Professional Documents
Culture Documents
substantial portion of the literature change in actuator command. The fuzzy The misconception in equating fuzzy
on fuzzy control deals with the use controller would execute control rules of control with nonlinear PID control has
of fuzzy rules to implement nonlinear pro- the form: if set-point error is positive big steered many engineers away from ex-
portional-integral-derivative (PID) type and error change is positive small, then ploiting the full potential of fuzzy logic in
control. As a result, many control engi- actuator output is negative big. When control applications. The purpose of this
neers are led to believe that using fuzzy used in this way, fuzzy control is not much article is to point out the many ways that
rules to implement nonlinear PID control different from conventional PID con- fuzzy logic can be used in control applica-
is the prime use of fuzzy logic for control. trol-it is solving the same set-point regu- tions beyond fuzzy PID control. In partic-
However, when we examine commercial lation problem addressed by PID control ular, the emphasis here is on the use of
products in which fuzzy control is said to and solving it in essentially the same way fuzzy logic to perform high-level control
be incorporated, we rarely see fuzzy logic as PID control, except that fuzzy control functions that fall outside the domain of
being used to implement nonlinear PID provides a nonlinear input/output map- conventional control methods. We will
(proportional-integral-derivative) con- ping. Hence, fuzzy control is often viewed examine industrial applications where
trol; fuzzy logic is used mostly to handle as a form of nonlinear PID control, and fuzzy logic was employed for supervisory
high-level control functions that tradi- control, for selecting discrete control ac-
comparisons of fuzzy control versus con-
tional control methods do not address. tions, for identifying the operating envi-
ventional PID control abound in litera-
This article discusses different ways ronment, and for evaluating controller
ture. Many engineers lose interest in
that fuzzy logic can be used in high-level performance.
fuzzy control after finding the perfor-
control functions. Specifically, we exam- mance improvement offered by fuzzy PID
ine the use of fuzzy logic for supervisory Supervisory Control
control cannot offset the increased com-
control, for selecting discrete control ac- Fuzzy control in the form of nonlinear
plexity in computation and controller tun-
tions, for identifying the operating envi- PID control has not found much accep-
ing, or after finding nonlinear PID control
ronment, and for evaluating controller tance in industry, because conventional
can be more efficiently implemented
performance. The purpose of this article is PID control is well entrenched, simpler,
to stimulate the use of fuzzy logic to pro- through other mechanisms (e.g., gain
scheduling or look-up table) than through low cost, and works satisfactorily for
vide new control functions that are out- most applications. For the instances
s i d e t h e d o m a i n of t r a d i t i o n a l fuzzy rules. Conventional PID control is
where fuzzy logic is applied to set-point
control-where fuzzy control is likely to well-established and can satisfy the per-
regulation, it is typically used in a
provide the greatest payoff. formance requirements of most set-point
high-level module that supervises a con-
regulation problems at minimal cost; ventional PID controller. Here we give
Introduction there is little incentive to switch from con- several examples to illustrate different
Much of fuzzy control research is fo- ventional PID control to a more complex, forms of fuzzy supervisory control.
cused on the set-point regulation problem, nonlinear form of PID control unless the The temperature controller produced
where the control objective is to drive a conventional controller is doing an unsat- by Yokogawa Electric [I] is a good exam-
process variable (e.g., motor shaft posi- isfactory job. Hence, the fuzzy PID con- ple of how fuzzy logic can be used for su-
tion, oven temperature) to a commanded troller that is ubiquitous in technical pervisory control. Temperature control
set-point. If one reads a paper on fuzzy literature rarely appears in actual com- usually involves processes that have a
control, chances are the paper will de- mercial applications; commercial appli- long time delay; for many processes, it is
scribe a fuzzy controller with set-point er- cations of fuzzy control are largely also imperative that the temperature does
ror and error change as its inputs, and the focused on high-level, task-oriented con- not overshoot the desired set-point. How-
output is an actuator command or a trol rather than set-point regulation. ever, it is difficult to avoid overshoot
trol shifting. Define Optimality Measure Fig. 6. Fuzzy membership functions char-
In addition to using fuzzy logic to in- How do we compare two system be- acterize the accelerator input and vari-
fer the operating environment, another haviors and judge which one is better? ance in accelerator inputproduced by the
interesting aspect of this application is This issue arises when a control engineer driver in different driving environments.
a
E Normal Carpet notch position associated with the optimal
Wood Floor, Mat outcome is then selected as the notch
F command.
Operating Time
In many control applications we know
how we want a system to behave but find it
Fig. 7. The,fuzzy vacuum cleaner from Matsushita infers the floor type from the rate of difficult to express the desired behavior in
decrease of dust. an analytic formula. Fuzzy logic is a power-
ful tool for expressing human preferences
optimality is the human designer, not the tion/deceleration is controlled by setting a and making the control system behavior ac-
numerical value produced by the analytic power lever and a brake lever at different curately reflect these preferences.
measure. Unfortunately, we tend to notch positions. Changing the notch posi-
choose an optimality measure based on tion frequently or in large increments cre- Conclusion
whether it has nice mathematical proper- ates an uncomfortable ride. In addition to There are many ways that fuzzy logic
ties, not based on whether it can express riding comfort, the controller must con- can be used in a control system to enhance
what is in the designers mind. sider safety, on-time arrival, energy con- capabilities and reduce operating
The use of fuzzy logic to express s u m p t i o n , and stopping the train The high payoff applications are us
optimality measures is perhaps the most accurately at a specified position along not in replacing a conventional PID con-
valuable benefit that fuzzy logic brings to the station platform. Optimizing train troller with a fuzzy PID controller, but in
control applications. Fuzzy membership control requires trading off between these using fuzzy logic at higher levels of con-
function is a natural framework for ex- multiple, often conflicting objectives. trol. The previous application examples
pressing the human designers conception The control method is based on pre- illustrate the wide range of opportunities
of the ideal system behavior and of how to dicting the outcome of each possible con- that exist and the many different ways that
measure closeness to the ideal behavior. trol action and then choosing the action fuzzy logic can be used to complement
Fuzzy logic can also provide smooth tran- that corresponds to the optimal outcome. conventional controllers. A designer who
sitions in the optimality measure to em- A simple simulation of the train dynamics wishes to exploit the full potential of
phasize certain control objectives as the is used to predict the resultant speed, stop- fuzzy logic must maintain a broad view of
operating condition changes. ping position, and time of arrival for each the different aspects of a control problem
The subway train control system de- possible choice of notch position. The and be creative in applying fuzzy logic
veloped by Hitachi [8] is an example optimality of each predicted outcome is where appropriate.
where fuzzy logic was applied to evaluate then rated by a set of fuzzy rules, taking The essence of fuzzy logic is that it lets
the optimality of control actions. For this into account factors such as the degree to you express whats on your mind. It is not
particular system, the trains accelera- which the train is on schedule, the degree surprising that most commercial applica-
tions of fuzzy control has been for high
level, task-onented control, where there
Stephen L. Chiu i s a Research Scientist at Rockwell Science are no standard analytic solutions but an-
ple human intuition.
Center. He received the B.S. degree in mechanical engmeer-
ing and nuclear engineering from U.C. Berkeley in 1983 and
the S.M. degree in mechanical engineering from M.I.T. in References
[ l ] Yokogawa, UT15RJT14 Digital Indicating
1985. He has conducted research in robotics, fuzzy control, Controllers, technical informatlon publication TI
and neuro-fuzzy systems. His current work is focused on 5B4A7-01E, Yokogawa Electnc, Tokyo, Japan,
neuro-fuzzy systems and on managing a smart motor devel- 1990.
opment project. He is presently an associate editor of the Jour- [2] S Chand, On-line, self-monitoring tuner for
nal ofhtelligent and Fuzzy Systems and a member of the board proportional integral derivative controllers,
of the North Amencaii Fuzzy Information Processing Society. He served previ- Proc. IEEE Conf Decision and Control, Brigh-
ously as an associate editor of IEEE Control Systems Magazine and as a co-guest ton, England, Dec 1991
editor of the Proceedings of IEEE in a special issue on fuzzy logic. [3] P Bonissone, V Badami, K Chiang, P
The author is with Rockwell Science Center, 1049 Camino Dos Rios, Thou- Khedkar, K Marcelle, and M Schutten, Indus-
sand Oaks, CA 91360 (slchiu@rsc.rockwell.com). Portions of this article were trial applications of fuzzy logic at General Elec-
published in the proceedings of the IEEE/IFIP International Conference on Infor- tnc, Proc IEEE, vol 83, no 3, pp 450-465,
mation Technology for Balanced Automation Systems, Prague, August 1998. March 1995
These portions are used with permission from Kluwer Academic Press. [4] T Froese, C von Altrock, and S Franke, Op-
timization of a water-treatment system with fuzzy
The Transistor
Turns 50! You can customizeProgram CC by incorporating
powerful macros and user-addedcommands. Ppogram CC
can easily link to other software.
The PROCEEDINGS OF THE IEEE will
proudly honor the invention of the
Transistor as it turns 50. Our January so much so that our
1998 issue will be devoted to all
aspects of this pivotal anniversary!
A chronicle of one of the most New Package The Classical Control package can now be
significant developments in the history purchased separately and may be upgraded to include Modern
of technology, this Special Issue is Control at any time. Our Complete Program CC includes both
destined to become a collectors item! Classical Control and Modern Control in one package.
S Y S T E M S
TECHNOLOGY
and receive 12 issues for $23.00
IEEE Member Price: $23.00 list Price 5375 00 INCORPORATED
Subrcriphon Older No 500501
Outside USA, Canada ond Mexicoplease odd 525 shippmg ond handlhg
Call 1 800 678-IEEE or 1 732 981-0060
9 E-mail: curtomer.rervice@ieee.org