)
no. of berths
1 2 3 4 5
no. of effective berths
1.00 1.75 2.25 2.45 2.50
10-sec Clearance g/C g/C g/C g/C g/C
Dwell Time/Stop 0.50 1.00 0.50 1.00 0.50 1.00 0.50 1.00 0.50 1.00
30 sec 60 75 105 131 135 169 147 184 150 188
60 seca 38 42 66 74 86 94 93 103 95 105
90 sec 28 30 49 52 63 68 69 74 70 75
120 sec 22 23 38 40 50 54 54 56 55 58
15-sec Clearance
Dwell Time/Stop
30 sec 50 67 88 117 112 151 122 164 125 168
60 seca 33 40 58 70 74 90 81 98 82 100
90 sec 25 28 44 49 56 63 61 69 62 70
120 sec 20 22 35 38 45 50 49 54 50 55
a
Typical bus stop (pm peak).
NOTE: Assumes R = 0.833.
SOURCE: Computed.
Figure 12-3. Bus stop capacity related to dwell times and loading positions (15-sec clearance between buses).
Thus, the number of people per hour that a bus route can carry 3. Passenger capacity of a bus route—The capacity of any
depends not only on the dwell times at the busiest stop, but also busway, bus terminal-approach system, downtown bus street or
on the distribution of boarding passengers along the line. bus lane will be governed by the number of passengers (a) boarding
The number of buses per hour that can pass the heaviest boarding and/or alighting at the heaviest stop or (b) traveling past the maxi-
point does not in itself determine the number of people per hour mum point (between stops), whichever is less. The sequence of
along the route. This can be illustrated as follows: analysis is as follows when the approach volumes of buses and
A dwell time of 60 sec per stop and 15-sec clearance time passengers are specified, and it is desired to estimate the required
results in 33 buses per hour for a single loading position. This number of berth positions:
corresponds to 2,000 people per hour on the bus route assuming
60 people per bus (g/C of 0.5). T The maximum load point demand establishes bus frequency
requirements in the corridor.
The 60-sec dwell time enables 20 people to board each bus T Bus service frequency and boarding volumes determine the
assuming a service time of 3 sec per passenger. This translates minimum headway per berth. (For planned systems, where no
into 1,200 people per berth per hour. boarding counts are available, the estimated percentage of passen-
Consequently, the 2,000 people/bus/hour can be achieved only gers boarding at the heaviest stop is a key parameter of total
if another 800 people board buses before the maximum load passenger capacity.)
point is reached. T The maximum bus frequency per berth depends on this mini-
2. Passenger capacity of a bus berth—The capacity of a bus mum headway.
berth in persons served per hour can be estimated from the follow- T Berth needs or stopping positions are derived from the re-
ing equations. They assume that loading conditions govern and quired bus frequency at the maximum load point and the maximum
that there are no traffic signal interruptions. Similar equations bus frequency that can load at the heaviest berth.
can be derived based on passenger interchange or alighting. For
The following equations show how the maximum load point
alighting, K replaces J in these equations.
and heaviest transit stops interrelate. They assume that loading
Maximum passengers per berth per hour, Q conditions govern. A similar set of equations would apply where
passengers alighting (or passenger interchange) dominate and de-
3,000 B termine capacities.
Q = f ′ B = R (3,600 B/(bB + tc) = (12-12)
bB + tc The capacity of a bus route at the maximum (peak) load point
is given by the following expression:
Effective berths required, Nb, to serve J passengers per hour
P=f×S (12-15)
J (bB + tc) bB + tc
Nb = J/Q = = (12-13)
R (3,600) B Rh′ where:
where R = 0.833. P = capacity of bus route past peak load point, in persons/hour;
Where traffic signals control bus movements, the following f = frequency of buses past the peak load point during peak
equations should be used: hour; and
S = passengers/bus.
3,600 BR
Q = f ′B = (g/C) (12-14a)
(tc + Bb (g/C)) Normally, P is derived based on the peak 15-min values for f
and S, and adjusted downward to an hourly basis by means of
J[tc + Bb (g/C)] a PHF.
Nb = J/Q = (12-14b)
(g/C) 3,600 BR
a. Uninterrupted flow, busway or bus terminal—The passengers
Since R equals 0.833, 3,000 replaces (3,600 R), and Q equals P at the maximum load point (maximum service volume) can be
maximum passengers per berth per hour; J equals number of pas- obtained as follows:
sengers boarding at heaviest stop, per hour (peak-hour flow rate);
K equals number of passengers alighting at the heaviest stop per T As a function of boarding passengers per bus at the busiest
hour, and other symbols are as described before. stop:
Table 12-21 contains illustrative calculations based on these equa-
tions for a single berth, assuming that loading conditions govern: 3,600 R NbS 3,000 NbS
P= = (12-16)
(bB + tc) bB + tc
a. Uninterrupted flow conditions (g/C = 1.00) would require 2
effective berths in the example shown. To achieve 2.0 effective T As a function of the proportion of the total passengers board-
berths, 3 physical berths would have to be provided, with a capacity ing at the busiest stop: The proportion of passengers boarding at
of 2.25 effective berths. the heaviest stop, X, equals B/S. Thus Eq. 12-16 becomes:
b. A 0.50 green/cycle would require 2.7 effective berths.
Providing 5 berths would result in 2.5 effective berths. With on- 3,600 NbR 3,000 Nb
P= = (12-17)
line stations in practice, 5 berths would be provided. To provide Xb + tc /S Xb + tc /S
sufficient capacity, loading times should be reduced by pre-
payment, rear-door loading and/or changes in stopping patterns,
and signal timing adjustments should be made. T As a function of the passenger capacity per berth:
12-24 urban streets
Table 12-21. Bus Berth Passenger Capacity Equations and Illustrative Examples—Boarding
Conditions Govern (R = 0.833)
case 1 uninterrrupted flow g/C = 1.00
variables equation (hourly rates) examples
Let: tc = 15 sec
b = 3 sec/pass.
B = 10 pass./bus
J = 1,340 boarding pass./hr
R = 0.833
Minimum headway at stop h′ = Bb + tc h′ = 10(3) + 15 = 45 sec
Maximum buses per berth per hour 3,000 f ′ = 3,000/45 = 67 buses/berth/hr
f ′ = R 3,600/h′ =
Bb + tc
Maximum passengers per berth per hour 3,000 B Q = 67(10) = 670 pass./berth/hr
Q = f′ B =
Bb + tc
Effective berths required to serve J J(Bb + tc) Nb = 1,340/670 = 2 berths
passengers per hour Nb = J/Q =
3,000 B
Bus frequency required to serve J f = f ′ Nb = J/B f = 2(67) = 134 buses
passengers per hour
case 2 g/C = 0.50
variables equation (hourly rates) examples
Same assumptions as Case 1
Minimum headway at stop h′ = 10(3) + 15 = 45 sec
1 2 115 + 10(3)(0.5)2 = 50
Maximum buses per berth per hour 3,600 R
f ′ = (0.5)
1C2
g 3,600 R
f′ =
g
tc + Bb
C
1 2
Maximum passengers per berth per hour Q = 50(10) = 500 pass./berth/hr
12
g 3,600
Q = f′ B = B
C g
tc + Bb
C
Effective berths required to serve J J[tc + Bb (g/C)] 1,340
passengers per hour Nb = J/Q = Nb = = 2.7 berths
(g/C) 3,600 BR 500
Bus-frequency required to serve J f = f ′ N b = J/B f = 50(2.7) = 135 buses

passengers per hour
SOURCE: Adapted from Refs. 4 and 34.
N bQ alighting service times required per passenger, and the clearance

P= =f×S (12-18)
X times between buses.
The following example shows how the equations can be ap-
T Number of berths at the busiest stop as a function of service plied. It is desired to find the total passengers that can be carried
volume at maximum load point: The number of effective berths past the maximum load point in an hour, based on the peak
at the busiest stop, Nb, to serve P passengers per hour is: 15-min flow rate. There is a 20-sec clearance between buses
(tc = 20), and 50 passenger buses with a load factor of 1.00
(S = 50). There is prepayment of fares and the ability to load
12 12
P (Xb + tc /S) P bXS + tc P bXS + tc
Nb = = = buses at a rate of 2.0 sec per passenger (b = 2.0). System design
3,600 R S 3,600 R S 3,000 anticipates that 50 percent of the total passengers will board at
(12-19) the maximum load point (X = 0.5).
Four off-line berths are provided, i.e., 3.25 effective berths.
Note that R = 0.833. The busway is grade separated in the central area, and there are
These equations indicate that the number of bus berths reno traffic signal interruptions.
quired at the heaviest stop or bus terminal varies directly with The number of people that can be carried on the system can
the total passengers to be served at that stop, the boarding and be estimated by applying Eq. 12-17.
(3,600) Nb (0.833) (3,000) (3.25) area with a g/C ratio of 0.5, then the number of people passing
P= =
Xb + tc (0.5)(2) + (20/50) the maximum load point would drop from 7,000 to 5,400 persons
= 6,964 = 7,000 persons/hr per hour. This result reflects the effects of traffic signal operations
and is obtained by applying Eq. 12-21b.
The actual hourly volume will be less since the capacity repre-
sents four times the peak 15-min flow rate. To calculate the hourly (0.833) (3,600) Nb (g/C)
P=
volume, a peak-hour factor, PHF, is used. In this example, if the [(g/C) Xb + tc /S]
PHF were 0.75, the hourly passenger volume carried would be
about 5,250 persons (i.e., 7,000 × 0.75). 3,000 (3.25) (0.5)
P= = 5,400
This represents level-of-service E, insofar as the movement of (0.5)(0.5)(2) + (20/50)
buses is concerned. The number of people that can be carried at
Applying a peak-hour factor of 0.75 would result in about 4,100
LOS D would be 0.90 (5,250) or about 4,700 persons per hour.
persons per hour.
b. Bus flow interrupted by traffic signals, arterial streets—The
c. Guidelines—The general expression for the maximum load
preceding equations should be modified as follows to account for
point passengers that can be carried for each effective berth at the
the reductive effect of traffic signals.
busiest stop provides a simple means of estimating system capac-
T As a function of boarding passengers per bus at the busiest ity. This equation is as follows:
stop: 3,600 (g/C) R 3,000 g/C
Pb = = (12-23)
3,600 Nb SgR 3,000 Nb Sg [Xb (g/C) + tc /S] [Xb (g/C) + tc /S]
P= = (12-20)
C [Bb (g/C) + tc] C [Bb (g/C) + tc] This equation is similar to Eq. 12-20 except that Pb is keyed to
T As a function of the proportion of the total volume boarding a single berth at the busiest stop, rather than to Nb effective berths.
at the heaviest stop: Typical values for key parameters that should be used with this
equation are as follows:
3,600 NbSgR 3,000 (NbSg)
P= = (12-21a) T Busway—prepayment of fares:
C [XbS (g/C) + tc] C [XbS (g/C) + tc]
or, by rearranging, 2 sec per passenger
3,600 Nb (g/C) R tc /S = 0.50 max
P= (12-21b) Peak-hour factor = 0.67 to 0.85
[Xb (g/C) + tc /S]
T Number of berths at the busiest stop: T Arterial street—pay on entering bus:
PC [Xb (g/C) + tc /S] 3 sec per passenger

Nb = tc /S = 0.30 to 0.40
(g) (3,600) R
Peak-hour factor = 0.67 to 0.85
PC [(g/C) bXS + tc]
Nb = (12-22) Table 12-22 gives values for Pb for uninterrupted flow conditions
(g) (3,600 S) R
(g/C = 1.0); it should be used for grade-separated busway
In the following example, assume that the busway in the conditions. Table 12-23 gives values for interrupted flow condi-
preceding example has signalized intersections in the central tions (g/C = 0.50) along city streets.
Table 12-22. Maximum Load Point Hourly Passengers Per Effective Berth at
Busiest Station—Uninterrupted Flow Conditions g/C = 1.00 (R = 0.833)
ratio: proportion of passengers boarding at busiest stop
(clearance between buses)/
(passengers per bus) 0.25 0.50 0.75 1.00
A. 2 Sec/Boarding Passenger
0.1 5,000 2,720 1,870 1,590
0.2 4,210 2,500 1,770 1,370
0.3 3,750 2,310 1,670 1,300
0.4 3,330 2,140 1,580 1,250
0.5 2,600 2,000 1,500 1,200
B. 3 Sec/Boarding Passenger
0.1 3,530 1,870 1,280 970
0.2 3,160 1,770 1,230 930
0.3 2,860 1,670 1,180 910
0.4 2,610 1,580 1,130 880
0.5 2,400 1,500 1,030 860
SOURCE: Computed from Eq. 12-23.
12-26 urban streets
Table 12-23. Maximum Load Point Hourly Passengers Per Effective Berth at Busiest Station—Interrupted
Flow Conditions g/C = 0.50 (R = 0.833)
ratio: proportion of passengers boarding at busiest stop
(clearance between buses)/
(passengers per bus) 0.25 0.50 0.75 1.00
A. 2 Sec/Boarding Passenger
0.1 4,280 2,500 1,770 1,370
0.2 3,330 2,140 1,580 1,250
0.3 2,720 1,870 1,420 1,150
0.4 2,310 1,670 1,310 1,075
0.5 2,000 1,500 1,200 1,000
B. 3 Sec/Boarding Passenger
0.1 3,160 1,770 1,220 930
0.2 2,610 1,580 1,130 880
0.3 2,220 1,420 1,050 830
0.4 1,930 1,320 980 790
0.5 1,720 1,200 920 750
SOURCE: Computed from Eq. 12-23.
Applications (2) Clearance time is 15 sec.

(3) Fifty percent of the peak load point passengers board at the
The following sections apply the methodology to busways, arte- heaviest stop.
rial streets, and bus terminals. They present key parameters and (4) Three loading berths are provided at the heaviest stop with
set forth ranges in capacities that are useful for planning, opera- loading and unloading areas separated.
tions, and design. (5) An adjustment factor of 0.75 is applied to all results to allow
for on-vehicle turbulence and schedule irregularity and variations
1. CBD busways—CBD busway capacity can be computed in dwell times at major bus stops. This R value corresponds to
from the preceding equations utilizing appropriate assumptions LOS D as shown in Table 12-17.
regarding type of bus used, maximum allowable bus loading, distri- (6) For D, a peak-hour factor of 0.67 is used to adjust from
bution of ridership among CBD stops, peak-hour factor, and type peak 15-min flow rates to full-hour volumes.
of berth. (7) Fares are prepaid (no fares collected on bus in the CBD);
the boarding time is 2 sec per passenger.
a. Bus use—The number of people per bus will depend on (1)
size of vehicles (about 50 seats/regular bus to 60 seats/articulated Table 12-25 gives the resulting average hourly passenger service
bus), and (2) operating policies with regard to standees. To provide volumes at the maximum load point for two types of stations and
an acceptable level of comfort for express bus commuters with a four types of bus operations. (Note that this reflects service level D.)
minimum nonstop run of 3 to 5 mi, the passenger load factor in Figure 12-4 shows how the door configuration and number of
the peak 15-min period should not exceed 1.00—i.e., there should berths increase the maximum load point capacity. The lower hori-
be a seat available for each passenger. Higher load factors are zontal scale applies to through-station operations where 50 percent
acceptable on local bus services. of all passengers board at the heaviest stop. The upper scale applies
b. Passenger distribution among CBD stops—A reasonable de- to a single-station situation where all riders board at the major
sign assumption is that 50 percent of the maximum load point stop. This figure can be used to estimate the number of passengers
volume is served at the heaviest CBD busway stop—assuming a per hour that can be accommodated by various numbers and types
minimum of three stops in the downtown area. (The Washington- of loading berths.
State Street subway station in Chicago accounts for about half of 2. Bus operations on city streets—The most common operating
all boarding passengers at the three downtown stops on the State condition for bus services is along downtown and arterial streets.
Street subway line.) Transit capacity estimates are complex in this setting, because:
c. Peak-hour factor—Peak-hour factors of 0.67 to 0.75 are rea-
sonable, depending on the location and type of operation. a. Buses must share the roadway with other vehicles, and are
d. Capacity guidelines—Illustrative busway capacity guidelines subject to interference from other elements of the traffic stream,
for central areas are given in Table 12-24 for a variety of bus such as traffic signals.
types and service conditions. The key assumptions are: b. The number and percentage of buses stopping at each bus
(1) The peak load point volume is limited to 50 passengers/bus stop depend on demand and operational factors such as ‘‘bunch-
for standard vehicles and 60 passengers/bus for articulated vehi- ing’’ of buses in platoons.
cles; this corresponds to a load factor of approximately 1.00, or c. Buses are subject to a variety of conflicts at intersections
level-of-service C, and provides a seat for all passengers. For other with pedestrians and turning vehicles, which add delay to the
load factors, multiply the values cited by the load factor transit system.
(the number of passengers/seat). d. Passenger loading and unloading take place both on the
Table 12-24. Illustrative Bus Capacity Guidelines for CBD Busways

loading condition
A B C D
station on-line off-line on-line off-line on-line off-line on-line off-line
Passengers boarding at heaviest stop
Number of passengers per bus 25 25 25 25 25 25 30 30
Boarding time per passenger, in seconds 2.0 2.0 1.2 1.2 0.7 0.7 0.5 0.5
Total boarding time, in secondsa 65 65 45 45 32.5 32.5 30 30
Berth use, in buses per hour
Maximum buses per hour per berthb 55 55 80 80 111 111 120 120
Use factor, for three berths 2.25 2.60 2.25 2.60 2.25 2.60 2.25 2.60
Total for all berths 124 143 180 208 250 289 270 312
Adjusted total for all berthsc 93 107 135 156 188 217 200 234
Passengers per hour—maximum load
point d
Peakd—flow rate (15 min × 4) 4,650 5,350 6,750 7,800 9,400 10,850 12,000 14,040
Averagee—peak hour 3,115 3,570 4,520 5,200 6,300 7,320 8,040 9,360
Loading condition A: Single door conventional bus, simultaneous loading and unloading
Loading condition B: Two-door conventional bus, both doors loading or double-stream doors simultaneously loading and unloading
Loading condition C: Four-door conventional bus, all double-stream doors loading
Loading condition D: Six-door articulated bus, all doors loading
a
Includes 15-sec bus clearance interval.
b
Computed based on 3,600 effective sec per hour (R = 1).
c
Adjusted by a factor of 0.75 to account for turbulence, schedule irregularities, and the like.
d
Assumes that 50 percent of all passengers board at heaviest stop.
e
Adjusted by a factor of 0.67 from peak flow rate.
SOURCE: Adapted from Ref. 5, p. 39, Table 8.
green and red signal, and hence capacities are less than for uninter- Table 12-25. Busway Service Volumes at Maximum
rupted flow conditions. Load Points (Passengers Per Hour)
Aside from bus berth capacity considerations, the capacity of on-line off-line
an arterial street for buses is influenced by the total capacity of type of operation stationse stationse
the street and the amount of other traffic present. Conventional busa
Arterial street bus passenger capacity at maximum load points One door available 3,100 3,550
Two doors availableb 4,500 5,200
should be estimated by using the general factors given in Table Four doors availablec 6,300 7,250
12-23. Alternatively, the dwell times given in Table 12-16 can Articulated
be used, by making assumptions regarding load factors and 60-passenger busd 8,050 9,350
passenger distribution. In both cases, a maximum of 2.5 effective a
Single door for loading.
b
berths should be used because on-line stopping conditions prevail. Double-door entrance or front and rear single doors with separate or negligi-
Table 12-17 then can be used to adjust for specific levels ble alighting.
c
Wide double-doors front and rear with separate or negligible alighting.
of service as desired. d
Six-door channels and separate or negligible alighting.
e
Three loading positions.
NOTE: Peak 15-minute flow rates would be 50 percent higher, assuming a
typical load factor of 0.67.
SOURCE: Summarized from Table 12-24.
Figure 12-4. Typical CBD busway line-

haul passenger volumes (flow rates).
(Source: Adapted from Ref. 5, p. 39)
12-28 urban streets
Table 12-26. Typical Arterial Street Bus Service Volumes at Maximum Load Point (Service Level E)
seated passengers 50 percent standees
approx. dwell time (50 persons/bus) approx. dwell time (75 persons/bus)
at busiest stop hourly volume at busiest stop hourly volume
condition (sec) flow rate (phf = 0.80) (sec) flow rate (phf = 0.80)
20 percent board at
busiest stop 30 6,250 5,000 45 7,500 6,000
25 percent board at
busiest stop 38 5,560 4,450 56 6,270 5,020
30 percent board at
busiest stop 45 5,000 4,000 68 5,770 4,620
40 percent board at
busiest stop 60 4,170 3,340 90 4,690 3,750
50 percent board at
busiest stop 75 3,560 2,850 112 3,950 3,160
a
Level-of-service E 5,000 4,000 7,500 6,000
(based on current
operating experience)
Table 12-12 CBD streets
Assumptions:
1. 15 Sec clearance between buses.

2. Clearance time (sec)/(pass./bus) = 0.3 for seated and 0.2 for standees.
3. 3-Sec service time per passenger.
4. g/C Ratio of 0.5.
5. All buses stop at busiest stop.
6. 2.5 Effective berths.
7. R factor = 0.833.
a
These values are given for comparison purposes.
Ranges in passenger service volumes are given in Table 12- These volumes could be adjusted downward to reflect specific
26 for 50-passenger transit buses. The table gives hourly flow service levels by the factors given in Table 12-17, i.e.,
rates and likely hourly passenger volumes for seated loads and
for 50 percent standees under varying assumptions regarding A 0.40
passenger distribution among stops and dwell times at the busiest B 0.60
stop. The salient figures, based on a PHF of 0.8 for service- C 0.80
level E, are: D 0.90
E 1.00
Maximum service volume for 5,000 to 6,000
dispersed loading conditions persons/hour 3. Bus stops—The number of bus berths at outlying stops should
reflect: (a) the number of buses that each stop should accommodate
simultaneously during the peak 15-min period; (b) maneuvering
Maximum service volume for 3,300 to 5,000 requirements of buses to enter and leave the stop; (c) minimum
typical CBD conditions (45 sec persons/hour clearance times between buses; (d) the type of stop; and (e) allow-
to 60 sec/stop) able queues.
Equation 12-10b can be used for any given berth configuration
and dwell time. Alternatively, maximum capacities can be obtained
Maximum service volume for 2,500–3,300
directly from Table 12-20. In both cases suitable reductions must
concentrated stop—CBD (75 sec persons/hour
be made to avoid unacceptably long queues. Accordingly, the fol-
or more/stop)
lowing guidelines are suggested.
Maximum service volume for 4,000–6,000 a. CBD stops—Levels-of-service C and D are acceptable. They
CBD conditions—planning persons/hour result in probabilities of 10 to 20 percent, respectively, that queues
method (Table 12-12) will develop beyond the bus stop (Table 12-17).
b. Outlying stops—Level-of-service B should be provided
wherever possible, especially when buses must pull into stops
The data given in Table 12-26 provide a realistic set of pa- from the traveled lane. This results in queues beyond bus stops
rameters for estimating service volumes of arterial streets. Be- only 2.5 percent of the time. Level-of-service C is, however,
cause of the many variables involved, it becomes difficult to acceptable.
select a single ‘‘number’’ for capacity. The level-of-service B criteria result in the following equation
for estimating bus berth requirements along arterial streets outside capacities, that is, the g/C ratio should be 1.0 in Eq. 12-24.
of the city center: However, clearance times should be adequate to assure reentry
into the main freeway lanes.
1,800 g/C 1,800 g/C Criteria for the spacing, location, and geometric design of bus
fd = = (12-24)
h′ (g/C) D + tc stops are given in several references (4, 8, 33). Such criteria must
be carefully applied to assure both good traffic and transit
where:
operations.
fd = maximum buses per hour per berth (for service level B);
h′ = minimum headway at stop; 4. Bus terminals—The design of a bus terminal or ‘‘transit
D = dwell time = passenger loading time; center’’ involves not only estimates of passenger service times of
buses that will use the center, but also a clear understanding of
tc = clearance time between buses;
how each bus route will operate. Therefore, such factors as sched-
g = green time; and
ule recovery times, driver relief times, and layovers to meet sched-
C = cycle length.
uled departure times become the key factors in establishing berth
For example, a 30-sec headway between buses (i.e., 20 sec stop, requirements and sizing the facility. In addition, good operating
practice suggests that each bus route, or geographically compatible
10 sec clearance), would result in 1,800/30 or 60 buses per hour
groups of routes, should have a separate loading position; this is
at a nonsignalized location.
essential to provide clarity for the passengers.
An alternate approach to bus berth requirements in outlying
Berth space requirements should recognize the specific type of
areas is to assume that buses arrive at random. Table 12-27 gives
carrier operations, fare collection practices, bus door configura-
the number of bus berths that should be provided based on the tions, passenger arrival patterns, amount of baggage, driver lay-
Poisson distribution, and allowing only a 5 percent chance that over-recovery times, terminal design, and berth configuration.
the berths will overload. Thus, it is a reasonable approximation of They should reflect both scheduled and actual peak period bus
level-of-service C. Emergent criteria for arterial (non-CBD) bus arrivals and departures, since intercity bus services regularly run
stop capacity are as follows: ‘‘extras’’ during the busiest seasonal travel periods.
Bus route and service patterns also influence berth requirements.
Passenger service times of 20 sec or less: one bus berth per Good operating practice calls for a maximum of two distinct routes
60 peak-hour buses (this is the typical radial arterial street (i.e., ‘‘services’’) per loading position.
condition). Berth space requirements at major bus terminals can be com-
Passenger service times of 30 to 40 sec: one bus berth per puted by the preceding equations. However, because passenger
30 peak-hour buses. service times represent only a small portion of the total time that
buses spend at a terminal, the equations will seriously overstate
c. Bus pullouts on exclusive roadways—Bus loading zones berth capacity unless the other key factors also are considered. It
on an exclusive roadway (pullout or turnout) within a freeway is essential to add the time needed for entering and leaving bus
right-of-way have capacities generally similar to those for curb- berths, schedule recovery, and driver relief. Bus service times also
side loading zones. Here again, the length of the stop and the may be increased to enable buses to meet scheduled departure
ability of buses to overtake other buses are important. Given times. Consequently, it may be necessary to add 5 min or more
similar loading facilities, differences reflect the length and capac- to computed clearance and dwell times for urban services.
ity of the roadway lane leading into and away from the stop. Typical urban transit and commuter bus capacities, based on
Uninterrupted flow conditions should be used to estimate stop operating experiences, suggest 8 to 10 buses per berth per hour.
Table 12-27. Berth Requirements at Bus Stops

no. of berths when
service time at stop is
peak-hour
bus flow headway 10 20 30 40 50 60
(buses/hr) (min) sec sec sec sec sec sec
15 4 1 1 1 1 1 1
30 2 1 1 1 1 1 2
45 1 1 2 2 2 2
60 1 1 1 2 2 2 2
75 1 2 2 3 3 3
90 1 2 3 3 4 4
105 1 2 3 3 4 4
1
120 ⁄2 1 2 3 3 5 5
150 2 3 3 4 5 5
1
180 ⁄3 2 3 4 5 6 6
NOTE: 95 percent probability that number of berths will not be overloaded. Assumes a Poisson distribution of bus arrivals.
SOURCE: Ref. 4.
12-30 urban streets
Intercity berth capacities are lower, in the range of 1 to 2 buses forth earlier. Guidelines presented in the previous section can
per hour. be used to estimate capacities. The following section summarizes
5. Increasing capacities—The person capacity of a busway, bus the pertinent features, planning guidelines, and potential benefits
lane or terminal depends heavily on the number of doors per bus, associated with various bus and high-occupancy vehicle priority
methods of fare collection, and concentrations of passengers at measures.
major stops. Consequently, bus system and bus stop capacities can
be increased by (a) increasing the number of major downtown (or
‘‘terminal’’) stations on a busway, or bus route, thereby decreasing Operational Overview
the number of boarding and alighting passengers at the heaviest
stop; (b) reducing the loading and unloading times for passengers Table 12-28 summarizes the state of the art of bus priority
through multiple doors on buses, prepayment, and/or selective sep- treatments as of January 1985. It groups treatments by type of
aration of loading-unloading; and (c) using larger buses (or where facility (freeways, arterial streets, and terminals), and within each
feasible higher load factors) to reduce the clearance interval time group it further classifies treatments by type of operation:
losses between successive vehicles.
1. Freeways, busways, reserved lanes and ramps.
a. Spreading stops—Where the number of buses to be accom- 2. Arterial streets, reserved lanes, bus streets, signal preference,
modated along a street exceeds the capacity of the busiest stops, and turn permissions.
routes may be separated into two groups of approximately equal 3. Terminals, central and outlying areas.
bus volumes. Separated stops can be provided for each group of
routes. This requires buses to be able to pass each other, and Most bus priority measures take the form of reserved bus lanes
land use patterns that make the dispersal of stops practical from on city streets, usually in the same direction as the general traffic
a passenger standpoint. In such cases, the total number of buses flow. However, the number of bus-only streets—such as State
that can be accommodated represents the sum of the capacities for Street in Chicago, Nicollet Mall in Minneapolis, and Chestnut
Street in Philadelphia—is increasing.
the stops in each group.
Busways and reserved lanes on freeways are mainly found or
b. Reducing dwell times—Dwell times can be reduced by (1)
are being proposed in larger American cities, usually with a large
prepayment of fares, (2) use of auxiliary personnel to allow
downtown employment and heavy peak-hour bus ridership. In the
rear-door fare collection and entry, (3) pay-as-you-leave fare
early 1980’s, a few medium-sized cities, such as Miami and Port-
collection on outbound trips, (4) removal of sidewalk obstructions
land, installed normal flow freeway bus and car pool lanes, but
at bus stops, (5) dispersal of downtown boarding points where
this tendency has subsided.
possible, and (6) platooning of buses. The Chicago Transit
Effective distribution of buses in central areas remains an impor-
Authority has been able to handle 45 to 50 buses in 15 min
tant challenge, and communities are giving this item increased
on the State St. Mall by operating buses in 3-bus platoons, and
attention. Freeway-related treatments generally provide good ac-
providing auxiliary rear door loading and fare collection in the
cess to the CBD perimeter, but do not substantially improve service
evening peak hours.
within the downtown core. Terminals are not always located near
In extreme cases, buses (or trains) cannot be unloaded or loaded
major employment concentrations and may require secondary dis-
at certain stops as rapidly as passengers accumulate (or before the
tribution. Because curb bus lanes are not always effective, there
next unit arrives). Thus, the headway that theoretically would be
have been several efforts to install contraflow bus lanes in down-
adequate for the demand volume as measured at the maximum
town areas.
load point cannot be delivered as line throughput. Such situations
Many bus priority measures have produced important passenger
can be alleviated by changing vehicle or stop configuration, using
benefits, especially those relating to freeways. Some have achieved
collectors to load rear doors, or having prepaid areas.
time savings of 5 to 30 min—savings that compare favorably with
those resulting from rail transit extensions or new systems.
Successful priority treatments are usually characterized by: (1)
BUS PRIORITY TREATMENTS an intensively developed downtown area with limited street capac-
ity and high all day parking costs, (2) a long-term reliance on
Over the past decade, much attention has been paid to expediting public transport, (3) highway capacity limitations on approaches
transit flow by providing various forms of priority treatment. Such to downtown, (4) major water barriers that limit road access to
treatments are aimed at improving schedule adherence and reduc- the CBD and channel bus flows, (5) fast nonstop bus runs for
ing travel times and delays for transit users. They may attract new considerable distances, (6) bus priorities on approaches to or across
riders, increase transit capacity, and/or improve the transit level water barriers, (7) special bus distribution within the CBD (often
of service. off-street terminals), and (8) active traffic management, mainte-
A growing number of cities have established exclusive bus lanes nance, operations, and enforcement programs (4).
and other bus priority measures to improve person-flow over city
streets and highways. Bus priority measures are an essential part
of transportation system management (TSM) programs that attempt Planning Considerations
to maximize transport system efficiency consistent with social,
economic, and environmental objectives. Planning and implementing bus priority measures requires:
Because buses may stop within priority lanes to pick up and (1) a reasonable concentration of bus services, (2) a high degree
discharge passengers, the ability of these lanes to carry people of bus and car congestion, (3) suitable street and road geometry,
will be affected by loading and unloading time requirements set and (4) community willingness to support public transport and
Table 12-28. Significant Examples of Bus Priority Treatments—United States and Canada (1984–1985)
type of treatment significant examples
1. Freeways
A. Busways
1. Busway on special right-of-way T Ottawa
T South PAT way, Pittsburgh
2. Busway in freeway median or right-of-way T Shirley Busway (I-95), Washington, D.C., areaa
T San Bernardino Busway, Los Angelesa
T Gulf Freeway, Houstonb
B. Reserved Lanes and Ramps
1. Bus preemption of freeway lanes (peak-hours) T Ottawa River Pkwy, Ottawa
2. Bus lanes on freeways, normal flow T U.S. 101, Marin County, Californiaa
T 9th St. Expressway, Washington, D.C.
T I-95, Miamia
T I-280, San Franciscoa
T Moanalua Freeway, Hawaiia
T Banfield Freeway, Portland, Ore.a
3. Bus lanes on freeways, contraflow T I-495, New Jersey
T Long Island Expressway, N.Y. City
T Gowanus Expressway, N.Y. City
T U.S. 101, Marin County
T North Freeway, Houston
4. Bus lane bypass of toll plaza T San Francisco–Oakland Bay Bridgea
5. Exclusive bus access to nonreserved freeway (or arterial) lanes T I-5 Seattle, Blue Street Express Bus Service & Ramp
T Braddock Ave., Pittsburgh
T O’Hare Field Connection to Kennedy Expressway, Chicago
6. Metered freeway ramps with bus bypass lanes T South Capitol St. Bridge, Washington, D.C.
T Various Freeways, Los Angeles, San Diego
T I-35 W, Minneapolis
7. Bus stops along freeway T Hollywood Freeway, Los Angeles
2. Arterial Streets
A. Reserved Lanes and Streets
1. Bus tunnels T Harvard Sq., Cambridge
T Providence, Rhode Island
2. Bus streets T Fifth & Sixth Streets, Portland, Ore.
T 10th Street, Washington, D.C.
T Nicollet Mall, Minneapolis
T State Street, Chicago
T State Street, Madison
T Chestnut Street, Philadelphia
T Granville Street, Vancouver
T Halsted and 63rd Streets, Chicago
T Fulton Street, Brooklyn, N.Y.
3. CBD bus lanes, normal flowc T Washington, D.C.
T Baltimore, Md.
T New York City, N.Y.
T San Francisco, California
T Rochester, N.Y. (Main Street)
T Ottawa, Ont.
4. Dual CBD bus lanes, normal flow T Madison Ave., N.Y. City
5. Arterial curb bus lanes, normal flowc T Hillside Ave., Queens, N.Y. City
T Connecticut Ave., Washington, D.C.
T Lincoln Ave., Denver
T Post, Sutter, Geary, O’Farrel St., San Francisco
T Eglinton Ave., Toronto
6. CBD median bus lanes T Canal St., New Orleans
7. Arterial median bus lanes T Broadway, Denver
T Barbour Blvd., Portlandd
T S. Dixie Highway, Miamia
Continued
12-32 urban streets
Table 12-28. Significant Examples of Bus Priority Treatments—United States and Canada (1984–1985)
Continued
type of treatment significant examples
8. CBD curb bus lane, contraflow T Spring St., Los Angeles
T Alamo Plaza, San Antonio
T Market St., Harrisburg
T Marquette, 2nd. Aves., Minneapolis
T Fifth Ave., Pittsburgh
T Madison, Washington, Adams, Jackson Streets, Chicago
9. Arterial curb bus lanes, contraflow T Ponce de Leon, Fernandez Juncos, San Juan
T College Ave., Indianapolis
T Kalanianole, Honolulu
B. Miscellaneous
1. Bus signal preemption T Barbour Blvd., Portland, Ore.
T Kent, Ohio
2. Special signal phases T Cermak Rd. at Kenton, Chicago
T Washington, D.C.
3. Special turn permissionc T Los Angeles
T Washington, D.C.
3. Terminals
A. Central Area Terminalsc T Midtown Terminal, N.Y. City
T Transbay Terminal, San Francisco
B. Outlying Transfer Terminalsc T Eglinton Ave., Toronto
T 95th, Dan Ryan Bus Bridge, Chicago
T River Road, Chicago
T Pentagon, Washington, D.C.
T Wilson, Toronto
C. Outlying Park-and-Ride Terminalsc T Route 3 on Lincoln Tunnel Approach at I-495 Contraflow Bus Lane,
New Jersey
a
Includes priority use by car-pools.
b
Under construction.
c
Selected examples.
d
Reversible lane.
SOURCE: Updated from Ref. 33.
to enforce regulations. There is little value in providing bus bus routing patterns; (j) bus passenger loading requirements
priority measures where service is poor, costly, or nonexistent; along curbs; and (k) community attitudes and resources.
where there are neither buses nor congestion; or where the Bus priority measures must fit real-world street systems. They
community has no desire to maintain and improve bus services must be reasonable, not only in how they improve bus service,
or to enforce bus lanes. but how they impact other traffic as well. Community acceptance
and support are essential, especially over the long run. Effective
1. Objectives—Planning calls for a realistic assessment of de-
enforcement and maintenance are also necessary elements in prior-
mands, costs, benefits, and impacts. The objective is to apply
ity treatments.
measures that (a) alleviate existing bus service deficiencies, (b)
achieve attractive and reliable bus service, (c) serve demonstrated Buses must be able to enter and leave priority lanes easily and
existing demands, (d) provide reserve capacity for future growths safely, and alternative routings must be available for potentially
in bus trips, (e) attract auto drivers, and (f) relate to long-range displaced automobile traffic. New problems should not be created,
transit improvement and downtown development programs, and nor should existing problems merely be transferred from one loca-
(g) have reasonable operating costs. tion to another.
2. Factors—Key factors include: (a) the intensity and growth Before any treatment is placed into effect, an a priori assessment
prospects of the city center; (b) the historic and potential future should be made of its benefits and effects. This is important to
reliance on public transport; (c) street width, configuration, provide a rational basis for implementing the treatment and to
continuity, and congestion; (d) the suitability of existing streets ensure good operations. A commitment also should be obtained
(and expressways) for express bus service; (e) bus operating from appropriate government agencies regarding enforcement and
speeds and service reliability in the city center; (f) availability maintenance. Unless enforcement is strict, frequent violations may
of alternate routes for displaced auto traffic; (g) locations of occur, undermining the benefits of the priority operations.
major employment centers in relation to bus routes; (h) goods Traffic management and bus priority studies of urban freeways
and service vehicle loading requirements; (i) express and local are, in reality, freeway operations studies. Demands, queues,
and densities, as well as speeds and volumes, should be clearly e. The benefits resulting from priority measures generally
identified. Various computer models may be used to investigate should be proportional to the amount of congestion before the
lane and ramp control strategies. measure was installed.
3. Principles—The following principles underly bus priority
planning: Guidelines for Specific Treatments
a. Bus priorities should be developed as an integrated system Specific criteria for introducing bus priority measures will
of treatments that improve bus speeds and schedule dependability. vary among cities. The illustrative planning and installation
b. Bus priority treatments should maximize person-flow and guidelines given in Table 12-29 are based on NCHRP studies
minimize person-delay over the long run. There should be a net (4) as updated by more recent research. They are expressed in
saving in the average travel time per person. terms of peak-hour buses and passengers, but they also identify
c. Priority measures should expedite bus service without ad- other relevant planning factors. Bus and passenger volumes
versely impacting general traffic flow. should be based on future ‘‘design year’’ demands to allow for
d. Costs should be reasonable in relation to existing and generated traffic. However base-year (existing) conditions should
potential demands and benefits. meet approximately 75 percent of the volume requirements.
Table 12-29. Summary of Illustrative Planning Guidelines for Bus Priority Treatments
general
applicability to: design-year conditions
range in range in
limited- minimum minimum one-way
local express planning one-way peak-hour bus
type of bus bus period in peak-hour passenger related land-use and
treatment service service years bus volumes volumes transportation factors
Freeway-Related
Busways on special right- x x 10 to 20 40 to 60 1,600 to 2,400 Urban population, 750,000; CBD
of-way employment, 50,000; 20-million sq ft
CBD floor space congestion in corridor;
save buses 1 min/mi or more.
Busways within freeway x 10 to 20 40 to 60 1,600 to 2,400 Freeways in corridor congestion in peak
right-of-way hour; save 1 min/mi or more.
Busways on railroad x x 5 to 10 40 to 60 1,600 to 2,400 Not well located in relation to service
right-of-way area. Stations required.
Freeway bus lanes, x 5 60 to 90 2,400 to 3,600 Applicable upstream from lane-drop. Bus
normal flow passenger time saving should exceed
other road user delays. Normally
achieved by adding a lane. Save buses 1
min/mi or more.
Freeway bus lanes, x 5 40 to 60 1,600 to 2,400 Freeways six or more lanes.
contraflow Imbalance in traffic volumes permits
level-of-service D in off-peak travel
directions. Save buses 1 min/mi.
Bus lane bypass at toll x 5 20 to 30 800 to 1,200 Adequate reservoir on approach to toll
plaza station.
Exclusive bus access x x 5 10 to 15 400 to 600
ramp to nonreserved
freeway or arterial lane
Bus bypass lane at x 5 10 to 15 400 to 600 Alternate surface street route available
metered freeway ramp for metered traffic. Express buses leave
freeways to make intermediate stops.
Bus stops along freeway x 5 5 to 10 50 to 100a Generally provided at surface street level
in conjunction with metered ramp.
Arterial-Related
Bus streets x x 5 to 10 20 to 30 800 to 1,200 Commercially oriented frontage.
CBD curb bus lanes, x 5 20 to 30 800 to 1,200 Commercially oriented frontage.
main street
Continued
12-34 urban streets
Table 12-29. Summary of Illustrative Planning Guidelines for Bus Priority Treatments
Continued
general
applicability to: design-year conditions
range in range in
limited- minimum minimum one-way
local express planning one-way peak-hour bus
type of bus bus period in peak-hour passenger related land-use and
treatment service service years bus volumes volumes transportation factors
Curb bus lanes, normal x 5 30 to 40 1,200 to 1,600 At least 2 lanes available for other
flow traffic in same direction.
Median bus lanes x x 5 60 to 90 2,400 to 3,600 At least 2 lanes available for other traffic
in same direction; ability to separate
vehicular turn conflicts from buses.
Contraflow bus lanes, x 5 20 to 30 800 to 1,200 Allow buses to proceed on normal route,
short segments turn around, or bypass congestion on
bridge approach.
Contraflow bus lanes, x x 5 40 to 60 1,000 to 2,400 At least 2 lanes available for other traffic
extended in opposite direction. Signal spacing
greater than 500-ft intervals.
Bus turnouts x 5 10 to 15 400 to 600 Points of major passenger loadings on
streets with more than 500 peak-hour
autos using curb lane.
Bus preemption of traffic x 1 to 5 10 to 15 400 to 600 Wherever not constrained by pedestrian
signals clearance or signal network constraints.
Special bus signals and x 1 to 5 5 to 10 200 to 400 At access points to bus lanes, busways,
signal phase, bus-actuated or terminals; or where special bus
turning movements must be
accommodated.
Special bus turn x 1 to 5 5 to 10 200 to 900 Wherever vehicular turn prohibitions are
provisions located along bus routes.
a
Boarding or alighting passengers in peak hour.
SOURCE: Ref. 4, p.28.
IV. APPLICATIONS AND SAMPLE PROBLEMS
Transportation engineers and planners encounter many prob- It is necessary to identify the controlling bottleneck along any
lems that involve transit operations and capacities. This section transit route, and to estimate the maximum frequency of service
contains sample problems that illustrate the use of the various at this point. The passengers per vehicle can be established from
charts, tables, equations, and procedures. It presents each problem field observations, projections, or system policy. Peak-hour load
in step-by-step detail, and it fully discusses the results. In factors should be estimated to relate peak 15-min periods to hourly
practice, many solutions would be shorter and less detailed. flows. Berth efficiency factors or, alternatively ‘‘unequal loading
factors’’ should be used to discount for the unequal use of a group
of buses or trains of rail cars, as appropriate.
GENERAL APPROACH Table 12-30 gives the various equations to be used, and shows
where each applies. Table 12-31 defines the basic capacity vari-
ables used. Table 12-32 identifies the application of each figure and
Transit capacity estimates require many assumptions regarding table; and Table 12-33 sets forth suggested planning parameters for
passenger distribution, service and dwell times, vehicle clearance use where local experience is unavailable.
and method of operation. It is essential to make reasonable assump-
tions regarding these factors because they have important effects 1. Equations 12-25a and 12-25b (renumbered Eqs. 12-2a and
on transit system capacity. 12-2b) identify the basic relationships from which other equations
Capacity of a transit stop or lane depends on the size and flow. The number of transit vehicles per hour per channel or stop
loading standards of vehicles, the minimum clearance time that can move past a critical point, assuming no signal interrup-
between buses or trains at stops, and passenger service times. tions, is expressed as
Passenger service times, in turn, depend on method of fare
collection, and door size and configuration. It is important to
recognize that these factors are largely determined by transit 3,600 R
cv = (12-25a)
system operating policy, and may vary from system to system. D + tc
R is used to adjust for irregularities in dwell times, arrival rates, 4. Where the distribution of passengers along a bus (or rail)
or for varying levels of service (LOS). Additional adjustments are route limits the number of vehicles that can get on at other points
made to compensate for the reductive effect of signal timing. R is along the line, then it is necessary to apply Eq. 12-16 or some
assumed as 0.833; thus 3,000 replaces the (3,600 R) in the equation. derivative of it, i.e., Eq. 12-20 or Eq. 12-21.
2. The number of vehicles that can pass through the heaviest 5. The number of effective bus berths at a stop can be estimated
boarding point is limited by the passengers that board and alight from Eq. 12-13, or Eq. 12-14. Factors then can be applied to
there. If these vehicles are able to be filled to their maximum seated estimate the actual number of berths that should be provided.
and/or standing loads, Eq. 12-25b, that is, applies directly, for each 6. Along arterial streets, where the curb lane is used by parked
effective loading position. cars, it is essential that bus stops are long enough to prevent buses
from backing out into the traffic lane. For this, and for other arterial
3,600 nSR
cp = nScv = (12-25b) street design purposes, Eq. 12-24 is used. In effect, the value 1,800
D + tc replaces the term (3,600 R) in Eq. 12-9a or Eq. 12-10b.
3. Where traffic signals are involved, the dwell time D is re- 7. On-street rail transit operation is similar to bus operations
duced by g/C and the entire expression is then reduced by g/C. In except for differing car lengths, seating configurations, and door
this case, Eq. 12-25c applies: arrangements. Estimates of passenger dwell times at a stop must
recognize the unequal loading among doors. Clearance times
3,600 nSR (g/C) should consider train length.
cp = (12-25c)
(g/C) D + tc
Table 12-30. Summary and Applications of Transit Capacity Equations

eq. no. equation application
12-1 cp = f ′ O1 + [(1,800 − 1.5 f ′) O2] Person capacity of a freeway lane.
3,600 R 3,600 R General equation—number of vehicles past critical
12-2a cv = =
h D + tc point, per channel or berth, uninterrupted flow
3,600 nS General equation—number of people past a critical

12-2b cp = nScv = R
D + tc point, per channel or berth, uninterrupted flow
(g/C) 3,600 nSR General equation—number of people past critical point,

12-2c cp = nScv =
(g/C) D + tc per channel or berth, flow interrupted by traffic signals
An Passengers per vehicle based on number of seats and

12-2d S i = sn +
Li square feet per standee
TL = (g/C) N (D + L) Time loss, seconds per hour, resulting to queues in

12-3
same lane as buses stopping for passengers
12-4 HV = (Peak 15-min volume)(4)(PHF) Determining hourly service volume
12-5a Trains Cars Seats Pass. Rail transit capacity, passengers per hour
P= × × ×
Hour Train Car Seat
OR,
12-5b Cars Seats Pass.

P= × ×
Hour Car Seat
Ft2 Ft2
@ Pass.
12-6 Trains Cars
P= × ×
Hour Train Car
3,600 R 3,600 R Buses per hour at critical stop (no interruptions),

12-7 f′ = = general equation
h′ D + tc
12-8a h′ = bB + tc Boarding Minimum headway at a bus stop
12-8b h′ = aA + tc Alighting
12-8c h′ = aA + bB + tc Two-way flow
Continued
12-36 urban streets
Table 12-30. Summary and Applications of Transit Capacity Equations

Continued
eq. no. equation application
Note: R is assumed as 0.833 in formulas that follow:
3,600 R 3,000 Boarding Maximum buses per berth per hour, uninterrupted flow;
12-9a f′ = = busway, terminal
bB + tc bB + tc
3,600 R 3,000 Alighting
12-9b f′ = =
aA + tc aA + tc
3,600 R 3,000 Two-way flow
12-9c f′ = =
aA + bB + tc aA + bB + tc
g Maximum buses per berth, signal interrupted per cycle
12-10a f ′c = per cycle (12-10a) per hour (12-10b)
tc + D(g/C)
(g/C)3,600 R (g/C) 3,000
12-10b f′ = = per hour
tc + D(g/C) tc + D(g/C)
(g/C) 3,000 (LOS Factor)i City street per hour at level-of-service i

12-11 cvi =
tc + D(g/C)
3,600 R B Max. boarding pass. per berth per hour, uninterrupted
12-12 Q= flow, busway, terminal
bB + tc
J(bB + tc) bB + tc bB + tc Number of effective berths to serve a given passenger
12-13 Nb = = = flow(s), uninterrupted flow, busway, terminal
(3,600)R(B) h′ R 0.833h′
3,600 R B Max. pass. per berth per hour with traffic signal
12-14a Q = (g/C)
tc + Bb(g/C) interruptions, city street
J[tc + Bb(g/C)] Number of effective berths to serve a given passenger
12-14b Nb =
(g/C) 3,600 R B flow with traffic signal interruptions, city street
Max. load point pass./hour based on bus frequency and
12-15 P=f×S
load factor
3,600 R NbS As a function of number Passenger capacity at max. load point, uninterrupted
12-16 P= of boarding passengers at flow, busway
bB + tc busiest stop
3,600 R Nb As a function of Passenger capacity at max. load point, uninterrupted

12-17 P= proportion of passengers flow, busway
Xb + (tc /S) boarding at busiest stop
N bQ As a function of Passenger capacity at max. load point, uninterrupted

12-18 P= passenger capacity per flow, busway
X berth
P(Xb + tc /S) Number of effective berths at busiest stop, uninterrupted
12-19 Nb = =
3,600 R flow
bXS + tc Keyed to pass volume at max. load point, busway, or

Nb = (P/S) terminal
3,600 R
3,600 R g NbS Function of no. of Passenger capacity at max. load point, signals interrupt
12-20 P= boarding passengers at flow (city street)
C[Bb(g/C) + tc] busiest stop
3,600 NbSgR Function of proportion of Passenger capacity at max. load point, signals interrupt
12-21a P= passengers boarding at flow (city street)
C[XbS(g/C) + tc] busiest stop
3,600 R Nb(g/C)
12-21b P=
[Xb(g/C) + tc /S]
PC[Xb(g/C) + tc /S] Number of effective berths at busiest stop, signals
12-22 Nb = interrupt flow, keyed to pass volume at max. load point
(g)(3,600) R
3,600 R(g/C) Line-haul passenger capacity at maximum load point
12-23 Pb = per effective berth, all applications/general equation
[Xb(g/C) + tc /S]
1,800(g/C) Design capacity of a bus stop, service-level B; stops
12-24 fd =
(g/C) D + tc along outlying arterial route, best applications
Table 12-31. Basic Transit Capacity Variables

symbol description
A Alighting passengers per bus measured in peak 15 min
An Net area available on a transit vehicle for standees
a Alighting service time per passenger, in seconds
B Boarding passengers per bus measured in peak 15 min
b Boarding service time per passenger, in seconds
C Cycle length, in seconds
c ′b Design capacity of a bus stop in buses per hour
cv Buses per hour per channel
cv(i) Buses per hour at level-of-service i
cp People per hour per channel
D Bus dwell time at bus stop, in seconds (time when doors open and bus is stopped)
f Bus frequency, in buses per hour (all routes using the facility), at maximum load point (if all buses stop at all stations,
f = (N)f ′)
f′ Maximum peak bus frequency at a berth, in buses per berth per hour
f ′c Bus frequency at a berth, in buses per cycle
f ′d Design bus frequency, in buses per berth per hour
g Green + yellow time per cycle
H Alighting passenger capacity per berth per hour
HV Hourly volume, vehicles or passengers in an hour
h Bus headway on a facility, in seconds, at maximum load point; for cars, h is the headway between successive vehicles, in
seconds
h′ Minimum bus headway at a berth, in seconds (h′ = 3,600/f ′)
J Passengers boarding at heaviest stop, per hour
K Passengers alighting at heaviest stop, per hour
L Additional time loss due to stopping, starting, and queuing, in seconds
Li Net square feet per standee for level-of-service i
N Buses per hour that stop at given location
Nb Number of effective berths at a bus station or stop (N = N′ × u)
N ′b Number of berth spaces provided in a multiberth station
n Number of vehicles per unit, i.e., cars per train
O1 Bus occupancy (in peak 15 min) along freeway (passengers per hour)
O2 Car occupancy (in peak 15 min) along freeway (passengers per car)
P Linehaul capacity of a bus facility, in persons per hour, past the maximum load point (hourly flow rate on maximum 15
min)
Pb Unit linehaul capacity of a bus facility in persons per hour, at the maximum load point, based on a single berth at the
busiest stop (hourly flow rate based on busiest 15 min)
PHF Peak-hour factor
Q Boarding passenger capacity per berth per hour
R Reductive factor to compensate for variations in dwell time or bus arrivals, also can be used to obtain levels of service
S Passengers on bus or rail car (varies with design and policy, may include seated passengers and standees)
Si Passengers/vehicle or passenger spaces/vehicle, for service level i
sn Seats per transit vehicle
T Total time at a stop = dwell time plus clearance time
TL Time loss, seconds per hour, resulting from buses blocking cars at a stop
tc Clearance time between successive buses, in seconds (time between closing of doors on first bus and opening of doors on
second bus)
u Berth utilization factor (an efficiency factor applied to the total number of berths to estimate realistic capacity of multiberth
stations (u = Nb /N′b)
X Proportion of maximum load point passengers that board at heaviest stop (X = J/P = B/S)
Y Proportion of maximum load point passengers that alight at heaviest stop (Y = K/P)
SOURCE: Adapted from Ref. 4, p. 41.
12-38 urban streets
Table 12-32. Summary and Application of Transit Capacity Figures and Tables
exhibit
number description application
Table 12-1 Peak-hour use of public transit by persons Informational
entering or leaving the central business district
Table 12-2 Important terms in transit capacity Informational
Table 12-3 Factors that influence transit capacity Informational
Figure 12-1 Example of freeway person capacity Informational
Figure 12-2 The two-dimensional nature of transit level of Informational
service
Table 12-4 Characteristics of transit vehicles Informational
Table 12-5 Levels of service for bus transit vehicles Informational
Table 12-6 Levels of service for rail transit vehicles Informational
Table 12-7 Typical space requirements for seated and Estimating the passengers on vehicles for
standing passengers varying seating configurations
Table 12-8 Passenger equivalency of urban buses at Adjustments in intersection capacity
Table 12-9 Passenger boarding and alighting times related Estimates of boarding and alighting coefficients
to service conditions
Table 12-10 Typical bus passenger boarding and alighting Estimates of boarding and alighting coefficients
service times for selected bus types and door
configurations
Table 12-11 Suggested bus flow service volumes for Planning estimates of bus service volumes on
planning purposes city streets
Table 12-12 Suggested bus passenger service volumes for Planning estimates of bus passenger service
planning purposes volumes
Table 12-13 Reported rail rapid transit peak-hour passenger Informational; analogy comparisons
volumes
Table 12-14 Reported light rail (street car) peak-hour Informational; analogy comparisons
passenger volumes (in peak direction)
Table 12-15 Typical rail transit capacities Estimate rail transit capacities and passenger
service volumes
Table 12-16 Estimated maximum capacity of bus stops Bus berth capacity and berth requirements
Table 12-17 Suggested levels of service for bus stops All level-of-service computations for design
purposes
Table 12-18 Typical service levels, single stop Detailed capacity data for 15-sec bus clearance,
60-sec dwell time
Table 12-19 Efficiency of multiple linear berths Capacity provided by more than one berth
Table 12-20 Estimated capacity of on-line bus stops Detailed bus stop by number of berths capacity.
Data for 10- and 15-sec clearance and 30-, 60-,
120-sec dwell times
Figure 12-3 Bus stop capacity related to dwell times Detailed bus stop and loading positions
capacity. Data for 15-sec clearance and 30-, 60-,
120-sec dwell times
Table 12-21 Bus berth passenger capacity equations and Informational
illustrative examples
6
Table 12-22 Maximum load point hourly passengers per
effective berth at the busiest station, Estimate number of berths for a given flow at max.
uninterrupted flow conditions load point. Also, estimate flow at max. load point for
Table 12-23 Maximum load point hourly passengers per a given number of berths
effective berth at the busiest station
Table 12-24 Illustrative bus capacity guidelines for CBD Informational
busways
Table 12-25 Busway service volumes at maximum load point Estimate passenger service volume at max. load
point for various types of operation
Figure 12-4 Typical CBD busway linehaul passenger Estimate berth requirements for given busway
volumes flow rates passenger flow and conversely
Table 12-26 Typical arterial street service volumes at Design and operations—estimate maximum
maximum load point passenger capacities and service volumes
Table 12-27 Berth requirements at bus stops (outlying Alt. approach to design of bus berths at
locations) outlying locations
Table 12-28 Significant examples of bus priority Informational
treatments—U.S. and Canada
Table 12-29 Summary of illustrative planning guidelines for Informational for planning decisions
bus priority treatments
Table 12-33. Guidelines for Application

1. Boarding Times per Passenger—Pay Fare on Vehicle (Single Vehicle)
Low-level platform—single door:
2.6 sec single coin fare
3.0 sec exact fare (general values)
3.5 sec exact fare (standees on bus)
2. Alighting Times Per Passenger—Low Level Platform
1.7 to 2.0 sec (use 2 sec)
3. Boarding and Alighting—Heavy Two-Way Passenger Flows
Through a single door:
1.2 (2 A + 3 B)
where: A = alighting passengers/bus
B = boarding passengers/bus
4. Clearance Between Successive Buses
15 sec—desirable minimum
20 sec—minimum for operations on high-speed roadway
5. Levels of Service—Buses
Approximate Probability
Proportion of E of Queue Forming
LOS R 3,600 R (LOS Index) Behind Bus Stop
A 0.400 1,200 0.40 <1
B 0.500 1,800 0.60 2.5
C 0.667 2,400 0.80 10
D 0.750 2,700 0.90 20
E-Capacity 0.833 3,000 1.00 30
E-Capacity 1.000 3,600 50
(Perfect conditions)
6. High-Level Platform
Boarding-prepayment 1.0 to 1.5 sec
Alighting-prepayment 1.0 to 1.5 sec
TYPES OF PROBLEMS an urban arterial, estimate the losses occurring to auto traffic
and the resulting PCE values corresponding to the operations
Many kinds of problems can be addressed by the transit capacity of buses making stops, using berths located either in a through
analysis procedures. A common problem from the perspective of lane (on-line) or in a separated area (off-line).
the transit agency is to determine how many vehicles are needed 4. Passenger service times—Estimate the passenger service
to carry a given number of riders and to see if these vehicles can times (a) at a stop and (b) along a bus route for various boarding
be accommodated at the major boarding points. The solution is and alighting characteristics, fare collection methods, and bus door
simple if only one transit line is involved, but it may become more configurations.
complex where several routes converge. Solution of this problem 5. Arterial street bus capacities and service levels planning
calls for establishing load factor criteria (i.e., persons per vehicle) applications—Estimate the level of service for a specified bus
and identifying dwell times, berth requirements, fare collection passenger volume along an arterial street.
practices, and bus stopping patterns in the central terminal area. 6. Bus berth capacity—Estimate the capacity of a bus berth for
Typical problems that can be solved by the procedures in this given passenger loading and unloading characteristics; alterna-
chapter include the following: tively estimate the number of berths needed for a given passenger
volume.
1. Person-flow—Using car and bus occupancies, estimate the
7. Bus terminal capacity—Estimate the number of loading posi-
total person-flow for an arterial street or freeway.
tions needed to accommodate given passenger and bus volumes
2. Person-capacity—Using observed car and bus occupancies,
consistent with operating criteria.
and the present mix of transit vehicles in the traffic flow, estimate
8. Bus system (route) capacity—Estimate the capacity of an
the total person-capacity.
arterial street or busway in passengers per hour. Alternatively,
3. Effect of buses on highway capacity—
estimate berth requirements at major stops to serve a specified
a. Freeway—For a freeway traffic lane carrying mixed traffic transit flow.
(automobiles, buses), estimate the capacity reduction resulting 9. Design capacity, arterial street bus stops—Estimate the
from buses and the passenger car equivalent (PCE) volume. number of berths needed to serve a given bus flow and dwell
b. Arterial street—For a lane carrying mixed traffic along time for design purposes.
12-40 urban streets
10. Rail transit capacity—Estimate the number of people per The person-capacity of the lane with bus and car traffic can be
hour that can be carried past the maximum load point for a speci- estimated by using Eq. 12-1:
fied train length and level of service (i.e., private right-of-way).
11. Light rail transit—Estimate the number of people per hour Person-capacity = [f1 × O1] + [(1,800 − 1.5 f1) × O2]
that can be carried past the maximum load point, with on-street
operations and traffic signal control. where:
These problems are mainly defined in terms of bus transit. How- f1 = number of buses/hour;
ever, many can also apply to light rail transit. The problems cited O1 = bus occupancy, 50 people/bus; and
are illustrated in the sample calculations that follow. O2 = car occupancy, 1.5 people/car.
Thus, the person-capacity of the shared lane is:

SAMPLE CALCULATIONS
[100 × 50] + [1.5 × (1,800 − (1.5) (100))]
= 5,000 + 2,475 = 7,475 people
Calculation 1—Person-Flow
1. Description—A given urban freeway carries 4,500 cars and The person-capacity of the two lanes is 2,700 + 7,475, or 10,175
50 buses in the peak hour. Sample vehicle occupancy counts people.
show 1.3 for cars and 50 for buses, respectively. Find the The effects of various bus volumes on the person-capacity of
person-flow. the shared freeway lane are given in Table 12-34.
2. Solution—The total flow represents the sum of the number 3. Comment—In some situations, such as a downtown street,
of people carried by each type of vehicle. Data can be tabulated with a bus lane, the person-capacity of the bus lane should be
as follows: estimated and added to that of the other lanes. Note that this
represents the maximum potential person-capacity, while the ex-
Veh/hr People/veh People/hr ample computed the person-capacity under prevailing or likely
Cars 4,500 1.3 5,850 conditions of flow.
Buses 50 50.0 2,500
Total 4,550 8,350
Percent Bus 1.1 29.9
Calculation 3—Effect of Buses on Freeway Capacity
The total person-flow is 8,350. Buses represent 1.1 percent 1. Description—Ninety buses operate in the peak direction of
of the total traffic and account for 29.9 percent of the total a four-lane freeway during the peak hour. The freeway also carries
person-flow. 3,400 passenger cars in this direction. Average occupancies are
40 persons/bus and 1.4 persons per car.
It is desired to find: (a) the equivalent peak hour, peak direction
Calculation 2—Person-Capacity
passenger car volume; (b) level of service, assuming 12-ft lanes,
1. Description—A four-lane urban freeway (two lanes in each no lateral obstructions, and 70-mph design speed; and (c) the total
direction) has a capacity of 1,800 passenger car equivalents per person-volume.
lane per hour. Car occupancy averages 1.5 people per car. It is 2. Solution—It is reasonable to assume that each bus is the
planned to initiate express bus service with 100 buses per hour, equivalent of 1.5 passenger vehicles. Therefore, 90 buses are the
and each bus is estimated to carry 50 people. The buses will be equivalent of 135 cars (90 × 1.5 = 135). The equivalent passenger
restricted to one lane. It is desired to find the one-way, peak- car volume is 3,400 plus 135, or 3,535. Service volumes for LOS
hour, person-capacity of the freeway. Each bus is assumed as 1.5 E range from 3,100 × PHF to 3,700 × PHF. If PHF = 0.90, the
equivalent passenger cars. volumes are 2,790 vph to 3,330 vph (see Chapter 3). This indicates
2. Solution—The person-capacity of the freeway lane where that the freeway is operating at LOS E.
no buses will operate is 1,800 × 1.5 or 2,700 people. The total person-volume is calculated as follows:
Table 12-34. Person-Capacity of a Freeway Lane for Varying Bus Volumes

buses
condition and (pce’s)
vehicle cap. buses 1 bus = no. of people people person
before buses (veh) 1.5 pce’s pass. cars by bus by car cap.
1 1,800 0 0 1,800 0 2,700 2,700
2 1,800 50 75 1,775 2,500 2,660 5,160
3 1,800 100 150 1,750 5,000 2,630 7,630
4 1,800 150 225 1,725 7,500 2,590 10,090
5 1,800 200 300 1,700 10,000 2,550 12,550
Buses: 90 at 40 people/bus = 3,600 (43 percent) 630

× 100 Percent = 35 Percent
Cars: 3,400 at 1.4 people/car = 4,760 (57 percent) 1,800
Total = 8,360 (100 percent)
and the capacity loss is:
Calculation 4—Effect of Buses on Arterials 0.35 × 750 = 262 pcph
1. Description—Sixty buses per hour operate along an arterial This results in a PCE value of:
street with an average dwell time of 15 sec per stop. Find the
reduction in available green time to the lane in which buses stop
PCE = 262/60 = 4.37
if (a) buses stop in the adjacent parking lane, and (b) buses stop
in the through-traffic lane. Assume that the capacity of the through Other lanes are not affected. Also note that buses stopping in a
lane is 1,500 cars per hour of green, and that the green/cycle, through lane have over 3 times the effect of buses stopping in a
g/C, time is 0.50, giving a capacity of 750 cars per hour. parking lane for this case.
What is the time loss per hour in each case? What percentage
of total lane capacity is required for bus operation? How can this
be translated into a PCE value? Calculation 5—Passenger Service Times (Bus Stop)
2. Solution—For case (a), buses stop in parking lane, the time
loss to a right-hand through lane when buses stop in the adjacent 1. Description—Field observations show that 15 passengers
parking lane is due to acceleration and deceleration of the bus board each bus and 5 alight at a given stop during the peak hour.
while entering and leaving the through lane. It has been noted in Assuming on-vehicle fare collection with an ‘‘exact fare’’ and a
the section entitled ‘‘Effects of Buses on Vehicular Capacity’’ that single door, find the passenger service and dwell times. If a rear
this loss averages 3 to 4 sec per bus. Using 4 sec, it follows that: door is available for alighting passengers, find the service time.
2. Solution—The passenger service dwell times can be esti-
Time loss/hour = 4 × 60 = 240 sec/hour mated by applying Eqs. 12-8 (a,b,c), as follows, using passenger
service rates of 3 sec per boarding passenger and 2 sec per alighting
As the g/C ratio is 0.5, the total green time/hour available to as drawn from the section on ‘‘Passenger Service Times’’ and
the through lanes is 0.5 × 3,600 = 1,800 sec/hr. The percent loss Tables 12-9 and 12-10:
in lane capacity may be expressed as:
Service Time Clearance
240
× 100 = 13.3 Percent
1,800 1. Single door h = Aa + bB + tc
(entering and
It results in a capacity loss of 100 passenger cars per hour: exiting) h = 15 (3) + 5 (2) + tc = 55 + tc
750 pcph × 0.133 = 100 pcph 2. Single door h = bB + tc
(entering only) h = 15 (3) + tc = 45 + tc
In that one lane the passenger car equivalent (PCE) for this condi-
tion represents the ratio of the pcph loss in capacity divided by
the number of buses/hr causing the loss, or: Thus, the passenger service times would be 55 sec for entering
and exiting through a single door, and 45 sec if a rear door is
PCE = 100/60 = 1.67
available for exiting. The clearance times normally include the
Note that the headway of each bus is, in effect, 4 sec as compared door opening and closing times, about 5 sec. Therefore, the total
with 2.4 for cars. Thus, 900 buses/hour would be the equivalent time spent at the stop for the two sets of conditions is 60 and
of 1,500 cars. Each bus, therefore, has the equivalency of 1,500/ 50 sec, respectively, when door opening and closing times are
900, or 1.67 cars. considered.
For case (b), buses stop in through lanes, the time loss for buses
stopping in a through lane is computed using Eq. 12-3:
TL = (g/C) × (N) × (D + L) Calculation 6—Passenger Service Times (Bus Routes)
where: 1. Description—The following values represent the number of

g/C = 0.50 (Given); passengers boarding and alighting each bus on a selected bus route:
N = 60 buses/hour (Given);
D = 15 sec/bus (Dwell time, Given); Stop No. → 1 2 3 4 5 6 7 8
L = 6 sec/bus (Loss time, avg. conditions assumed); and Alighting Pass. 0 2 2 5 8 15 25 10
TL = (0.50) (60) (15 + 6) = 630 sec/hour. (A)
Boarding Pass. 20 10 10 15 10 1 1 0
Then, the percent reduction in lane capacity is: (B)
12-42 urban streets
Passengers board and alight through a single door. A $0.50 by the transit agency, are shown in Table 12-35. The 1985 data
exact fare is used. Compute the dwell time at each stop. What is are based on actual schedules, while the 2000 data are based on
the total dwell time for the route? Consider the effects of lost time a forecast of growth of 60 percent for local bus service and 100
due to opening and closing doors. percent for freeway bus service.
2. Solution—From Tables 12-9 and 12-10, the average boarding In 1985, 22 local buses and 16 express buses would use the
time per passenger for the conditions given would be 2.6 to 3.0 Center in the peak direction compared with some 10 local buses
sec (use b = 2.8 sec), and the typical alighting time is a = 1.7 sec. and 6 express buses in the off-peak direction. By 2000, some 35
For boarding and alighting through a single door, the dwell time local buses and 32 express buses would use the Center in the peak
is given by: direction, while some 16 local buses and 12 express buses would
use the Center in the off-peak direction.
aA + bB Bus berths would be assigned according to principal ‘‘geograph-
ical’’ destinations.
Thus, for each bus stop: Bus dwell times at the Transit Center would approximate 5 min
Stop per bus for buses passing through the Center and 8 min per bus
1 0 (1.7) + 20 (2.8) = 56.0 sec for buses that begin and end trips there. These dwell times compare
2 2 (1.7) + 10 (2.8) = 31.4 sec with about a 3-min passenger service time needed to fill an empty
3 2 (1.7) + 10 (2.8) = 31.4 sec bus to seated capacity, assuming that exact fares are paid on the
4 5 (1.7) + 15 (2.8) = 50.5 sec bus.
Stop 2. Solution—Estimated berth requirements for 1985 and 2000
5 8 (1.7) + 10 (2.8) = 41.6 sec are given in Tables 12-36 and 12-37. The berths were estimated
6 15 (1.7) + 1 (2.8) = 28.3 sec as follows:
7 25 (1.7) + 1 (2.8) = 45.3 sec a. The bus routes were grouped by geographic destination in 3
8 10 (1.7) + 0 (2.8) = 17.0 sec categories.
Total Time 301.5 sec b. The ‘‘capacity’’ of each type of service was obtained by the
equation f = 60/D, where D was the specified dwell time, in minutes
The time lost in opening and closing doors would amount to including clearance. Thus, a 5-min dwell time could accommodate
another (8 × 5) or 40 sec. Thus, the total time lost at stops would 12 buses/berth/hour; an 8-min dwell time, 7.5.
be 341.5 sec, or almost 6 min. c. The number of inbound berths for the AM peak hour was
Note that because of the heavy passenger interchange at stops computed by dividing the number of buses by the berth capacity.
4 and 5, one could increase these time values about 20 percent Thus, for lines 42 and 68, in 1985, 12 buses would need 12/7.5
(i.e., 0.20 (50.5 + 41.6)). This would add 18.3 sec, resulting in a or 1.6 berths; this number was rounded up to 2.
total dwell time of eactly 6 min. d. The bus lines that start at the center would need only inbound
berths. The other bus services would need an equal number of
outbound berths to accommodate PM peak hour bus flows, and
to ensure that each major geographic destination would have its
Calculation 7—Planning Applications, Downtown Street, Level specified own boundary area.
of Service e. The total berth requirements represent the sum of the inbound
and outbound berths. As a result, 10 loading positions would be
1. Description—Field observations show that a CBD street car- needed for 1985 conditions; and 13 loading positions for 2000.
ries 4,500 passengers in 80 buses, during the peak hour, based on Ideally 15 loading positions should be provided to account for
peak 15-min flow rates. At what level of service does this street growth and traffic fluctuations within the peak hour.
operate?
2. Solution—The approximate level of service can be estimated Note that 38 inbound buses with a berth capacity of 10 buses/
from Table 12-11 or Table 12-12. The 80 buses per hour produce berth/hour would require only 4 inbound loading positions in 1985
level-of-service D, verging on level-of-service E, in terms of bus if routes were not separated geographically. However, this is not
flow. Referring to Table 12-12, level-of-service D from a passen- advisable when one considers clarity to the riding public, so that
ger perspective has a passenger volume range of 4,000 to 5,000, 6 berths are to be anticipated based on the grouping shown in
based on 80 buses per hour. Thus, the bus routes operate at level- Table 12-35.
of-service D from both the traffic flow and passenger standpoint.
Note that the lower half of Table 12-11 and Table 12-12, per- Calculation 9—Berth Capacity for Loading
taining to downtown streets, was used in making this broad plan-
ning assessment. 1. Description—A rail-bus interchange (intermodal terminal) is
planned for two urban bus lines.
Passengers pay a ‘‘single-coin’’ fare, and enter via the front
door. Each bus has a seating capacity of 50 people, and is equipped
Calculation 8—Bus Terminal (Transit Center) with single-width doors. It is assumed that loading would occur
through the front door, and unloading through the rear. It is desired
1. Description—It is desired to estimate ‘‘base year’’ 1985, and to determine the berths needed, assuming a minimum clearance
‘‘design year’’ 2000, berth requirements for an outlying transit time of 15 sec between buses. Bus frequency on line 1 is 20 buses/
center. hour, and 30 buses/hour on line 2.
The bus lines serving the proposed transit center, as identified 2. Solution—This problem can be analyzed by applying the
Table 12-35. Anticipated Peak-Hour Buses at Transit Center

peak
direction off-peak direction type of
route 1985 (2000) 1985 (2000) service
Local
Service
42 Holman 8 13 Terminating
Crosstown
68 Brays Bayou 4 6 Terminating
Crosstown
76 Lockwood 4 6 6 Through
Crosstown 4
77 MLK Limited 6 10 6 10 Through
Subtotal Local 22 35 10 16
Expressway
Service
242 Clear Lake 3 6 Through
Park & Ride
245 Edgewood 3 6 Through
Park & Ride
250 Hobby 2 4 Through
Park & Ride
255 Fuqua 4 8 Through
Park & Ride
41 Garden Villas 2 4 Through
Limited
147 Sagemont 2 4 Through
Express
Off Peak Direction 12
All Lines 6
Subtotal Express 16 32 6 12
TOTAL 38 67 16 28
SOURCE: Adapted from Herbert Levinson and Texas Transportation Institute: Conceptual Planning and Design,
Lockwood Transit Center, March 1983.
Table 12-36. Bus Berth Requirements, Year-1985

inbound
dwell buses/ buses am max. (outbound)
time/bus berth/ peak hour berths needed
service (minutes) hour (From Tab. inbound for pm peak total
bus line type (assumed) f = 60/D 12-35) berths hour berths
Local Service
42–68
Holman Crosstown Start 8 min 7.5 12 2 — 2
76
Crosstown Through 5 min 12 4 1 1 2
77
MLK Limited Through 5 min 12 6 1 1 2
Subtotal 22 4 2 6
Freeway
Expressway Lines Through 5 min 12 16 2 2 4
To City Center am
(From City Center pm)
TOTAL 38 6 4 10
12-44 urban streets
Table 12-37. Bus Berth Requirements, Year-2000

inbound
buses am max. outbound
dwell peak hour berths needed
service time/bus (from tab. inbound for pm total
bus line type (minutes) 12-34) berths peak hour berths
42–68
Holman Crosstown Start 8 min 19 3 0 3
76
Crosstown Through 5 min 6 1 1 2
77
MLK Limited Through 5 min 10 1 1 2
Subtotal 35 5 2 7
Expressway Lines Through 5 min 32 3 3 6
To City Center am
(From City Center pm)
TOTAL 67 8 5 13
procedures for estimating berth requirements, assuming unin- J1 = 50 × 20 = 1,000 pass./hr

terrupted flow (Eq. 12-11). J2 = 50 × 30 = 1,500 pass./hr
In this example, each route is analyzed separately. Because both 1,000 [150 + 15] 103(165)
bus lines operate on short headways, and would continue after N1 = = 3 = 1.10 (Use 1 berth)
(3,600) 50 (0.833) 10 (180)(0.833)
receiving and discharging passengers, no allowance is made for 1,500 [150 + 15] 165
schedule recovery or layover; such provisions may be needed in N2 = = = 1.67 (Use 2 berths)
(3,600) 50 (0.833) 120 (0.833)
practice and would have to be added to the dwell times.
The number of berths required for a given passenger volume
can be computed from the following Eq. 12-13: During the peak 15 or 20 min, buses will probably load to their
‘‘design’’ or ‘‘crush’’ capacity. In this short period (a) dwell times
J (bB + tc) bB + tc will increase, and/or (b) clearance times between buses will de-
Nb = =
(3,600) BR 0.833 h′ crease. The berths needed to accommodate loads of 75 to 80 pas-
sengers per bus are determined as follows:
where:
Nb = number of effective berths; bB + C 3 (80) + 15
N1 = = = 1.70 (Use 2 berths—case 1)
J = total number of passengers to be served per hour; h1 (0.833) 180 (0.833)
B = number of boarding passengers/bus; bB + C 3 (80) + 15
b = dwell time per boarding passenger; N2 = = = 2.55 (Use 3 berths—case 2)
h2 (0.833) 120 (0.833)
tc = clearance time per bus in seconds;
h′ = headway between buses, in seconds; and
R = 0.833. Note also that the 75 passengers per bus in the peak 15 min as
compared with 50 for the entire hour indicates a peak-hour factor
Substituting the values of B = 50 passengers per bus, tc = 15 of 0.67. Also note that the use of the R factor to reduce queuing
sec and b = 3 sec (exact fare); and headways of 180 sec for line does not change the berth requirements that would otherwise be
1 and 120 sec for line 2, the number of berths becomes: needed.
3 (50) + 15 165
Line 1: N1 = = = 1.10 (Use 1 berth)
180 (0.833) 180 (0.833) Calculation 10—Bus Berth Unloading
3 (50) + 15 165
Line 2: N2 = = = 1.67 (Use 2 berths)
120 (0.833) 120 (0.833) 1. Description—A facility is being built in an outlying area
to facilitate transfer between feeder buses and a rail rapid transit
Note that using the alternative form of the equation, line. It is assumed that buses will enter the facility on 1-min
headways and that each bus will discharge 50 passengers. This
J (bB + tc) corresponds to a total passenger flow of (50 people/bus) × (60
Nb =
3,600 BR buses/hour) = 3,000 people/hour. Clearance time required for
one vehicle to manuever out of the berth and for another to Calculation 11—Berth Capacity for Loading at
enter it is assumed as 20 sec. Major Stops
It is desired to know the number of unloading berths that should
be provided assuming the following bus configurations: 1. Description—It is desired to estimate the capacity of a bus
line where 10 people board each bus, passenger service time is 3
T Single-width door, one door used. sec per passenger, and clearance time is 15 sec per bus. It is
T Single-width door, two doors used. assumed that boarding conditions govern. The signal timing along
the street has a g/C ratio of 0.45.
2. Solution—The number of berths required for a given passen- 2. Solution—The problem may be analyzed in detail by use of
ger volume can be computed from the variation of Eq. 12-13, Eq. 12-14a:
which also applies to alighting:
3,600 BR
J (aA + tc) aA + tc Q = (g/C)
Nb = = tc + Bb (g/C)
(3,600) AR h′ (0.833)
where:
where:
g/C = green time per cycle, 0.45;
N = number of effective berths;
tc = clearance between buses, 15 sec;
A = number of alighting passengers per bus = 50;
B = boarding passengers per bus, 10;
tc = clearance time per bus = 20 sec;
b = passenger service time, 3 sec/pass.; and
J = total passengers per hour to be served = 3,000;
R = 0.833.
a = dwell time per alighting passenger = (1.7 and 0.9 sec);
h′ = headway between buses arriving at station; and
R = 0.833. Substituting gives:
This equation is similar to Eq. 12-13, except that unloading

115 + (10) (3) (0.45)2 = 0.45 1 2
(0.833) 3600 (10) (0.833) 3600
rather than loading passenger flows and coefficients are used. Note Q = 0.45
15 + 13.5
that the uninterrupted flow equation is used since the unloading = 568(0.833) = 473
will not be affected by traffic signal delay.
Substituting yields: The number of buses per hour would be 473/10 or 47.
An approximate solution may be obtained from Table 12-20
3,000 [a (50) + 20] a (50) + 20 using a g/C ratio of 0.5. Table 12-20 shows that for a g/C ratio
Nb = =
3,600 (50) (0.833) 60 (0.833) of 0.5, 30-sec dwell time per stop (10 pass. × 3 sec/pass.), and 15-
sec clearance that 50 buses per hour could be accommodated. This
The appropriate alighting service time factors are obtained from translates into 500 people. The difference between 473 and 500
Tables 12-9 and 12-10 as follows (note that Table 12-33 suggests results from the use of a 0.50 g/C ratio rather than 0.45.
117 to 210 sec): Note that if there were no signal delays, 670 passengers per
hour on 67 buses could be accommodated. In this case one could
T Single width door, 1 door used: a = 1.7 sec use Eq. 12-14a with g/C = 1.00 or Eq. 12-12 directly:
T Single width door, 2 doors used: a = 0.9 sec
R 3,600 b (3,600) (10) (0.833)
Q= = = 666 pass. per hour
1.7 (50) + 20 bB + tc 3 (10) + 15
Case 1: N1 =
60 (0.833)
= 2.1(Use 3 berths, although 2 would suffice) Since 10 passengers board per bus, some 67 buses could be
accommodated.
0.9 (50) + 20
Case 2: N2 = = 1.30 (Use 2 berths)
60 (0.833)
In practice, allowance should be made for: (a) some buses car- Calculation 12—Arterial Street Capacity
rying full or standing loads during part of the peak hour, (b) buses
operating at closer headways during parts of the hour, and (c) 1. Description—A central business district ‘‘bus-only street’’
imbalanced use of doors. provides 4 loading positions at the busiest stop. There is a 15-
One approach is to assume that all buses would operate with sec clearance between buses and a maximum of 75 passengers
standees for design purposes. Berth requirements, assuming 75 per bus past the maximum load point, during the peak 15 min.
persons per bus would be 2.46/0.833, or 3 berths assuming avail- An exact fare pay-as-you-enter system is used, with entry through
ability of both doors for passenger discharge. a single door. Rear doors of buses are used for passenger exit.
Given this condition which recognizes the likelihood of peak A g/C ratio of 0.52 is assumed. Field studies show that
15-min flow rates that are 25 percent greater, it is desirable to 25 percent of the passengers at the maximum load point board
provide 3 unloading berths. at the major stop and that the peak-hour load factor is 0.80.
12-46 urban streets
It is desired to estimate the hourly passenger volumes and f = P/S

bus frequency at the maximum load point.
where:
2. Solution—The number of people that can be carried past the
maximum load point can be estimated from Eq. 12-21b: f = bus frequency at maximum load point;
P = demand at maximum load point, in passengers per peak 15
3,600 Nb (g/C) R min; and
P=
Xb (g/C) + tc /S S = passenger capacity of bus (seated + standing).
where:
Therefore: f = 2,000/75 = 26.7 buses per peak 15 min.
g/C = green time per cycle, 0.52; The number of berths can be computed from Eq. 12-19, because
tc = clearance between buses, 15 sec; uninterrupted flow conditions can be assumed.
b = service time per passenger, 3 sec;
S = pass./bus at maximum load point, 75; Xb + tc /S
Nb = P
P = pass./hour (flow rate) at max. load point; 3,600 R
Nb = number of effective berths, max. = 2.5 (Table 12-19, noting
where:
4 loading positions provided);
X = proportion of passengers at maximum load point boarding P = persons per hour (flow rate) = (1,500 × 4) = 6,000;
at busiest stop, 0.25; and Nb = number of effective berths;
R = reductive factor for queuing = 0.833. S = bus capacity (seated + standing) = 75;
tc = clearance between buses = 15 sec;
Substituting gives: b = boarding time per passenger = 2.0 sec;
X = proportion of maximum load point passengers who board
3,600 (2.5) (0.52) (0.833) 4,680 (0.833) at heaviest stop (X = 750/1,500 as given); and
P= = = 6,607
[(0.25) (3) (0.52) + 15/75] [0.39 + 0.20] R = 0.833.
This represents the flow rate during the peak 15 min. Adjusting Therefore, in this example:
by the PHF of 0.80 gives 5,286 passengers at the maximum load
1 2
(0.5 × 2) + 15/75 (1.2)
point during the entire hour. N = 6,000 = 6,000
3,600 (0.833) 3,600 (0.833)
The 5,286 passengers at 75 passengers/bus would result in 70
= 2.0 Effective berths
buses/hour. If this service frequency were maintained for the entire
hour, it would result in 60 passengers per bus. This is probably Thus, 2 effective berths should be provided. Allowing for berth
more realistic than reducing the service frequency to maintain 75 ‘‘inefficiencies,’’ 3 loading positions should be provided (Table
persons per bus during the entire 60-min period. 12-19). This corresponds to a cumulative capacity of 2.25 berths
The number of people passing the maximum load point also for ‘‘on-line’’ stations and 2.60 berths for ‘‘off-line’’ linear
can be estimated using Table 12-23, assuming a g/C ratio of stations.
0.50. In using these exhibits, a value of 0.20 (i.e., 15/75) is
used for the clearance time to passenger per bus ratio. They
result in 2,610 passengers passing the maximum load point for
each effective berth. This corresponds to 6,525 passengers per
hour (flow rate) for 2.5 berths, or 5,220 when the peak-hour
factor is applied. This approximation is sufficiently accurate for
most planning purposes.
Calculation 13—CBD Busway
1. Description—A central business district busway serves 2,000

people past the maximum load point in the peak 15 min. The
heaviest stop has a 15-min boarding volume of 1,000 people. It is
Calculation 14—Arterial Bus Turnout
desired to determine (a) the bus frequency, and (b) the number of
berths required to accommodate the boarding passenger volume. 1. Description—It is planned to build bus turnouts along an
It is assumed that ‘‘schedule design’’ bus volumes are 75 persons/ artery. Observations show that bus dwell times approximate 45
bus at the maximum load point, clearance time between buses at sec and clearance time 15 sec. The peak-hour factor is 0.67.
each stop is 15 sec, and a pay-as-you-leave fare system is used in It is desired to find the desired number of buses per hour that
the downtown area. can use the turnout, assuming that stopped buses will not back up
2. Solution—Tables 12-9 and 12-10 give a range of 1.5 to 2.5 onto traffic.
sec per passenger through a single door, pay-as-you-leave. A value 2. Solution—To provide for buses backing out onto traffic
of 2.0 sec per passenger will be used. the turnout should be adequate 95 to 97.5 percent of the time.
The number of buses per hour can be determined from Eq. 12- The corresponding R value is 0.5 (LOS B). This results in Eq.
15, stated as: 12-24.
1,800 1,800 The calculations are shown below.

f′ = = = 30 Buses/hour (flow rate)
D + tc 45 + 15
a. Train clearance times: (1) minimum spacing between trains—
Applying the peak-hour factor of 0.67 results in 20 buses per estimated at 20 sec; (2) time for train to clear stop (station)—
hour. equals (length of train)/(average speed).
Note that the maximum service volume at LOS B would occur The train length is 75 × 2 or 150 ft. Assuming the train acceler-
if 5 loading positions are provided. Applying the berth efficiency ates from rest to 15 mph (22 ft/sec), the average speed is 11 ft per
factor 2.5 to the 20 buses would result in a service volume of 50 sec. Therefore 150/11 or about 14 sec is needed for clearance.
buses over the hour. In practice, one might accept a greater proba- Total clearance, therefore, is 34 sec.
bility of queue formation by providing fewer bays. Alternatively, b. The maximum number of transit units per hour can be ob-
fare collection procedures could be improved to reduce the tained from Eq. 12-10b, or Eq. 12-2a, adjusted for the g/C ratio
dwell times. of 0.50. For on-street operations, R is equal to 0.833.
g/C (3,600 R) (0.50) (3,600) (0.833) 1,500

cv = = =
Calculation 15—Rail Rapid Transit tc + D (g/C) 34 + 60 (0.50) 64
1. Description—A rail rapid transit line operates twenty 8-car = 23.4 units/hour, say 23 units.
trains per track per hour. Scheduled loads average 2.0 passengers c. Passengers per train values can be estimated in two ways:
per seat. How many people can the line carry? Cars are 75 ft long (1) Table 12-4 shows LRV’s having a crush load of 400 to 460
and can seat 75 people. passengers per pair of cars, and a maximum schedule load of 180
2. Solution—This number of people per hour per track can be to 190 passengers per car; (2) Table 12-6 shows maximum sched-
estimated by applying Eqs. 12-5 or 12-6. For instance, applying ule loads ranging from 3.3 to 3.9 persons per sq ft. Assuming a
Eq. 12-5, 75 × 8.8-ft car, this corresponds to 170 to 200 passengers per car.
Trains Cars Seats Pass. Selecting the midpoint, 3.6 sq ft per passenger, results in 185
Passengers per hour = × × × persons per car.
Hour Train Cars Seat
= 20 × 8 × 75 × 2.0 = 24,000 persons/hr Using the 185 persons per car gives a capacity of 185 × 2 or
370 persons per 2-car train.
d. Passenger capacity is computed as follows. The passengers
Calculation 16—Light Rail Transit on City Street per hour reflects the product of the passengers per train and the
trains per hour. This gives 370 × 23 or 8,500 passengers per hour
1. Description—A light rail transit line operates within a city (rounded).
street median through signalized intersections. Service is provided Note that the passenger capacity could be computed directly
by 2-car trains, with each car about 75 ft long. The g/C time is from Eq. 12-2c.
0.50 and the passenger dwell times are 60 sec. How many people
per hour can the trains carry? (g/C) 3,600 nSR
cp = nScv =
2. Solution—Estimating the passenger capacity of the line re- (g/C) D + tc
quires three intermediate calculations. These are:
where S = 185 and n = 2.
a. The train clearance times including: (1) minimum separation
between trains, and (2) time for a train to clear the stop. This capacity can be realized, if there are at least two major
b. The maximum number of trains per hour. stops, prepayment of fares, and at least two sets of double-width
c. The number of passengers that each train can carry. doors on each car available for boarding passengers.
REFERENCES
4. Levinson, H.S., Adams, C.L., and Hoey, W.F., ‘‘Bus Use of

1. Levinson, H.S., Characteristics of Urban Transportation De- Highways—Planning and Design Guidelines.’’ NCHRP Re-
mand. Prepared for Urban Mass Transportation Administra- port 155 (1975) 161 pp.
tion, Federal Highway Administration, Washington, D.C. 5. Hoey, W.F., and Levinson, H.S., ‘‘Bus Capacity Analysis.’’
(1977). Transportation Research Record 546, Transportation Re-
2. Levinson, H.S., et al., ‘‘Bus Use of Highways—State of search Board, Washington, D.C. (1975).
the Art.’’ NCHRP Report 143 (1973) 406 pp. 6. Annual Reports. Board of Supervising Engineers, Chicago
3. Rainville, W.S., Homburger, W.S., and Hyde, D.C., ‘‘Pre- Traction 1912, Chicago, Illinois (1916).
liminary Progress Report of Transit Subcommittee, Committee 7. Pushkarev, B.S., Zupan, J.M., Cumella, R., Urban Rail in
of Highway Capacity.’’ HRB Proc., Vol. 40, Highway Re- America: A Regional Plan Association Book. Indiana Univer-
search Board, Washington, D.C. sity Press, Bloomington, Ind. (1982).
12-48 urban streets
8. Homburger, W.S., Ed., Transportation and Traffic Engi- 23. Szasz, D., et al., COMONOR, ‘‘Coordinated Bus Convey.’’
neering Handbook. Second Edition, Prentice-Hall Inc., Engle- CET Technical Bulletin No. 9 (São Paulo: Companhia de En-
wood Cliffs, N.J. (1982). genharia de Trafego) (1978) p. 80.
9. Vuchic, V.R., Urban Public Transportation: Systems and 24. Capacities and Limitations of Urban Transportation Modes.
Technology. Prentice-Hall Inc., Englewood Cliffs, N.J. (1981). An Informational Report, Institute of Traffic Engineers, Wash-
10. Scheel, W., and Foote, J.E., ‘‘Bus Operation in Single Lane ington, D.C. (May 1965).
Platoons and Their Ventilation Needs for Operation in Tun- 25. Vuchic, V., Day, F., and Anderson, B., ‘‘Theoretical and
nels.’’ Research Publication GMR-808, General Motors Re- Practical Capacities of Transit Modes.’’ Intersociety Commit-
search Laboratories, Warren, Michigan (1962). tee on Transportation, Atlanta, Ga. (July 14–18, 1975).
11. Scheel, J.W., and Foote, J.E., ‘‘Comparison of Experimental 26. Pushkarev, B., and Zupan, J., Where Rail Transit Works.
Results with Estimated Single Lane Bus Flows Through a Regional Plan Association, New York, N.Y. (1978).
Series of Stations Along a Private Busway.’’ Research Publi- 27. MBTA Patronage and Capacity Statistics (1971).
28. Bruce Campell Associates for MBTA, Surface Car Operations
cation GMR-888, General Motors Research Laboratories,
Study—Beacon Street Green Line in Brookline and Boston.
Warren, Michigan (1969).
Prepared for the MBTA, Boston, Mass. (1969).
12. Soberman, R.M., and Hazard, H.A., Ed., Canadian Transit
29. Daimant, E.L., Light Rail Transit: State of the Art Review.
Handbook. University of Toronto and York University, Joint
DeLeuw, Cather & Co., Report No. DOT UT 50009 (1976).
Program in Transportation, Toronto, Canada (Jan. 1980), 30. Homburger, W.S., ‘‘Notes on Transit System Characteris-
Chapter 7, Transit Capacity. tics.’’ University of California, Institute of Transportation
13. Hodgkins, E.A., ‘‘Effect of Buses on Freeway Capacity.’’ Studies, Information Circular 40, Berkeley, Calif. (1975).
Highway Research Record 59, Transportation Research Board, 31. Downes, D.P., The Effect of an Additional Transit Lane on
Washington, D.C. (1965). Bus Travel Times. Thesis, Yale Bureau of Highway Traffic,
14. Canty, E.T., ‘‘Stimulation and Demonstration of Innovative Yale University (1959).
Transit Systems.’’ Research Publication GMR-1400, Research 32. Hayashida, K., Fujita, A., Humayasu, T., and Lum, G.,
Laboratories of General Motors Corporation, Warren, Mich. ‘‘Street Capacity for Buses in the Honolulu Central Business
(1973). District.’’ Transportation Research Record 699, Transporta-
15. Crowley, K.W., ‘‘Analysis of Car-Bus Relationships in the tion Research Board, Washington, D.C. (1979).
Lincoln Tunnel.’’ Traffic Eng., Vol. 63, No. 12, Institute of 33. ‘‘Bus Route and Schedule Planning Guidelines.’’ NCHRP
Transportation Engineers, Washington, D.C. (Sept. 1963). Synthesis of Highway Practice 69 (1980) 99 pp.
16. Blake, H.W., and Jackson, W., Electric Railway Transporta- 34. ‘‘Interim Materials in Highway Capacity.’’ Transportation
tion, McGraw-Hill Book Company, New York, N.Y. (1924). Research Circular 212, Transportation Research Board,
17. Cuntill, M.A., and Watts, P.F., ‘‘Bus Boarding and Alight- Washington, D.C. (1980).
ing Times.’’ Great Britain Transport and Road Research Labo- 35. Levinson, H.S., and Texas Transportation Institute, Concep-
ratory, Crowthorne, England, Report LR 521 (1973). tual Planning and Design, Lockwood Transit Center (Mar.
18. Levinson, H.S., ‘‘Analyzing Transit Travel Time Perfor- 1983).
mance.’’ Transportation Research Record 915, Transportation 36. ‘‘Highway Capacity Manual.’’ HRB Special Report 87, Trans-
Research Board, Washington, D.C. (1983). portation Research Board, Washington, D.C. (1965).
19. Kraft, W.H., An Analysis of the Passenger Vehicle Interface 37. Draft Alternatives Analysis Procedures and Technical Guide-
of Street Transit Systems with Applications to Design Optimi- lines: Appendix A Estimating of Transit Supply Parameters.
Urban Mass Transportation Administration (1980).
zation. Doctoral Dissertation, New Jersey Institute of Technol-
38. Transportation and Parking for Tomorrow’s Cities. Wilbur
ogy, Newark, N.J. (Sept. 1975).
Smith & Associates, New Haven, Conn. (1966).
20. Kraft, W.H., and Eng-Wong, P., Passenger Service Time
39. Jacobs, M., Skinner, R.E., and Lemer, A.C., ‘‘Transit Proj-
Characteristics of Street Transit Systems. Compendium of ect Planning Guidance—Estimation of Transit Supply Param-
Technical Papers, Institute of Transportation Engineers, 47th eters,’’ UMTA-MA-09-9015-85-01 (Oct. 1984).
Annual Meeting, Mexico City, Mexico (Oct. 2–6, 1977). 40. Lang, A.S., and Soberman, R.M., Urban Rail Transit: Its
21. Levinson, H.S., INET Transit Travel Times Analysis. Pre- Economics and Technology. Massachusetts Institute of Tech-
pared for Urban Mass Transportation Administration, Wash- nology Press, Cambridge, Mass. (1964).
ington, D.C. (Apr. 1982). 41. Rothery, R., Silver, R., Herman, R., and Torner, C.,
22. Milwaukee Central Area Transit Distribution System. Barton ‘‘Analysis of Experiments on Single-Lane Bus Flow.’’ Opera-
Aschman Associates, Chicago, Illinois. tions Research, Vol. 12, No. 6 (Nov.–Dec. 1964).
APPENDIX I
BUS CAPACITY EXPERIENCE
Table I.12.1. Reported Theoretical Bus Lane Capacities

average
bus stop average equivalent
buses headway spacing bus speed passengers
facility or source per hour (sec) (ft) (mph) per houra
Uninterrupted Flow
G.M. Proving Grounds:
Uninterrupted Flow
(Initial Studies) 1,450b 2.5 No Stops 33 72,500
Highway Capacity Manual, 1985
Freeway: Level-of-Service D 1,060 3.4 No Stops 40–47 53,000
Level-of-Service C 780 4.6 No Stops 48–50 39,000
Freeway: Level-of-Service D 940 3.8 No Stops 33 47,000
Level-of-Service C 690 5.2 No Stops 40–50 34,500
G.M. Proving Grounds:
c
6-Bus Platoons, 30-sec On-Line Stops 400 0.3 mile 15 20,000
City Streets
Arterial Streets—25-sec Loading
Random Arrival (Approximate LOS C) 72 50 Not Cited Not Cited 3,600
Toronto Transit Commission
(Planning Criteria) 60 60 500–600 ft 10 3,000
a
Equivalent passenger volume assumes 50 passengers per bus.
b
Ref. 41; subsequent studies have reported bus volumes of 900 to 1,000 vehicles per lane per hour; these are consistent with reported flows.
c
2.4 sec within the platoon with a platoon every 54 sec on the average.
SOURCE: Compiled from various bus-use studies.
12-50 urban streets
Table I.12-2. Observed Peak-Hour Bus Volumes on Streets and Freeways

average
bus stop average
buses headway spacing bus speed pass. per
facility or source per hour (sec) (ft) (mph) hour remarks
Freeway or Busway
Lincoln Tunnel 735 4.9 No Stops 30 32,560 Connects to Midtown
Uninterrupted Flow bus terminal
I-495 (New Jersey) 485 7.3 No Stops 30–40 21,600
Exclusive Bus Lane,
Uninterrupted Flow
San Francisco– 350 10.3 No Stops 30–40 13,000 Pre-BART connects
Oakland Bay Bridge to Transbay terminal
Shirley Highway 200 18.0 No Stops 35(Freeway) 10,000 900-ft stop spacing
Busway, Wash., D.C. in CBD
Bus-Only Mall
State Street, Chicago 180 20.0 400 0–5 9,000 Based on peak 15-
min rate
Portland, 5th at 6th Ave. 180 20.0 NA 5–10 9,000
Arterial Street
Michigan Ave., Chicago 228 15.0 NA NA 11,400 Some multiple lane
use, 5-min rate
Madison Ave., N.Y.C. 200 6 18.0 1,000 NA 10,000 Two exclusive
bus lanes
Hillside Ave., N.Y.C. 170 21.0 530 Not Cited 8,500a Multiple lane use
with lightly
patronized stops
14th Street, Wash., D.C. 160 23.0 900 5–12 8,000 Approach to CBD
Market St., Philadelphia 150 24.0 300–600 5–10 6,100–9,900 Multiple lanes—
Pre-Chestnut St. mall
K Street, Wash., D.C. 130 28.0 500 5–8 6,500 Pre-Metro
Main St., Rochester 80 45.0 1,000 5 4,000 Some platooning
at stops
Downtown Streets with 80–120 30.0–45.0 500 5–10 4,500–6,000a
Stops (Various Cities)
a
Estimated, assuming 50 passengers per bus; (1 ft = 0.305 m; 1 mph = 1.6 kph).
SOURCE: Compiled from various bus-use studies—1972–1978 conditions. Summarized in Ref. 34.
Table I.12-3. Observed Bus Volumes on Urban Limited Access Facilities—Peak Direction of Flow,
1972–1976 Conditions
vehicles percent
per hour passengers carrieda carried
facility area bus auto bus auto total by bus
Lincoln Tunnel New York 735 3,200 32,560 5,065 37,625 85.5
Bay Bridge, Post BART, San Francisco San Francisco 200 8,700 8,900 16,000 24,900 35.7
Bay Bridge, Pre-BART, San Francisco Oakland 327 8,115 13,000 10,400 23,400 55.5
Shirley Highway (I-95) Wash., D.C. 200 3,600 10,000 5,000 15,000 67.0
Gowanus Expressway New York 106 2,900 5,300 4,350 9,650 54.9 (1976)
Ben Franklin Bridge Philadelphia 137 4,490 5,065 5,620 10,685 47.5
Long Island Expressway New York 89 2,710 3,560 4,100 7,660 46.5
Memorial Bridge Wash., D.C. 100 3,690 4,020 6,650 10,670 37.6
Lions Gate Bridge Vancouver, BC 45 3,300 2,000 4,600 6,600 30.2
Schuylkill Expressway Philadelphia 78 5,300 2,800 6,650 9,450 29.5
Southeast Expressway Boston 65 4,200 2,450 6,000 8,450 29.0
I-71 Cleveland 35 3,200 1,850 4,500 6,350 29.0
Golden Gate Bridge San Francisco 80 6,650 3,750 9,250 13,000 28.8
San Bernardino Freeway Los Angeles 70 6,800 3,500 10,000 13,500 25.9
South Capitol St. Bridge Wash., D.C. 32 3,335 1,920 5,000 6,920 27.7
George Washington Bridge New York 108 9,440 4,245 13,215 17,460 24.3
14th St. Bridge Wash., D.C. 79 6,565 3,295 10,425 13,720 24.0
North Lake Shore Drive Chicago 80 9,500 4,000 14,200 18,200 22.0
John C. Lodge Freeway Detroit 40 4,950 1,800 6,920 8,720 20.6
North Central Expressway Dallas 32 4,000 1,200 5,600 6,800 17.5
Bayshore Freeway San Francisco 35 6,800 2,270 10,880 13,150 17.3
South Lake Shore Drive Chicago 24 5,700 1,400 8,000 9,400 14.9
I-5 Seattle 47 9,800 2,300 13,700 16,000 14.4
Hollywood Expressway Los Angeles 36 7,650 1,755 10,500 12,255 14.4
North Expressway Atlanta 24 4,550 1,070 6,380 7,450 14.4
East Memorial Shoreway Cleveland 24 5,800 1,250 8,100 9,350 13.3
Memorial Drive Houston 11 2,250 500 3,380 3,880 12.9
Stevenson Expressway Chicago 16 4,600 840 6,900 7,740 10.9
Harbor Freeway Los Angeles 23 7,200 1,050 10,000 11,050 9.5
I-45N Houston 19 6,450 875 9,550 10,425 8.4
I-35W Minneapolis, 13 4,950 585 6,900 7,485 7.8
St. Paul
US 59 Houston 13 6,900 600 10,300 10,900 5.5
I-45S Houston 11 6,000 505 9,000 9,505 5.3
I-10W Houston 8 5,870 370 8,800 9,170 4.0
Jones Falls Expressway Baltimore 3 2,780 125 3,900 4,025 3.1
Chrysler Freeway Detroit 4 5,550 180 7,750 7,930 2.3
a
Involves assumption in some cases as to car or bus occupancy.
12-52 urban streets
Table I.12-4. Peak-Hour Bus Volumes on Urban Arterials, Ranked by Percentage of Total Passengers Carried
by Bus, in Dominant Direction of Flow Under 1972–1976 Conditions
vehicles per hour passengers carrieda percent
carried
arterial location city bus auto total bus auto total by bus
Nicollet Mall Minneapolis 64 0 64 2,900 0 2,900 100.0
Market St. (East of Broad) Philadelphia 143b 465 608 8,300 695 8,995 92.5
b
State St. at Madison Chicago 151 465 616 6,100 660 6,760 90.0
b
Hillside Ave. New York 170 630 800 8,500 950 9,450 90.0
Pennsylvania Ave. at Washington, D.C. 120 600 720 6,000 900 6,900 87.0
Seventh St.
Market St. at Van Ness San Francisco 155b 1,200 1,355 9,900 1,550 11,450 86.5
Main St. at Fourth St. Los Angeles 115 720 835 5,850 1,100 6,950 84.0
Main St. at Harwood St. Dallas 100 635 735 4,400 900 5,300 83.0
Hill St. at Seventh St. Los Angeles 109 800 909 5,250 1,200 6,450 81.5
Broad St. at Hunter St. Atlanta 48 290 338 1,920 435 2,355 81.5
Seventh St. at Main St. Los Angeles 91 705 796 4,500 1,050 5,550 81.0
Forbes Ave. at Wood St. Pittsburgh 47 400 447 2,300 560 2,860 79.5
Fifth Ave. at Smithfield Pittsburgh 47 420 467 2,300 590 2,890 79.5
Liberty St. at Sixth Ave. Pittsburgh 66 650 716 3,250 910 4,160 78.2
K St. N.W. at 13th St. Washington, D.C. 130 1,300 1,430 6,500 1,950 8,450 77.0
Eye St. at 13th St. Washington, D.C. 104 1,100 1,204 5,200 1,600 6,800 76.5
Smithfield St. at Fifth Ave. Pittsburgh 50 550 600 2,450 770 3,220 76.0
Thirteenth St. at F St. Washington, D.C. 101 1,050 1,151 5,000 1,600 6,600 75.8
Broadway at Sixth St. Los Angeles 78 850 928 4,000 1,390 5,390 74.5
Adams Street Bridge Chicago 107 785 892 3,425 1,220 4,645 73.7
Granville St. at Georgia Vancouver 70 900 970 3,150 1,200 4,350 72.5
Wisconsin Ave. Milwaukee 78 935 1,013 3,100 1,200 4,300 72.0
Chestnut St. at 12th St. Philadelphia 67 890 957 3,350 1,350 4,700 71.5
State St. at Roosevelt Chicago 72 670 742 2,305 935 3,240 71.4
Washington St. at Wacker Chicago 108 1,100 1,208 3,800 1,540 5,340 71.4
Wood St. at Forsyth Ave. Pittsburgh 55 800 855 2,700 1,120 3,820 70.8
Seventh St. at Washington, D.C. 80 1,150 1,230 4,000 1,720 5,720 70.0
Pennsylvania Ave.
Main St. at Pratt Hartford 75 625 700 1,875 815 2,690 70.0
Jackson Blvd. Bridge Chicago 88 845 933 2,815 1,325 4,140 68.0
Sixth Ave. at Smithfield Pittsburgh 33 560 593 1,620 780 2,400 67.6
Eglinton Ave. at Bathurst Toronto 80 1,200 1,280 3,300 1,700 5,000 66.0
Elm St. at Harwood Dallas 80 1,345 1,425 3,500 1,880 5,380 65.2
Sacramento St. San Francisco 25 410 435 1,000 535 1,535 65.0
Constitution Ave. at 15th Washington, D.C. 120 2,200 2,320 6,000 3,300 9,300 64.5
St.
Spring St. at Seventh St. Los Angeles 111 1,500 1,611 4,450 2,500 6,950 64.0
Sixteenth St. at Florida Washington, D.C. 80 1,500 1,580 4,000 2,250 6,250 64.0
Ave.
Continued
Table I.12-4. Continued

vehicles per hour passengers carrieda percent
carried
arterial location city bus auto total bus auto total by bus
Fourteenth St. at Washington, D.C. 80 1,550 1,630 4,000 2,350 6,350 63.0
Constitution Ave.
Connecticut Ave. at Washington, D.C. 90 1,800 1,890 4,500 2,700 7,200 62.5
Cathedral Ave.
Walnut at 15th St. Philadelphia 48 960 1,008 2,400 1,450 3,850 62.5
Commerce St. at St. Paul Dallas 72 1,415 1,487 3,300 2,120 5,420 61.0
St.
Sheridan Rd. at Hollywood Chicago 32 500 532 1,100 700 1,800 61.0
Ave.
Michigan Ave. at Chicago 77 770 847 1,815 1,210 3,025 60.0
Roosevelt Rd.
Asylum St. at Main St. Hartford 35 450 485 875 585 1,460 60.0
Michigan Ave. Bridge Chicago 116 1,590 1,706 3,580 2,390 5,970 60.0
(Upper Level)
Sutter St. San Francisco 63 1,300 1,363 2,500 1,700 4,200 59.5
Madison Ave. at 42nd St. New York 96 2,400 2,496 4,800 3,600 8,400 57.1
Second Ave. at 42nd St. New York 110 2,800 2,910 5,500 4,200 9,700 56.8
First Ave. at 44th St. New York 110 2,800 2,910 5,500 4,200 9,700 56.8
Sixth Ave. at Figueroa St. Los Angeles 29 965 994 1,875 1,430 3,305 56.7
Georgia Ave. at Granville Vancouver 45 1,200 1,245 2,000 1,600 3,600 55.5
Clay St. San Francisco 26 650 676 1,050 850 1,900 55.3
Ninth St. at Market St. Philadelphia 22 600 622 1,100 900 2,000 55.0
Second Ave. North Birmingham, Ala. 44 1,400 1,444 2,300 1,950 4,250 54.0
Grand Ave. at Temple St. Los Angeles 24 855 879 1,400 1,215 2,615 53.5
Geary St. San Francisco 43 1,250 1,293 1,720 1,630 3,350 51.4
Howard St. at Fayette St. Baltimore 30 470 500 790 755 1,545 51.0
Marietta at Spring St. Atlanta 35 1,050 1,085 1,400 1,580 2,980 47.0
Peachtree St. at Ellis St. Atlanta 55 1,700 1,755 2,200 2,550 4,750 46.5
Tyron St. Charlotte, N.C. 40 1,150 1,190 1,200 1,700 2,900 41.4
Eighth St. at Los Angeles Los Angeles 30 1,155 1,185 1,290 1,835 3,130 41.3
St.
O’Farrell St. San Francisco 27 1,200 1,227 1,080 1,550 2,630 41.2
Trade St. Charlotte, N.C. 30 1,030 1,000 1,000 1,500 2,500 40.0
Pratt St. at Paca St. Baltimore 64 2,390 2,454 2,215 3,825 6,040 36.7
Charles St. at Madison St. Baltimore 33 1,915 1,948 1,480 3,060 4,540 32.6
Lombard St. at Greene St. Baltimore 42 1,750 1,792 1,335 2,800 4,135 32.0
Eleventh St. Bridge Washington, D.C. 54 4,120 4,174 2,870 7,735 10,605 27.1
Cathedral St. at Eager St. Baltimore 36 1,545 1,581 880 2,470 3,350 26.3
St. Paul St. at Preston St. Baltimore 45 2,815 2,860 1,375 4,505 5,880 23.4
Calvert St. at Lexington St. Baltimore 39 2,645 2,684 1,185 4,230 5,415 21.9
a
Data involve assumptions in some cases as to auto or bus occupancy.
b
Buses operate in more than one lane.
12-54 urban streets
Table I.12-5. Observed Bus Volumes on Urban Arterials—Update, Peak Direction of Flow (1978–1984)
metropolitan vehicles per hr passengers carried % carried

facility area bus auto total bus auto total by bus
Main St. Rochester 80 700 780 4,000 1,050 5,050 79.2
Madison Ave. N.Y. City 200a 1,500 1,700 10,000 2,000 12,000 83.3
Temple St. New Haven 35 575 610 1,380 750 2,130 64.8
W Chapel St. New Haven 25 745 770 1,000 970 1,970 50.8
Church St. New Haven 33 1,400 1,433 1,320 1,820 3,140 42.0
14th St. Wash., D.C. 160b 1,480 1,640 6,400 1,930 8,330 76.8
a
Estimated, based on 700 buses in 5 hr. This is in dual bus lane.
b
Pre-Metro.
SOURCE: Herbert S. Levinson.
Table I.12-6. Observed Passengers at Major Bus Terminals

george washington greyhound bus transbay bus
port authority bridge bus terminal, clark terminal, san
bus terminal, terminal, and randolph francisco,
new york, n.y. new york, n.y. sts., chicago, ill.f calif.
Development Costsb $58,000,000.00 $15,000,000.00 $8,000,000.00 $11,000,000.00
Type of Bus Service Commuter and Commuter and Mainly Intercity Intercity and
Intercity Intercity Commuter
Date Completed 1950 1963 1952 1960
Number of Bus Levels 3 2 1 1
Number of Bus Loading Docks 184c 43 30 37
Contiguous Transportation Subway, Local Bus, Subway, Local Bus Subway, Local Bus, Streetcar and
Facilities Auto Parking Curb Parking Bus, Auto
Parking
Direct Ramp Connections Lincoln Tunnel George Washington Garvey St. and San Francisco–
Bridge Wacker Dr. Oakland Bay
Bridge
Number of Passengersa Daily 105,500 20,000 — 44,000, 35,000d,
22,000e
Peak Hr 32,600 4,200 10,000 13,000, 16,000d,
10,000e
Number of Daily 3,350 850 — 1,150d, 900e
Buses Peak Hr 730 108 — 400d, 250e
Average Bus Daily 27.4 23.5 — 20.0
Occupancy Peak Hr 44.1 39.0 — 37.2
Avg. No. of Buses Daily 18.2 19.6 — 24.3–31.1
Per Dock Peak Hr 4.0 2.5 — 6.8–10.8
Avg. Bus Layover Daily 1.32 1.22 — 0.40
Time in Hours Peak Hr 0.25 0.4 — 0.16
Ancillary Land Uses Retail convenience Retail convenience Retail convenience Retail
goods, restaurants goods, restaurants goods and offices convenience
over station goods
Remarks Saves buses 30 min Located over Cross Designed to allow Prior to 1960
over previous Bronx Expressway office building over Key System
operations station trains used
terminal
a
One-direction-only bus volumes.
b
Data on maintenance costs and revenues unavailable.
c
Before expansion.
d
Before BART.
e
After BART.
f
This terminal is being replaced, 1985.
Table I.12-7. Observed Peak Bus Berth Volumes and Flow Rates at Bus Terminals
peak hour
buses loading
terminal and city (one-way) berths buses/berth
Eglinton, Torontob,d 250 13 19.2
Trans Bay, San Francisco (Pre-BART) 400 37 10.8
Jefferson Park, Chicago 140 14 10.0
69th St. and Ryan, Chicago 40 4 10.0
69th St., Philadelphiaa 90 10 9.0
Southwest, Washington 80 10 8.0
Dixie, Cincinnati 48 6 8.0
Wilson Subway, Torontod 136 18 7.6
Trans Bay, San Francisco (Post-BART) 250 37 6.8
95th St. and Ryan, Chicago 106 22 4.8
McKeesport, Pittsburgh 30 7 4.3
Midtown, New Yorkc 730 184 4.0
George Washington Bridge, New York 108 43 2.5
a
Includes buses and streetcars.
b
Before Yonge St. subway extension.
c
Includes 26 intercity bus bays; before terminal expansion.
d
Free transfer to subway.
APPENDIX II
RAIL CAPACITY EXPERIENCE
This appendix contains information on actual observed rail tran- Using representative values of the parameters, the results for
sit values, summarized in the first two tables. Table II.12-3 con- the several relations are computed in the same table. More
tains an enumeration of several theoretical relations for the detailed information on specific transit vehicle characteristics and
minimum headway between trains, extracted from the literature. capacities is contained in Ref. 39.
Table II.12-1. Observed Peak-Hour Passenger Volumes on Streetcar and LRT Lines—Europe
trains cars passengers passengers
transit system per per headway in peak per train
type city year hour hour seconds direction (rounded)
Streetcars in Street Vienna 1937 180 540 20 26,200a 150
Stuttgart 1930 160 480 23 23,000a 140
Hamburg 1957 120 300 30 19,200a 160
Hong Kong 1978 96 38 38 8,000a 83
Melbourne 1978 89 NA 40 4,400 49
Dusseldorf 1975b 90 NA 40 NA
Hannover 1975b 80 NA 45 NA
Cologne 1975b 60 NA 60 NA
Belgrade 1978 51 79 46 4,200 60
Köln 1978 32 NA 113 6,500 203
LRT-Tunnels Köln 1978 48 48 75 10,000 208
Partial or Full Hannover 1978 38 76 95 10,830a
Signal Control Stuttgart 1975b 38 NA 95 NA
a
Estimated by Ref. 9.
b
Estimated herein.
SOURCE: Refs. 8, 9, 29.
12-56 urban streets
Table II.12-2. Rapid Transit Car and Train Capacities

total seated
length width area seated passengers maximum passengers/
(ft) (ft) (ft2) passengers schedule crush cars/train train
New IRT 51.33 8.79 451.2 44 140 180 10–11 440–484
York
IND 60.50 10.0 605 50 180 220 10 500
City
Transit R-44 225
75.00 10.0 750.0 72–76 225 8 576–608
Authority R-46 290
Port Authority
of N.Y. and N.J. 51.25 9 473.0 42 140 200 7 294
(PATH)
Chicago Transit
48.25 9.33 450.1 c.50 125 135 8 400
Authority
Philadelphia (SEPTA)
Broad St. 67.50 10.00 675.0 67 NA 281 (est.) 6 450
Market St. 55.33 9.08 502.4 55 115 200 8 (est.) 440
Massachusetts
Bay Transportation
Authority
Blue Line 48.75 8.58 418.3 48 125 191 4 192
Orange Line 55.31 9.28 513.3 54 175 240 4 216
Red Line 69.81 10.35 722.5 63 208 275 4 252
New Jersey (PATCO) 67.83 10.12 686.4 80 100 200 8 640
Toronto Transit
Commission
1962–1975 74.76 10.33 772.3 84 230 310 6 504
1953–1958 57.00 10.33 588.8 62 174 233 8 496
Bay Area
75.00 10.5 787.5 72 144 216 8 576
Rapid Transit
Montreal Urban
Community 56.42 8.25 465.5 39 157 208 29 351
Transit Commission
Greater Cleveland
Regional Transit
Authority
Airporter 70.25 10.41 731.3 80 120 140 4 320
Other 48.75 10.33 403.6 54 100 197 6 324
Washington
Metropolitan
75.00 10.15 761.2 80 175 240 6 480
Area Transit
Authority
Continued
Table II.12-2. Continued

total passengers/ seated total passengers/ ft2/ ft2/total
train passengers/ foot of length seated passengers
design crush foot of length schedule crush passengers schedule crush
New IRT 1,400 1,800 0.86 2.72 3.51 10.2 3.22 2.50
York
IND 1,800 2,200 0.83 2.97 3.64 12.1 3.36 2.75
City
Transit R-44
1,800 2,240 0.96–1.01 3.00 3.73 9.9–10.1 3.33 2.67
Authority R-46
Port Authority
of N.Y. and N.J. 980 1,400 0.82 2.73 3.90 11.3 3.37 2.36
(PATH)
Chicago Transit
1,000 1,480 1.03 2.59 3.83 9.0 3.60 2.43
Authority
Philadelphia (SEPTA)
Broad St. NA 1,686 0.99 NA 4.16 10.1 NA 2.40
Market St. 920 1,600 0.99 2.07 3.61 9.1 4.37 2.51
Massachusetts Bay
Transportation
Authority
Blue Line 500 764 0.98 2.56 3.91 8.7 3.34 2.19
Orange Line 700 960 0.98 3.16 4.34 9.5 2.93 2.14
Red Line 832 1,100 0.90 2.98 3.94 11.4 3.47 2.62
New Jersey (PATCO) 800 1,600 1.01 1.47 2.95 8.6 6.68 3.43
Toronto Transit
Commission
1962–1975 1,380 1,860 1.12 3.08 4.14 9.2 3.36 2.49
1953–1958 1,392 1,864 1.09 3.05 4.09 9.5 3.38 2.52
Bay Area
1,152 1,728 0.96 1.92 2.88 10.9 5.47 3.64
Rapid Transit
Montreal Urban
Community 1,413 1,872 0.69 2.78 3.69 11.9 2.96 2.23
Transit Commission
Greater Cleveland
Regional Transit
Authority
Airporter 480 560 1.14 1.71 1.99 9.1 6.09 5.22
Other 600 1,182 1.11 2.05 4.04 9.3 5.04 2.55
Washington
Metropolitan
1,050 1,440 1.07 2.33 3.20 9.52 4.35 3.17
Area Transit
Authority
SOURCE: Computed by Herbert S. Levinson from data obtained from Roster of North American Rapid Transit Cars 1945–1976, American Public Transit
Association. Schedule and crush load data are based on information received from APTA.
12-58 urban streets
Table II.12-3. Theoretical Rail Rapid Transit Equations

A. Equation
1. Lang and Soberman, 1980a 3. Vuchic, 1981c
h = ts + nL1/V + V/2a + 5.05 V/2bn (1) h = ts + tr + nL1/V + V(k + 1)/2bn + √2 nL1/a (4)
2. Rice, 1977b h = ts + tr + nL1/V + √2 nL1 b1/a(a + b1) + V/b2 (5)
If maximum speed is not reached, b1 = b2

h = ts + tr + nL1/V + V(1/bn + 1/2be + √2(D + nL1)/a (2) b1, b2 = bn
b1, b2 = be
If maximum speed is reached, (Note: excludes safety factor)
h = ts + tr + 2 nL1/V + V(1/bn + 1/2be + 1/2a) + D/V (3)
B. Symbols
h = minimum headway between trains, in sec;
tr = reaction time, in sec, for driver response;
ts = dwell time, in sec, in station;
k = safety factor;
L = length of train = nL1, where: n = no. of cars and L1 = length/car;
V = maximum approach speed, ft/sec;
a = acceleration rate from stop, ft/sec/sec;
b1 = braking rate of lead train, ft/sec/sec;
b2 = braking rate of following car, ft/sec/sec;
bn = normal braking rate, ft/sec/sec;
be = emergency braking rate, ft/sec/sec; and
D = ‘‘run-out’’ distance, ft.
C. Typical Values
English S.I.U.
ts ............................................................................................................. 20–60 sec................................................................... 20–60 sec
tr ............................................................................................................. 3.0 sec........................................................................ 3.0 sec
k.............................................................................................................. 1.5 .............................................................................. 1.5
L = nL1 ................................................................................................... 300–600 ft ................................................................. 91.5–183 m
V ............................................................................................................. 20–30 mph................................................................. 32.2–48.3 kph
29.3–40.0 ft/sec 8.9–13.4 mps
a ............................................................................................................. 2.0 mph/sec................................................................ 0.9 m/sec/sec
2.9 ft/sec/sec
bn ............................................................................................................ 2.9 mph/sec................................................................ 1.3 m/sec/sec
4.3 ft/sec/sec
be ............................................................................................................ 6.7 mph/sec................................................................ 3.0 m/sec/sec
9.8 ft/sec/sec
D............................................................................................................. 150 ft ......................................................................... 45.7 m
D. Results of Computations for:

30 mph (13.4 m/sec)
600 ft (183-m train)
Equation
1. h = ts + 47.13
2. h = ts + 47.30 D = 0 ft
49.74 D = 150 ft
3. h = ts + 50.29 D = 0 ft
53.70 D = 150 ft
4. h = ts + 49.71
5. h = ts + 42.47
For 30 mph and 600-ft long trains, the headway is: 50-sec plus station dwell time
For 60-sec station dwell times, this results in a headway of 110 sec or 33 trains per hour.
a
Lang, A. S., and Soberman, R. M., Urban Rail Transit: Its Economics and Technology. Massachusetts Institute of Technology Press, Cambridge, Mass. (1964).
b
Rice, P., ‘‘Practical Urban Railway Capacity—A World Review.’’ Proc. Seventh International Symposium on Transportation and Traffic Theory. Sasaki, T.,
and Yamaoka, T., 1977, Kyoto, Japan, Institute of System Science Research.
c
Vuchic, V. R., Urban Public Transportation. Systems and Technology. Prentice Hall Inc., Englewood Cliffs, N.J. (1981).
APPENDIX III
EXAMPLES OF BOARDING AND ALIGHTING TIME
Table III.12-1. Typical CBD Service Times per Passenger

seconds per passenger
am midday pm
Philadelphia, 1977
Chestnut St. Transitway 2.5 to 2.8 2.4 to 3.7 2.5 to 3.5
Walnut Street 2.5 3.6 2.9
Minneapolis, 1977
Nicollet Mall 2.3 to 2.5 2.3 to 3.6 3.8 to 4.3
Other Streets 1.4 to 1.7 1.9 to 3.8 1.3 to 4.4
(Second, Marquette)
New Haven 1979–1980
15 Locations 2.9 to 3.1a
2 Locations 3.2 to 3.4a
12 Locations 1.0 to 2.0b
3 Locations 2.1 to 2.5b
a
Boarding
b
Alighting
SOURCE: H. S. Levinson, ‘‘INET Transit Travel Times Analysis.’’ Final report prepared for UMTA, April 1982.
Table III.12-2. Observed Rail Transit Station Dwell Times, 1980

dwell
line time location time (sec) remarks
Single Observations
Lexington Ave. Expr., N.Y. City am Peak 42nd St. 77
Lexington Ave. Local, N.Y. City am Peak 42nd St. 90 50 sec: Pass. dwell time;
40 sec: Wait for expr.
Evanston Express, Chicago pm Peak Howard St. 65 Major transfer
pm Peak Randolph Wells 47
Red Line, Boston pm Peak Park St. 60
Green Line (LRV), Boston pm Peak Park St. 95 Crowded car
Line Observations
time mean stand. dev
Lexington Ave. Expr., N.Y. City pm Peak 53 17
Evanston Express, Chicago pm Peak 42 14
Green Line (LRV), Boston pm Peak 58 24
Milwaukee, Chicago Post pm Peak 19 6
Note: This is a sampling.
SOURCE: H. S. Levinson, ‘‘INET Transit Travel Times Analysis.’’ Final report prepared for UMTA, April 1982.
12-60 urban streets
Table III.12-3. Bus Boarding and Alighting Times in Selected Urban Areas
boarding and boarding and
alighting fare alighting
location bus type method fare scheme collection relationshipa
Louisville, Ky. One-man Alighting only Flat fare Driver T = 1.8 + 1.1 F
One-man Boarding only Flat fare Driver T = −0.1 + 2.6 N
One-man Simultaneous Flat fare Driver T = 1.8 + 1.0 F + 2
−0.02 FN
London Two-man Consecutive Graduated Conductor T = 1.3 + 1.5 (N + F)
One-man Consecutive Graduated Driver T = 8 + 6.9 N + 1.4 F
One-man Simultaneous Flat fare
Single coin Mechanical T = 7 + 2.0N
Two coin Mechanical T = 5.7 + 3.3N b
Toronto One-man Simultaneous Zonal Fare Box T = 1.7N, T = 1.25F
T = 1.4(N + F)
Copenhagen One-man Simultaneous Flat fare Split entryc T = 2.2N
Dublin Two-man Consecutive Graduated Conductor T = 1.4 (N + F)
One-man Consecutive Graduated Driver T = 6.5 N + 3.0 F
France:
Bordeaux One-man Simultaneous Flat fare Driver T = 15 + 3 N
Toulouse One-man Simultaneous Flat fare Driver T = 11 + 4.6 N
Paris One-man Simultaneous Graduated Driver T = 4+5N
Two-man Simultaneous Graduated Conductor T = 2.3 N
a
T = stop time, in sec; N = number of passengers boarding; F = number of passengers alighting.
b
In peak time, T = 5.7 + 5.0 N in off-peak time.
c
Driver and machine.
Table III.12-4. Means and Variances of Observed Passenger Service Time Distributions
direction doors time in seconds coefficient of

location of flow bus type on bus mean variance variation (%)
Montreal, Canada Boarding Can. Car 2 2.097 0.727 40.67
Montreal, Canada Boarding GMC 2 2.034 0.834 44.89
New Brunswick, N.J. Alighting GMC 1 1.972 1.045 51.83
New Brunswick, N.J. Boarding GMC 1 3.471 3.499 53.90
San Diego, Calif. Alighting GMC 2 1.472 0.403 43.34
San Diego, Calif. Boarding GMC 2 2.180 0.868 42.75
SOURCE: Ref. 19.
chapter 13
PEDESTRIANS
CONTENTS
i. introduction .......................................................................................................................................................................... 13-1

Pedestrian Capacity Terminology ......................................................................................................................................... 13-3
Principles of Pedestrian Flow ............................................................................................................................................... 13-3
Pedestrian Speed-Density Relationships........................................................................................................................... 13-3
Flow-Density Relationships .............................................................................................................................................. 13-3
Speed-Flow Relationships ................................................................................................................................................. 13-4
Speed-Space Relationships................................................................................................................................................ 13-4
Effective Walkway Width..................................................................................................................................................... 13-4
Pedestrian Type and Trip Purposes ...................................................................................................................................... 13-6
ii. methodology.......................................................................................................................................................................... 13-7
Levels of Service in Walkways ............................................................................................................................................ 13-7
Walkway Level-of-Service Criteria .................................................................................................................................. 13-8
Effect of Pedestrian Platoons ............................................................................................................................................ 13-10
Levels of Service in Queuing Areas..................................................................................................................................... 13-11
Application of Criteria .......................................................................................................................................................... 13-12
Street Corners .................................................................................................................................................................... 13-12
Crosswalks ......................................................................................................................................................................... 13-14
iii. procedures for application and sample calculations ................................................................................................. 13-14
Analysis Procedures for Walkways ...................................................................................................................................... 13-14
Computational Steps.......................................................................................................................................................... 13-14
Sample Calculation............................................................................................................................................................ 13-14
Analysis Procedures for Street Corners and Crosswalks..................................................................................................... 13-16
Street Corner Analysis (Computational Steps and Sample Calculation) ........................................................................ 13-18
Crosswalk Analysis (Computational Steps and Sample Calculation) ............................................................................. 13-22
Estimating the Decrement to Crosswalk LOS Due to Right-Turning Vehicles ............................................................. 13-26
iv. references .............................................................................................................................................................................. 13-26
appendix i. Worksheets for Use in Analysis of Walkways, Crosswalks, and Street Corners ............................................. 13-26
I. INTRODUCTION
The purpose of this chapter is to describe the basic principles centrated pedestrian movement occurs at public events, in and near
of pedestrian traffic flow, and to present a general framework transit terminals, high-rise buildings, department stores, theaters,
and procedures for the analysis of pedestrian facilities. The scope stadia, parking garages, and other major traffic generators. Pedes-
is limited to sidewalks, crosswalks, and street corners, but the trian safety, trip patterns, and convenience are also a necessary
analysis techniques can be applied to other pedestrian facilities. consideration in all multimodal traffic and transportation studies.
The chapter includes examples illustrating several typical Table 13-1 presents some high pedestrian volumes observed in
applications. several major urban centers.
Pedestrian activity can be a major component in urban street The concentration of pedestrian activity at street corners and
capacity analysis, and pedestrian characteristics are an important crosswalks makes them critical traffic links for both sidewalk
factor in the design and operation of transportation systems. Con- and street networks. An overloaded corner or crosswalk not
13-1
13-2 urban streets
Table 13-1. Observed Pedestrian Flow Rates in Urban Areas*
walkway avg. flow rates peak flow rates for periods

width for full hour less than 1 hour
location time (ft) ped/min ped/min/ft ped/min ped/min/ft
Boston
Washington St (1960) 12–1 PM 7.0 53 7.6 — —
Chicago
CTA (1976) PM — — 5.2 — —
State St/Wash (1960) 12–1 PM 25.0 112 4.5 — —
State St/Wash (1972) 4–5 PM 25.0 93 3.7 — —
State St/Wash (1939) 12–1 PM 25.0 206 8.2 — —
State St/Mad (1929) — 25.0 342 13.7 471 18.8
(15 min)
State St/Mad (1929) — 20.0 287 14.4 368 18.4
(15 min)
Soldiers Fld (1940) — 21.5 202 9.4 298 13.9
(1 min)
Dyche Stadium (1940) — 10.0 114 11.4 167 16.7
(5 min)
Los Angeles 125
Broadway (1940) — 18.0 — — (12 min) 6.9
Des Moines and
Ames, Iowa
Veteran’s Aud. (1975) 10 PM 8.2 — — — 20.0
(5 min)
22.2
(1 min)
College Creek 12 Nn 6.0 — — — 22.3
Footbridge (5 min)
(1975) 31.8
(1 min)
CY Stephens 4:40 PM 7.5 — — — 31.9
Auditorium (5 min)
(1975) 39.2
(1 min)
Iowa State Univ. 1 PM 2.8 — — — 28.7
Armory (1 min)
New York City
Madison Av (1969) 12–1 PM 13.0 167 12.8 — —
Fifth Av (1969) 12–1 PM 22.5 250 11.1 — —
Lexington Av (1969) 12–1 PM 12.0 100 8.3 — —
Eighth Av (1969) PM 15.0 167 11.1 — —
42nd Street (1969) PM 20.0 105 5.3 — —
Port Authority Bus PM — — 25.0 — —
Terminal (1965)
Washington, D.C.
7th St SW (1968) PM 10.0 42 4.2 — —
F Street NW (1981) PM 15.0 19 1.3 — —
Seattle
CBD (1976) PM — — — — 9.6
San Francisco
CBD (1976) PM — — — — 10.8
Winnepeg
CBD Street (1980) 3–4 PM 17.0 74 4.4 — —
* Compiled by H. Levinson ad R. Roess from:
1. Chicago Loop Pedestrian Movement Study, City of Chicago, Chicago, Ill., 1973.
2. Pushkarev, B., and Zupan, J., Urban Space for Pedestrians, Regional Plan Association, New York, N.Y., 1976.
3. Traffic Circulation and Parking Plan-CBD Urban Renewal Area-Boston, Mass., Barton-Aschman Associates, 1968.
4. ‘‘Traffic Characteristics,’’ Traffic and Transportation Engineering Handbook, Institute of Transportation Engineers, Prentice-Hall, Englewood Cliffs, N.J., 1976.
5. ‘‘Characteristics and Service Requirements of Pedestrians and Pedestrian Facilities,’’ Informational Report, ITE Journal, Institute of Transportation Engineers,
Washington, D.C., May 1976.
6. Carstens, R., and Ring, S., ‘‘Pedestrian Capacity of Shelter Entrances,’’ Technical Note, Traffic Engineering, Institute of Transportation Engineers, Washington,
D.C., December 1970.
pedestrians 13-3
only affects pedestrian convenience, but can delay vehicle turning used for vehicular flow are the freedom to choose desired speeds
movements, thereby reducing the capacity of the intersection and and to bypass others. Other measures more specially related to
connecting streets. pedestrian flow include the ability to cross a pedestrian traffic
The principles of pedestrian flow analysis are similar to those stream, to walk in the reverse direction of a major pedestrian
used for vehicular flow. The fundamental relationships among flow, and to generally maneuver without conflicts and changes in
speed, volume, and density are similar. As the volume and density walking speed or gait.
of a pedestrian stream increases from free-flow to more crowded Additional environmental factors which contribute to the walk-
conditions, speed and ease of movement decreases. When the pe- ing experience, and therefore to perceived level of service, are
destrian density exceeds a critical level, volume and speed become the comfort, convenience, safety, security, and economy of the
erratic and rapidly decline. walkway system.
Pedestrian flow on sidewalks is affected by reductions in effec-
tive walkway width caused by various items of street ‘‘fur- 1. Comfort factors include weather protection, climate control,
niture,’’ such as parking meters, light standards, mail boxes, arcades, transit shelters, and other pedestrian amenities.
and trash cans, and by interruptions to flow caused by traffic 2. Convenience factors include walking distances, pathway di-
signals. The traffic signal cycle also results in queues of waiting rectness, grades, sidewalk ramps, directional signing, directory
pedestrians at street corners, which decreases corner circulation maps, and other features making pedestrian travel easy and
capacity and concentrates crossing pedestrians into denser uncomplicated.
platoons. 3. Safety is provided by separation of pedestrian from vehicular
The level-of-service (LOS) concept, first used to define relative traffic, horizontally in malls and other vehicle-free areas, and
degrees of convenience on highways, is also applicable to vertically using overpasses and underpasses. Traffic control de-
pedestrian facilities. With this concept, such convenience factors vices can provide for time separation of pedestrian and vehicular
as the ability to select walking speeds, bypass slower pedestrians, traffic.
and avoid conflicts with others are related to pedestrian density and 4. Security features include lighting, open lines of sight, and
volume. The concept can also be applied to degrees of crowding in the degree and type of street activity.
queuing areas, such as sidewalk corners, transit platforms, and 5. Economy aspect relates to the user costs associated with
other waiting areas. travel delays and inconvenience, and to the rental value and retail
The following sections define pedestrian traffic terminology, development as influenced by pedestrian environment.
develop the principles of pedestrian flow, present the concept of
pedestrian level of service, and provide detailed analysis proce- These supplemental factors can have an important effect on
dures for use. the pedestrian perception of the overall quality of the street
environment. While auto users have reasonable control over
most of these factors, the pedestrian has virtually no control
PEDESTRIAN CAPACITY TERMINOLOGY over them. Although the bulk of this chapter emphasizes level-
of-service analysis, which relates primarily to pedestrian flow
Pedestrian analysis uses some familiar traffic terms, as well as measures, such as speed and space, these environmental factors
others not used elsewhere in the manual. The following listing should always be considered because they can greatly influence
defines the major terms used throughout this chapter: pedestrian activity.
1. Pedestrian speed is the average pedestrian walking speed,
generally expressed in units of feet per second.
2. Pedestrian flow rate is the number of pedestrians passing a Pedestrian Speed-Density Relationships
point per unit time, expressed as pedestrians per 15 minutes or
pedestrians per minute; ‘‘point’’ refers to a perpendicular line of The fundamental relationship between speed, density, and vol-
sight across the width of a walkway. ume for pedestrian flow is analogous to vehicular flow. As volume
3. Unit width flow is the average flow of pedestrians per unit and density increase, pedestrian speed declines. As density in-
of effective walkway width, expressed as pedestrians per minute creases, and pedestrian space decreases, the degree of mobility
per foot. afforded the individual pedestrian declines, as does the average
4. Platoon refers to a number of pedestrians walking together speed of the pedestrian stream.
in a group, usually involuntarily, because of signal control and Figure 13-1 shows the relationship between speed and density
other factors. for a variety of pedestrian classes as determined by four re-
5. Pedestrian density is the average number of pedestrians searchers, including two European sources. The density term,
per unit of area within a walkway or queuing area, expressed as when used to describe pedestrian streams and specified in per-
pedestrians per square foot. sons per square foot, will have small values, generally under
6. Pedestrian space is the average area provided for each 0.50.
pedestrian in a walkway or queuing area, expressed in terms of
square feet per pedestrian; this is the inverse of density, but is a
more practical unit for the analysis of pedestrian facilities.
Flow-Density Relationships
PRINCIPLES OF PEDESTRIAN FLOW The relationship between density, speed, and flow for pedes-
trians is of the same form as for vehicular traffic streams, that
The qualitative measures of pedestrian flow similar to those is:
13-4 urban streets
Figure 13-1. Relationships between pedestrian speed and density.

(Source: Ref. 2)
Figure 13-2. Relationships between pedestrian flow and space.

Flow = Speed × Density (Source: Ref. 2)
v=S×D (13-1)
where flow is expressed as pedestrians per minute per foot, speed

is expressed as feet per minute, and density is expressed as pedestri-
ans per square foot.
The flow variable used in this expression is the ‘‘unit width
flow’’ defined earlier. An alternative and more useful expression
can be developed using the reciprocal of density, or space, as
follows:
Flow = Speed/Space
v = S/M (13-2)
The basic relationship between flow and space, as recorded by

several researchers, is illustrated in Figure 13-2.
The conditions at maximum flow are of interest because this
represents the capacity of the walkway facility. From Figure 13- Figure 13-3. Relationships between pedestrian speed and flow.
2, it is apparent that all observations of maximum unit flow fall (Source: Ref. 2)
within a very narrow range of density—that is, with the average
space per pedestrian varying between 5 and 9 sq ft/ped. Even with other pedestrians. When a critical level of crowding occurs,
the outer range of these observations indicates that maximum flow movement becomes more difficult, and both flow and speed
occurs at this density, although the actual flow in this study is decline.
considerably higher than the others. As space is reduced to less
than 5 sq ft/ped, the flow rate declines precipitously. All movement
Speed-Space Relationships
effectively stops at the minimum space allocation of
2 to 4 sq ft/ped. Figure 13-4 further confirms the relationships of walking speed
These relationships show that pedestrian traffic can be evalu- and available space, and suggests some points of demarcation that
ated qualitatively by using level-of-service concepts similar to can be used to develop level-of-service criteria. The outer range
vehicular traffic analysis. At flows near capacity, an average of of observations shown on Figure 13-4 indicates that at an average
5 to 9 sq ft/ped is required for each moving pedestrian. However, space of about 15 sq ft/ped, even the slowest pedestrians cannot
at this level of flow, the limited area available restricts pedestrian achieve their desired walking speed. Faster pedestrians wishing to
speed and the pedestrian’s freedom to maneuver within the walk at speeds up to 350 ft/min are not able to achieve such speeds
pedestrian stream. until average space is 40 sq ft/ped or more. The space values of
15 and 40 sq ft/ped become critical points in defining level-of-
Speed-Flow Relationships service boundaries, as is illustrated in the ‘‘Methodology’’ section
of this chapter.
Figure 13-3 illustrates the relationship between pedestrian
speed and flow. These curves, similar to vehicular flow curves, EFFECTIVE WALKWAY WIDTH
show that when there are few pedestrians on a walkway (low
flow levels), space is available to choose higher walking speeds. The concept of a pedestrian ‘‘lane’’ has sometimes been used
As flow increases, speeds decline because of closer interactions to analyze pedestrian flow, comparable to the analysis of a
pedestrians 13-5
To avoid interference while passing each other, two pedestrians

should each have at least 2.5 ft of walkway width, as observed
by Oeding and Pushkarev (2). Pedestrians who know each other
and are walking close together will each occupy a width of 2
ft, 2 in., a distance at which there is considerable likelihood of
contact due to body sway. Lateral spacing less than this occurs
only in the most crowded of situations.
The term ‘‘clear walkway width’’ is related to the portion of
a walkway that can be effectively used for pedestrian movements.
Moving pedestrians will shy away from the curb, and will not
press closely against building walls. Therefore, unused space
must be subtracted when determining pedestrian LOS. Further,
a strip preempted by pedestrians standing near a building (as
Figure 13-4. Relationships between pedestrian speed and space.
(Source: Ref. 2) in window shopping) and/or near physical obstructions such as
light poles, mail boxes, and parking meters, should also be
excluded.
The degree to which point obstructions (poles, signs, hy-
drants) influence pedestrian movement and reduce effective walk-
highway lane. The ‘‘lane’’ should not be used in pedestrian way width is not extensively documented. While a single such
analysis, because photographic studies have shown that pedestri- obstruction would not reduce the effective width of an entire
ans do not walk in organized lanes. The ‘‘lane’’ concept is walkway, it would have such an effect in the immediate vicinity
meaningful only in determining how many persons can walk of the obstruction.
abreast on a given walkway width, as in the case of determining A list of typical obstructions and the estimated width of
the minimum sidewalk width to permit two pedestrians to walkways which they preempt is provided in Table 13-2. Figure
conveniently pass by each other. 13-5 shows the width of walkway preempted by curbs, buildings,
Figure 13-5. Preemption of walkway width. (Source: Adapted from Ref. 4)

13-6 urban streets
Table 13-2. Fixed Obstacle Width Adjustment Factors for Walkways*

obstacle approx. width preempted (ft)a
Street Furniture
Light Poles 2.5–3.5
Traffic Signal Poles and Boxes 3.0–4.0
Fire Alarm Boxes 2.5–3.5
Fire Hydrants 2.5–3.0
Traffic Signs 2.0–2.5
Parking Meters 2.0
Mail Boxes (1.7 ft by 1.7 ft) 3.2–3.7
Telephone Booths (2.7 ft by 2.7 ft) 4.0
Waste Baskets 3.0
Benches 5.0
Public Underground Access
Subway Stairs 5.5–7.0
Subway Ventilation Gratings (raised) 6.0+
Transformer Vault Ventilation Gratings (raised) 5.0+
Landscaping
Trees 2.0–4.0
Planting Boxes 5.0
Commercial Uses
Newsstands 4.0–13.0
Vending Stands variable
Advertising Displays variable
Store Displays variable
Sidewalk Cafes (two rows of tables) variable, try 7.0
Building Protrusions
Columns 2.5–3.0
Stoops 2.0–6.0
Cellar Doors 5.0–7.0
Standpipe Connections 1.0
Awning Poles 2.5
Truck Docks (trucks protruding) variable
Garage Entrance/Exit variable
Driveways variable
* To account for the avoidance distance normally occurring between pedestrians and obstacles, an additional 1.0 to 1.5 ft must be added to the preemption width
for individual obstacles.
a
Curb to edge of object, or building face to edge of object.
SOURCE: Ref. 2.
or fixed objects. Figure 13-5 may be used as a guideline when Pedestrians going to and from work, using the same facilities
specific walkway configurations are not available. day after day, exhibit higher walking speeds than shoppers. This
has been shown in Figure 13-1. Older or very young persons
PEDESTRIAN TYPE AND TRIP PURPOSES
will tend to walk at a slower gait than other groups. Shoppers
The analysis of pedestrian flow is generally based on mean, or not only tend to walk slower than commuters, but may decrease
average, walking speeds of groups of pedestrians. Within any the effective walkway width by stopping to window shop. Thus,
group, or among groups, there can be considerable differences in applying the techniques and numerical data in this chapter,
in flow characteristics due to trip purposes, land use, type of group, the analyst should adjust for pedestrian behavior which deviates
age, and other factors. Figure 13-6 shows a typical distribution of from the regular patterns represented in the basic speed, volume
free-flow walking speeds. and density curves.
pedestrians 13-7
Figure 13-6. Typical free-flow

walkway speed distribution.
(Source: Ref. 3)
II. METHODOLOGY
LEVELS OF SERVICE IN WALKWAYS below the 35- to 40-sq ft/ped range. Above that level, Fruin
states that the probability of ‘‘stopping or breaking the normal
The criteria for various levels of service (LOS) for pedestrian walking gait’’ is reduced to zero. Below 15 sq ft/ped, virtually
flow are based on subjective measures that may be somewhat every crossing movement encounters a conflict. Similarly, the
imprecise. However, it is possible to define ranges of space per ability to pass slower pedestrians is unimpaired above 35 sq ft/
pedestrian, flow rates, and speeds which can be used to develop ped, but becomes progressively more difficult as space alloca-
quality of flow criteria. tions drop to 18 sq ft/ped, a point at which passing becomes
Speed is an important level-of-service criterion because it can virtually impossible.
be easily observed and measured, and because it is a descriptor Another level-of-service indicator is the ability to maintain
of the service pedestrians perceive. At speeds of 150 ft/min or flow in the minor direction in opposition to a major pedestrian
less, most pedestrians resort to an unnatural ‘‘shuffling’’ gait. flow. Here the quantitative evidence is somewhat less precise. For
Figure 13-4 shows that this speed corresponds to a space per pedestrian streams of roughly equal flow in each direction, there
pedestrian in the range of 6 to 8 sq ft/ped. At 15 sq ft/ped or less, is little reduction in the capacity of the walkway compared with
even the slowest walkers are forced to slow down (shown by the one-way flow, because the directional streams tend to separate
cross-hatching in Figure 13-4). The fastest walkers cannot reach and occupy a proportional share of the walkway. However, if the
their chosen speed of 350 ft/min until areas are over 40 sq ft/ped. bidirectional split is 90-10, and space is 10 sq ft/ped, capacity
Further, from Figure 13-2, it is evident that these three space reductions of about 15 percent have been observed. This reduction
values, 6, 15, and 40 sq ft/ped, correspond approximately to the is a consequence of the inability of the minority flow to utilize a
maximum flow at capacity, two-thirds of capacity, and one-third proportional share of the walkway.
of capacity, respectively. Photographic studies show that pedestrian movement on
There are other significant indicators of service levels. For sidewalks is affected by the presence of other pedestrians, even
example, the ability of the pedestrian to cross a pedestrian stream at areas above 40 sq ft/ped. At 60 sq ft/ped, pedestrians have been
is shown by Fruin (3) in Figure 13-7 to be impaired at areas observed walking in a ‘‘checkerboard’’ pattern, rather than
13-8 urban streets
Figure 13-7. Cross-flow traffic—

probability of conflict. (Source:
Ref. 3)
Table 13-3. Pedestrian Level of Service on Walkways*

expected flows and speeds
level vol/cap
of space ave. speed, S flow rate, v ratio,
service (sq ft/ped) (ft/min) (ped/min/ft) v/c
A ≥ 130 ≥ 260 ≤ 2 ≤ 0.08
B ≥ 40 ≥ 250 ≤ 7 ≤ 0.28
C ≥ 24 ≥ 240 ≤ 10 ≤ 0.40
D ≥ 15 ≥ 225 ≤ 15 ≤ 0.60
E ≥ 6 ≥ 150 ≤ 25 ≤ 1.00
F <6 < 150 —Variable—
* Average conditions for 15 min.
directly behind or alongside each other. These same observations Graphic illustrations and descriptions of walkway levels of
suggest that up to 100 sq ft/ped are required before completely service are shown in Figure 13-8.
free movement occurs without conflicts, and that at 130 sq ft/ It should be noted that the pedestrian LOS, according to the
ped, individual pedestrians are no longer influenced by others (5). criteria of Table 13-3, is quite good in most areas, as the high
Bunching or ‘‘platooning’’ does not completely disappear until pedestrian flows required for the poorer levels generally occur
space is about 500 sq ft/ped or higher. only in and around major activity centers. In most areas, the design
of walkways is based on the minimum widths required for volun-
tary pedestrian groups to pass each other and similar factors, rather
Walkway Level-of-Service Criteria than on the flow rate.
The LOS criteria apply to pedestrian flow and the space
Table 13-3 shows the criteria for pedestrian level of service. provided for that flow. Pedestrian facilities may also include
The primary measure of effectiveness used in defining pedestrian extensive space intended to enhance the general environment
level of service is space, the inverse of density. Mean speed and that is not used or intended to handle basic pedestrian move-
flow rate are shown as supplementary criteria. Capacity is taken ments. When analyzing pedestrian flow rates per unit width of
to be 25 ped/min/ft, a representative value from Figures 13-2 walkway, such space should not be included. Thus, pedestrian
and 13-3. space intended to provide for window shopping, browsing, or
pedestrians 13-9
Figure 13-8. Illustration of walkway levels of service.

13-10 urban streets
simply sitting or standing in informal groups should not be Short-term fluctuations are present in most unregulated pedes-
considered to be part of the effective walkway width. trian traffic flows because of random arrivals of pedestrians. On
It should also be emphasized that the level-of-service criteria sidewalks, these random fluctuations are further exaggerated by
of Table 13-3 are based on the assumption that pedestrians the interruption of flow and queue formation caused by traffic
distribute themselves uniformly throughout the effective walk- signals. Transit facilities can create added surges in demand by
way width. Pedestrian flow is subject to wide variability on a releasing large groups of pedestrians in short-time intervals,
minute-by-minute basis, and the analyst must consider the effects followed by pauses during which no flow occurs. Until they
of platooning as described in the next section. disperse, pedestrians in these types of groups move together as
a platoon. Platoons can also form if passing is impeded because
Effect of Pedestrian Platoons
of insufficient space, and faster pedestrians slow down behind
slower walkers.
The average flow rates at different levels of service are of It is important for the analyst to determine if platooning or
limited usefulness unless reasonable time intervals are specified. other traffic patterns alter the underlying assumptions of average
Figure 13-9 illustrates that ‘‘average flow rates’’ can be misleading. flow in LOS calculations, and to make appropriate adjustments
The data shown are for two locations in Lower Manhattan, but where necessary.
the pattern is generally characteristic of many concentrated CBD In walkway sections having pronounced platooning effects,
locations. The maximum 15-min flow rates average 1.4 and 1.9 the duration and magnitude of these variations in demand should
ped/min/ft of effective walkway width during the periods measured. be established. This is done by timing and counting these
However, Figure 13-9 shows that flow during a 1-min interval can short-term surges in demand. The magnitude and frequency of
be more than double the rate in another, particularly at relatively occurrence of the platoons would then be compared to the longer
low flows. Even during the peak 15-min period, incremental varia- term 15-min average flow to provide a more accurate view of
tions of 50 to 100 percent frequently occur from one minute to LOS conditions on the walkway segment.
the next.
The scatter diagram shown in Figure 13-10 indicates the platoon
Depending on traffic patterns, it is clear that a facility de-
signed for average flow can afford lower quality of flow for a flow rate (i.e., the rate of flow within platoons of pedestrians) in
proportion of the pedestrian traffic using it. However, it is extrava- comparison to the average flow rate for 58 data periods of 5- to
gant to design for extreme peak 1-min flows which occur only 1 6-min duration. The dashed line approximates the upper limit of
percent or 2 percent of the time. A relevant time period must platoon flow observations.
therefore be determined through closer evaluation of the short- The mathematical expression of this line relating maximum pla-
term fluctuations of pedestrian flow. toon flow rates to average flow rates is:
Figure 13-9. Minute-by-minute variations in pedestrian flow. (Source: Ref. 2)

pedestrians 13-11
Figure 13-10. Relationship between platoon flow and average flow. (Source: Ref. 2)
Platoon Flow = Average Flow + 4 LEVELS OF SERVICE IN QUEUING AREAS

vp = v +4 (13-3)
The concept of using the average space available to pedestrians
where both flows are expressed as pedestrians per minute per as a walkway level-of-service measure can also be applied to
foot. This equation is valid for flows greater than 0.5 ped/min/ queuing or waiting areas. In such areas, the pedestrian stands tem-
ft. For lower flows, consult Figure 13-10 directly. porarily, while waiting to be served. The LOS of the waiting area
The form that this equation takes—a constant increment is related to the average space available to each pedestrian and the
added to the average flow—shows that platooning has a relatively degree of mobility allowed. In dense standing crowds, there is
greater impact at low volumes than at high volumes. This pattern little room to move, but limited circulation is possible as average
is logical, because gaps between platoons tend to fill up as flow space per pedestrian is increased.
increases. The equation can be used in general analyses where Level-of-service descriptions for standing spaces based on av-
specific platooning data are not available. erage pedestrian space, personal comfort, and degrees of internal
Although the magnitude and frequency of platoons should be mobility are shown on Figure 13-11. Standing areas in the LOS
verified by field studies, the LOS occurring in platoons is E category of 2–3 sq ft/ped are experienced only in the most
generally one level poorer than that determined by average flow crowded elevators or transit vehicles. LOS D, at 3–7 sq ft/ped,
criteria, except for some cases of LOS E, which encompasses more typically exists where there is crowding, but where some
a broad range of pedestrian flow rates. The selection of an internal maneuverability is still present. This commonly occurs
appropriate design objective to accommodate either average flows at sidewalk corners where a large group of pedestrians is waiting
over a longer period, or the surges in demand occurring in to cross. Waiting areas where more space is required for cir-
platoons, depends on an evaluation of pedestrian convenience, culation, such as theater lobbies and transit platforms, also
available space, costs, and policy considerations. require a higher LOS.
13-12 urban streets
Figure 13-11. Levels of service for queuing areas. (Source: Ref. 3)
APPLICATION OF CRITERIA Street Corners
The application of these LOS criteria is relatively straight- The street corner is a more complex problem than the mid-
forward for walkways and waiting areas, as indicated in the block situation, involving intersecting sidewalk flows, pedes-
previous sections. Two remaining pedestrian facilities of interest, trians crossing the street, and others queued waiting for the
however, present more complicated situations: street corners signal to change. Because of the concentration of these activities,
and crosswalks. Each of these is briefly discussed in the following the corner is often the critical link in the pedestrian sidewalk
sections. network. An overloaded street corner can also affect vehicular
pedestrians 13-13
operations by requiring added green crossing time or by delaying field study and possible remedial measures. Corrective measures
turning movements. There are two different types of pedestrian could include sidewalk widenings, vehicle-turning restrictions,
area requirements at corners: and/or changes in signal timing. Identifying problem areas is a
primary objective of using LOS as an analytic tool.
Corners function as a ‘‘time-space’’ zone, with waiting pe-
1. Circulation area—Needed to accommodate (a) pedestrians destrians requiring less standing space, but occupying the corner
crossing during the green signal phase, (b) those moving to join for longer periods of time, and moving or circulating pedestrians
the red phase queue, and (c) those moving between the adjoining requiring more space, but occupying the corner for only a few
sidewalks, but not crossing the street.
seconds. The total time-space available for these activities is sim-
2. Hold area—Needed to accommodate standing pedestrians
ply the net area of the corner in square feet multiplied by the time
waiting during the red signal phase.
of the analysis period. The analytical problem is allocation of this
time-space in ways that provide a reasonable corner LOS for both
Precise analysis of pedestrian activity at corners is difficult waiting and moving pedestrians.
because of the many combinations of movements that are pos- The method assumes that standing pedestrians waiting for the
sible, as is illustrated in Figure 13-12. Each of the four direc- signal to change form a ‘‘competitive queue,’’ in which each
tional movements into the corner may proceed straight ahead, pedestrian occupies 5 sq ft/ped. This assumes midrange LOS D
or may turn left or right. This makes accurate collection of field conditions within the queue, typical of many urban situations, and
data at busy intersections an almost impossible task. Methods simplifies computational procedures. The average time moving
for determining approximate LOS of street corners using more pedestrians occupy the corner, typically in the range of 3 to 5
typically available crossing count data are given in the ‘‘Pro- sec, is also assumed. This assumption of the travel time along
cedures for Application’’ section of this chapter. the path of the longest dimension of a corner is actually
The methodology is relatively straightforward and is adequate conservative, as many pedestrians ‘‘short cut’’ corner edges,
to establish problem locations which may require more detailed reducing their time-space requirements.
Figure 13-12. Pedestrian movements at a street corner.

13-14 urban streets
Crosswalks in the demand for the space. Division of demand into the available
time-space produces the space per moving pedestrian available
Pedestrian flow characteristics in crosswalks are similar to those during the green phase. This area can be compared with LOS
on sidewalks, with the basic relationships of speed, density, space, criteria.
and flow consistent with observed values for uninterrupted flow However, there is a brief maximum flow or surge condition
during the WALK phase which must be examined. This occurs
on walkways. However, traffic signals control movement on the
when the two lead platoons from opposite corners, formed dur-
crosswalk, collecting pedestrians into denser platoons, and altering
ing the waiting phase, are simultaneously in the crosswalk.
the normal distribution of walking speeds. Average walking speed
Excessive pedestrian flows during this surge could cause pedestri-
in crosswalks is frequently taken to be 4.5 ft/sec.
ans to drift out of the marked crosswalk area, potentially endan-
Level-of-service concepts developed primarily for movement of gering them.
pedestrians on walkways can be applied to crosswalk analysis, Neither the average nor the maximum estimate of crosswalk
but signal timing and the effects of turning vehicles during the LOS accounts for the effects of turning vehicles during the
pedestrian green phase can alter the underlying assumptions of the pedestrian crossing phase. Rough estimates of pedestrian LOS
LOS analysis. Where crosswalk analyses show low pedestrian degradation by turning vehicles can be made by assuming a
LOS, vehicle-turning restrictions must be seriously considered. vehicle swept path area and time in the crosswalk (time-space)
Like corners, the crosswalk can also be analyzed as a time- decrement for each turning vehicle. An example of this is shown
space zone. The available time-space is the product of the WALK in the ‘‘Procedures for Application’’ section of this chapter. It
phase time less a platoon start-up time, assumed to be 3 sec should be noted, however, that the nature of pedestrian-vehicle
herein, and the area of the crosswalk in square feet. The product interactions in the crosswalk may be greatly influenced by local
of pedestrian crossing flow and the average crossing time results right-of-way practices.
III. PROCEDURES FOR APPLICATION AND SAMPLE CALCULATIONS
In this chapter procedures for application and sample cal- T Peak 15-min pedestrian count, VP15, in peds/15 min.
culations are presented as a cohesive unit. Since procedures for T Total walkway width, WT, in ft
analysis of walkways, street corners, and crosswalks are all rela-
tively unique, illustrative calculations are shown with each proce- T Identification of obstacles in the walkway
dural presentation. 2. The effective width of the walkway, WE, must be determined
by subtracting any unusable width from the total walkway width,
WT. Table 13-2 and Figure 13-5 can be used to estimate the unus-
ANALYSIS PROCEDURES FOR WALKWAYS able portion of walkway width.
3. The pedestrian unit flow rate, in ped/min/ft, is computed
Computations for walkways are based on peak 15-min pe- as:
destrian counts. A midblock walkway should be counted for
v = Vp15/15WE
several different time periods during the day to establish vari-
ances in directional flows. For new locations or to analyze future 4. The rate of flow within platoons may be estimated as:
conditions, forecasts of the flows must be made. Methods of
forecasting pedestrian trip volumes and pedestrian trip gener- vp = v + 4
ation rates for various types of land uses are contained in
Ref. 8. 5. Levels of service for average or platoon conditions are
found by comparing these flow rates to the criteria of Table
13-3.
Computational Steps
The methodology requires a specific sequence of computations Sample Calculation

which is presented below. Figure 13-13 is a worksheet which may
be used in summarizing these computations. 1. Description—A given sidewalk segment on Third Street
has a peak 15-min pedestrian flow of 1,250 ped/15 min. The
1. Preliminary data needed to conduct an analysis include 14-ft sidewalk has a curb on one side and stores with window
the following. For existing cases, field studies would be made shopping displays on the other. There are no other sidewalk
to collect the information; for future cases, forecasts of demand obstructions. At what LOS does the sidewalk operate, on the
and probable designs would be assumed: average and within platoons?
pedestrians 13-15
Figure 13-13. Worksheet for walkway analysis.
2. Solution—The total sidewalk width of 14 ft must be re- The average unit width flow rate is computed as:
duced to account for unused ‘‘buffer’’ areas at the curb and
building line. From Figure 13-5 the curb buffer is 1.5 ft, and v = Vp15/15WE
the building buffer (with window shopping assumed) is 3.0 ft.
Thus, the effective walkway width is 14.0 − 1.5 − 3.0 = 9.5 v = 1,250/(15 × 9.5) = 8.8 ped/min/ft
ft, and it is this figure that is used to determine the average
and platoon flow rates. The rate of flow within platoons may then be estimated as:
13-16 urban streets
vp = v + 4 ANALYSIS PROCEDURES FOR STREET CORNERS

v = 8.8 + 4 = 12.8 ped/min/ft AND CROSSWALKS
Table 13-3 is entered with these flow values to estimate the As noted previously, the analysis of street corners requires
level of service. The LOS for average conditions is C, while the consideration of the amount of circulation area available for
LOS within platoons is estimated to be D. pedestrians moving through the corner, and the amount of hold-
These computations can be summarized on the walkway analysis ing area required for standing pedestrians waiting to cross the
worksheet, as illustrated in Figure 13-14. street. Figure 13-15 illustrates the geometrics of a typical street
Figure 13-14. Illustration of solution to walkway problem.

pedestrians 13-17
Figure 13-15. Intersection corner geometrics and pedestrian movements.
corner, and also the directional flow variables which will be The analysis of street corners and crosswalks is based on a
used in subsequent LOS analyses. comparison of available time and space to pedestrian demand. The
Figures 13-16 and 13-17 show the two signal phase conditions product of time and space, i.e., time-space, is the critical parameter
which are analyzed in both corner and crosswalk computations. for consideration, because physical design limits available space
Condition 1 is the minor street crossing phase during the major and signalization controls available time.
street green, with pedestrians held in a queue on the major street In order to simplify the presentation and application of the time-
side during the minor street red phase. Condition 2 is the major space analysis approach, the development of relationships (equa-
street crossing phase, with pedestrians crossing during the minor tions) is presented in parallel with the solution of a sample calcula-
street green, and held in a queue on the minor street side by tion. Worksheets are illustrated in Figures 13-18 and 13-19 for
the major street red phase. crosswalk and street corner calculations respectively.
When making street corner computations, it is advisable to refer The sample calculation illustrated in the analysis of street cor-
to Figures 13-15, 13-16, and 13-17 for graphic illustrations of the ners and crosswalks is as follows:
various parameters used.
The point of maximum pedestrian queuing and minimum 1. Description—The sidewalks at a major and minor street in-
available circulation space on the corner occurs just before tersection are each 16 ft wide, with a corner radius of 20 ft. The
signal phase change. At this time, there is an average flow of roadway width for the major street is 46 ft; and for the minor
outbound pedestrians leaving the corner, a more concentrated street, 28 ft. The signal cycle length, C, is 80 sec with a two-phase
platoon of inbound pedestrians approaching from the opposite split of 48 sec of green plus amber, Gmj, for the major street (60
side of the street, and an average flow joining the pedestrian percent) and 32 sec of green plus amber, Gmi, for the minor street
queue waiting to cross at the signal change. At this same (40 percent). The 15-min peak period pedestrian crossing and side-
time, there are also pedestrians moving between the intersecting walk counts are shown below. Refer to Figure 13-15 for a graphic
sidewalks, not crossing the street. definition of flows.
13-18 urban streets
Peak 15-Min Average Flow Average Flow T The average LOS for pedestrians crossing the minor and
Pedestrian Rate Per Cycle major streets.
Flow Count (ped/min) (ped/cycle) T The decrement in average crosswalk pedestrian LOS due
vci 540 36 48 to five turning vehicles per cycle on the major street cross-
vco 300 20 27 ing.
vdi 450 30 40 Procedures for analysis of street corners and crosswalks are
vdo 240 16 21 presented in a step-by-step fashion, along with the solution of the
va,b 225 15 20 sample calculation.
Totals 1,755 117 156
Street Corner Analysis (Computational Steps and
Sample Calculation)
Note that flow rates in pedestrians per minute are rounded to
the nearest integer. Pedestrians per cycle are computed by mul-
tiplying pedestrians per minute by the signal cycle length (in
seconds) divided by 60 sec. For this calculation, the multiplier Step 1—Determine Total Available Time-Space
is 80 sec/60 sec = 1.33. Pedestrians per cycle are also rounded to
the nearest integer. The total time-space available in the intersection corner for
2. Find— circulation and queuing, for an analysis period of t minutes, is
the product of the net corner area, A, and the time t. For street
T The average LOS for pedestrian circulation at the street corner and crosswalk analysis, t is taken to be one signal cycle
corner during a typical peak-period signal cycle. and is, therefore, equal to the cycle length, C. The net corner
Figure 13-16. Intersection corner Condition 1—minor street crossing.

pedestrians 13-19
Figure 13-17. Intersection corner Condition 2—major street crossing.
area is found by multiplying the intersecting sidewalk widths, average pedestrian holding times, Qtco and Qtdo, of persons waiting
Wa and Wb, and deducting the area lost due to the corner radius to use crosswalks C and D, respectively, are 1/2 the product of
and any obstructions. Then, assuming there are no obstructions the outbound flows during a signal cycle (vco and vdo, in ped/cycle),
in the corner area: the proportion of cycle that these flows are held up, and their
holding time based on the red signal phase:
A = WaWb − 0.215 R2 (13-4)
For Condition 1, the minor street crossing, which occurs during
TS = A × C/60 (13-5) the major street WALK or green phase:
where: Qtdo = [vdo × (Rmi /C) × (Rmi /2)]/60 (13-6)
A = area of the street corner, in sq ft; For Condition 2, the major street crossing, which occurs during
Wa = width of the sidewalk a, in ft; the minor street WALK or green phase:
Wb = width of sidewalk b, in ft; Qtco = [vco × (Rmj /C) × (Rmj /2)]/60 (13-7)
R = radius of corner curb, in ft;
C = cycle length, in sec; and where:
TS = total time-space available, in sq ft-min.
Qtdo = total time spent by pedestrians waiting to cross the major
street during one signal cycle, in ped-min;
For the sample calculation described earlier, the following val- Qtco = total time spent by pedestrians waiting to cross the minor
ues may be computed: street during one signal cycle, in ped-min;
vdo = the number of pedestrians per cycle crossing the minor
A = (16 × 16) − 0.215 (202) = 170 sq ft
street, in ped/cycle;
TS = 170 × 80/60 = 227 sq ft-min
vco = the number of pedestrians per cycle crossing the major
Step 2—Compute Holding Area Waiting Times street;
Rmi = the minor street red phase, or the DON’T WALK phase where
If uniform arrivals are assumed at the crossing queues, the pedestrian signals exist, in sec;
13-20 urban streets
Rmj = the major street red phase, or the DON’T WALK phase where duration of R/2 sec. The division by 60 converts time from
pedestrian signals exist, in sec; and seconds to minutes.
C = cycle length, in sec. For the sample calculations, the following values are com-
puted:
The term R/C is used to estimate the number of pedestrians

per cycle that must wait for the green indication. The number Qtco = [27 × 0.40 × 32/2]/60 = 2.9 ped-min
is estimated as v × R/C. Assuming that arrivals are uniformly
distributed, each pedestrian that waits does so for an average Qtdo = [21 × 0.60 × 48/2]/60 = 5.0 ped-min
Figure 13-18. Worksheet for crosswalk analysis.

pedestrians 13-21
Step 3—Determine Holding Area Time-Space where TSh equals the total time-space holding area requirements
Requirements for the intersection, in sq ft-min.
For the sample calculation, the following value is determined:
The holding area needs of waiting pedestrians are the product
of the total waiting times determined in Step 2 (Qtdo and Qtco) TSh = 5 (5.0 + 2.9) = 39.5, say 40 sq ft-min
and the average area used by a waiting pedestrian, which is
taken to be 5 sq ft/ped for a competitive queue. Then:
TSh = 5 (Qtdo + Qtco) (13-7)
Figure 13-19. Worksheet for street corner analysis.

13-22 urban streets
Step 4—Determine the Net Corner Time-Space Step 8—Determine the Corner Level of Service
Available for Circulation
The corner LOS is found by comparing the pedestrian area
The total time-space available for circulation is the total module, M, to the criteria found in Table 13-3. Values below
intersection time-space minus that used for holding waiting LOS C indicate a potential problem that should be the subject
pedestrians; or: of further field study and possible remedial actions, which could
include changes in the signal timing, prohibition of vehicle-
TSc = TS − TSh (13-8) turning movements, sidewalk widening, and removal of sidewalk
obstructions.
where TSc equals the total time-space available for circulating From Table 13-3, for a pedestrian area module of 18.0 sq ft/
pedestrians, in sq ft-min. ped, the LOS for the sample calculation is found to be D. The
need for further field study and possible remedial action is
For the sample calculation: indicated.
Figure 13-20 illustrates the solution of the sample calculation
TSc = 227 − 40 = 187 sq ft-min on the street corner worksheet.
Crosswalk Analysis (Computational Steps and Sample

Calculation)
Step 5—Determine the Total Number of
Circulating Pedestrians Per Cycle
Analysis procedures for crosswalks use the same basic principles
of accounting for time-space. The procedure is explained in the
The number of pedestrians which must use the available circula-
following steps.
tion time-space during each cycle is the sum of all pedestrian
flows; each flow is expressed in units of ped/cycle:
vc = vci + vco + vdi + vdo + va,b (13-9)

Step 1—Determine the Total Available Time-Space
where vc equals total number of circulating pedestrians, in ped/
The total time-space available in the crosswalk during one signal
cycle.
cycle is the product of the crosswalk area and the WALK interval
For the sample calculation: for the crosswalk. Where pedestrian signals are not present, the
green time minus 3 sec is substituted for WALK time. Note that in
vc = 48 + 27 + 40 + 21 + 20 = 156 ped computing crosswalk area, the effect of the corner radius is not
considered. Then:
Step 6—Determine the Total Circulation Time
Utilized by Circulating Pedestrians Aw = W × L (13-12)
TSw = Aw × Gw/60 (13-13)
The time that pedestrians consume while walking through the
corner area is taken as the product of the total circulation volume where:
and an assumed average circulation time of 4 sec, or:
Aw = area of the crosswalk, in sq ft;
tc = vc × 4/60 (13-10) W = width of the crosswalk, in ft;
where tc equals the total circulation time, in ped-min. L = length of the crosswalk, in ft;
For the sample calculation: TSw = total time-space available in the crosswalk during one
signal cycle, in sq ft-min; and
tc = 156 × 4/60 = 10.4 ped-min Gw = WALK interval, in sec.
Then, for Crosswalk C in the illustrative calculation:
A = 16 × 28 = 448 sq ft
Step 7—Determine the Circulation Area Per
Pedestrian TSw = 448 × (48 − 3)/60 = 336 sq ft-min
The circulation area per pedestrian is referred to as the ‘‘pedes- and for Crosswalk D:
trian area module,’’ and given the symbol, M. It is computed as A = 16 × 46 = 736 sq ft
the net time-space available for circulation, TSc, divided by the
total circulation time, tc: TSw = 736 × (32 − 3)/60 = 356 sq ft-min
M = TSc /tc (13-11)
For the sample calculation: Step 2—Determine the Average Crossing Times
M = 187/10.4 = 18.0 sq ft/ped The average time a pedestrian occupies each crosswalk is
pedestrians 13-23
Figure 13-20. Worksheet for street corner analysis of sample calculation.

13-24 urban streets
obtained by dividing the length of the crosswalk (street width) by For Crosswalk C:
the assumed walking speed. Average walking speed in crosswalks
is taken to be 4.5 ft/sec. Then: M = 336/7.8 = 43 sq ft/ped (LOS B, Table 13-3)
tw = L/4.5 (13-14)
and for Crosswalk D:
where:
M = 356/10.4 = 34 sq ft/ped (LOS C, Table 13-3)
tw = average time spent by pedestrian in the crosswalk, in
sec; and
L = length of the crosswalk, in ft.
Step 5—Determine the Level of Service for the
Then, for the sample calculation Crosswalk C: Maximum Surge Condition
tw = 28/4.5 = 6.2 sec Step 4 yields an analysis of conditions that are average for
the WALK interval. The point at which the maximum number
and for Crosswalk D: of pedestrians are in the crosswalk should also be examined.
This occurs when the lead pedestrians in opposing crossing
tw = 46/4.5 = 10.2 sec platoons reach the opposite corner. The area module for the
surge condition is the area of the crosswalk divided by the
maximum number of pedestrians in the crosswalk. Crosswalk
flows in pedestrians per minute (NOT the ped/cycle units which
Step 3—Determine the Total Crosswalk Occupancy have been used for other analysis steps) are multiplied by the
Time DON’T WALK interval + the crossing time, tw. The DON’T WALK
interval is used to estimate the number of pedestrians queued
The total crosswalk occupancy time is the product of the when the WALK interval is given, and the crossing time is added
average crossing time and the number of pedestrians using the to estimate the number of new arriving pedestrians during the
crosswalk during one signal cycle. Then: period that the queued pedestrians cross the street. Where pedes-
trian signals are not present, the red interval + 3 sec is substituted
Tw = (vi + vo) tw /60 (13-15)
for the DON’T WALK interval. Then:
where: Vm = (vi + vo) (Rw + tw)/60 (13-17)
M = A/Vm (13-18)
Tw = total crosswalk occupancy time, in ped-min;
where:
vi = incoming pedestrian volume for the subject crosswalk, in
ped/cycle; and
Vm = maximum number of pedestrians occupying crosswalk;
vo = outgoing pedestrian volume for the subject crosswalk, in
vi = incoming crosswalk volume, in ped/min;
ped/cycle.
vo = outgoing crosswalk volume, in ped/min; and
Rw = DON’T WALK interval, in sec.
For the sample calculation, Crosswalk C:
For the sample calculation, Crosswalk C:
Tw = (48 + 27) 6.2/60 = 7.8 ped-min
Vm = (36 + 20) × (32 + 3 + 6.2)/60 = 38.5 ped
M = 448/38.5 = 11.6 sq ft/ped (LOS E, Table 13-3)
Tw = (40 + 21) 10.2/60 = 10.4 ped-min
Vm = (30 + 16) × (48 + 3 + 10.2)/60 = 46.9, say 47 ped

Step 4—Determine the Average Circulation Space per Pedestrian
and the Average Level of Service M = 736/47 = 15.7 sq ft/ped (LOS D, Table 13-3)
The average circulation space provided for each pedestrian is
determined by dividing the time-space available for crossing by Note that the surge LOS is worse than the average LOS, particu-
the total occupancy time. This yields the average area module larly for Crosswalk C, where the value fell from B for average
provided for each pedestrian, which is related to level of service conditions to E for surge conditions. This emphasizes the need to
by the criteria of Table 13-3. consider both conditions.
Figure 13-21 shows the worksheet for the sample calculation
M = TSw /Tw (13-16) discussed herein.
pedestrians 13-25
Figure 13-21. Worksheet for crosswalk analysis of sample calculation.

13-26 urban streets
Estimating the Decrement to Crosswalk LOS Due decrement to available time-space would be: 10.7 × 5 = 54 sq
to Right-Turning Vehicles ft/min.
For the major street crossing (Crosswalk D), the total available
The time-space method allows for an approximate estimate time-space was computed to be 356 sq ft-min. Deducting 54 sq
to be made of the effect of turning vehicles on the average ft-min, only 302 sq ft-min remain for pedestrian use. The pedes-
LOS for pedestrians crossing during a given green phase. This trian space module is now recomputed using this figure in Eq.
is done by assuming an average area occupancy of a vehicle 13-16:
in the crosswalk, based on the product of vehicle swept-path
and crosswalk widths, and an estimate of the time that the M = 302/10.4 = 29 sq ft/ped (LOS C, Table 13-3)
vehicle preempts this space. The swept-path for most vehicles In this case, the decrement has not caused a reduction in the
may be estimated at an average of 8 ft, and it is assumed that LOS, although the area per pedestrian is clearly reduced. This
a vehicle occupies the crosswalk for a period of 5 sec. is an indication that the crosswalk can handle both the pedestrian
For the sample calculation, each turning vehicle will preempt: demands and the turning vehicle demands without experiencing
a capacity or delay problem. Where the decrement causes a
[8 ft × 16 ft (crosswalk width) × 5 sec]/60 = 10.7 sq ft-min/veh
significant decline in LOS, particularly where LOS F would
If 5 vehicles turn during an average green phase, the total result, further field studies and remedial action should be pursued.
IV. REFERENCES
The basic pedestrian characteristics used in this chapter were Draft Pedestrian Planning Procedures Manual, Vols. I–III,
presented in Transportation Research Board Circular 212 (1). Pion- RTKL Associates, Baltimore, Md. (1977).
eering references of great interest were authored by Pushkarev and 5. Hall, D., The Hidden Dimension. Doubleday and Co., New
Zupan (2) and Fruin (3). References 4 through 8 offer additional York, N.Y. (1966).
information for interested users of this manual. 6. Virkler, M., and Guell, D., ‘‘Pedestrian Crossing Time Re-
quirements at Intersections.’’ Transportation Research Record
1. ‘‘Interim Materials on Highway Capacity.’’ Transportation Re- 959, Transportation Research Board, Washington, D.C. (1984)
search Board Circular 212, Transportation Research Board, pp. 47–51.
Washington, D.C. (1980). 7. Fruin, J., and Benz, G., ‘‘Pedestrian Time-Space Concept for
2. Pushkarev, B., and Zupan, J., Urban Space for Pedestrians. Analyzing Corners and Crosswalks.’’ Transportation Research
MIT Press, Cambridge, Mass. (1975). Record 959, Transportation Research Board, Washington, D.C.
3. Fruin, J., Pedestrian Planning and Design. Metropolitan Asso- (1984) pp. 18–24.
ciation of Urban Designers and Environmental Planners, New 8. Kagan, L., et al., ‘‘A Pedestrian Planning Procedures Man-
York, N.Y. (1971). ual.’’ Vols. I–II, FHWA Report Nos. RD-78-45, RD-78-46, RD-
4. Feasibility Analysis and Design Concepts and Criteria for 79-47, Federal Highway Administration, Washington, D.C.
Communitywide Separated Pedestrian Networks. Phase III, (Nov. 1978).
APPENDIX I
WORKSHEETS FOR USE IN ANALYSIS OF WALKWAYS, CROSSWALKS, AND
STREET CORNERS
WORKSHEETS
Walkway Analysis Worksheet.............................................................................................................................................................. 13-27

Crosswalk Analysis Worksheet ............................................................................................................................................................ 13-28
Street Corner Analysis Worksheet ....................................................................................................................................................... 13-29
pedestrians 13-27
13-28 urban streets
pedestrians 13-29
chapter 14
BICYCLES
CONTENTS
i. introduction .......................................................................................................................................................................... 14-1

ii. methodology and procedures for application .............................................................................................................. 14-2
Impacts on Intersection Capacity.......................................................................................................................................... 14-2
Passenger-Car Equivalents for Bicycles ........................................................................................................................... 14-2
Effect of Bicycles on Right-Turning Vehicles................................................................................................................. 14-2
Left-Turning Bicycles from Bike Lanes........................................................................................................................... 14-3
Effects of Bicycles on Roadway Segments Between Intersections..................................................................................... 14-3
Bicycle Facilities ................................................................................................................................................................... 14-3
iii. sample calculations ............................................................................................................................................................ 14-3
Calculation 1—Passenger-Car Equivalents .......................................................................................................................... 14-3
Calculation 2—Left-Turn Impacts on a Multilane Approach.............................................................................................. 14-4
Calculation 3—Impacts of a Bike Lane on Right-Turning Vehicles .................................................................................. 14-4
iv. references .............................................................................................................................................................................. 14-4
I. INTRODUCTION
A bicycle is defined as a vehicle having two tandem wheels, While the state of knowledge concerning specific impacts of
propelled solely by human power, upon which any person or per- bicycles on the capacity and level of service of highway facilities
sons may ride. is not advanced, this chapter presents some insights and procedures
Bicycles make up a small percentage of the traffic stream at for approximately analyzing the effects of bicycles in the traffic
most locations in North America. Nevertheless, there are many stream. It also presents approximate information on the capacity
locations where the impact of bicycles on the vehicular traffic of various types of bicycle facilities. Specifically, this chapter
stream is noticeable. Many cities have initiated extensive programs addresses the following aspects of bicycle capacity:
to provide facilities for bicycles in the form of designated bicycle
1. The impacts of bicycle presence on intersection capacity.
lanes on streets and highways and bikeways with physically sepa-
2. The impacts of bicycle presence on roadway segments be-
rated rights-of-way. The use of bicycles as a regular means of
tween intersections.
personal transportation has increased, particularly in warm cli-
3. The capacity of designated bicycle facilities.
mates. The bicycle is a popular mode in and around many univer-
sity campuses, and is an attractive alternative in congested city The sections that follow detail these types of analyses, and
areas where vehicular traffic is difficult. illustrate their use with sample calculations.
14-1
14-2 urban streets
II. METHODOLOGY AND PROCEDURES FOR APPLICATION
IMPACTS ON INTERSECTION CAPACITY Table 14-1. Passenger-Car Equivalent for Bicycles
bicycle lane width (ft)

Bicycles affect the capacity and operating conditions at intersec- movement < 11 11-14 > 14
tions in two principal ways:
Opposed 1.2 0.5 0.0
1. Where bicycles share a lane with other vehicles, they utilize Unopposed 1.0 0.2 0.0
a portion of the lane’s capacity. This effect is accounted for by
assigning an appropriate ‘‘passenger-car equivalent’’ (pce) for
each bicycle.
2. Where vehicles execute turning movements through a con-
flicting bicycle stream, they encounter opposition in addition to
Effect of Bicycles on Right-Turning Vehicles
that normally presented by opposing vehicle streams and pedestri-
ans. The intersection analysis techniques of Chapters 9 and 10
should be modified to account for this conflict. At intersections where a curb bicycle lane is provided, right-
turning vehicles encounter not only a conflicting pedestrian flow,
but a conflicting bicycle flow as well. Figure 14-1 illustrates
these conflicts.
Passenger-Car Equivalents for Bicycles
Table 14-1 presents the recommended values of passenger-car

equivalents for bicycles. The equivalent varies with lane width
and depends on whether the bicycle movement in question is ‘‘op-
posed’’ or ‘‘unopposed.’’
A bicycle moving straight through an intersection, encountering
no significant interference from vehicles or pedestrians, is consid-
ered to be unopposed. A left-turning bicycle must cross an oppos-
ing vehicular flow on two-way streets, and would be considered
to be opposed. Right-turning bicycles may or may not encounter
significant pedestrian interference, and could be classified as either
opposed or unopposed. Where the conflicting crosswalk flow ex-
ceeds 100 peds/hour, it is recommended that right-turning bicycles
be considered opposed.
As indicated in Table 14-1, the impact of bicycles sharing
vehicular lanes increases as lane width decreases. When lane
widths are 14 ft or greater, bicycles tend to use a portion of the
lane as a bike lane, and have little impact on vehicular flow. It
should also be noted that these factors are conservative, as they
assume that most bicyclists move through the intersection on the
green signal.
Table 14-1 is used as follows. The number of bicycles (segre-
gated by type of movement) is multiplied by the appropriate
passenger-car equivalent values. The result is added to the
vehicular volume, yielding a total equivalent vehicular volume
which is used in subsequent computations. Consider a signalized
intersection with a vehicular volume of 500 vph which shares
a 10-ft lane with a bicycle volume of 100 bicycles/hour, one- Figure 14-1. Illustration of right-turn conflicts with bicycles and
half of which are opposed. pedestrians.
Then:
Equivalent volume = 500 + 100(0.5)(1.2)

+ 100(0.5)(1.0) Where such conflicts exist, right-turning vehicles experience
= 500 + 60 + 50 = 610 vph considerably more friction than in situations where no bike lane
exists. Table 9-11 in Chapter 9, ‘‘Signalized Intersections,’’ gives
where 1.2 and 1.0 are the passenger-car equivalent values for adjustment factors used in correcting for the impact of pedes-
opposed and unopposed bicycle movements selected from Table trian interference on right-turn saturation flow. Where a bicycle
14-1. Further computations would proceed using a volume of lane exists, it is recommended that this table be entered with
610 vph in the procedures of Chapter 9, ‘‘Signalized Intersec- total number of pedestrians plus bicycles which interfere with the
tions.’’ subject right-turn movement. Thus, if a right-turn movement
bicycles 14-3
must cross a pedestrian flow of 100 peds/hour and a bicycle ignated by striping, signing, and pavement markings for the prefer-
flow of 150 bicycles/hour, Table 9-11 would be entered as if ential or exclusive use of bicyclists.
the conflicting pedestrian flow were 100 + 150 = 250 peds/ 2. Bike path—A bikeway physically separated from motor-
hour. ized vehicular traffic by an open space or barrier, either within
Where bicycles share a vehicular lane, it is not necessary to the highway right-of-way or within an independent right-of-
include this adjustment because the approach volume is already way.
inflated to account for bicycle presence. Where shared-lane width
is 14 ft or greater, however, it was assumed that bicycles separate There is not a great deal of information available concerning
into the right portion of the lane, using it essentially as a bike the capacity of such facilities. Planning and design criteria for
lane. In such cases, their impact on right-turning vehicles should bicycle facilities are available from a number of sources (2–5),
be considered as indicated in this section. including the Transportation and Traffic Engineering Handbook
(6). A summary of available data was compiled from Ref. 2, and
is presented in Table 14-2.
Left-Turning Bicycles from Bike Lanes Reference 3 cites the capacity of a bicycle facility as 0.22 bicy-
cles per second per foot of bikeway. This is equivalent to 2,376
Bicycles turning left out of a bike lane must mix with other bicycles/hour for a 3-ft bikeway.
vehicles as they approach the intersection and execute the left-turn
maneuver. An appropriate passenger-car equivalent value
is selected from Table 14-1 and added to the vehicular volume in
the leftmost lane. The passenger-car equivalent value for bicycles
is also added to the volume in each lane the bicycles must cross Table 14-2. Reported One-Way and Two-Way High
Volumes of Bicycle Facilities
in transferring from the right-hand bike lane to the leftmost traf-
fic lane. range of reported
capacities
type of facility no. of lanesa (bicycles/hour)
EFFECTS OF BICYCLES ON ROADWAY SEGMENTS
BETWEEN INTERSECTIONS One Way Bike Lane 1 1,700–2,530
or Path
There is little existing data or information on the impacts of Two Way Bike Path 1 850–1,000
bicycles on capacity or operating conditions between intersec- 2 500–2,000
tions. Bicycles are not expected to have any impact on flow a
Lane widths 3–4 ft/lane
where curb-lane widths exceed 14 ft. Where bicycle volumes SOURCE: Adapted from Refs. 2 and 6
are less than 50/hour, impacts are also believed to be negligible,
except where lanes are narrow (≤11 ft).
One study (1) has indicated that vehicular intersection ap-
proach speeds are reduced by approximately 2.5 mph when It should be noted that the wide variation of reported high
bicycles are present in an adjacent bike lane. volumes reflects a similarly wide range in environmental condi-
tions, skill and familiarity of cyclists, and specific geometric
features of the facilities reported. Bikeway capacity is also rarely
BICYCLE FACILITIES observed in practice, as demand levels are generally well below
the capacity of the facility. Indeed, the planning and design
Bicycle facilities separated from vehicular traffic can be pro- documents referenced previously all emphasize the need to have
vided in two basic forms: bicycle facilities that provide sufficient capacity to allow good-
to-excellent operating conditions if they are to be successful in
1. Bike lane—A portion of a roadway which has been des- encouraging bike use.
III. SAMPLE CALCULATIONS
CALCULATION 1—PASSENGER-CAR a flow of 110 peds/hour. Convert the approach volume to an equiv-
EQUIVALENTS
alent which accounts for the effects of bicycles.
1. Description—An intersection approach with one 12-ft lane 2. Solution—Both left-turning and right-turning bicycles are
has a vehicular demand of 500 vph. It is shared by 50 bicycles/ considered to be ‘‘opposed.’’ From Table 14-1, the following pas-
hour, 10 of which turn left and 15 of which turn right across senger-car equivalent values are found:
14-4 urban streets
1 Through bicycle = 0.2 pce proach lanes. Thus, any additional analysis would proceed using
1 Left-turning bicycle = 0.5 pce the following adjusted approach volumes:
1 Right-turning bicycle = 0.5 pce
Left lane: 250 + 60 = 310 vph
The total equivalent demand volume on the intersection ap- Center lane: 350 + 60 = 410 vph
proach may then be expressed as: Right lane: 220 + 60 = 280 vph
Equivalent volume = 500 + 25(0.2) + 10(0.5) + 15(0.5) Note that the equivalents are added to each lane that is crossed
= 500 + 5 + 5 + 7.5 by bicycles transferring from the bike lane to the leftmost
= 517.5, SAY 518 vph traffic lane.
Note that this is not the final conclusion of the analysis of

the intersection in question. If the intersection were signalized, CALCULATION 3—IMPACTS OF A BIKE LANE ON
RIGHT-TURNING VEHICLES
the analysis would proceed using the procedures of Chapter 9,
but with a demand volume for 518 vph on the subject approach 1. Description—A single-lane approach at a signalized intersec-
rather than 500 vph, which is the actual vehicular demand tion is adjacent to a curb bike lane carrying 400 bicycles/hour.
volume. If the intersection were unsignalized, the procedures of What right-turn adjustment factor would be selected if right-turn-
Chapter 10 would be applied to complete the analysis. ing vehicles also interfere with a pedestrian flow of 200 pedestri-
ans/hour? Right turns make up 20 percent of the total volume in
the single lane.
CALCULATION 2—LEFT-TURN IMPACTS ON A
2. Solution—Right-turn adjustment factors for right turns at
MULTILANE APPROACH
signalized intersections are selected from Table 9-11 (Ch. 9).
1. Description—An intersection approach has three traffic Single-lane approaches are represented by Case 7 in that table.
lanes and a right-hand curb bicycle lane. The three lanes have A factor would normally be selected for 200 pedestrians/hour
the following approach volumes: left lane, 250 vph; center lane, and 20 percent right turns, yielding an adjustment factor of
350 vph; right lane, 220 vph. There are 50 bicycles/hour exe- 0.86, which is applied to the saturation flow rate for the ap-
cuting left turns. How should the vehicular volumes be adjusted proach.
to reflect the impact of these bicycles? Traffic lanes are 11 ft Where a bicycle lane is present, however, the factor is selected
wide. as if the pedestrian volume were the total of pedestrians and bicy-
2. Solution—From Table 14-1, each bicycle has an equivalent cles. Thus, a factor is selected for 200 + 400 = 600 pedestrians
of 1.2 (opposed, 11-ft lanes). Thus, the 50 left-turning bicycles/ and 20 percent right turns. This factor would be 0.82. Thus, the
hour are equivalent to 50 × 1.2 = 60 vph. These passenger-car presence of the bicycle reduces the capacity of the single-lane
equivalents should now be added to the volume in all three ap- approach by 0.86 − 0.82 = 0.04, or 4 percent.
IV. REFERENCES
1. Opiela, K., Khasulis, S., and Datta, T., ‘‘Determination of 4. ‘‘Pedestrian and Bicycle Accommodation and Projects.’’ Fed-
the Characteristics of Bicycle Traffic at Urban Intersections.’’ eral Register, Vol. 49, No. 57, Federal Highway Administra-
Transportation Research Record 743, Transportation Research tion, Washington, D.C. (Mar. 22, 1984).
Board (1980). 5. Guide to Development of New Bicycle Facilities, American
2. Bikeway Planning Criteria and Guidelines. Institute of Traffic Association of State Highway and Transportation Officials,
and Transportation Engineering, University of California at Los Washington, D.C. (1981).
Angeles (1972). 6. King, C., and Harkens, W., ‘‘Geometric Design.’’ Trans-
3. Safety and Locational Criteria for Bicycle Facilities. Users portation and Traffic Engineering Handbook, Institute of
Manual, Vol. 2, Federal Highway Administration, Washington, Transportation Engineers, Prentice-Hall, Englewood Cliffs,
D.C. (1976). N.J. (1976).
APPENDIX A
GLOSSARY
Adjustment factor—A multiplicative factor that adjusts a capacity Average travel time—The average time spent by vehicles travers-
or service flow rate from one representing an ideal or base ing a highway segment of given length, including all stopped-
condition to one representing a prevailing condition. time delay, in seconds per vehicle or minutes per vehicle.
Alighting time—Time for a passenger to leave a transit vehicle, Balanced operation—An operating condition in a weaving area
expressed as the time per passenger or total time for all in which both weaving and nonweaving vehicles achieve the
passengers. same level of service.
Annual average daily traffic—The total volume passing a point Basic freeway section—A segment of freeway facility on which
or segment of a highway facility in both directions for one operations are unaffected by weaving, diverging, or merging
year divided by the number of days in the year. maneuvers.
Approach—A set of lanes accommodating all left-turn, through, Berth—A position for a bus to pick up and discharge passengers,
and right-turn movements arriving at an intersection from a including curb bus stops and other types of boarding and
given direction. discharge facilities.
Arterial—Signalized streets that serve primarily through traffic Bicycle—A vehicle having two wheels tandem, propelled solely
and provide access to abutting properties as a secondary func- by human power, upon which any person or persons
tion, having signal spacings of 2 mi or less and turn move- may ride.
ments at intersections that usually do not exceed 20 percent Bike lane—A portion of a roadway that has been designated by
of total traffic. striping, signing, and pavement markings for the preferential
Arterial class—A categorization of arterials involving functional or exclusive use of bicycles.
and design categories and free-flow speed. Bike path—A bikeway physically separated from motorized traf-
Arterial section—The aggregation of a sequence of consecutive
fic by an open space or barrier, either within the highway
arterial segments of comparable length and characteristics.
right-of-way or within an independent right-of-way.
Arterial segment—A one-way length of arterial from one signal
Bikeway—Any road, path, or way that in some manner is specifi-
to the next, including the downstream signalized intersection
cally designated as being open to bicycle travel, regardless
but not the upstream signalized intersection.
of whether such facilities are designated for the exclusive use
Average approach delay—Average stopped-time delay at a sig-
of bicyclists or are to be shared with other vehicles.
nalized intersection plus average time lost because of deceler-
Boarding time—The time for a passenger to board a transit vehi-
ation to and acceleration from a stop, generally estimated as
cle, expressed as time per passenger or total time for all
1.3 times the average stopped time delay.
passengers; a function of fare collection procedures.
Average running speed—The average speed of a traffic stream
Bus—A heavy vehicle involved in the transport of passengers on
computed as the length of a highway segment divided by the
a for-hire, charter, or franchised transit basis.
average running time of vehicles traversing the segment, in
miles per hour. Bus lane—A lane restricted to bus usage by special regulations
Average running time—The average time vehicles are in motion and markings.
while traversing a highway segment of given length, exclud- Busway—A right-of-way restricted to bus usage by physical sepa-
ing stopped-time delay, in seconds per vehicle or minutes per ration from other traffic lanes.
vehicle. Capacity—The maximum rate of flow at which persons or vehi-
Average stopped-time delay—The total time vehicles are stopped cles can be reasonably expected to traverse a point or uniform
in an intersection approach or lane group during a specified segment of a lane or roadway during a specified time period
time interval divided by the volume departing from the ap- under prevailing roadway, traffic, and control conditions, usu-
proach or lane group during the same time period, in seconds ally expressed as vehicles per hour or persons per hour.
per vehicle. Change interval—The ‘‘yellow’’ plus ‘‘all red’’ intervals that
Average total delay—The total additional travel time experienced occur between phases of a traffic signal to provide for clear-
by drivers, passengers, or pedestrians as a result of control ance of the intersection before conflicting movements are
measures and interaction with other users of the facility di- released.
vided by the volume departing from the corresponding cross Circulation area—The portion of a sidewalk street corner used
section of the facility. by moving pedestrians passing through the area, in
Average travel speed—The average speed of a traffic stream square feet.
computed as the length of a highway segment divided by the Clearance lost time—The portion of the time between signal
average travel time of vehicles traversing the segment, in phases during which an intersection is not used by any traffic
miles per hour. movement, in seconds.
A-1 Updated October 1994
A-2 appendix a
Clearance time—The minimum possible time interval between Design hour factor—Proportion of 24-hr volume occurring during
the departure of one bus from a bus berth and the entrance the design hour for a given location or area.
of another. Direction design hour volume—The traffic volume for the design
Collector street—Surface street providing land access and traffic hour in the peak direction of flow, usually a forecast of the
circulation within residential, commercial, and industrial relevant peak-hour volume, in vehicles per hour.
areas. Direct ramp—A ramp roadway on which vehicles turn only in
Composite grade—A series of adjacent grades along a highway the direction of their intended directional change, that is, a
having a cumulative effect on operations that is more severe ramp providing a left-turn connection that does not require
than if each grade were considered separately. vehicles to turn right or vice versa.
Conflicting approach—The approach at approximately 90 de- Diverge—A movement in which a single lane of traffic separates
grees to the subject approach at an all-way stop-controlled into two separate lanes without the aid of traffic control
(AWSC) intersection. devices.
Conflicting traffic volume—The volume of traffic that conflicts Downstream—The direction toward which traffic is flowing.
with a specific movement at an unsignalized intersection. Downtown street—Surface facilities primarily providing access
Constrained operation—An operating condition in a weaving to abutting lands in a downtown area.
area in which, because of geometric constraints, weaving ve- Dwell time—The time that a transit vehicle is stopped in a berth
hicles are unable to occupy as large a portion of available for the purposes of boarding or discharging passengers.
lanes as required to achieve balanced operation. Effective green time—The time allocated for a given traffic move-
Control conditions—Prevailing conditions concerning traffic con- ment (green plus yellow) at a signalized intersection less the
trols and regulations in effect for a given segment of street start-up and clearance lost times for the movement.
or highway, including the type, phasing, and timing of traffic Effective red time—The time during which a given traffic move-
signals; stop signs; lane use and turn controls; and similar ment or set of movements is directed to stop; cycle length
measures. minus effective green time.
Crawl speed—The maximum sustained speed that can be main- Effective walkway width—The width of a walkway that is usable
tained by a specified type of vehicles on a constant upgrade by pedestrians; the total walkway width minus the width of
of a given percent, in miles per hour. unusable buffer zones at the curb and building line and other
Critical density—The density at which capacity occurs for a given portions unusable because of obstacles and obstructions in
facility, usually expressed as vehicles per mile per lane. the walkway, in feet.
Critical gap—The minimum time interval between vehicles in a Flow ratio—The ratio of actual flow rate to the saturation flow
major traffic stream that permits side-street vehicle at a rate for a given lane group at a signalized intersection.
stop-controlled approach to enter the intersection under pre- Follow-up time—The time span between the departure of one
vailing traffic and roadway conditions, in seconds. vehicle from the minor street and the departure of the next
Critical speed—The speed at which capacity occurs for a given vehicle using the same gap under a condition of continuous
facility, usually expressed as miles per hour. queueing, in seconds.
Critical v/c ratio—The proportion of available intersection capac- Free-flow speed—(1) The theoretical speed of traffic when density
ity used by vehicles in critical lane groups. is zero, that is, when no vehicles are present; (2) the average
Crosswalk—The marked crossing area for pedestrians crossing speed of vehicles over an arterial segment not close to signal-
the street at an intersection or designated midblock location. ized intersections under conditions of low volume.
Crown line—A lane marking that directly connects the nose of Freeway—A multilane divided highway having a minimum of
the entry gore area to the nose of the exit gore area in a two lanes for exclusive use of traffic in each direction and
weaving section. full control of access and egress.
Crush capacity—The maximum number of passengers who can Freeway surveillance—A system in which freeway operations
physically be accommodated on a transit vehicle. are monitored and controlled in real time.
Cycle—Any complete sequence of signal indications. Fully actuated control—Signal control of an intersection in which
Cycle length—The total time for a signal to complete one cycle. the occurrence and length of every phase are controlled by
Delay—Additional travel time experienced by a driver, passenger, actuations of vehicle detectors placed on each approach to
or pedestrian beyond what would reasonably be desired for the intersection.
a given trip. Functional category—A type of arterial defined by the type of
Demand volume—The traffic volume expected to desire service traffic service provided.
past a point or segment of the highway system at some future Gore area—The area located immediately between the left edge
time, or the traffic currently arriving or desiring service past of a ramp pavement and the right edge of the roadway pave-
such a point, usually expressed as vehicles per hour. ment at a merge or diverge area.
Density—The number of vehicles occupying a given length of Green ratio—The ratio of the effective green time for a given
lane or roadway averaged over time, usually expressed as movement at a signalized intersection to the cycle length.
vehicles per mile or vehicles per mile per lane. Green time—The actual length of the green indication for a given
Design analysis—A usage of capacity analysis procedures to de- movement at a signalized intersection.
termine the size (number of lanes) required on a given seg- Headway—The time between two successive vehicles in a traffic
ment of a facility in order to provide a specified level of lane as they pass a point on the roadway, measured from
service. front bumper to front bumper, in seconds.
Design category—A type of arterial defined by geometric features Heavy vehicle—Any vehicle with more than four wheels touching
and roadside environment. the pavements during normal operation.
glossary and symbols A-3
High-occupancy-vehicle lane—A lane of a freeway reserved for Merge—A movement in which two separate lanes of traffic com-
the use of vehicles with more than a preset number of occu- bine to form a single lane without the aid of traffic signals
pants; such vehicles often include buses, taxis, and carpools. or other right-of-way controls.
Ideal conditions—Characteristics for a given type of facility that Mountainous terrain—Any combination of horizontal and verti-
are assumed to be the best possible from the point of view cal alignment causing heavy vehicles to operate at crawl
of capacity, that is, characteristics that if further improved speeds for significant distances or at frequent intervals.
would not result in increased capacity. Movement capacity—The capacity of a specific movement at
Impedance—The effect of congestion in higher-priority move- a stop-controlled intersection approach, assuming that the
ments at a stop-controlled approach on lower-priority move- movement has exclusive use of a separate lane, in passenger
ments, which reduces the capacity of lower-priority move- cars per hour.
ments. Multilane highway—A highway with at least two lanes for the
Interrupted flow—A category of traffic facilities having traffic exclusive use of traffic in each direction, with no or partial
control of access, that may have periodic interruptions to flow
signals, stop signs, or other fixed causes of periodic delay or
at signalized intersections.
interruption to the traffic stream; examples include intersec-
Nonweaving flows—Traffic movements in a weaving area not
tions and arterials.
actually engaged in weaving movements.
Interval—A period of time in a signal cycle during which all
No-passing zone—A segment of a two-lane, two-way highway
signal indications remain constant.
along which passing is prohibited in one or both directions.
Jam density—The density at which congestion becomes so severe One-sided weaving section—A weaving area in which vehicles
that all movement of persons or vehicles stops, usually ex- entering the highway approach from the same side of the
pressed as vehicles per mile (per lane) or pedestrians per roadway as exiting vehicles depart from it.
square foot. Operational analysis—A use of capacity analysis to determine
Lane 1—The highway lane adjacent to the shoulder. the prevailing level of service on an existing or projected
Lane balance—A condition at a diverge point where the number facility, with known or projected traffic, roadway, and control
of lanes leaving the diverge is equal to the number of lanes conditions.
approaching it plus one. Opposing approach—The approach approximately 180 degrees
Lane group—A set of lanes on an intersection approach that has opposite the subject approach at an all-way stop-controlled
been established for separate capacity and level-of-service (AWSC) intersection.
analysis. Passenger car equivalent—The number of passenger cars that
Level of service—A qualitative measure describing operational are displaced by a single heavy vehicle of a particular type
conditions within a traffic stream, generally described in under prevailing roadway, traffic, and control conditions.
terms of such factors as speed and travel time, freedom to Passenger service time—The time required for a passenger to
maneuver, traffic interruptions, comfort and convenience, board or alight from a transit vehicle, in seconds per
and safety. passenger.
Level terrain—Any combination of horizontal and vertical align- Passing sight distance—The visibility distance required to allow
ments that permits heavy vehicles to maintain approximately drivers to execute safe passing maneuvers in the opposing
the same speed as passenger cars; this generally includes short traffic lane of a two-lane, two-way highway.
grades of no more than 1 to 2 percent. Peak-hour factor—The hourly volume during the maximum vol-
Load factor—The number of passengers occupying a transit vehi- ume hour of the day divided by the peak 15-min rate of flow
cle divided by the number of seats on the vehicle. within the peak hour; a measure of traffic demand fluctuation
Loop ramp—A ramp serving a left-turn movement by requiring within the peak hour.
Pedestrian—An individual traveling on foot.
vehicles to execute that movement by turning right; typically,
Pedestrian area module—The space provided per pedestrian on
a 90-degree left turn is accomplished by making a 270-degree
a pedestrian facility, expressed as square feet per pedes-
right turn.
trian; space.
Lost time—Time during which the intersection is not effectively
Pedestrian flow rate—The number of persons passing a point per
used by any movement; clearance lost time plus start-up
unit time, usually expressed as pedestrians per 15 min or
lost time. pedestrians per minute.
Major weaving section—A weaving area having at least three Pedestrian speed—The average walking speed of pedestrians, in
entry and exit legs with two or more lanes. feet per second.
Maximum load point—The section of a transit line that has the Permitted turns—Left or right turns at a signalized intersection
highest passenger demand during a specified time interval. that are made against an opposing or conflicting vehicular or
Maximum service flow rate—The highest 15-min rate of flow pedestrian flow.
that can be accommodated on a highway facility under ideal Person-capacity—The maximum number of persons who can be
conditions while maintaining the operating characteristics for carried past a given point on a highway or transit right-of-
a stated level of service, expressed as passenger cars per hour way during a given time period under specified operating
per lane. conditions without unreasonable delay, hazard, or restriction,
Measures of effectiveness—Parameters describing the quality of in persons per hour.
service provided by a traffic facility to drivers, passengers, Person level of service—The quality of service offered the passen-
or pedestrians; examples include speed, density, delay, and ger within a transit vehicle, as determined by the available
similar measures. space per passenger.
A-4 appendix a
Phase—The part of the signal cycle allocated to any combination Roadway conditions—Geometric characteristics of a street or
of traffic movements receiving the right-of-way simultane- highway, including the type of facility, number and width of
ously during one or more intervals. lanes (by direction), shoulder widths and lateral clearances,
Planning analysis—A use of capacity analysis procedures to esti- design speed, and horizontal and vertical alignments.
mate the number of lanes required by a facility in order to Rolling terrain—Any combination of horizontal and vertical
provide for a specified level of service based on approximate alignments causing heavy vehicles to reduce their speed sub-
and general planning data in the early stages of project stantially below that of passenger cars but not causing heavy
development. vehicles to operate at crawl speeds for any significant amount
Platoon—A group of vehicles or pedestrians traveling together as of time.
a group, either voluntarily or involuntarily because of signal Saturation flow rate—The equivalent hourly rate at which vehi-
control, geometrics, or other factors. cles can traverse an intersection approach under prevailing
Platoon flow rate—The rate of flow of vehicles or pedestrians conditions, assuming that the green signal is available at all
within a platoon. times and no lost times are experienced, in vehicles per hour
Potential capacity—The capacity of a specific movement at a of green or vehicles per hour of green per lane.
stop-controlled intersection approach, assuming that it is un- Saturation headway—The average headway between passenger
impeded by other movements and has exclusive use of a cars in a stable moving queue as they pass through a signal-
separate lane, in passenger cars per hour. ized intersection, in seconds.
Pretimed control—Traffic signal control in which the cycle Seat capacity—The number of seats on a transit vehicle.
length, phase plan, and phase times are preset and are repeated Service flow rate—The maximum hourly rate at which persons
continuously according to the preset plan. or vehicles can be reasonably expected to traverse a point of
Productive capacity—A measure of transit efficiency or perfor- uniform section of a lane or roadway during a given time
mance; the product of passenger capacity and speed along a period (usually 15 min) under prevailing roadway, traffic, and
section of a transit line. control conditions while maintaining a designated level of
Protected turns—Left or right turns at a signalized intersection service, expressed as vehicles per hour or vehicles per hour
made with no opposing or conflicting vehicular or pedes- per lane.
trian flow. Shared-lane capacity—The capacity of a lane at an unsignalized
Queue—A line of vehicles or persons waiting to be served by the intersection that is shared by two or three movements, in
system in which the rate of flow from the front of the queue passenger cars per hour.
determines the average speed within the queue. Slowly mov- Space—The average area provided for pedestrians in a moving
ing vehicles or people joining the rear of the queue are usually pedestrian stream or pedestrian queue, in square feet per
considered a part of the queue. The internal queue dynamics pedestrian.
may involve a series of starts and stops. A faster-moving line Space mean speed—The average speed of the traffic stream com-
of vehicles is often referred to as a moving queue or a platoon. puted as the length of the highway segment divided by the
Ramp—A short segment of roadway serving as a connection be- average travel time of vehicles to traverse the segment; aver-
tween two traffic facilities; usually services flow in one direc- age travel speed; in miles per hour.
tion only. Spacing—The distance between two successive vehicles in a traf-
Ramp control—A system in which the entry of vehicles onto a fic lane measured from front bumper to front bumper, in feet.
limited access facility from a ramp is metered by a traffic Speed—A rate of motion expressed as distance per unit time.
signal; the signal allows one vehicle to enter on each green Standees—The number of passengers standing in a transit vehicle.
indication or green flash. Start-up lost time—Additional time consumed by the first few
Ramp-freeway junction—The roadway area over which an on- vehicles in a queue at a signalized intersection above and
or off-ramp joins the mainline of a freeway. beyond the saturation headway because of the need to react
Ramp junction—A short segment of highway along which vehi- to the initiation of the green phase and to accelerate to ambient
cles transfer from an on-ramp to the main roadway or from speed, in seconds.
the main roadway to an off-ramp. Street corner—The area encompassed within the intersection of
Ramp-street junction—The roadway area over which an on- or two sidewalks.
off-ramp joins with a surface street or arterial. Subject approach—The approach under study at two-way and
Ramp-weave section—A weaving area formed by a one-lane on- all-way stop-controlled (AWSC) intersections.
ramp followed by a one-lane off-ramp where the two are Three-lane highway—A highway having a three-lane cross sec-
joined by a continuous auxiliary lane. tion; the third lane (center) may be used in a variety of ways
Rate of flow—The equivalent hourly rate at which vehicles or including as a passing lane, a two-way left-turn lane, or a
persons pass a point on a lane, roadway, or other trafficway climbing lane.
for a period of time less than 1 hr; computed as the number Time mean speed—The arithmetic average of individual vehicle
of persons or vehicles passing the point divided by the time speeds passing a point on a roadway or lane, in miles per hour.
interval in which they pass (in hours); expressed as vehicles Traffic conditions—The distribution of vehicle types in the traffic
or persons per hour. stream, directional distribution of traffic, lane use distribution
Recreational vehicle—A heavy vehicle, generally operated by a of traffic, and type of driver population on a given facility.
private motorist, engaged in the transportation of recreational Truck—A heavy vehicle engaged primarily in the transport of
equipment or facilities; examples include campers, boat trail- goods and materials or in the delivery of services other than
ers, and motorcycle trailers. public transportation.
Turnout—A short section of a lane added to a two-lane, two-way Volume—The number of persons or vehicles passing a point on
highway for the purpose of allowing slow-moving vehicles a lane, roadway, or other trafficway during some time interval,
to leave the main roadway and stop to allow faster vehicles often taken to be 1 hr, expressed in vehicles.
to pass. Walkway—A facility provided for pedestrian movement and seg-
Two-lane highway—A roadway having a two-lane cross section regated from vehicular traffic by a curb or provided on a
with one lane for each direction of flow, on which passing separate right-of-way.
maneuvers must be made in the opposing lane. Weaving area—A length of highway over which traffic streams
Two-sided weaving section—A weaving area in which vehicles cross each other’s path without the aid of traffic signals over
entering the highway approach on the right and vehicles de- a length of highway, doing so through lane-changing maneu-
parting the highway depart on the left, or vice versa; weaving vers; formed between merge and diverge points, as well as
vehicles must essentially cross the mainline highway flow. between on-ramps and off-ramps on limited access facilities.
Two-way left-turn lane—The center lane on a three-lane or multi- Weaving configuration—The organization and continuity of
lane highway that is used continuously for vehicles turning
lanes in a weaving area; determines lane-changing character-
left in either direction of flow at midblock locations.
istics in the weaving area.
Unconstrained operation—An operating condition in a weaving
Weaving diagram—A schematic drawing of flows in a weaving
area where geometric constraints do not limit the ability of
area used as an aid to analysis.
weaving vehicles to achieve balanced operation.
Uninterrupted flow—A category of facilities having no fixed Weaving flows—Traffic movements in a weaving area actually
causes of delay or interruption external to the traffic stream; engaged in weaving movements.
examples of such facilities include freeways and unsignalized Weaving length—The length of a weaving area measured from a
sections of multilane and two-lane rural highways. point at the entrance gore where the right edge of the shoulder
Unit width flow rate—The pedestrian rate of flow expressed as highway lane and the left edge of the ramp are separated by
pedestrians per minute per foot of walkway or crosswalk 2 ft to a point at the gore where the lane edges are separated
width. by 12 ft, expressed in feet.
Unsignalized intersection—Any intersection not controlled by Work zone—An area of a highway in which maintenance and
traffic signals. construction operations are taking place that impinge on the
v/c ratio—The ratio of demand flow rate to capacity for a traffic number of lanes available to moving traffic or affect the oper-
facility. ational characteristics of traffic flowing through the area.
SYMBOLS
A .................... total area of a pedestrian facility, or portion thereof, cml ................... movement capacity of the left-turn movement in
sq ft; also the average number of alighting passen- the shared lane, pcph
gers per bus during a peak 15-minute period, pas- cmt ................... movement capacity of the through movement in
sengers/bus the shared lane, pcph
a ..................... alighting service time per passenger discharging cmr ................... movement capacity of the right-turn movement in
from a transit vehicle, sec the shared lane, pcph
Ac ................... circulation area of a pedestrian facility, sq ft cpi ................... potential capacity for movement i at an unsignal-
Ah ................... holding area of a pedestrian facility, sq ft ized intersection, pcph
An ................... net area available on a bus for standees, sq ft cSH .................. shared-lane capacity at an unsignalized intersec-
Aw ................... crosswalk area, sq ft tion, pcph
AADT ............. average annual daily traffic, veh/day cT .................... capacity of climbing lane under prevailing condi-
ART SPD ...... average travel speed on an arterial segment, mph tions, vph
B .................... average number of boarding passengers per bus cv .................... maximum number of buses per hour per channel
during a peak 15-min period, passengers/bus or bus berth at level of service i
b ..................... boarding service time per passenger entering a
cvi ................... maximum number of buses per hour per channel
transit vehicle, sec
or bus berth at level of service i
C .................... cycle length, sec
c ..................... capacity, vph D .................... density, pc/mi/ln, veh/mi/ln, or pedestrians/sq ft;
c′b ................... design capacity of a bus stop, buses/hr also the directional distribution factor used in con-
cI .................... approximate capacity of a multilane intersection verting AADT to DDHV; also approach delay on
approach, vph an arterial intersection approach, sec; also bus
dwell time at a bus stop, sec
ci .................... capacity of lane group i at a signalized intersec-
tion, vph d ..................... average stopped-time delay per vehicle, unadjusted
cj .................... capacity per lane for a freeway multilane highway for arrival type, sec/veh
under ideal conditions, for design speed j, pcphpl d1 .................... first-term delay, accounting for uniform delay,
cLT ................... left-turn capacity at a signalized intersection, vph sec/veh
cmi ................... movement capacity for movement i at an unsignal- d2 .................... second-term delay, accounting for incremental de-
ized intersection, pcph lay over and above uniform delay, sec/veh
A-6 appendix a
dA ................... average stopped-time delay for Approach A at a Gp ................... minimum pedestrian green phase at a signalized
signalized intersection, sec/veh intersection, sec
di .................... average stopped-time delay for lane group i at a gq .................... portion of a green phase during which left turns
signalized intersection, sec/veh are blocked by the clearance of an opposing queue
dI .................... average stopped-time delay for a signalized inter- of vehicles, sec
section, sec/veh gu .................... portion of a green phase during which left turns
Dd ................... distance to a downstream adjacent ramp, ft are not blocked by the clearance of an opposing
Du ................... distance to an upstream adjacent ramp, ft queue of vehicles, sec
Dj ................... jam density, pc/mi/ln or veh/mi/ln G/C ................ ratio of green time to cycle length
DDHV ............ directional design hour traffic, vph g/C ................. ratio of effective green time to cycle length
E .................... passenger car equivalent for a standard mix of ve- h ..................... saturation headway, sec/veh; also average headway
hicles on a specific grade on a two-lane, two-way of a transit facility at its maximum load point, sec
rural highway
h′ .................... minimum bus headway at a bus stop or berth, sec
EB ................... passenger car equivalent for buses
H .................... alighting passenger capacity per berth per hour
EHV ................. passenger car equivalent for the prevailing mix of
heavy vehicles on a two-lane, two-way rural Hv ................... hourly transit volume, passengers/hr
highway HV .................. heavy vehicle
EL ................... through passenger car equivalent for left turns at i ...................... movements of Rank 1 at unsignalized intersections
a signalized intersection Ip .................... impedance factor for the effect of specific grades
Eo ................... passenger car equivalent for a standard mix of ve- on the operation of passenger cars on a two-lane,
hicles on a level section of two-lane, two-way rural two-way highway grade
highway j ...................... movements of Rank 2 at unsignalized intersections
ER ................... passenger car equivalent for recreational vehicles J ..................... number of passengers boarding a bus line at the
ET ................... passenger car equivalent for trucks heaviest stop, passengers/hr
f ...................... bus frequency at the maximum load point, k ..................... movements of Rank 3 at unsignalized intersections
buses/hr K .................... ratio of design hour traffic to AADT; also number
f ′ .................... maximum bus frequency at a berth, buses/hr of passengers alighting a bus line at the heaviest
fa ..................... area type adjustment factor stop, passengers/hr
fbb .................... bus blockage adjustment factor l ...................... movements of Rank 4 at unsignalized intersections
f ′c ................... bus frequency at a berth, buses/cycle L ..................... length of weaving area, ft; also lost time per cycle
f ′d ................... design bus frequency, buses/berth/hr at a signalized intersection, sec; also additional lost
fd ..................... directional distribution factor time due to buses stopping, starting, and queueing
fE .................... multilane highway type and environment adjust- near bus stops, sec
ment factor Lh ................... length of a weaving area, in hundreds of ft
fHV ................... heavy vehicle adjustment factor Li .................... net square feet per standee for level of service i
fg ..................... grade adjustment factor for passenger cars on spe- Lo ................... number of lanes on the opposing approach
cific grades on a two-lane, two-way rural highway Ls .................... number of lanes on the subject approach
and for all vehicles at a signalized intersection Lt .................... length of a queue upstream of a work zone, ft
fLT ................... left-turn adjustment factor l1 ..................... start-up lost time, sec
fm .................... adjustment factor for permitted left turns in a l2 ..................... clearance lost time, sec
shared or exclusive lane at a signalized intersec- LT .................. left turn
tion—applied only to flow in the lane from which LTpc ................ proportion of the volume on the conflicting ap-
left turns are made proaches turning left
fp ..................... driver population adjustment factor; parking condi- LTpo ................ proportion of the volume on the opposing approach
tion adjustment factor for signalized intersections turning left
fRT ................... right-turn adjustment factor M .................... pedestrian space, sq ft/ped
fs ..................... opposing flow saturation factor, used in estimating MSF ............... maximum per lane service flow rate for a given
the left-turn adjustment factor for permitted left level of service, pcphpl
turns at a signalized intersection N .................... number of lanes on a facility or in a lane group
fw .................... lane width and lateral clearance adjustment factor at an intersection, generally in one direction; also
fx ..................... capacity adjustment factor for movement x that number of the vehicles in a signalized intersection
accounts for the impeding effects of higher-ranked queue at which start-up lost times no longer exist;
movements also number of buses per hour stopping at a given
gf .................... initial portion of a green phase during which bus stop, veh
through vehicles may move in a shared left-turn NB ................... number of local buses stopping at an intersection
through lane until the arrival of the first left-turning to pick up or discharge passengers, buses/hr
vehicle, sec Nb ................... number of effective berths at a bus stop
Gi ................... green time for phase i at a signalized intersec- N′b .................. number of berths provided at a multiberth bus
tion, sec station
gi .................... effective green time for phase i at a signalized Nm ................... number of parking maneuvers per hour within 250
intersection, sec ft of an intersection
O1 ................... bus occupancy (during peak 15 min) along a free- s ..................... saturation flow rate under prevailing conditions,
way, passengers/hr vphg or vphgpl
O2 ................... car occupancy (during peak 15 min) along a free- sn .................... seats per transit vehicle
way, passengers/hr so .................... saturation flow rate under ideal conditions, pcphgpl
P .................... line-haul capacity of a bus facility past the maxi- sop ................... saturation flow rate in an opposing lane group at
mum load point, person/hr a signalized intersection
p′ .................... adjustment to the major-street left-turn, minor- S ..................... average travel speed, mph; also average pedestrian
street impedance factor speed, fps
P″ ................... probability of a queue-free state for the conflicting Sc .................... critical speed, the speed at which capacity oc-
major-street left-turning traffic curs, mph
Pb ................... unit line-haul capacity of a bus facility, assuming Sw ................... average travel speed of weaving vehicles in a
a single berth at the heaviest stop, persons/hr weaving area, mph
PB ................... proportion of buses in the traffic stream Snw .................. average travel speed of nonweaving vehicles in a
PHV ................. proportion of heavy vehicles on a two-lane, two- weaving area, mph
way highway grade SF .................. service flow rate for a given level of service, vph
PL ................... proportion of left turns in a lane from which left SFL ................ service flow rate per lane, vphpl
turns are made at a signalized intersection T ..................... total time spent by a bus in a berth or stop, the
PLT .................. proportion of left turns at a signalized intersection sum of dwell time plus clearance time, sec; also
PLTo ................. proportion of left turns in the opposing flow at a analysis time period, in hours (for a 15-min analy-
signalized intersection sis period, use T = 0.25) (for unsignalized
Pox .................. probability that conflicting movement x at unsig- intersections)
nalized intersections will operate in queue-free tc ..................... clearance time between successive buses at a bus
state berth, sec
P*ox ................ factor indicating the probability there will be Tg ................... critical gap for movement i at an unsignalized in-
no queue in the shared lane for major-street move- tersection, sec
ments of Rank 1 and 4 for a minor-street move- tf ..................... follow-up gap, in sec
ment x TH .................. through vehicle
PP ................... proportion of passenger cars in the traffic stream ti ..................... start-up lost time for the ith vehicle in a queue at
PR ................... proportion of recreational vehicles in the traffic a signalized intersection; also travel time for the
stream ith vehicle traversing a highway section, sec
TL ................... time loss resulting from a bus blocking other vehi-
PRT .................. proportion of right turns at a signalized intersection
cles at an intersection, sec
PT ................... proportion of trucks in the traffic stream
TS ................... time-space in a pedestrian area, ped-min
PT/HV ................ proportion of trucks among heavy vehicles on a
TSc .................. circulation time-space in a pedestrian area, ped-
two-lane, two-way rural highway grade
min
PEDS ............. number of pedestrians per hour conflicting with a
TSh ................. holding time-space in a pedestrian area, ped-min
given right-turn movement at a signalized
tw .................... average time a pedestrian spends in a crosswalk,
intersection
sec/ped
PF .................. progression factor
U .................... lane utilization factor; also bus berth utilization
PHF ............... peak-hour factor factor
PVG ............... proportion of vehicles arriving during the green V .................... hourly volume, vph or vphpl
phase at a signalized intersection v ..................... rate of flow, vph, vphpl, or pedestrians/min/ft of
PTG ............... G/C ratio walkway
Q .................... maximum number of passengers per berth per hour va .................... total flow rate on a signalized intersection ap-
Qi ................... number of pedestrians in a holding area, for flow proach, vph
i during one signal cycle, peds Vcy .................. volume of traffic in conflicting stream y, in vph
Qt ................... number of vehicles in a queue upstream of a work vg .................... lane group flow rate at a signalized intersection,
zone, veh unadjusted for lane utilization, vph
Qu ................... total time spent in a holding area by pedestrians in V1 ................... volume in lane 1 (shoulder lane) of a freeway im-
flow i during one signal cycle, pedestrian-minutes mediately upstream of a ramp junction
R .................... weaving ratio; also reductive factor used to com- Vc ................... volume in a diverging lane immediately upstream
pensate for variations in bus dwell times in transit of a major diverge point, vph
analysis vc .................... flow rate at which capacity occurs on a two-lane,
r ..................... corner radius, ft two-way rural highway grade
ri ..................... length of red for phase i, sec Vcx .................. conflicting volume for movement x at an unsignal-
Rp ................... platoon ratio ized intersection, vph
RT .................. right turn Vd ................... total diverge volume, vph
RTpc ................ proportion of the volume on the conflicting ap- vd .................... total diverge rate of flow, vph
proaches turning right Vf .................... total freeway volume at a ramp junction, vph
RTpo ................ proportion of the volume on the opposing approach Vl .................... volume or flow rate of the left-turn movement in
turning right the shared lane, in pcph
A-8 appendix a
vLE .................. equivalent left-turn flow rate at a signalized inter- vw1 .................. weaving flow rate with the larger numeric value
section, in through passenger cars/hr among the two weaving flows, vph
vLT ................... left-turn flow rate, vph vw2 .................. weaving flow rate with the smaller numeric value
Vm ................... total merge volume, vph; also maximum number among the two weaving flows, vph
of pedestrians occupying a crosswalk during one vw .................... total weaving flow rate in a weaving area, vph
signal cycle, peds Vx ................... volume for movement x, vph
vm ................... total merge rate of flow, vph V15 .................. volume per 15-minute interval
vM ................... mainline flow in a signalized intersection ap- V1–12 ................ volume for various movements at an unsignalized
proach, vph intersection, vph
vo .................... opposing flow rate at a signalized intersection, vph v1–12 ................. volume or rate of flow for various movements at
vp .................... movement flow rate during peak 15-min. at a sig- an unsignalized intersection, pcph
nalized intersection, vph; also pedestrian flow rate VR .................. volume ratio
in platoons, peds/min/ft WB .................. buffer width on a walkway, not usable by moving
Vpo .................. proportion of the intersection volume on the oppos- pedestrians, ft
ing approach WE .................. effective walkway width, ft
Vps .................. proportion of the intersection volume on the sub- WT .................. total walkway width, ft
ject approach X .................... proportion of maximum load point passengers that
Vr .................... total ramp volume, vph; also volume or flow rate board at the heaviest stop in a transit line
of the right-turn movement in the shared lane, pcph Xc ................... critical flow rate-to-capacity ratio for a signalized
vr .................... total ramp rate of flow, pvh intersection
vRT .................. right-turn flow rate, vph Xi .................... flow rate-to-capacity ratio for lane group i at a
Vt .................... volume or flow rate of the through movement in signalized intersection
the shared lane, in pcph Y ..................... proportion of maximum load point passengers that
Vu ................... total volume on an adjacent upstream ramp, vph alight at the heaviest stop in a transit line; also
flow ratio

INDEX
All locators in this index refer to chapter number and page number separated by a hyphen, with beginning and ending page numbers connected by an
arrow (e.g., 9-15 refers to page 15 of Chapter 9; 9-15c17 refers to pages 15 through 17 of Chapter 9). The greatest detail will be found under headings
that correspond to the chapter headings. A knowledge of the chapter number and name will help establish the context of the index entry:
1 = Introduction, Concepts, and Applications 6 = Freeway Systems 11 = Arterial Streets

2 = Traffic Characteristics 7 = Multilane Rural and Suburban Highways 12 = Transit Capacity
3 = Basic Freeway Sections 8 = Two-Lane Highways 13 = Pedestrians
4 = Weaving Areas 9 = Signalized Intersections 14 = Bicycles
5 = Ramps and Ramp Junctions 10 = Unsignalized Intersections A = Appendix (Glossary)
For example, locators beginning with 11 will treat the topic from the standpoint of arterials and locators beginning with A refer to definitions in
Appendix A (Glossary), i.e., A-2 refers to page 2 of the Appendix A, where the term is defined.
The letters following the locators are abbreviations, which are explained in the legend at the bottom of each page of the index. Although many illustrations
have their own internal number in the text, all locators in this index refer to chapter and page number (e.g., 11-29,t refers to a table that appears on
page 29 of Chapter 11; 11-29c31,e,p refers to pages 29 through 31 of Chapter 11, which includes at least one equation and one photo). A list of figures,
photographs, and tables organized by their internal number appears immediately following the Contents at the beginning of this manual.
AADT see annual average daily traffic (AADT) characteristics 11-2 c3

adjustment factors A-1 comparison
arterials 11-10 c12,t with collector streets 11-2
basic freeway sections 3-15 c16, 3-21 c22 with downtown streets 11-2
capacity 1-8,t with multilane highways 7-2, 11-2
heavy vehicles 3-15 delay 11-3
interchange density 3-22 average control delay 11-9,e
lanes 3-21 c22 average stopped time delay A-1
multilane highways 7-10 c14,t incremental delay 11-9 c11
signalized intersections 9-12 c25, 9-28 c29,t intersection delay 11-6, 11-9, 11-13, 11-18,f, 11-21, 11-24 c25,
two-lane highways 8-8 c13 11-32, 11-35, 11-38, 11-42
walkways 13-6 t progression adjustment factor 11-10 c12,t
all-way stop-controlled (AWSC) intersections 10-59 c81 residual demand delay 1-10
capacity 10-62 c63, 10-69, 10-78 c79,t uniform delay 11-10 c11,t
delay 10-64, 10-69, 10-78 c79,t upstream signals 11-11
headway 2-9, 10-59 c60,f, 10-68, 10-78 c80,t field data 11-4, 11-23 c25
intersection approaches 10-60 c62,f flow rates 11-2 c3, 11-21 c23, 11-28,f
conflicting approach 10-60,f, 10-63 c64,f,t, A-2 free-flow speed 11-3 c4,t, 11-6 c7, 11-9, 11-16, 11-23 c27, A-2
multilane 10-64 c67,t lane distribution 11-2, 11-11, 11-16
opposing approach 10-60,f, A-3 level-of-service 11-2, 11-4 c5,f,t, 11-13 c14,f, 11-15 c39
single lane 10-63 medians 11-16
level-of-service 10-67,t, 10-78 c80,t peak-hour factor 11-16
sample calculations planning
multilane, four-way intersections 10-77 c80 analysis 11-15 c17
single lane, four-way intersection 10-75 c77 applications 11-2, 11-29 c32
T intersections 10-77 platoons 11-10 c11
volume 10-78,f, 10-79 c80 running time 11-6 c9,e,t, 11-20
worksheets 10-67 c75 sample calculations
annual average daily traffic (AADT) 2-17, 2-23,t, 3-25, A-1 arterial classification 11-17
multilane highways 7-19 arterial level-of-service 11-43
two-lane highways 8-13 c14,e,t southbound 11-17 c20
annual vehicle miles of travel 2-14,f northbound 11-20, 11-27,f
arterials 11-1 c44, A-1 effect of traffic flow rates on 11-20 c23
arterial class A-1 effect of traffic flow rates and length on 11-23, 11-28 c29, f,t
design category 11-6 c8,p,t, 11-17, A-2 arterial with large signal spacings 11-25 c29
functional category , 11-6 c8,t, 11-17, A-2 evaluation based on field data 11-23 c25, 11-44
arterial sections 11-6 c8,t, A-1 planning application
arterial segments 11-6 c8,t, 11-9,f, A-1 determining level-of-service 11-29 c31
bus transit 12-10,e, 12-13 c14,t, 12-25, 12-28,t, 12-41, 12-45, 12-50t, determining volumes 11-31 c32
12-52 c54,t stop control 11-32 c35
capacity 11-2 two-lane arterial 11-35 c39
A=Appendix/Glossary; e=equation/formula; f=figure/chart; p=photo/illus; t=table Updated December 1997
1
2 index
arterials (continued) right turning vehicles across curb bike lanes 14-2,f
saturation flow rates 11-16 effect on roadway segments between intersections 14-3
speed 11-3 opposed vs unopposed traffic 14-2 c3
average running speed 11-3, A-1 passenger car equivalents 14-2 c4,t
average travel speed 11-3 c4, 11-12, 11-20, 11-23,f, 11-27 c28,f, sample calculations
11-31,f, 11-34,f, 11-40,f impacts of a bike lane on right turning vehicles 14-4
free-flow speed 11-3 c4,t, 11-6 c7, 11-9, 11-23 c27 left turn impacts on a multilane approach 14-4
speed profiles 11-15,f, 11-23,f, 11-27,f, 11-31,f, 11-34,f, 11-40,f passenger car equivalents 14-3 c4
time-space trajectories 11-3,f bike lanes 14-2 c3,f, A-1
traffic signals 11-3, 11-25 c32 bicycle vehicle conflicts at intersections 14-2 c3
adjustment factors 11-10 c11,e,t capacity 14-3,t
fully actuated control 11-11 c12, A-2 lane width 14-2
green ratio 11-16, A-2 shared lanes 14-2 c3
spacings 11-25 c29 bike paths 14-3, A-1
pretimed control 11-12 capacity 14-3,t
travel time studies 11-6, 11-44 bikeways A-1
test-car method 11-40 bottlenecks 6-8 c9,f
turns 11-16 breakdown conditions 6-6 c7,f
volume 2-11, 2-15 c16,t, 11-31 c32 bus berths 12-11, 12-18 c29,e,t, 12-42 c45,t, 12-55,t
worksheets see also bus stops; bus terminals
intersection delay estimates 11-13, 11-18, 11-21, 11-24 c25, 11-32, berth A-1
11-35, 11-38, 11-42 bus lanes 12-10 c11, 12-13, 12-49,t, A-1
level-of-service 11-14, 11-19, 11-22, 11-26, 11-30, 11-33, 11-36, bus roadways see busways
11-39, 11-43 bus stops 12-10 c11, 12-18 c22,e,t, 12-28 c30
travel time 11-44 see also bus berths
average annual daily traffic see annual average daily traffic (AADT)
bus terminals 12-11, 12-29, 12-42, 12-54 c55,t
AWSC see all-way stop-controlled (AWSC) intersections
see also bus berths
peak hour 12-43,t
bus transit 12-1 c14, 12-18 c55, 12-59 c60
B see also bus berths; bus lanes; bus stops; bus terminals; bus
turnouts; buses; busways; transit; transit capacity
barriers 6-12,t arterials 12-10,e, 12-13 c14,t, 12-25, 12-28,t, 12-41, 12-45, 12-50t,
basic freeway sections 3-1 c40, 6-2 12-52 c54,t
adjustment factors 3-15c16, 3-21 c22 boarding and alighting service time 12-12 c13,t, 12-19,e, 12-41 c42,
basic freeway segments 3-1, 3-23, A-1 12-59 c60,t
capacity 3-1 c2
alighting time A-1
composite grades 3-37 c39,f
boarding time A-1
congested flow 3-4 c5,f
clearance time 12-3,t, 12-11, 12-20,t, A-2
crawl speed 3-16, A-2
dwell time 12-3,t, 12-10 c12, 12-30, A-2
density 3-6 c8, 3-22,t
passenger service time 12-44, A-3
design analysis 3-23
time loss 12-10,e
design speed 3-7
capacity 12-9 c14, 12-18 c29,t, 12-34 c39,t, 12-49 c55,t
driver population 3-8, 3-19
passenger capacity 12-4 c6
extended general freeway segments 3-15 c16,t, 6-2
person capacity 12-3 c4,t, 12-10, 12-14, 12-40, A-3
free-flow speed 3-4 c5,f, 3-11,t, 3-19 c20
productive capacity 12-3,t, 12-7, A-4
heavy vehicle factors 3-15 c19,t, 3-37,t
HOV facilities 3-25 downtown streets 12-13 c14,t, 12-26 c28,t, 12-42
ideal conditions 3-3 fare collection procedures 12-12, 12-25
lane width 3-4 c5, 3-21,t field data 12-39
lateral clearance 3-4 c5, 3-21,t flow rates 12-27,f
level-of-service 3-8 c14,f,p, 3-22, 3-26 c38 effect on vehicle capacity 12-10,e
methodology 3-8 c22 interrupted flow 12-3,t, 12-25
operational analysis 3-23 person flow 12-40
peak-hour factor 3-15 uninterrupted flow 12-3,t, 12-19, 12-23,e
planning analysis 3-23, 3-25 freeways 12-4,e, 12-10, 12-29, 12-40, 12-50,t
queue discharge 3-2, 3-4 headway 12-10 c11, 12-19,e
roadway conditions 3-2 level-of-service 12-7 c9,f,t, 12-13 c14,t, 12-20 c21,e,t, 12-39,t
speed-flow relationships 3-3 c4,f person level of service 12-3,t, A-3
vehicle equivalents 3-7,p, 3-16 c19,t, 3-37 c40,f load factors A-3
worksheets loading standards 12-8,t
basic freeway sections 3-14,f, 3-40,f maximum load point 12-3,t, 12-25,t, A-3
four-lane freeways 3-27, 3-33 peak hour 12-14, 12-50 c54,t
six-or-more-lane freeways 3-29, 3-31, 3-35 priority treatments 12-30 c34,t
bicycles 14-1 c4, A-1 routes 12-41 c42
see also bike lanes; bike paths; bikeways signalized intersections 9-15,t
effect on intersection capacity 14-2 c4 terminology 12-3,t
left turning bicycles from curb bike lanes 14-3 c4 bus turnouts 12-46 c47
Updated December 1997 A=Appendix/Glossary; e=equation/formula; f=figure/chart; p=photo/illus; t=table

index 3
buses 1-6, 12-7 c8, A-1 D

see also heavy vehicles
characteristics 12-8,t DDHV see direction design hour volume (DDHV)
crush capacity 12-3,t, A-2 delay A-2
doors 12-12 see also travel time
grades 7-13 c14,t, 8-9,t AWSC intersections 10-64, 10-69, 10-78 c80,t
passenger car equivalents 7-12 c13, 12-10 c11,t arterials 11-3, 11-10 c13
seat capacity 12-3,t, A-4 average approach delay A-1
standees 12-3,t, A-4 average total delay A-1
busways 12-10, 12-26 c27,f,t, 12-46, 12-50,t, A-1 control delay 9-7,t, 9-27,e, 9-29,e, 9-117 c120, 9-136
average control delay 10-31, 11-9,e
incremental delay 9-27 c30,e, 11-9 c11
interrupted flow facilities 2-9
C intersection delay 11-6, 11-9, 11-13, 11-18,f, 11-21, 11-24 c25, 11-32,
11-35, 11-38, 11-42
calculations see sample calculations lost time 2-7 c8,f, 2-31 c33, 9-2 c5,f, A-3
capacity 1-3 c8,t, 2-10 c36, A-1 signalized intersections 9-5 c7,t, 9-27 c30,e,t, 9-117 c120, 9-138 c143
adjustment factors 1-7 c8,t stopped time delay 2-9, 9-7
AWSC intersections 10-62 c63, 10-69, 10-78 c79,t average stopped time delay A-1
basic freeway sections 3-1 c8,f two-lane highways 8-2 c4, 8-29,f
bike paths 14-3,t TWSC intersections 10-3, 10-22 c25,f, 10-31, 10-50,f, 10-52 c55,t,
bus transit 12-9 c14, 12-18 c29,t, 12-34 c39,e,t, 12-49 c55,t 10-58,t
by facility type 2-34,t uniform delay 9-27 c28,e, 9-30, 9-48 c50,f, 9-132, 11-10 c11,t
field data 1-10 c11 density A-2
freeway 3-1 c2, 6-9 c13,f basic freeway sections 3-6 c8, 3-22,t
critical density A-2
HOV lanes 6-15
jam density A-3
work zones 6-9 c13,f
multilane highways 7-5, 7-8,f
ideal conditions 1-5
pedestrians 13-3
movement capacity 10-5 c6, A-3
ramp influence areas 5-8,t, 5-13,e
multilane highways 7-5, 7-16
uninterrupted flow facilities 2-4 c6,e
passenger capacity 12-4 c6
density-flow relationships 2-29 c30
crush capacity 12-3,t, A-2
multilane highways 7-3 c5,f, 7-14,e, 7-34,f
person capacity 1-4, 12-3 c4, 12-40, A-3
pedestrians 13-3 c4,f
arterials 12-4, 12-13 c14,t
design analysis 1-9 c10,t, A-2
downtown streets 12-13 c14,t basic freeway sections 3-23
freeway lanes 12-4 c5,e,f, 12-10 freeways 6-2 c6,e,f
person capacity vs vehicle capacity 12-2 multilane highways 7-16 c18, 7-24 c30,f
potential capacity A-4 ramps 6-2 c6
precision 1-10 signalized intersections 9-97 c98,f,t
productive capacity 12-3,t, 12-7, A-4 two-lane highways 8-17
rail transit 12-14 c18,e,t, 12-47, 12-55 c58 weaving areas 4-16 c17, 4-19
ramp-freeway terminals 5-3 c7,f,t design hour factors 2-16 c18, A-2
ramp roadways 5-11 c14,t design speed 3-7
roundabouts 10-84,e,f, 10-86 direction design hour volume (DDHV) 2-20,e, A-2
seat capacity A-4 basic freeway sections 3-23, 3-25
shared lane capacity A-4 multilane highways 7-19,e
signalized intersections 9-3 c8,e, 9-44, 9-66, 9-75 c76, 9-78, 9-83, 9-130 directional distribution 1-6, 2-19 c20,e, 2-23,t
two-lane highways 8-2, 8-4, 8-6, 8-11 pedestrians 13-7
TWSC intersections 10-3, 10-6,e, 10-15 c16,e, 10-28 c31,f, 10-48 c49,f, two-lane highways 8-4, 8-6, 8-8 c11,t
10-51,t, 10-54,t divided highways 7-10,t, 7-21 c24,e
vehicle capacity 1-3 see also medians; undivided highways
walkways 13-4, 13-8 downgrades see under grade segments
weaving areas 4-8,t downtown streets A-2
work zones 6-10 c14,f,t bus transit 12-13 c14,t, 12-26 c28,t, 12-42
CBD streets see downtown streets compared with arterials 11-2
central business district (CBD) streets see downtown streets person capacity 12-4
climbing lanes 8-11, 8-20, 8-26 driver population 3-8, 3-19
collector streets A-2
compared with arterials 11-2 E
composite grades 3-37 c39,f
extended general freeway segments 3-15 c16,t
conflicting traffic 10-4, 10-7 c10,f, 10-30, 10-50,f
conflicting traffic volume , A-2 F
congested flow 3-4 c5,f
control conditions 1-6 c8,t, A-2 field data
crawl speed 1-5 arterials 11-4, 11-23 c25
crosswalks see under walkways capacity 1-10 c11

4 index
field data (continued) follow-up time A-2

flared approaches 10-4, 10-21,f, 10-31, 10-58,t TWSC intersections 2-8 c9, 2-33, 10-3 c5
multilane highways 7-9, 7-16 general terrain segments 1-5
signalized intersections 9-9 c12,f, 9-31 c32, 9-105, 9-117 c123 basic freeway segments 3-15 c16
flow-density relationships see density-flow relationships level terrain 1-5, 3-16, 7-12, A-3
flow rates 2-10 c24 two-lane highways 8-5 c9,t, 8-14 c15, 8-21 c23
see also saturation flow rates; service flow rates grade segments 1-6
arterials 11-2 c3, 11-21 c23, 11-28,f basic freeway sections 3-15 c16, 3-23 c24
bus transit 12-27,f composite grades 3-19, 3-37, A-2
flow ratio 9-6, A-2 downgrades 3-18,t, 6-2
freeways 3-14,e mountainous terrain 1-5, 3-16,t, 6-2, 7-12, A-3
multilane highways 7-11,e, 7-32 c34,f multilane highways 7-12 c14
pedestrians 13-2 c4,e,f,t, 13-7, 13-9 c11,e,f, 13-18 rolling terrain 1-5, 3-16,t, 6-2, 7-12, A-4
platoon flow rates A-4 two-lane highways 8-4 c6,t, 8-8 c13,e,t, 8-15 c17, 8-23 c25
ramp influence areas 5-3 c4,e,f, 5-9 c14
upgrades 3-16 c18,t, 6-2
rates of flow A-4
grades 9-9 c10,f, 9-14,t, 9-17
signalized intersections 9-5 c6
uninterrupted flow facilities 2-3, 2-5 c6,e
unit width flow rates 13-3, A-5
weaving areas 4-8 c10,e,t H
flow space relationships
pedestrians 13-4,f headway 2-30 c31,f, A-2
flow speed relationships see speed flow relationships see also platoons
free-flow speed A-2 AWSC intersections 2-9, 10-59 c60,f, 10-68, 10-78 c80,t
see also speed bus transit 12-10 c11, 12-19,e
arterials 11-3 c4,t, 11-6 c7, 11-9, 11-16, 11-23 c27 saturation headway A-4
basic freeway sections 3-4 c5, 3-11,t, 3-19 c20 signalized intersections 2-7,f
multilane highways 7-3, 7-5 c7 spacing A-4
freeway surveillance see surveillance and control under freeways uninterrupted flow facilities 2-5
freeway systems see freeways heavy vehicles 1-5 c6, A-2
freeway-to-freeway ramps 5-2 see also passenger car equivalents (PCE)
freeways 3-1, 6-2 c16, A-2 basic freeway sections 3-7, 3-15 c19, 3-37 c38
see also basic freeway sections; ramps and ramp junctions; buses 1-6, 3-17 c18,t
weaving areas crawl speed 1-5, 3-16, A-2
bottlenecks 3-2, 6-8 c9,f multilane highways 7-7, 7-12 c14,e,t
breakdown conditions 6-6 c7,f recreational vehicles 1-6, 3-18,t
bus transit 12-4,e, 12-10 c11, 12-29, 12-40, 12-49 c50,t signalized intersections 9-10, 9-14,t, 9-17
components 3-1c2,f trucks 1-6, 3-17 c18,t, 3-38
design analysis 6-2 c6,e,f two-lane highways 8-8 c9,t, 8-11,e, 8-13,f, 8-30,f
HOV lanes 6-14 c15,f, A-3 high occupancy vehicle (HOV) lanes A-3
incidents 6-9,f basic freeway sections 3-3, 3-25
lane distribution 2-10, 2-14,t, 2-20 c21 freeways 6-14 c15,f
level-of-service 6-6,f Highway Capacity Manual
operational analysis 6-6 development, scope, organization 1-1 c3,t
passenger car equivalents 12-10
HOV lanes see high occupancy vehicle (HOV) lanes
person capacity per lane 12-4 c5,e,f, 12-10, 12-40,t, 12-49,t
sample calculations
design problem 6-2 c6,f
HOV lane impact analysis 6-14 c15,f I
queue analysis for a work zone 6-12 c13,f
speed 2-26c27,f ideal conditions A-3
speed-flow relationships 2-28 c29,f basic freeway sections 3-3
surveillance and control 6-7 c9 capacity 1-5
freeway surveillance A-2 intersection approaches 1-5
ramp metering 6-8,f level-of-service 1-5
volume 2-10, 2-12 c14,t multilane highways 7-4
weather effects 2-21 c22, 6-13 two-lane highways 8-4
weaving areas 4-2 uninterrupted flow facilities 1-5
work zones 6-9 c13,f impedance 8-11,e, 10-3, 10-10 c15,e,f, 10-30, 10-41, 10-46, 10-48 c49,t,
capacity 6-9 c13,f,t 10-51,t, 10-54,t, A-3
concrete barriers 6-12,t incidents 6-9,f
lane narrowing 6-10 intelligent vehicle highway systems (IVHS) see intelligent transportation
queuing 6-11 c13,f systems (ITS)
shoulder use 6-10 intelligent transportation systems (ITS) 1-7
work crew at site 6-9 c10,f interrupted flow facilities 1-2,t, 1-6 c8,t, 2-6 c9
see also signalized intersections; unsignalized intersections
G interrupted flow A-3
gap acceptance intersection approaches 10-33
critical gap A-2 approach A-1

index 5
intersection approaches (continued) freeways 6-6,f

average approach delay A-1 levels A-F 1-4
AWSC intersections 10-60 c67,f,t multilane highways 7-7 c11,f,t, 7-36,f
conflicting approach A-2 person level of service A-3
ideal conditions 1-5 rail transit 12-8 c9,t, 12-18,t
impedance 10-3, 10-11 c15,e,f, 10-30, 10-48 c49,t, 10-51,t, 10-54,t, A-3 ramp-freeway junctions 5-7 c8,t, 5-9
lane groups A-3 signalized intersections 9-5 c7,t, 9-27 c31, 9-47, 9-67 c68, 9-75, 9-77,
signalized intersections 9-5 c6,e, 9-12 c13,f,t, 9-19,e, 9-22 c27,f,t 9-79, 9-84 c85, 9-90, 9-131
subject approach A-4 two-lane highways 8-2, 8-5 c7,t, 8-9,t, 8-11 c13, 8-23 c25
intersections see signalized intersections; unsignalized intersections TWSC intersections 10-25, 10-31, 10-50,f, 10-54 c55,t, 10-58,t
ITS see intelligent transportation systems (ITS) walkways 13-3, 13-7 c16,f,t, 13-18, 13-24,e, 13-26
weaving areas 4-9,e,t, 4-11
level terrain see under general terrain segments
L light rail transit see under rail transit
LOS see level-of-service (LOS)
lanes lost time 2-7 c8,f, 2-31 c33, 9-2 c6,f, 9-108, A-3
acceleration lanes 5-2 c3 clearance lost time A-1
climbing lanes 8-11, 8-20, 8-26 start-up lost time 2-7, A-4
deceleration lanes 5-2 c3
HOV lanes 3-25, 6-14 c15,f, A-3
lane 1 5-2, A-3
M
lane balance A-3
lane changing 4-2 c4,t, 8-2 measures of effectiveness 1-4 c5,t, A-3
lane distribution 1-6 medians
arterials 11-2, 11-11, 11-16 arterials 11-16
freeways 2-10, 2-14,t, 2-20 c21 multilane highways 7-6, 7-10,t
multilane highways 7-14 models 1-9
ramp influence areas 5-3 c4 motor vehicle registrations 2-11,f
signalized intersections 9-12 c13,t, 9-54, 9-72, 9-87, 9-89, 9-92 c95 mountainous terrain see under grade segments
speed 2-27 c28,t multilane highways 7-1 c37, A-3
volume 2-20 c21 access points 7-6, 7-10 c11,t
lane groups 9-5 c6,e, 9-12 c13,f,t, 9-22 c27,f,t, A-3 adjustment factors 7-10 c14,t
arrival type 11-10 c11,e,f annual average daily traffic 7-19
lane utilization 9-12 c13,e,t
capacity analysis 7-5, 7-16 c20
lane width
characteristics 7-2 c3,p, 7-11,p
basic freeway sections 3-4 c5, 3-21,t
compared with arterials 7-2, 11-2
bicycles 14-2
compared with freeways 7-2
multilane highways 7-6, 7-10,t
compared with two-lane highways 7-2
signalized intersections 9-14,t, 9-17
density 7-5, 7-8,f
two-lane highways 8-8 c9,t, 8-11
density-flow relationships 7-3 c5,f, 7-14,e, 7-34,f
passing lanes 8-18 c20,f,t
design analysis 7-16 c18, 7-24 c30,f
ramp-freeway junctions 5-2 c8, 5-9 c14
divided highways 7-10,t, 7-21 c24,e
reversible lanes 8-20
field data 7-9, 7-16
shared lanes 9-20,f, 10-25, 10-31, 10-55,t, 10-57,t
free-flow speed 7-3, 7-5 c7
three-lane highways A-4
grade segments 7-12 c14,t
expansion of two-lane highways 8-18
multilane highways 7-19 heavy vehicle factors 7-7, 7-12 c14,e,t
two-way left-turn lanes A-5 ideal conditions 7-4
multilane highways 7-6 lane distribution 7-14
two-lane highways 8-20 lane width 7-6, 7-10,t
TWSC intersections 10-25, 10-29 c31, 10-50,f, 10-54 c55,t, 10-57,t lateral clearance 7-6, 7-10,t
weaving lanes 4-2 c4,f,t, 4-7 c8,t level-of-service 7-7 c11,f,t, 7-36,f
lateral clearance medians 7-6, 7-10,t
basic freeway sections 3-4 c5, 3-21,t methodology 7-7 c14
multilane highways 7-6, 7-10,t operational analysis 7-15 c16, 7-21 c24
left-hand ramps 5-12, 5-24, 5-26,f passenger car equivalents 7-12 c14,t
left turning vehicles peak-hour factor 7-7, 7-12
bicycles 14-3 c4 planning analysis 7-18 c20,t, 7-30
signalized intersections 9-13, 9-18 c22, 9-32 c33, 9-48 c49,f, 9-74,f, sample calculations
9-97 c98, 9-103 design analysis of an existing multilane roadway 7-24 c30,f
level of analysis 1-9 c10,t, 8-2, 9-8 design of a multilane highway 7-24, 7-27,f
see also design analysis; operational analysis; planning analysis operational analysis
level-of-service (LOS) 1-3 c5,t, A-3 divided highway 7-21, 7-25 c26,f
arterials 11-2, 11-4 c5,f,t, 11-13 c39 undivided highway 7-21, 7-23,f
AWSC intersections 10-67,t, 10-78 c80,t planning analysis for a new roadway 7-29,f
basic freeway sections 3-8 c13,f,p, 3-22, 3-26 c38, 6-2 c6,e,f segments 7-12,t, 7-14
bus transit 12-7 c9,f,t, 12-13 c14,t, 12-20 c21,e,t, 12-39,t service flow rates 7-19,t

6 index
multilane highways (continued) safety 13-3

signalized intersections 7-19 security 13-3
speed enforcement 7-6 flow rates 13-2 c4,e,f,t, 13-7, 13-9 c11,e,f, 13-18
speed-flow relationships 2-29, 7-4,f, 7-8,f, 7-15,f, 7-32 c34,f pedestrian flow rates A-3
three-lane highways 7-19 signalized intersections 13-17
two-way left-turn lanes 7-6, A-5 undivided highways 7-10,t, 7-21,e unit width flow rates 13-3, A-5
uninterrupted-flow 7-2 c3 flow space relationships 13-4,f
volume 2-10, 2-15,t, 7-9, 7-14 measurement 13-3
worksheets methodology 13-7 c14
operational and design analysis 7-15 c18, 7-20, 7-24 c28, 7-35 peak-hour factor 13-2,t
planning analysis 7-18 c19, 7-29, 7-36 platoons 13-3, 13-10 c11,e,f, 13-14 c16
multiple weaving areas 4-11 c12,f, 4-17 c18,f, 6-2 c6,f queuing 13-11, 13-19,e
sample calculations
analysis procedures for walkways 13-14 c16
N crosswalk analysis 13-22 c26
street corner analysis 13-18 c22,f
no passing zones 8-4, 8-10,t, A-3 signalized intersections 9-9 c11,e, 13-13 c14, 13-17, 13-22
space 13-3 c4, 13-7 c10,f,t, A-4
pedestrian area module A-3
O speed 13-3, 13-6 c8,f
pedestrian speed A-3
off-ramps see under ramp-freeway junctions speed density relationships 13-3 c4,f
on-ramps see under ramp-freeway junctions speed flow relationships 13-4,f
operational analysis 1-9 c10,t, A-3 speed space relationships 13-4 c5,f
basic freeway sections 3-23 trip purpose 13-6
freeways 6-6 worksheets
multilane highways 7-15 c16, 7-21 c24 crosswalk analysis 13-20, 13-25, 13-28
ramp-freeway junctions 5-2,f, 5-9 c14,f street corner analysis 13-21, 13-23, 13-29
signalized intersections 9-8 c30, 9-34 c49, 9-60 c83 walkway analysis 13-15 c16, 13-27
two-lane highways 8-2, 8-6 c17 planning analysis 1-9 c10,t, A-4
weaving areas 4-9 c12 arterials 11-2, 11-15 c17, 11-29 c32
basic freeway sections 3-23, 3-25
multilane highways 7-18 c20,t, 7-30
signalized intersections 9-8, 9-12,t, 9-31 c33, 9-50 c57, 9-80 c87, 9-91,
P
9-133 c135
parking 9-10,f, 9-12,t, 9-15,t, 9-17, 9-32 two-lane highways 8-2, 8-13 c14, 8-17, 8-26 c27
passenger car equivalents (PCE) A-3 platoons A-4
see also heavy vehicles see also headway
basic freeway sections 3-7,p, 3-16 c19,t, 3-37 c40,f arterials 11-10,e,t
bicycles 14-2 c4,t pedestrians 13-3, 13-10 c11,e,t, 13-14 c16
buses 7-12 c13, 12-10 c11,t platoon flow rates A-4
freeways 12-10 platoon ratio 9-11,e,t, 11-10 c11,e,t, A-4
multilane highways 7-12 c14,t signalized intersections 9-10 c11
two-lane highways 8-8 c9,t, 8-11 c12,e,t two-lane highways 8-3, 8-9
passenger cars 2-24,f TWSC intersections 10-3, 10-16 c20,f, 10-30, 10-53,t
passing lanes 8-18 c20,f,t public transportation see transit
passing sight distance 8-4, 8-18, A-3
PCE see passenger car equivalents (PCE)
peak-hour factor 2-3,e, 2-16 c18, A-3 Q
arterials 11-16
basic freeway sections 3-15 queues A-4
multilane highways 7-7, 7-12 basic freeway sections 3-2, 3-4
pedestrians 13-2,t pedestrians 13-11, 13-19,e
signalized intersections 9-12,e, 9-33 signalized intersections 9-19, 9-30 c31,f, 9-76 c80, 9-103, 9-106,e,
transit 12-2 c3,t, 12-14 c18,e,t, 12-43,t, 12-50 c55,t 9-110 c111,f, 9-117, 9-138 c143
two-lane highways 8-7,t TWSC intersections 10-4, 10-16 c18, 10-21 c22,f, 10-25 c27,f, 10-31,
pedestrians 13-1 c29, A-3 10-50,f, 10-53 c55,t, 10-58,t
see also walkways walkways 13-11
and right turning bicycles 14-2 c3 work zones 6-11 c13,f
cross flow traffic 13-7 c8,f
density 13-3
density flow relationships 13-3 c4,f R
directional distribution 13-7
environmental factors rail transit
comfort 13-3 see also transit; transit capacity
convenience 13-3 capacity 12-14 c18,e,t, 12-55 c58
economic factors 13-3 crush capacity 12-3,t, A-3

index 7
rail transit (continued) on-ramp off-ramp pair on eight-lane freeway 5-19 c23,f
capacity (continued) two-lane on-ramp on six-lane freeway 5-23,e
dwell time 12-59,t worksheets
productive capacity 12-7, A-4 consecutive off-ramps 5-17 c18
space requirement 12-9,t four-lane freeways 5-15
vehicle capacity 12-6,t, 12-8 c9,e,t isolated on-ramp 5-15
characteristics 12-8, 12-14 c16 left-side on-ramp 5-26
headway 12-15 on-ramp off-ramp pair on eight-lane freeway 5-20 c21
level-of-service 12-8 c9,t, 12-18,t off-ramp on ten-lane freeway 5-25
light rail transit 12-16 c17,t, 12-47, 12-55,t six-lane freeway 5-22
peak-hour factor 12-14 c18,e,t, 12-55,t ten-lane freeway 5-25
rapid transit 12-15 c17,t, 12-47, 12-56 c58,e,t two-lane on-ramps 5-22
street cars 12-16,t, 12-55,t rapid transit see under rail transit
terminology 12-3,t rates of flow see flow rates
ramp-freeway junctions 5-1 c27,e,f,t, A-4 recreational vehicles 1-6, 3-18,t, 7-13,t, 8-9,t, A-4
see also ramps and ramp junctions see also heavy vehicles
capacity 5-3 c7,t right-turn-on-red 9-13 c14, 9-18
characteristics 5-2,f right turning vehicles 13-26, 14-2,f
density 5-8,t, 5-13,e roadway conditions 1-5 c6, 1-8,t, A-4
diverge 5-2, 5-4 c7,t, A-2 adjustment factors 1-8,t
lane balance A-3 basic freeway sections 3-2
major diverge sites 5-13,f signalized intersection, 9-5, 9-9 c10,f, 9-97 c98
flow rates 5-2 c7,e,f rolling terrain see under grade segments
ten-lane freeway sections 5-12,t, 5-23 c24,f roundabouts 10-81 c89
freeway-to-freeway ramps 5-2 capacity 10-84,e,f, 10-86
lanes 5-2 characteristics 10-81 c82,f
acceleration lanes 5-3 sample calculations 10-87 c89
deceleration lanes 5-3 worksheets 10-84 c86
distribution 5-3 c4 running time
lane 1 5-2 arterials 11-6 c9,e,t, 11-20
lane additions 5-11 c12 average running time A-1
lane drops 5-11 c12 rural highways
length 5-2, 5-12 multilane highways 7-1 c37, A-3
level-of-service 5-7 c8,t two-lane highways 8-1 c33, A-5
merge 5-2, 5-4 c7,e,f,t, A-3 rural interstate travel by vehicle type 2-14,f
major merge sites 5-1 c13,f
methodology 5-3 c8
models 5-3 c8,e,f,t, 5-8,t, 5-13,e
off-ramps 5-2 c3,f, 5-9 c10, 5-11, 5-23 c24,t S
left-hand off-ramps 5 - 1 2
right-hand off-ramps 5-4,e, 5-6,e,f sample calculations
two-lane off-ramps 5-11,f, 5-14 Note: For greater detail, see the subheading sample calcula-
on-ramp off-ramp pair 5-19,e,f tions under the main headings for the topics listed below:
on-ramps 5-2 c3,e,f, 5-8, 5-14 c15,f arterials 11-17 c29, 11-43 c44
left-hand on-ramps 5-12, 5-24 c26,f AWSC intersections 10-75 c80
right-hand on-ramps 5-4 c5,e,f, 5-9 c11 basic freeway sections 3-23 c33
two-lane on-ramps 5-9 c11,f, 5-14, 5-23 bicycles 14-3 c4
speed 5-8,e,t freeways 6-2 c6, 6-12 c15,f
ramp junctions 3-1, 4-2, 5-1, A-4 multilane highways 7-21 c30
ramp roadways 5-2, 5-14,t pedestrians 13-14 c26
direct ramps A-2 ramps and ramp junctions 5-14 c26,e,f
ramp-street junctions 5-1, 5-14, A-4 roundabouts 10-87 c89
ramp weave sections A-4 signalized intersections 9-60 c96
weaving areas 4-3,f, 4-13 c14,e transit capacity 12-40 c47
ramps and ramp junctions 5-1 c27, 6-2 c6,f, 6-8,f, A-4 two-lane highways 8-21 c27
see also ramp-freeway junctions; ramp roadways; ramp-street TWSC intersections 10-46 c58
junctions weaving areas 4-12 c20,e,f,t
components 5-2 saturation flow rates A-4
design analysis 6-2 c6 arterials 11-16
loop ramps A-3 flow ratio A-2
ramp control 5-12, A-4 signalized intersections 2-7 c8,e,f, 2-31 c33,t, 9-5 c6, 9-9,f, 9-14 c22,
metering 6-8,f 9-38 c43, 9-63,f, 9-68, 9-71,f, 9-82, 9-121 c123, 9-127, 9-137
sample calculations 5-14 c26 service flow rates 1-4, A-4
consecutive off-ramps on six-lane freeway 5-16 c18,f maximum service flow rates A-3
isolated on-ramp 5-14 c16,f multilane highways 7-19,t
left-side on-ramp 5-24,e, 5-26,f precision 1-10
off-ramp on ten-lane freeway 5-23 c24,f signalized intersections 9-8, 9-58 c59,f, 9-89

8 index
service flow rates (continued) default values 9-12,t, 9-33

two-lane highways 8-8 c9,e, 8-12,e, 8-14,e, 8-21 c25 platoon ratio 9-11,e,t
shoulder use 6-10, 6-12 platoons 9-10 c11
shoulders 8-8 c9,t, 8-11, 8-18 c19,p queues 9-19, 9-30 c31,f, 9-76 c80, 9-103, 9-106,e, 9-110 c111,f, 9-117,
sidewalks see under walkways 9-138 c143
signalized intersections 2-6 c8,f, 9-1 c143 right-turn-on-red 9-13 c14, 9-18
see also traffic signals; all-way stop-controlled (AWSC) inter- roadway conditions 9-5
sections; two-way stop-controlled (TWSC) intersections; round- geometric conditions 9-9 c10,f, 9-97 c98
abouts sample calculations 9-60 c96
adjustment factors 9-12 c25, 9-28 c29,e,t operational analysis
area type (CBD, other) 9-15,t, 9-17 c18 determining v/c and service flow rates 9-88 c89, 9-96,t
bus transit 9-15,t, 9-17, 12-6, 12-10 c11,e, 12-19 c20,e, 12-23,e, 12-25,e existing pretimed two phase signal 9-60 c69
capacity 9-3 c8,e multiphase actuated signal 9-78 c85
delay 9-5 c7, 9-27 c30,e,t, 9-117 c120, 9-138 c143 three-phase pretimed signal 9-69 c78
adjustment factors 9-28,e planning analysis
average approach delay A-1 intersection with multilane approaches 9-84 c90
control delay 9-7,t, 9-27,e, 9-29,e, 9-117 c120, 9-136 intersection with single-lane approaches 9-86 c88, 9-91 c95,f
incremental delay 9-27 c30,e saturation flow rates 2-7 c8,e,f, 2-31 c33,t, 9-5 c6, 9-21, A-4
measurement methods 9-117 c120 direct measurement 9-121 c123, 9-137
residual demand delay 9-29 saturation flow rates module 9-9,f, 9-14 c22, 9-38 c43, 9-63,f, 9-71,f,
stopped delay 9-7 9-127
uniform delay 9-27 c28,e, 9-30, 9-48 c50,f, 9-132 service flow rates 9-8, 9-58 c59,f, 9-89
design 9-97 c98,f,t signalization conditions 9-5, 9-9 c10,f
effect of bicycles on vehicle flow 14-2 c4 through-car equivalents 9-20,f
left turning bicycles from curb bike lanes 14-2 c3 traffic conditions 9-5, 9-9 c10,f
right turning vehicles across bike lanes 14-2,f turns
field data 9-9 c12,f, 9-31 c32, 9-105, 9-117 c123 left turns 9-13, 9-18 c22, 9-32 c33, 9-49,f, 9-74,f, 9-97 c98
flow rate 9-5 c6 basic model for multilane approaches and exclusive-permitted left-
flow ratio 9-6, A-2 turn lanes 9-19 c22
degree of saturation 9-6, 9-102 basic model for simple permitted left turn lanes 9-18 c21
volume-to-capacity ratio 9-6,e sneakers 9-19, 9-48, 9-103
grades 9-9 c10,f, 9-14,t, 9-17 basic model for single-lane approaches opposed by single-lane ap-
headway 2-7,f proaches 9-20 c21
saturation headway A-4 special cases for permitted left turns 9-21
heavy vehicle factors 9-10, 9-14,t, 9-17 not opposed turns 9-3
lane groups 9-5 c6,e, 9-12 c13,f,t, 9-22 c27,f,t, A-3 permitted turns 9-3, 9-23,f, 9-35,f, 9-39, 9-41 c42, 9-54,t, 9-64 c65,
arrival types 9-10 c12,t 9-78 c79,f, 9-83,f, 9-128 c129, A-3
critical lane groups 9-6,e, 9-22 c27,f protected turns 9-3, 9-35,f, 9-54,t, A-4
critical paths 9-23 c27,f right turns (factors) 9-16,t
shared lanes 9-19,e right turns (formulas) 9-15,t
lane distribution 9-12 c13,t v/c ratio 9-2, 9-5 c7,e, 9-22 c23, 9-30, 9-58 c59,f, 9-89, 9-102,e
lane utilization 9-12 c13,t critical v/c ratio 9-6,e, 9-23, 9-31, 9-102,e, A-2
lane volume 9-87,f, 9-89,f, 9-92 c95,f, 9-134 v/s ratio 9-6, 9-23, 9-26
lane width 9-14,t, 9-17 volume 9-9,f, 9-12 c24, 9-37 c39,f
level-of-service 9-5 c7,t, 9-27 c31, 9-138 worksheets 9-35,f
criteria 9-6 c7,f capacity analysis module 9-44, 9-66, 9-75, 9-83, 9-130
LOS module 9-9,f, 9-27 c31 for permitted-plus-protected phasing 9-78
lost time 2-7 c8,f, 2-31 c33, 9-2, 9-4 c6,e,f, 9-108, A-3 for protected-only phasing 9-76
clearance lost time A-1 field saturation flow rates study 9-123, 9-137
start-up lost time 2-7, A-4 input module 9-36, 9-61, 9-70, 9-81, 9-125
multilane highways 7-19 intersection control delay 9-118, 9-120
operational analysis 9-8 c30, 9-9,f, 9-12,t, 9-34 c49, 9-60 c83 lane volume 9-54, 9-72, 9-87, 9-92
alternative computations 9-58 c59,f with EB right turn lane 9-93
capacity analysis module 9-9,f, 9-23 c27, 9-43 c45, 9-130 with geometric modifications 9-89
input module 9-8 c12, 9-34 c37, 9-125 with NB and SB left turn lanes 9-95
default values 9-12,t with NB and SB split phase operation 9-94
LOS module 9-9,f, 9-27 c30, 9-45 c48, 9-131 LOS module 9-47, 9-67, 9-68, 9-75, 9-84, 9-90, 9-131
adjustment factors 9-28 c29 for permitted-plus-protected phasing 9-79
saturation flow rate module 9-9,f, 9-14 c22, 9-38 c43, 9-127 for protected-only phasing 9-77
adjustment factors 9-14 c22 with no lane utilization factor 9-68
volume adjustment module 9-9,f, 9-12 c14, 9-38 c39, 9-126 with revised signal timing 9-85
adjustment factors 9-12 c14 with timing modifications 9-69
worksheet information flow 9-35,f planning method input 9-51, 9-54, 9-86, 9-91, 9-133
parking 9-10,f,t, 9-12,t, 9-15,t, 9-17, 9-32 lane volume 9-52, 9-134
peak-hour factor 9-12,e, 9-33 signal operations 9-53, 9-135
pedestrians 9-11,e, 13-14, 13-17, 13-22 saturation flow rates module 9-40, 9-63, 9-68, 9-71, 9-82, 9-127
planning analysis 9-8, 9-12,t, 9-31 c33, 9-50 c57, 9-84 c87, 9-133 c135 field study 9-123, 9-137

index 9
signalized intersections (continued) three-lane highways A-4

worksheets (continued) expansion of two-lane highways 8-18 c21
signal operations 9-54, 9-73, 9-88 multilane highways 7-19
with geometric modifications 9-90 through-car equivalent 9-20,f
with NB and SB left-turn lanes 9-95 toll roads 3-1
supplemental for permitted left turns 9-41, 9-83 see also freeways
multilane approach 9-41, 9-64, 9-128 traffic characteristics 2-1 c36
permitted-plus-protected phasing 9-78 traffic composition 2-21
single-lane approach 9-42, 9-65, 9-129 see also heavy vehicles; passenger car equivalents
supplemental left-turn 9-74 traffic conditions 1-6, 1-8,t, 2-1 c36, A-4
supplemental uniform delay 9-76, 9-85 capacity 2-10 c36
for left turns with primary and secondary phases 9-49, 9-132 interrupted flow facilities 2-6 c9
for permitted-plus-protected phasing 9-79 signalized intersections 9-5, 9-9 c10,f
traffic-actuated control input data 9-104 speed 2-24 c28
traffic-actuated timing computations 9-112 two-lane highways 8-17 c18
volume adjustment module 9-38, 9-62, 9-71, 9-82, 9-126 uninterrupted flow facilities 2-2 c6
simple weaving areas 4-9 c11,f volume 2-10 c24
space flow relationships see flow space relationships traffic flow 2-1 c36
space speed relationships see speed space relationships see also interrupted flow; uninterrupted flow
speed 2-9, 2-24 c28, A-4 downstream A-2
arterials 11-3 c4, 11-12, 11-15,f, 11-20, 11-23,f, 11-27 c28,f, 11-31,f, models 1-9
11-34,f, 11-40,f traffic priority ranks 10-5 c6,f
average running speed 2-4, 11-3, A-1 traffic signals 1-6 c7, 11-3, 11-10 c11
average travel speed 2-3 c4,e, 8-2, 11-3, 11-12, 11-20, A-1 see also signalized intersections
crawl speed 1-5, 3-16, A-2 and bus flow 12-11,t, 12-19
critical speed A-2 and crosswalks 13-13 c14
free-flow speed A-2
controller type 9-2 c3, 9-28, 9-98 c99
arterials 11-3 c4,t, 11-6 c7, 11-9, 11-23 c27
fully actuated operations 9-2, 9-98 c99
basic freeway sections 3-4 c5, 3-11,t, 3-19 c20
fully actuated control 11-11 c12, A-2
multilane highways 7-3, 7-5 c7
pretimed operations 9-2, 9-98 c99, 9-101 c102
freeways 2-26 c27,f
pretimed control A-4
lane distribution 2-27 c28,t
semiactuated operations 9-3, 9-98 c99, 9-112 c115
pedestrians 13-3, 13-6 c8,f
coordination 9-32
ramp influence areas 5-8,t
cycle 9-2, A-2
space mean speed A-4
change interval 9-2, A-1
spot speed 2-25,f,t
temporal variations cycle length 9-2, 9-33, A-2
day vs night 2-27 c28,t green ratio 9-2, 11-16, A-2
hourly 2-26 c27,f interval 9-2, A-3
time mean speed A-4 phase 9-2
trends 2-24 effective green time 9-2, 9-4 c5,e,f, 9-19 c22, A-2
two-lane highways 8-3, 8-6,t, 8-9,e, 8-12 c13,e,f effective red time 9-2, 9-4,e, A-2
weaving speed 4-6 c7,e,t green time 9-2 c5,f, 9-19 c31, 9-101 c103,e,f, 9-106 c107, A-2
speed density relationships 13-3 c4,f lost time 9-2, 9-4 c5
speed-flow relationships signal design 9-11, 9-33 c34, 9-98 c116
basic freeway sections 3-3 c4,f signal phasing 9-3 c5,f, 9-11, 9-59,f, 9-85,f, 9-98 c116
freeways 2-28 c29,f multiphase control 9-78 c84, 9-99 c101, 9-109 c115
multilane highways 2-29, 7-4,f, 7-8,f, 7-15,f, 7-32 c34,f pretimed control 9-11, 9-29,t, 9-60 c78, 11-12
pedestrians 13-4,f protected-plus-permitted phasing 9-5,f, 9-33
two-lane highways 2-29 c30,f, 8-4,f, 8-9,e, 8-12,e, 8-29,f split-phase operations 9-33
speed limits 7-6 three-phase control 9-69 c78
speed space relationships 13-4 c5,f timing plan 9-3, 9-11, 9-22, 9-32 c34, 9-101 c103, 9-106
stop-controlled intersections see all-way stop-controlled (AWSC) inter- traffic-actuated control 9-11, 9-60, 9-100 c101, 9-103 c112,
sections; two-way stop-controlled (TWSC) intersections 9-115 c116
stop signs 1-7 two-phase control 9-99 c100,f, 9-107 c108
stopped time see under delay spacings 11-25 c29
street cars see rail transit transit 12-1 c60
street corners see under walkways see also bus transit; rail transit; transit capacity
streets see arterials; collector streets; downtown streets passenger space requirements 12-9,e,t
suburban arterials see arterials peak travel 12- c3,t, 12-14,e
suburban highways 7-1 c37, A-3 transit capacity 12-2 c3, 12-6,e,t, 12-9 c60,e,t
see also bus transit; rail transit
T passenger capacity 12-4
crush capacity 12-3,t, A-2
temporal variations 2-12 c19, 2-26 c28,f,t load factors A-3
terrain 3-16,t, 6-2 maximum load point A-3
see also general terrain segments; grade segments person capacity 12-2, 12-4

10 index
transit capacity (continued) heavy vehicle factors 8-8 c9,t, 8-11,e, 8-13,f, 8-30,f
sample calculations ideal conditions 8-4
arterial bus turnout 12-46 c47 impedance factors 8-11,e
arterial street capacity 12-45 c46 intersections 8-20
berth capacity 12-42 c45 lane changing 8-2
bus berth unloading 12-44 c45 lane width 8-8 c9,t, 8-11
bus terminal (transit center) 12-42 c44,t level of analysis 8-2
CBD busway 12-46 level-of-service 8-2, 8-5 c7,t, 8-9,t, 8-11 c13, 8-23 c25
effect of buses methodology 8-5 c14
on arterials 12-41 no passing zones 8-4, 8-10,t, A-3
on freeway capacity 12-40 c41 operational analysis 8-2, 8-6 c17
light rail transit on city street 12-47 passenger car equivalents 8-8 c9,t, 8-11 c12,e,t
passenger service times 12-41 c42 passing 8-2
person capacity (four lane urban highway) 12-40,t passing lanes 8-18 c20,f,t
person flow (urban freeway) 12-40 passing sight distance 8-4, 8-18, A-3
planning applications, downtown street, LOS 12-42 peak-hour factor 8-7,t
rail rapid transit 12-47 planning analysis 8-2, 8-13 c14, 8-17, 8-26 c27
seat capacity A-4 platoons 8-3, 8-9
terminology 12-3,t reversible lanes 8-20
vehicle capacity 12-2 sample calculations
travel 2-14,f consideration of a climbing lane 8-26
travel time LOS and capacity of a specific grade 8-24 c25
see also delay LOS for a general terrain segment 8-23
average running time A-1 planning applications 8-26 c27
average total delay A-1 service flow rates
average travel time A-1 for a general terrain segment 8-21 c23
travel time studies 11-6, 11-40, 11-44 for a specific grade 8-23 c24
trucks , 1-6, A-4 service flow rates 8-8 c9,e, 8-12,e, 8-14,e, 8-21 c25
see also heavy vehicles shoulders 8-8 c9,t, 8-11, 8-18 c19,p
basic freeway segments 3-17 c18,t, 3-38 speed 8-2 c3, 8-6,t, 8-9,e, 8-12 c13,e,f
multilane highways 7-12 c14,t speed flow relationships 2-29 c30,f, 8-4,f, 8-9,e, 8-12,e, 8-29,f
truck performance curves 3-38,t traffic conditions 8-17 c18
two-lane highways 8-9,t, 8-13,f, 8-30,f turnouts 8-20 c21,t, A-5
turnouts 8-20 c21,t, A-5 two-way left-turn lanes 8-20
see also bus turnouts v/c ratio 8-5,t, 8-8, 8-10 c11,t
turns volume 2-11, 2-15,t
see also left turning vehicles; right turning vehicles; two-way worksheets
left-turn lanes general terrain segments 8-15, 8-22, 8-31
arterials 11-16 specific grades 8-16, 8-25, 8-32 c33
not opposed turns 9-3 two-lane rural highways see two-lane highways
permitted turns 9-3, 9-23,f, 9-35,f, 9-39, 9-41 c42, 9-54,t, 9-64 c65, two-way left-turn lanes 7-6, 8-20, A-5
9-78 c79,f, 9-83,f, 9-128 c129, A-3 two-way stop-controlled (TWSC) intersections 10-1 c58
permitted left turns 9-18 c21 capacity 10-3, 10-6,e, 10-15 c16,e, 10-28 c31,f, 10-48 c49,f, 10-51,t,
protected turns 9-3, 9-35, 9-54,t, A-4 10-54,t
signalized intersections 9-13, 9-18 c22, 9-32 c33, 9-48 c49, 9-74,f, conflicting traffic 10-4, 10-7 c10,f, 10-30, 10-50,f
9-97 c98, 9-103 conflicting traffic volume A-2
split-phase operation 9-33, 9-94 delay 10-3, 10-22 c25,f, 10-50,f, 10-52 c55,t, 10-58,t
two-lane highways 8-1 c33, A-5 average control delay 10-31
adjustment factors 8-8 c13 average total delay A-1
annual average daily traffic 8-13 c14,e,t flared approach 10-4, 10-21,f, 10-31, 10-58,t
capacity 8-2, 8-4, 8-6, 8-11 follow-up time 10-3, 10-10 c11,t, 10-30
characteristics 8-2 c3,p gap acceptance 2-8 c9, 2-33, 10-3 c5, 10-30
climbing lanes 8-11, 8-20, 8-26 critical gap size 10-10 c11,t
delay 8-2 c5,f, 8-29,f two-stage gap acceptance 10-20 c21,f, 10-31, 10-51,f, 10-56 c57,t
design analysis 8-17 impedance effects 10-3, 10-10 c15,e,f, 10-30, 10-48 c49, 10-51,t, 10-54,t
directional distribution 8-4, 8-6, 8-8 c11,t lanes 10-29 c30, 10-50,f, 10-54,t
expansion to three-lane highways 8-18 shared lanes 10-25, 10-31, 10-55,t, 10-57,t
field data 8-8 level-of-service 10-25, 10-31, 10-50,f, 10-54 c55,t, 10-58,t
four-lane sections 8-21 platoons 10-3, 10-16 c20,f, 10-30, 10-53,t
functions queuing 10-4, 10-16 c18, 10-21 c22,f, 10-25 c27,f, 10-31, 10-50,f,
accessibility 8-2 10-53 c55,t, 10-58,t
connectors 8-2 sample calculations
mobility 8-2 four leg intersection 10-50 c58,e,t
general terrain segments 8-5 c9,t, 8-14 c15, 8-21 c23 T intersection 10-46 c49,e,t
grade segments 8-4 c6,t, 8-8 c13,e,t, 8-15 c17, 8-23 c25 traffic priority ranks 10-5 c6,f
downgrades 8-5 upstream signals 10-16 c20,f,t, 10-52 c53,t
upgrades 8-5 c6,t, 8-9 c11,t volume 10-4, 10-29, 10-34,f, 10-47,f, 10-56,f

index 11
two-way stop-controlled (TWSC) intersections (continued) lane distribution 2-20 c21

warrants 10-27 by vehicle type 2-23,t
worksheets temporal variations 2-12 c19
basic intersection 10-33 daily 2-15 c16, 2-19,f, 2-26 c27,f
capacity 10-41 hourly 2-16, 2-21 c22,f
critical gap 10-36 design hour factors 2-16 c18, A-2
delay 10-46 peak-hour factors 2-16 c18, A-3
flared approaches 10-45 seasonal/monthly 2-15, 2-17 c18,f
follow-up time 10-36 trends 2-9 c28
gap acceptance 10-42c43 two-lane highways 2-11, 2-15,t
impedance 10-41, 10-46 TWSC intersections 10-4, 10-29, 10-34,f, 10-47,f, 10-56,f
shared lanes 10-44 weather effects 2-21 c22
site characteristics 10-35 volume capacity ratio see v/c ratio
upstream signals 10-37 c40 volume speed ratio see v/s ratio
volume 10-34
TWSC see two-way stop-controlled (TWSC) intersections
U W
undivided highways 7-10,t, 7-21,e walkways 13-4, 13-7 c27, A-5

see also multilane highways see also pedestrians
uninterrupted flow facilities 1-2,t, 1-8,t, A-5 capacity 13-4, 13-8
see also freeways crosswalks 9-10, 13-14, 13-16, 13-18 c19,f, 13-22 c25, A-2
density 2-4 c6,e,f crossing time 9-11,e, 13-22 c24
flow rates 2-3, 2-5 c6,e,f occupancy time 13-24,e
headway 2-5,e right turning vehicles 13-26
ideal conditions 1-5 traffic signals 13-13 c14
peak-hour factor 2-3,e level-of-service 13-3, 13-7 c16,f,t, 13-18, 13-24,e, 13-26
spacing 2-5 pedestrian area module A-3
speed 2-3 c4, 2-6,f queuing areas 13-11
average running speed 2-4, A-1 sidewalks 13-14 c16
average travel speed 2-3 c4,e, A-1 street corners 13-12 c13,f, 13-16 c22,e,f, A-4
measurement 2-4 circulation area 13-13, 13-16, 13-22,e, A-1
space mean speed 2-4,f, A-4 geometrics 13-17,f
time mean speed 2-4,f, A-4 holding area 13-13, 13-16
traffic conditions 2-2 c6 width 13-4 c6
uninterrupted flow A-5 adjustment factors 13-6,t
volume 2-2 c3 clear walkway width 13-5,f
unsignalized intersections 2-8 c9, 2-33, 10-1 c89, A-5 effective walkway width 13-4 c6,f, A-2
see also all-way stop-controlled (AWSC) intersections; two-
time space zones 13-13 c14, 13-19,e, 13-21 c22,e
way stop-controlled (TWSC) intersections
warrants 10-27
upgrades see under grade segments
weather effects 2-21 c22, 6-13
urban arterials see arterials
weaving areas 3-1, 4-2 c20,e,f,t, 6-2 c6,e,f, A-5
upstream signals 10-16 c20,f,t, 10-52 c53,t
configuration 4-2 c4,t
types A, B, C 4-2 c4,f,t, 4-6 c9,e,t, 4-19,t
weaving configuration A-5
crown line A-2
V design analysis 4-16 c17,e
diagrams 4-2 c4,f, 4-9 c11,f, 6-2 c6,f
v/c ratio A-5
weaving diagrams A-5
critical v/c ratio 9-6,e, 9-23, 9-31, 9-102,e, A-2
flow rates 4-8 c10,e,t
signalized intersections 9-2, 9-5 c7,e, 9-22 c23, 9-30, 9-58 c59,f, 9-89,
9-102,e gore areas 4-2 c3,f, A-2
two-lane highways 8-5,t, 8-8, 8-10 c11,t length 4-2,f, 4-8,t
v/s ratio 9-6, 9-23, 9-26 weaving length A-5
vehicle equivalents see passenger car equivalents level of service 4-9,t, 4-11
volume 2-2c3, 2-10 c24, A-5 major weaving section 4-3 c4,f, 4-12,e,f, A-3
arterials 2-11, 2-15 c16,t, 11-31 c32 methodology 4-6 c9
AWSC intersections 10-78 c80,f multiple weaving areas 4-2,f, 4-11 c12,f, 4-17 c19,e,t, 6-2 c6,e,f
demand volume A-2 nonweaving flows 4-2, A-3
freeways 2-10, 2-12 c14,t operational analysis 4-9 c12,e,f
multilane highways 2-10, 2-15,t, 7-9, 7-14 balanced operation A-1
signalized intersections 9-9,f, 9-12 c14, 9-38 c39,f, 9-62, 9-71, 9-82, constrained vs unconstrained operation 4-5 c8,e,t, 4-10,e,f
9-126 constrained operation 4-5, 4-14 c16,e,f, A-2
spatial distribution 2-19 c21 unconstrained operation 4-5, 4-13,e, A-5
directional distribution 2-19 c20, 2-23,t parameters 4-5t

12 index
weaving areas (continued) worksheets

ramp-weave sections 4-3,f, 4-13 c14,e,f, A-4 Note: For greater detail, see the subheading worksheets under
one-sided weaving section 4-3,f, A-3 the main headings for the topics listed below:
two-sided weaving section 4-3,f, A-5 all-way stop-controlled (AWSC) intersections 10-6 c75
sample calculations 4-12 c20,e,f,t arterials 11-13 c14, 11-18 c19, 11-21 c22, 11-24 c26, 11-30, 11-32 c33,
analysis of a major weaving area 4-12 c13,e,f 11-35 c36, 11-38 c39, 11-42 c44
analysis of a ramp-weave section 4-13 c14,e basic freeway sections 3-14,f, 3-27, 3-29, 3-31, 3-35, 3-40,f
constrained operation 4-14 c16,e,f multilane highways 7-15 c20, 7-24 c29, 7-35 c36
design application 4-16 c17,e pedestrians 13-15, 13-20 c29
multiple weaving area 4-17 c19,e,f ramps and ramp junctions 5-15 c26
sensitivity analysis with design application 4-19 c20 roundabouts 10-84 c86
simple weaving areas 4-2,f, 4-9 c11,e,f signalized intersections 9-35 c36, 9-38, 9-40 c42, 9-44, 9-47, 9-49,
speed 4-6 c7,e,t 9-51 c54, 9-61 c73, 9-75 c79, 9-81 c87, 9-89 c95, 9-104, 9-112, 9-118,
weaving flows 4-2, 4-8,t, A-5 9-120, 9-123, 9-125 c135, 9-137
width 4-4 two-lane highways 8-21 c27
work zones 6-9 c13,f,t, A-5 two-way stop-controlled (TWSC) intersections 10-33 c36

Highway Capacity Manual 3rd Edition

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Highway Capacity Manual 3rd Edition

Uploaded by

Copyright:

Available Formats

HIGHWAY

TRANSPORTATION RESEARCH BOARD

Updated December 1997

iii Updated December 1997

For the Committee on Highway Capacity and Quality of Service

Updated December 1997

TRB COMMITTEE ON HIGHWAY CAPACITY AND QUALITY OF SERVICE

Other Committee Members During 1985 Manual Preparation Period

vii Updated December 1997

John L. Schlaefli, TRACOR, Inc.

Committee Members as of December 1, 1994

Committee Members as of December 31, 1997

James L. Powell, DeLeuw, Cather & Company

Subcommittee Members as of February 1, 1985

Subcommittee Members as of December 1, 1994

Kenneth G. Courage, University of Florida

Subcommittee Members as of December 31, 1997

Subcommittee on General Concepts and Definitions

Subcommittee on Freeways and Multilane Highways

Subcommittee on Interchange Ramp Terminals

Subcommittee on Signalized Intersections

Bruce W. Robinson, Kittelson & Associates, Inc.

Subcommittee on Unsignalized Intersections

Subcommittee on Urban and Suburban Arterials

Subcommittee on Two-Lane Roads

Updated December 1997

Subcommittee on Transit Systems

Subcommittee on Planning Applications

Subcommittee on Pedestrians and Bicycles

Subcommittee on User Liaison and Interpretations

Subcommittee on Research Coordination

Updated December 1997

Updated December 1997

INTRODUCTION, CONCEPTS, AND APPLICATIONS

i. introduction .......................................................................................................................................................................... 1-1

1-1 Updated December 1997

The various adjustments necessary to account for less-than-ideal

Updated December 1997

Table 1-3. Adjustment Factors Used for Analyses

Updated December 1997

MODELS OF TRAFFIC FLOW Table 1-4. Analysis Techniques

Updated December 1997

INTRODUCTION .................................................................................................................................................................................... 2-2

PART A. BASIC VARIABLES OF TRAFFIC FLOW ......................................................................................................................... 2-2

PART B. OBSERVED VALUES ........................................................................................................................................................... 2-9

2-1 Updated October 1994

Density-Flow Relationships .................................................................................................................................................. 2-29

REFERENCES ......................................................................................................................................................................................... 2-34

PART A. BASIC VARIABLES OF TRAFFIC FLOW

Consider the following travel times observed for vehicles trav-

D = 1,000 vph/50 mph = 20 vpm

II. INTERRUPTED FLOW

The saturation flow rate represents the number of vehicles per

Figure 2-4. Concept of saturation flow rate and lost time.

N more complex with an increasing number of phases in a signal

PART B. OBSERVED VALUES

I. NATIONAL ROADWAY TRAFFIC TRENDS

II. VOLUMES AND FLOW RATES

Figure 2-5. Motor vehicle registrations.

RURAL TWO-WAY, TWO-LANE HIGHWAYS URBAN ARTERIALS

Updated October 1994

III. VOLUME CHARACTERISTICS

Up/Down % LOS ___ Table 3-1