1
Exploring Human Hand Capabilities into Embedded
Multifingered Object Manipulation
Honghai Liu, Senior Member, IEEE
Abstract—This paper provides a comprehensive computational
account of hand-centred research, which is principles, method-
ologies and practical issues behind human hands, robot hands,
rehabilitation hands, prosthetic hands and their applications.
In order to help readers understand hand-centred research,
this paper presents recent scientific findings and technologies
including human hand analysis and synthesis, hand motion
capture, recognition algorithms and applications, it serves the
purpose of how to transfer human hand manipulation skills to
related hand-centred applications in a computational context.
The concluding discussion assesses the progress thus far and
outlines some research challenges and future directions, and
solution to which is essential to achieve the goals of human hand
manipulation skill transfer. It is expected that the survey will
also provide profound insights into an in-depth understanding of
realtime hand-centred algorithms, human perception-action and
potential hand-centred healthcare solutions.
Index Terms—Motion recognition, human hand modeling, mo-
tion capturing, multifingered robot manipulation and cognitive
robotics.
I. I NTRODUCTION
I T is evident that the dexterity and multi-purposed ma-
nipulation property of the human hand inspires cross-
disciplinary research and application of robotics and artificial
intelligence. It has been driven by the dream of developing an
artificial hand with the human hand’s properties [1]. However,
five decades on, it is clear that the priority of the dream coming
true has been given to computational hand models [2], [3]. It
is the primary challenge of how to transfer human hand capa-
bilities into multi-fingered object manipulation, especially, in
a realtime context. It underlies advanced artificial intelligence
robotics and its related disciplines and applications.
Recent innovations in motor technology and robotics have
achieved impressive results in the hardware of robotic hands
such as the Southampton Hand, DLR hands, Robonout hand,
Barret hand, DEKA hand, Emolution hand, Shadow hand
and iLimb hand [4]. Especially, the ACT hand [5] has not
only the same kinematics but also the similar anatomical
structure with the human hand, providing a good start for
the new generation of anatomical robotic hands. However,
anatomically correct robotic hand is still a far way to go
H. Liu is with Intelligent Systems and Biomedical Robotics Group, School
of Creative Technologies, University of Portsmouth, England, PO1 2DJ UK.
Email: honghai.liu@port.ac.uk
The author would like to acknowledge the projects under grant No
EP/G041377/1 funded by Engineering and Physical Science Research Council,
grant No IJP08/R2 by the Royal Society.
Copyright (c) 2009 IEEE. Personal use of this material is permitted.
However, permission to use this material for any other purposes must be
obtained from the IEEE by sending a request to pubs-permissions@ieee.org.
due to the lack of appropriate sensory systems, unsolved
human-robot interaction problems, mysterious neuroscience
issues, etc. In recent decades, due to significant innovations
in multifingered robot hands and mature algorithms in robot
planning, priority has been given to multifingered robot ob-
ject manipulation. As the hardware of multifingered robotic
systems has developed, there have been parallel advances in
three most important engineering challenges for the robotics
community, namely, the optimal manipulation synthesis prob-
lem, the real-time grasping force optimisation problem, and
coordinated manipulation with finger gaiting [6], [7]. Stable
multifingered robot manipulation is determined by engineering
criteria, that is, force closures. Computer scientists have made
significant advances in computational intelligence for robot
manipulation. Gomez et al. [8] developed an adaptive learning
mechanism, which allows a tendon driven robotic hand to
explore its own movement possibilities, to interact with objects
of different shapes, sizes and materials and learn how to
grasp and manipulate them. It is recognised that the state
of the art in hardware aspects of artificial hands confirms
that hardware platforms have capabilities of accommodating
advanced computational models for adapting the human hand
manipulation skills though it is still a challenge to operate
them in a real-time context [9]–[11].
However, the manipulation systems of the robotic hands are
hardcoded to handle specific objects by their corresponding
robot hands. It is evident that robot hand control and optimi-
sation problems are very difficult to resolve in mathematical
terms, however humans can solve their hand manipulation
tasks easily using skills and experiences. Object manipulation
algorithms are needed that have human-like manipulation
capabilities, are independent of robot hand hardware and are
conducted in realtime term. Hence, the main challenge that
researchers now face is how to enable robot hands to use what
can be learned from human hands, to manipulate objects, with
the same degree of skill and delicacy as human hands in a
reasonably fast manner. For instance, given the locations and
shapes of a cup by off-the-shelf image processing algorithms,
a robotic hand is required, inspired by human hand biological
capabilities, to reach and manipulate the cup by continuously
responding with appropriate shape configuration and force
distribution among the fingers and palm.
Transferring human manipulation skills to artificial hands
involves modeling and understanding the human hand mo-
tion capabilities, and advanced multifingered manipulation
planning and control systems, manipulation abilities, sensory
perception and motion algorithms of artificial hand systems.
It connects user manipulation commands/intentions to the

motion of a highly articulated and constrained human hand,
containing 27 bones, giving it roughly 27 degrees of freedom
or ways of moving. Chella et. al. [12] confirmed that prior
knowledge should be introduced in order to achieve fast and
reusable learning in behavioural features, and integrate it in
the overall knowledge base of a system. It serves the pur-
pose of fulfilling requirements such as reusability, scalability,
explainability and software architecture. Carrozza et. al. [13]
designed artificial hand systems for dexterous manipulation;
they showed that human-like functionality could be achieved
even if the structure of the system is not completely biologi-
cally inspired. Learning from human hand motions is preferred
for human-robot skill transfer in that, unlike teleoperation-
related methods, it provides non-contact skill transfer from
human motions to robot motions by a paradigm that can
endow artefacts with the ability for skill growth and life-long
adaptation without detailed programming [14]. In principle,
not only does the provide a natural, user-friendly means of
implicitly programming the robot, but it also makes the learn-
ing problem become significantly more tractable by separating
redundancies from important characteristics of a task.
This paper presents a survey of most recent work in the
procedure of the human hand skill transfer to artificial hand
systems, the detailed modules is illustrated in Fig. 1, namely,
hand analysis and synthesis, hand motion capture, hand skill
transfer and hand-centred applications. Though a variety of
problems of the multifingered artificial hand systems have
been addressed, it is evident that the research communities and
practitioners require a unified version of recent improvements
of exploring human hand capabilities into multi-fingered robot
manipulation with an emphasis on computational models of
hand motion recognition. Not only does this paper provide
a comprehensive computational account of hand-centred re-
search, but it also assesses the progress thus far and outlines
some research challenges and future directions, and solution to
which is essential to achieve the goals of human hand manip-
ulation skill transfer. Though computational issues related to
neuroscience, health sciences and developmental robotics are
not addressed in this paper in order to focus on the computa-
tional aspects of hand manipulation skills transfer to artificial
hands. The remainder of this paper is organized as follows:
Section II presents the human hand anatomy and summarizes
hand motion capturing devices; it is also devoted to hand
gesture capturing based on glove, vision and electromyography
(EMG). Section III discusses various recognition methods,
with particular emphasis on Hidden Markov Models (HMM),
Finite State Machine (FSM), and Connectionist approach.
Section IV overviews the hand-centred application, and the
last section concludes this paper and indicates some existing
challenges and future research possibilities.
II. T HE H UMAN H AND M ODELING
Every day human hands perform a huge amount of dex-
terous grasps to fetch, move and use different tools instinctly
due to the innate sense for goal attainment and sensorimotor
control. However, everyday tasks in human environments are
relatively difficult for multi-fingered artificial hands mainly
2
Fig. 1. A schematic of exploring human hand capabilities into multifingered
artificial manipulation
due to the lack of appropriate sensor systems, unsolved
problems involving with the human-robot interaction (HRI)
and neuroscience, etc. Though artificial hands may perform
stronger and faster grasps than the human hand, the high
dimensionality makes it hard to embedded program and ma-
nipulate human-like robotic hand for dexterous grasps as
humans do.
A. The Human Hand
The human hand has complex kinematics and highly artic-
ulated mechanism with 27 degrees of freedom (DOF), i.e., 4
in each finger, 3 for extension and flexion and 1 for abduction
and adduction; the thumb is more complicated and has 5
DOF, and 6 DOF for the rotation and translation of the wrist.
Natural anatomical restrictions subject to the muscle-tendon
controlling mechanism enable the human hand capability of
super dexterity and powerfully usage of a wide range of tools.
Of the 1000 or so different functions we perform daily with
the 19 bones in each hand as shown in Fig. 2. The wrist itself
contains eight small bones called carpals. The carpals join
with the two forearm bones, the radius and ulna, forming
the wrist joint. Further into the palm, the carpals connect
to the metacarpals. There are five metacarpals forming the
palm of the hand. One metacarpal connects to each finger and
thumb. Small bone shafts called phalanges line up to form each
finger and thumb. The main knuckle joints are formed by the
connections of the phalanges to the metacarpals. These joints
are called the metacarpophalangeal joints (MCP joints). The
MCP joints work like a hinge when you bend and straighten
your fingers and thumb. The three phalanges in each finger
are separated by two joints, called interphalangeal joints (IP
joints). The one closest to the MCP joint (knuckle) is called
the proximal IP joint (PIP joint). The joint near the end of
the finger is called the distal IP joint (DIP joint). The joints

Fig. 2. Anatomical structure of the human hand
of the hand, fingers, and thumb are covered on the ends with
articular cartilage to absorb shock and provide an extremely
smooth surface to facilitate motion.
The human hand is most profound system having capability
of performing more complicated, dexterous tasks than any ex-
isting systems does. In terms of artificial intelligence robotics,
all existing artificial hands can only mimic how the human
hand does in a rough grade, far from reflecting its inherent
motion capabilities, let alone to in-depth understanding of its
connection to the human brian. Though there is a substantially
growing interest in related research in neuroscience, health
science and cognitive robotics [15].
B. Hand Motion Capturing
Sensory information such as hand position, angle, force and
their related parameters are required available to build up a
computational model of the human hand, further transfer the
model to hand-centred applications. That complexity and dex-
terity of the human hand and its involvement in various tools
handling, makes it challenging to handle tradeoffs between
sufficiently acquired data representing hand manipulation ca-
pabilities and precision of hand computational models.
1) Hand Gesture Capturing: Hand gestures usually are
represented by combination of the parameters captured by
glove-based systems, marker-based capturing systems and
conventionally clinical measurement tools. Vision based hand
capturing systems are classified as an individual group par-
tially due to its widely usage and promising application
future. A comprehensive review is presented to intensively
summarize the historical development of gloved-based systems
and their application [2], though there are very few innovation
in dataglove systems, an update is provided to focus on hand
capturing oriented glove systems [16] in Table I. Maker-based
capturing systems are widely developed and employed for
hand motion capturing in order to compensate the drawbacks
of glove-based systems. For example, surface markers are
3
Fig. 3. Hand gesture capturing based on surface markers [17]
used to capture wrist, metacarpal arch, fingers and thumb
movements as shown in Fig. 3. Note that, however, hand
capturing systems in this category are inconvenient to use
since they have to always cling to hands and time-consuming
calibration.
2) Vision Based Capturing: Computer vision-based human
motion capture intensively reviews related capturing devices
and techniques before 2001 [18], though its overview is
focused on a taxonomy of system functionalities broken down
into initialisation, tracking, pose estimation and recognition.
Priority of human motion capturing, especially hand motion,
is moved to capturing devices based on multi-cameras, 3-
D cameras and innovative devices such as kinetic [19]. It
is evident that the latter two types of capturing methods
suffer from low precision. Multiple cameras and depth cam-
eras have been arranged for developing motion recognition
algorithms subject to individual applications, public available
multi-camera databases are summarized in [19], for instance,
HumanEva-I dataset contains 7 calibrated video sequences
that are synchronized with 3D body poses obtained from
a motion capture system. Additionally, the Kinect’s depth
sensor consists of an infrared laser projector combined with a
monochrome CMOS sensor. It captures video data in 3D under
a variety of ambient light conditions. Since the human hand
is highly articulated and constrained, vision based capturing
devices are used to achieve both 2D and 3D hand gesture
models.
3) Haptics Capturing: Haptics or touch feedback refers to
provide touch feedback to the human hands or artificial hands
in this paper. The data glove and camera can get sufficient
information of human hand joint angles or positions, but they
are incapable of capturing the force of human manipulation.
Tactile sensors are employed to collect data and recognize
tactile discrimination of identifying the differences in texture,
temperature and shapes of objects through touch alone [20].
Usually haptic sensory information can be captured by either
use of force displays such as PHANToM to represent actual
friction force or changing friction properties of the contact
surface. Of a variety of tactile sensors very few can be applied
to provide accurate hand motion, Finger TPS system is widely
adapted with positive feedbacks as shown in Fig. 4. Each
FingerTPS system supports two hands with up to 6 sensors
per hand, it includes the sensors, wrist-mounted interconnect
harness, rechargeable wireless interface module, and USB
Bluetooth transceiver. In addition, Tekscan Inc. provides a
wide range of FlexiForce sensors which can be tailored into
different shapes subject to individual applications.

TABLE I
A DVANCED D ATA G LOVE P RODUCTS S ELECTED TO H IGHLIGHT THE S TATE OF THE A RT
Device Technology
DG5-VHand accelerometer and piezo-
resistive
5DT Glove 14 fiber optic
X-IST Glove piezo-resistive, pressure
sensor and accelerometer
CyberGlove II piezo-resistive
Humanglove hall-effect sensors
Shapehand glove bend sensors
Fig. 4. Finger tactile pressure sensor
4) EMG Signal Capturing: Electromyography (EMG) is
a technique involving testing, evaluating and recording the
activation signal of muscles including non-invasive captured
signal, i.e., sEMG, and invasive captured signal, ie., iEMG,
which can be used to indirect estimate manipulation force a
human hand applies. sEMG signals have been used mainly for
interacting with machines due to its representation of global
muscle activity. For instance, hand gestures are captured using
sEMG sensors in order to evaluate and record physiologic
properties of muscles at rest and while contracting [21]. On the
other hand, recent innovation in the reliable and implantable
electrodes, iEMG has gained more interests in the myoelec-
tric control. Additionally, Delsys Inc. provides a commercial
surface EMG sensors which provides more reliable signals
than conventional surface EMG sensors as shown in Fig.
5. It is worthwhile noting that there are inherent difficulties
in deriving a general model of the relationship between the
recorded EMG and muscle output force in humans performing
static and dynamic contractions.
III. H AND S KILL T RANSFER
Generally speaking, hand skill transfer is the process of
learning a new skill using aspects of other mastered hand-
centred manipulation skills. In the context of this paper, hand
skill transfer is defined as the act of learning/retrieving a
skill for object manipulation of artificial hands based on the
4
Sensors and locations Precision Speed
6 (a 3 axes accelerometer in wrist; one 10 bit 25 Hz
bend sensor per finger )
14 (2 sensors per finger and abduction 8 bit Minimum
sensors between fingers) 75 Hz
14 (Sensor selection: 4 Bend sensor 10 bit 60 Hz
lenghts, 2 pressure sensor sizes, 1 two-
axis accelerometer)
22 (three flexion sensors per finger, four 8 bit minimum
abduction sensors, a palm-arch sensor, 90 Hz
and sensors to measure flexion and ab-
duction)
20/22 (three sensors per finger and the 0.4 deg 50 Hz
dbduction sensors between fingers)
40 ( flexions and adductions of wrist, n.a. 100 Hz
fingers and thumb) maxi-
mum
Fig. 5. Surface EMG sensors [22]
knowledge/demonstration learned from human hand motion
capabilities. The skill transfer can be considered as a three-
stage process: a) It is to construct a knowledge base, con-
sisting of human hand motion primitives and manipulation
scenarios, from human hand manipulation demonstration or
dataset captured from hand movements; b) given a specific
hand manipulation scenario, human hand skill is applied to
artificial hands; c) software or hardware solution to bridge
the gap between constructed knowledge base and adaptation
of retrieved hand skills into artificial hands [23]. Thanks
to engineering robotics, it is relatively mature for adapting
motion trajectories into artificial hands [24]. It is evident
that it is a long-standing problem for bridge the gap of
transferring hand skills to artificial hands, hardcoded methods
are probably the only solution at an ad hoc basis in practice.
The problem has inherent connection with symbolic grounding
in a psychological context. Priority of this section is given to
the core technique of how to construct motion knowledge base,
i.e., how to recognize hand motion.
A. Hand Motion Recognition
It is evident that recognizing hand motion is challenging,
mainly, due to coupled temporal and spatial characteristics
[25]. Human hand motion can be categorized into static
gestures, dynamic gestures and in-hand manipulation. The
former two hand motions refer to hand poses and continuous
hand motion. The later involves manipulating an object within

a 11 a 22 a 33 a 44
a 12 a 23 a 34
S1 S2 S3 S4
Start
State
a 13 a 24 a 35
V1 V2 V3
Fig. 6. Grasp recognition of five states and four observations HMM [26]
one hand, in which the fingers and thumb are used to best
position the object for the activity. Due to the difficulty of in-
hand manipulation recognition, there are very few research
reported. It is even more challenging to recognize in-hand
motion in a real-time context. A set of algorithms had been
developed for this purpose and achieved it at a suboptimal
realtime basis [26]. It is evident that difficulties confronting
hand motion recognition can be solved efficiently by fusion
of multiple information sources such as tactile sensing and
vision.
Hand gesture recognition, in general, is composed of proba-
bilistic graphical models, neural networks, finite state automa-
ton, rule-based reasoning and ad hoc methods. Probabilistic
graphical models such as hidden Markov models (HMM) has
demonstrated its high potential in hand gesture recognition,
an example is given in Fig. 6. A survey for general gesture
recognition can be found at [25]. In order to provide readers a
comparative overview of hand gesture recognition for the pur-
pose of hand skill transfer, motion recognition algorithms are
organised in terms of capture devices, namely marker/vision
based sensors, tactile sensors and EMG-based capture devices,
in this paper. Note that recognition algorithms mimic the
process how humans recognize hand gestures in principle:
hand gestures are naturally recognized firstly by segmenting
the gesture sequences in terms of the start and end spatial-
temporal position points of a hand gesture sequence; then
humans compare the specified partitions with ‘experience’
in one’s mind, and determine the grasp types. According to
state of the art in hand motion recognition, this paper is
focused on recognition of static and dynamic gestures, in
theory, the algorithms should work for in-hand manipulation
recognition as well though with rather poor recognition rate
due to complexity of in-hand motion.
B. Marker based Gesture Recognition
Marker based capture devices usually consist of glove-based
systems and capture systems such as Vicon motion capture.
Its corresponding recognition methods can be categorized into
model-based methods and model-free methods. Model based
methods can be decomposed into those with/without hand
motion constraints.
Model-free approaches such as neural networks, rule-based
methods, etc. have been effective for dealing with hand dex-
terity and uncertainty occurred by a wide range of factors
5
a 55 such as high dimension raw sensory data. Glovetalk project
firstly successfully demonstrated a neural network interface
mapping hand gestures to sign language [27]; hand gesture
a 45
S5 recognition was achieved with recognition rate of 90.5% based
End on Self-Growing and Self-Organized Neural Gas network [28];
state support vector machine was also introduced to recognize hand
gestures with competitive results; a comparison of classifi-
cation methods of grasp recognition was intensively studied
V4 [29], it introduced a systematic approach for the evaluation
of classification techniques for recognizing grasps performed
with a data glove. In particular, it distinguishes between 6
settings that make different assumptions about the user groups
and objects to be grasped. On the other hand, fuzzy systems
have been outstandingly successful in the recent years. For
instance, Rainer Palm et.al. [30] employed CyberGlove to
record the human grasp data to capture grasp primitives and
modeled hand trajectories of grasp primitives by T-S fuzzy
modelling, and the research results had partially implemented
into prothetic hands.
Model-based methods are defined by matching the true ge-
ometry of the human hand, trade-off has to be handled properly
regarding to hand models and hand motion constraints in order
to achieve suitable combination of recognition accuracy and
algorithm efficiency. For instance, Sudderth et. al. introduced
a kinematic model in which nodes correspond to rigid bodies
and edges to joints in comparison with deformable point
distribution hand model [31]. It is accepted that a whole hand
kinematical structure can not be presented without constrains,
the primary advantage of considering the motion constraints
among fingers and finger joints is to greatly reduce the size or
dimensions of hand gesture search space. Motion constraints
are investigated for different reduction of hand DoFs and
learning from a large and representative set of training samples
[32]: a)limits of finger motions for static constraints; b) limits
of dynamic constraints imposed on joints during motion and
limits in performing natural motion.
Furthermore, Chua et al. reduced 27 DoFs of the human
hand model to 12 DoFs by analyzing the hand constraints in
terms of eight different types, these eight constraints could
be defined as “weak constraints” in both static analysis and
dynamic analysis [33]. One of the drawbacks is that the these
constraints could lead to invalid solution. In addition, a graph-
based approach was presented to understand the meaning
of hand gestures by associating dynamic hand gestures with
known concepts and relevant knowledge. Hand gestures are
understood by the comparison between concepts represented
by sets of hand-gesture elements and existing knowledge in
the form of conceptual graphs combining graph morphisms
and projection techniques [34].
It is evident that marker based methods offer much faster
and more precise recognition than vision based motion recog-
nition. It would be also interesting to see whether or not mo-
tion constraints could assist model-free approaches to achieve
better performance for both gesture recognition accuracy and
efficiency. Misusing hand motion constraints could lead to
rather poor performance since a large number of hand motion
constraints are difficult and computationally expensive being
represented in closed forms.

C. Vision based Motion Recognition
Hand recognition is composed of two stages in this section:
hand motion tracking and recognition. The motion tracking
is to separate hand motion from contextual background; the
recognition algorithms in section III-B, in theory, can be
directly applied to vision based hand motion recognition. The
vision based recognition is organized in terms of static models
and dynamic models. The two types of methods are overlapped
with generative and discriminative models of a more generally
accepted recognition terms.
Static methods are dominated by template matching and
its variants. 2D, 3D and depth-involved static hand models
are categorized in terms of multiple factors such as camera
types, camera viewpoints and occlusion, etc [35]. 2D hand
models usually depend on the extracted image parameters,
which are derived from the hand image properties including
contours and edges, image moments, image eigenvectors and
other properties. Eigenvalues indicating the hand width and
length was extracted in [36] to build a hand gesture recognition
system for real time America Sign Language in unconstrained
environments, while Haar-like features are extracted from 2D
hand images to recognize hand gestures. Additionally, other
features such as ’bunch graph’ in [37], ’conducting feature
points’ and ’grid of neurons’ were also extracted from 2D
image frames. Note that 2D hand models are robust to self-
occlusion since they extracted no 3D features and directly
compare the 2D image properties between the input images
and the registered ones, but images without shadowed fin-
gers can enhance the validity of 2D hand models [38]. 3D
hand models may lead to the computational complexity due
to inverse kinematics and 3D reconstruction. For instance,
reduction of hand DoFs was employed to efficiently estimate
3D hand postures from a set of eight 2D projected feature
points. On the other hand, static hand model has been in-
tensively studied using multiple cameras [39]. Not only do
employing multiple cameras overcome the difficulties such as
self/object occlusion, but it also provide a wider range of poses
and achieve more reliable 3D hand configuration. Combining
multiple view information usually refers to: a) group all the
shape contexts from all the images together before clustering
to build the histograms; b) estimate the pose from each view
individually and combine the results at a high level using
graphical models; c) the information is fused in hybrid terms of
the other two multi-view combination. Additionally, infrared
cameras are used to retrieve depth information in 3D hand
reconstruction via techniques such as stereo triangulation,
sheet of light triangulation, coded aperture, etc.
Dynamic hand models involving spatial-temporal features
can be generated by a range of methods such as probabilistic
graphical models, finite state automaton, etc. [40], an example
is given as in Fig. 7. It is practically feasible to adapt static
hand models into dynamic models by introducing temporal pa-
rameters into hand modelling. Probabilistic graphical models,
e.g., HMM, and finite state automaton are dominant methods
for dynamic models. HMMs was proposed to model dynamic
process in nature. Selected HMM-based research had been
reported as follows [19]. A recognition strategy was proposed
6
Fig. 7. Graphical model based hand gesture recognition [43]
by combining static shape recognition, Kalman filter based
hand tracking and a HMM based temporal characterization
scheme; A pseudo 3D Hidden Markov Model was introduced
to recognize hand motion, which gained higher rate of recog-
nition than a 2D HMM. [41]. Note that HMMs limited for
handling three or more independent processes efficiently. To
alleviate this problem, researchers had generalized HMMs into
dynamic Bayesian networks. Dynamic Bayesian networks are
directed graphical models of a stochastic process, representing
the hidden and observed states having complex interdepen-
dencies, which can be efficiently represented by the structure
of the directed graphical models. On the other hand, finite
state automaton is a model of behavior composed of a finite
number of states, transitions between those states and involved
actions. A finite state automaton can be represented by a state
transition diagram, in which the continuous stream sensory
data of a gesture are represented as a sequence of states. For
example, Yeasin et al. employed a finite state automaton for
automatically interpret a gesture accurately and also avoids
the computationally intensive task of image sequence warping
[42]. Note that using vision sensors only such as conventional
cameras usually lead to intensive computation, marker-based
systems enhance precision and efficiency of gesture recogni-
tion.
D. Haptics based Motion Recognition
The human hand utilizes the distributed receptors both on
the surface and inside of the finger, where the stimuli changes
with the finger movement, to acquire rich information on the
manipulated object. Haptic devices have been introduced to
capture such information in the dynamic interaction, which
can also be used as the unique/additional motion feature for
discriminating the hand motions [44]. Haptic devices have
also been integrated with data gloves to collect multiform
information such as haptic data glove and SlipGlove [45].
These gloves with haptic-IO capability provide the motion
information of human hands and enhance the capturing of
human hand skills. Kondo et al. used contact state transition
to recognize complex hand motions captured by Cyberglove
and a tactile sensor sheet, Nitta BIG-MAT quarter, attached to
the lateral side of the cylindrical object [46]. Kawasaki et al.
employed hand motions consisting of contact points, grasped
force, hand and object positions to explore the maximized
manipulability a robotic hand by scaling the virtual hand
model [44].
On the other hand, the contact interaction between the finger
tip and the object surface associates the different manipulated
objects with different hand motions. Some researchers have
studied the object recognition with complicated static/dynamic

tactile pattern obtained from multi-contact, finally to differen-
tiate the manipulations [47]. Dynamic haptic pattern refers to
the time serial of the haptic sensing change and is capable
of providing more haptic information than static haptic pat-
tern during the interactions between the hand and objects.
Watanabe et al. utilized tactile spatiotemporal differential
information with a soft tactile sensor array attached to a
robot hand to classify object shapes [48]. While Hosoda et al.
described the learning of robust haptic recognition of a bionic
hand using regrasping strategy and neural network, through
the dynamic interaction with the objects [49] .
One of the challenges remain in haptics based recognition,
however, is how to model hand friction and its effects. Existing
approaches are based on point-contact friction models, either
the point friction model, cone friction model or compensating
friction via control strategy. Both models are idealistic for
practical robot manipulation in that the contact between an
object and the corresponding robot hand is a surface instead
of a point. Force distributed sensors are introduced to capture
the magnitude of the force applied on a surface, but most
of them are incapable of achieving the force directions [50].
The area and the complexity of the hand interacting space
is limited due to the small size and resolution of the haptic
sensors. Additionally, locality of the haptic sensor is another
challenge, which depends heavily on the contact conditions
[51]. Although compliant joints and soft finger tips [49] have
been proposed to simulate the hand soft tissues, it is still an
open problem to imitate the non-linearly viscoelastic property
possessed by the finger soft tissues including inner skin and
subcutaneous tissue.
E. EMG based Hand Motion Recognition
There are two types of electromyogram: the intramuscu-
lar EMG, nEMG, and surface EMG, sEMG. The former
involves inserting a needle electrode through the skin into
the muscle whose electrical activity is to be measured; the
latter refers placing the electrodes on the skin overlying the
muscle to detect the electrical activity of the muscle. nEMG
is predominantly used to evaluate motor unit function in
clinical neurophysiology. nEMG can provide focal recordings
from deep muscles and independent signals relatively free
of crosstalk. Due to the improvement of the reliable and
implantable electrodes, the use of nEMG for human hand
movement studies has been more explored. Farrell et al.
[52] presented that the intramuscular electrodes has the same
performance as surface electrodes on pattern classification
accuracy for prosthesis control. Ernest et al. [53] proved that
a selective iEMG recording is representative of the applied
grasping force and can potentially be suitable for proportional
control of prosthetic devices.
sEMG signals have been used as a dominant method of
interaction with machines [54]. In an EMG-based interaction
system, hand gestures are captured using sEMG sensors which
evaluate and record physiologic properties of muscles at rest
and while contracting [21]. Various classification methodolo-
gies have been proposed for processing and discriminating
sEMG signals for hand motions. As a computation technique
7
that evolved from mathematical models of neurons and sys-
tems of neurons, neural network is becoming one of the
most useful methods. Other neural network based classification
algorithms include log-linearized Gaussian mixture network,
probabilistic neural network, fuzzy mean max neural network
and radial basis function artificial neural network [55]. The
statistic classifiers such as HMM, Gaussian mixture model,
support vector machine have also been used intensively in
sEMG recognition. A few studies has compared several dif-
ferent methods [56], e.g., Claudio et al. [57] reported support
vector machine achieved a higher recognition in compari-
son with neural networks and locally weighted projection
regression, while Liu proposed Cascaded Kernel Learning
Machine which has been compared to other classifers such
as k-nearest neighbour, multilayer neural network and support
vector machine [58]. However, none of them has explained
why the performance is enhanced. In addition, there is a lack of
consideration of sEMG’s uncertainties such as non-stationary
nature, muscle wasting, electrode position, different subjects
and temperature impact. Muscle wasting or muscle fatigue
can be considered as the decrease in the force generating
capacity of a muscle and has been evidenced in numerous
studies. For the same hand motion, muscle fatigue results in
a different sEMG signal which may cause a failure of the
recognition method. Electrode position is also critical for the
valid sEMG signal and leads to estimates of sEMG variables
that are different from those obtained in other nearby locations
[59]. Temperature is additionally proved to have an important
effect on the nerve conduction velocities and muscle actions
[60]. These uncertainties need more consideration in extracting
sEMG’s features which are determinant to the performance of
classifiers.
IV. H AND -C ENTRED A PPLICATIONS
In recent decades, with the developments and innovations
in motor technology and robotics, exciting fruit has been seen
in the design of physically artificial hands aiming to improve
the flexibility of robotized systems. Artificial hands can be
generally categorized into mechanical grippers, robotic hands,
rehabilitation hands and prosthetic hands. The mechanical
grippers, which are usually different from the mechanism
of the human hand, have been widely used in industrial
applications for fast and effectively grasping and handling
a limited set of known objects [61], [62]. They are usually
designed for a specific task, executing a preprogrammed mo-
tion trajectory and featuring low anthropomorphism and low
manipulation capability [63]. Robotic hand and prosthetic hand
are anthropomorphically designed to mimic the performance
of human hands, targeting to learn human hand skills with
an adaption in dynamic unstructured environment and even
be competent for the work which human is incapable of. A
survey on artificial hands focusing on manipulative dexterity,
grasp robustness and human operability can be found at [64],
an attempt to summarize up-to-date applications of artificial
hands is provided as below.
Various anthropomorphic robotic hands had been developed
or improved in the past ten years. Not only does the anthro-
pomorphic design make the skill transfer from human hand

(a) (b) (c)
Fig. 8. Robotic hands:(a) DLR-HIT hand; (b) Robonaut Hand; (c) Shadow
hand C6M
to robotic hand easier, but it also intend to equip the artificial
hands with the natural heritage of adaption in human living
environment. The newly developed or improved robotic hand
have more haptic sensors/feedbacks, such as DLR-HIT and
GIFU III hand with 6-axis force sensor and Shadow hand
with force and temperature sensor on the finger tip [65], [66].
They are relatively smaller and lighter but more powerful,
e.g., “Smart motor” actuation system has been introduced
to Shadow hand C6M instead of the pneumatic air muscle
actuation system, and integrated force and position control
electronics, motor drive electronics, motor, gearbox, force
sensing and communications into a compact unit. The appear-
ance has also been improved, for example, the bionic hand
in [49] has soft skin with distributed receptors. These robotic
hands can be generally categorized in two types: dexterous
hands and under-actuated hands in terms of the number of
actuators of a finger. Dexterous hands has multiple actuators
for each finger, which enables the robotic hand to control every
degree of freedom by individual actuator. For example in the
DLR-HIT hand has four fingers with four joints and three
actuators each as show in Fig. 8. However, the large amount of
actuators controlling DoFs makes the automatic determination
of their movement very difficulty, and the high dimensional
search space for a valid joint path makes the computational
cost extremely high [67]. Under-actuated hands have been
developed as one of the solutions for such problem, utilizing
less actuators in one finger than its degree of freedom. Such a
mechanism is aimed to decrease the number of active degrees
of freedom by means of connected differential mechanisms
in the system [68]. It is still a bottleneck problem of how to
achieve a proper mechanical design ensuring both the hand
effectiveness and efficiency.
(a) (b) (c)
Fig. 9. Prosthetic hands: (a) Myohand; (b) Ilimb; (c) Bebionic Hand
As one of the first application fields envisaged for artificial
8
anthropomorphic hands, for obvious aesthetic as well as func-
tional reason [63], prosthetic hand has gain intensive attention
in the past decade. The available commercial prosthetic hands
mainly include Myohand from Otto Bock Ltd, I-limb from
Touch Bionics [69] and Bebionic hand [70] in Fig. 9. They are
usually controlled by sEMG signals extracted from the users’
residual functional muscles. The users need a period of time
to get used to the system, trying to generate the recognizable
muscle contractions. So far only a few simple but robust
motions, such as opening and closure, have been applied in the
systems, though i-Limb can generate different hand gestures
by different combination of the active actuators. Changeable
force can be carried out to grasp objects with different weights
using limited actuators and surface electrodes. Though the con-
trol is simple and robust, a large amount of the amputees do not
use their prosthetic hand regularly mainly due to the weight,
appearance/cosmetics and functionality [71]. In addition, there
are no haptic or proprioceptive feedback to the subject in these
prosthetic hands. Tasks are carried out automatically by the
pre-programmed or pre-mapped desired trajectory triggered by
the continuous results of the EMG classifier. The only sensory
feedback is the user’s direct vision, in which the user can
stop or reset the task if it is not successful [72]. Besides the
commercial prosthetic hands, significant examples can also be
seen in the research applications: Cyberhand, DEKA arm, AR
hand and so on. Since it is evident that force feedback system
has a wide variety of effects on the users [73], force sensors
have been widely applied to prosthetic hands by means of
vibrotactile or electrotactiles stimulation. For example, a force
resistive sensor has been integrated in the index finger and the
thumb of FLUIDHAND III respectively, whose signals are
used to stimulate a vibration motor attached to users’ skin [74].
The user can sense the grasping force through the strength of
the vibration.
It should be noted that it is relatively mature applications
of artificial hands in virtual reality and gaming in comparison
with physically artificial hands, for which one of the main
reasons is that animating virtual hands has a lot less constraints
such as mechanical constraints. Vision-based methods have
been proposed to estimate joint locations and creases on the
palmar surface based on extracted features and analysis of sur-
face anatomy [75]. It is expected that simulating natural hand
motion and producing more realistic hand animation would
be of a great help to physically artificial hand manipulation
and synthesis of the sign languages. Since the human hand
is an important interface with complex shape and movement,
it is also expected that the use of an individualized rather
than generic hand representation can increase the sense of
immersion and in some cases may lead to more effortless and
accurate interaction with the virtual world.
V. C ONCLUDING R EMARKS
Programming multifingered artificial hands to solve com-
plex manipulation tasks has remained an elusive goal. The
complexity and unpredictability of the interactions of multiple
effectors with objects is an important reason for this difficulty.
The paper has reviewed the cycle of hand skill transfer from

analyzing the human hands to hand-centred applications. The
primary challenge that researchers now confront is how to
enable artificial hands to use what can be learned from human
hands, to manipulate objects, with the same degree of skill
and delicacy as human hands. The challenging issues in hand-
centred research can be summarized as: a) there is lack of
generalized framework which dynamically merges hybrid rep-
resentations for hand motion description rather than employing
hardcoded methods. For instance, how to autonomously inter-
pret the semantic meanings of a hand motion; b) there does
not exist a scheme which provides hand sensory feedback to
artificial hands. For instance, it is required a sensor or model
which generates necessary sensory feedback information for
haptics perception understanding, view-invariant hand motion
of artificial hands, e.g. a feasible slip model; c) existing
algorithms fail to recognize and model human manipulation in-
tention due to a variety of uncertainties, e.g., quality of sensory
information, individual manipulation habits and clinical issues;
d) It is evident that reliable interfacing/implanting into the
peripheral sensory nervous system and contextual information
such as environmental models is missing for further bridging
the gap in artificial hands and human/environment interaction;
e) feasible embedded algorithms is crucial, in terms of sen-
sory hand/context information fusion, to have artificial hands
operational and functional in human environments [9], [76]. It
is expected that this paper has provided a relatively unified
feasibility account for problems, challenges and technical
specifications for hand-centred research, considering all the
major disciplines involved, namely, the human hand analysis
and synthesis, hand motion capture, hand skill transfer and
hand-centred applications. This account of the-state-of-the-
art presented has also provided partial insights into an in-
depth understanding of human perception-action and potential
healthcare solutions.
R EFERENCES
[1] N. Palastanga, D. Field, and R. Soames, “Anatomy and Human Move-
ment - Structure and Function, Fifth Edition,” Elsevier Press, 2006.
[2] L. Dipietro, A. Sabatini, and P. Dario, “A Survey of Glove-Based
Systems and Their Applications,” IEEE Transactions on Systems, Man
and Cybernetics, Part C, vol. 38, no. 4, pp. 461–482, 2008.
[3] B. Argall and A. Billard, “A Survey of Tactile Human-Robot Inter-
actions,” Robotics and Autonomous Systems, vol. 58, pp. 1159–1176,
2010.
[4] A. Muzumdar, “Powered upper limb prostheses,” Springer, p. 208, 2004.
[5] Y. Matsuoka, P. Afshar, and M. Oh, “On the design of robotic hands for
brain–machine interface,” Neurosurgical Focus, vol. 20, no. 5, pp. 1–9,
2006.
[6] X. Zhu and H. Ding, “Computation of force-closure grasps: An iterative
algorithm,” IEEE Transactions on Robotics, vol. 22, no. 1, pp. 172–179,
2006.
[7] ——, “An efficient algorithm for grasp synthesis and fixture layout
design in discrete domain,” IEEE Transactions on Robotics, vol. 23,
no. 1, pp. 157–163, 2007.
[8] G. Gomez, A. Hernandez, P. Eggenberger Hotz, and R. Pfeifer, “An
adaptive learning mechanism for teaching a robot to grasp,” Proc.
International Symposium on Adaptive Motion of Animals and Machines,
pp. 1–8, 2005.
[9] A. Malinowski and H. Yu, “Comparison of embedded system design
for industrial application,” IEEE Transactions on Industrial Informatics,
vol. 7, no. 2, pp. 244–254, 2011.
[10] S. Fischmeister and P. Lam, “Time-aware instrumentation of embedded
software,” IEEE Transactions on Industrial Informatics, vol. 6, no. 4,
pp. 652–663, 2010.
9
[11] A. Quagli, D. Fontanelli, D. Greco, L. Palopoli, and A. Bicchi, “Design
of embedded controllers based on anytime computing,” IEEE Transac-
tions on Industrial Informatics, vol. 6, no. 4, pp. 492–502, 2010.
[12] A. Chella, H. Džindo, I. Infantino, and I. Macaluso, “A posture sequence
learning system for an anthropomorphic robotic hand,” Robotics and
Autonomous Systems, vol. 47, no. 2-3, pp. 143–152, 2004.
[13] M. Carrozza, G. Cappiello, S. Micera, B. Edin, L. Beccai, and C. Cipri-
ani, “Design of a cybernetic hand for perception and action,” Biological
cybernetics, vol. 95, no. 6, pp. 629–644, 2006.
[14] S. Calinon, F. Guenter, and A. Billard, “On learning, representing, and
generalizing a task in a humanoid robot,” IEEE Transactions on Systems,
Man and Cybernetics, Part B, vol. 37, no. 2, pp. 286–298, 2007.
[15] G. R. Bradberry, T.J. and J. Contreras-Vidal, “Reconstructing Three-
Dimensional Hand Movements from Non-Invasive Electroencephalo-
graphic Signals,” Journal of Neuroscience, in press.
[16] R. Gentner and J. Classen, “Development and evaluation of a low-cost
sensor glove for assessment of human finger movements in neurophys-
iological settings,” Journal of Neuroscience Methods, vol. 178, no. 1,
pp. 138–147, 2009.
[17] C. Metcalf, S. Notley, P. Chappell, J. Burridge, and V. Yule, “Validation
and Application of a Computational Model for Wrist Movements Us-
ing Surface Markers,” IEEE Transactions on BIomedical Engineering,
vol. 55, no. 3, pp. 1199–1210, 2008.
[18] T. Moeslund and E. Granum, “A survey of computer vision-based human
motion capture,” Computer Vision and Image Understanding, vol. 81,
no. 3, pp. 231–268, 2001.
[19] X. Ji and H. Liu, “Advances in view-invariant human motion analysis:
A review,” IEEE Transactions on Systems, Man and Cybernetics, Part
C, vol. 40, no. 1, pp. 13–24, 2010.
[20] C. Metcalf and S. Notley, “Modified Kinematic Technique for Measuring
Pathological Hyperextension and Hypermobility of the Interphalangeal
Joints,” IEEE Transactions on Biomedical Engineering, in press.
[21] C. Fleischer and G. Hommel, “Calibration of an EMG-Based Body
Model with six Muscles to control a Leg Exoskeleton,” Proc. Interna-
tional Conference on Robotics and Automation, pp. 2514–2519, 2007.
[22] “Delsys inc.” http://www.delsys.com.
[23] H. Liu, “A Fuzzy Qualitative Framework for Connecting Robot Qual-
itative and Quantitative Representations,” IEEE Transactions on Fuzzy
Systems, vol. 16, no. 6, pp. 1522–1530, 2008.
[24] T. Yoshikawa, “Multifingered robot hands: Control for grasping and
manipulation,” Annual Reviews in Control, vol. 34, pp. 199–208, 2010.
[25] S. Mitra and T. Acharya, “Gesture Recognition: A Survey,” IEEE
transactions on Systems, Man and Cybernetics, Part C, vol. 37, no. 3,
pp. 311–324, 2007.
[26] Z. Ju and H. Liu, “A unified fuzzy framework for human hand motion
recognition,” IEEE Transactions on Fuzzy Systems, in press.
[27] S. Fels and G. Hinton, “Glove-talkii - a neural-network interface which
maps gestures to parallel format speech synthesizer controls,” IEEE
Transactions Neural Networks, vol. 9, no. 1, pp. 205–212, 1998.
[28] E. Stergiopoulou and N. Papamarkos, “Hand gesture recognition using
a neural network shape fitting technique,” Engineering Applications of
Artificial Intelligence, vol. 22, pp. 1141–1158, 2009.
[29] G. Heumer, H. Amor, and B. Jung, “Grasp recognition for uncalibrated
data gloves: A machine learning approach,” PRESENCE: Teleoperators
and Virtual Environments, vol. 17, no. 2, pp. 121–142, 2008.
[30] R. Palm, B. Iliev, and B. Kadmiry, “Recognition of human grasps by
time-clustering and fuzzy modeling,” Robotics and Autonomous Systems,
vol. 57, no. 5, pp. 484–495, 2009.
[31] E. Sudderth, M. Mandel, W. Freeman, and A. Willsky, “Visual Hand
Tracking Using Nonparametric Belief Propagation,” Proc. International
Conference on Computer Vision and Pattern Recognition, pp. 189–189,
2004.
[32] J. Lin and T. Wu, “Modeling the Constraints of Human Hand Motion,”
Urbana, vol. 51, no. 61, pp. 801–809.
[33] C. Chua, H. Guan, and Y. Ho, “Model-based 3D hand posture estimation
from a single 2D image,” Image and Vision Computing, vol. 20, no. 3,
pp. 191–202, 2002.
[34] B. Miners, O. Basir, and M. Kamel, “Understanding hand gestures using
approximate graph matching,” IEEE Transactions on Systems, Man and
Cybernetics, Part A, vol. 35, no. 2, pp. 239–248, 2005.
[35] C. Chan and H. Liu, “Fuzzy qualitative human motion analysis,” IEEE
Transactions on Fuzzy Systems, vol. 17, no. 4, pp. 851–862, 2009.
[36] N. Binh, E. Shuichi, and T. Ejima, “Real-Time Hand Tracking and Ges-
ture Recognition System,” Proc. International Conference on Graphics,
Vision and Image, pp. 362–368, 2005.

[37] J. Triesch and C. von der Malsburg, “Classification of hand postures
against complex backgrounds using elastic graph matching,” Image and
Vision Computing, vol. 20, no. 13-14, pp. 937–943, 2002.
[38] S. Ge, Y. Yang, and T. Lee, “Hand gesture recognition and tracking based
on distributed locally linear embedding,” Image and Vision Computing,
vol. 26, no. 12, pp. 1607–1620, 2008.
[39] B. Stenger, A. Thayananthan, P. Torr, and R. Cipolla, “Estimating 3D
hand pose using hierarchical multi-label classification,” Image and Vision
Computing, vol. 25, no. 12, pp. 1885–1894, 2007.
[40] A. Just and S. Marcel, “A comparative study of two state-of-the-art
sequence processing techniques for hand gesture recognition,” Computer
Vision and Image Understanding, vol. 113, no. 4, pp. 532–543, 2009.
[41] N. Binh and T. Ejima, “Real-Time hand Gesture Recognition Using
Pseudo 3-D Hidden Markov Model,” Proc. International Conference on
Cognitive Informatics, vol. 2, pp. 1–8, 2006.
[42] M. Yeasin and S. Chaudhuri, “Visual understanding of dynamic hand
gestures,” Pattern Recognition, vol. 33, no. 11, pp. 1805–1817, 2000.
[43] J. Lin, “Visual hand tracking and gesture analysis,” Ph.D. Dissertation,
University of Illinois at Urbana-Champaign, 2004.
[44] H. Kawasaki, T. Furukawa, S. Ueki, and T. Mouri, “Virtual Robot
Teaching Based on Motion Analysis and Hand Manipulability for Multi-
Fingered Robot,” Journal of Advanced Mechanical Design, Systems, and
Manufacturing, vol. 3, no. 1, pp. 1–12, 2009.
[45] Z. Wang, J. Yuan, and M. Buss, “Modelling of human haptic skill: A
framework and preliminary results,” Proc. 17th IFAC World Congress,
pp. 1–8, 2008.
[46] M. Kondo, J. Ueda, and T. Ogasawara, “Recognition of in-hand manipu-
lation using contact state transition for multifingered robot hand control,”
Robotics and Autonomous Systems, vol. 56, no. 1, pp. 66–81, 2008.
[47] N. Gorges, S. Navarro, D. Goger, and H. Worn, “Haptic object
recognition using passive joints and haptic key features,” Proc. IEEE
International Conference on Robotics and Automation, pp. 2349–2355,
2010.
[48] K. Watanabe, K. Ohkubo, S. Ichikawa, and F. Hara, “Classification
of Prism Object Shapes Utilizing Tactile Spatiotemporal Differential
Information Obtained from Grasping by Single-Finger Robot Hand
with Soft Tactile Sensor Array,” Journal of Robotics and Mechatronics,
vol. 19, no. 1, pp. 85–96, 2007.
[49] K. Hosoda and T. Iwase, “Robust haptic recognition by anthropomorphic
bionic hand through dynamic interaction,” Proc. IEEE/RSJ International
Conference on Intelligent Robots and Systems, pp. 1236–1241, 2010.
[50] K. Sato, K. Kamiyama, N. Kawakami, and S. Tachi, “Finger-Shaped
GelForce: Sensor for Measuring Surface Traction Fields for Robotic
Hand,” IEEE Transactions on Haptics, vol. 3, no. 1, pp. 37–47, 2010.
[51] S. Takamuku, A. Fukuda, and K. Hosoda, “Repetitive grasping with
anthropomorphic skin-covered hand enables robust haptic recognition,”
Proc. IEEE/RSJ International Conference on Intelligent Robots and
Systems, pp. 3212–3217, 2008.
[52] T. Farrell and R. Weir, “A comparison of the effects of electrode im-
plantation and targeting on pattern classification accuracy for prosthesis
control.” IEEE Transactions on Biomedical Engineering, vol. 55, no. 9,
pp. 2198–2211, 2008.
[53] E. N. Kamavuako, D. Farina, K. Yoshida, and W. Jensen, “Relationship
between grasping force and features of single-channel intramuscular emg
signals,” Journal of Neuroscience Methods, vol. 185, no. 1, pp. 143–150,
2009.
[54] K. Wheeler, M. Chang, K. Knuth, N. Center, I. Div, and C. Moffett Field,
“Gesture-based control and EMG decomposition,” IEEE Transactions
on Systems, Man and Cybernetics, Part C, vol. 36, no. 4, pp. 503–514,
2006.
[55] F. Mobasser and K. Hashtrudi-Zaad, “A method for online estimation
of human arm dynamics,” Proc. IEEE International Conference on
Engineering in Medicine and Biology Society, vol. 1, pp. 2412–2416,
2006.
[56] S. Kawano, D. Okumura, H. Tamura, H. Tanaka, and K. Tanno,
“Online learning method using support vector machine for surface-
electromyogram recognition,” Artificial Life and Robotics, vol. 13, no. 2,
pp. 483–487, 2009.
[57] C. Castellini and P. van der Smagt, “Surface EMG in advanced hand
prosthetics,” Biological Cybernetics, vol. 100, no. 1, pp. 35–47, 2009.
[58] Y. Liu, H. Huang, and C. Weng, “Recognition of electromyographic sig-
nals using cascaded kernel learning machine,” IEEE/ASME Transactions
on Mechatronics, vol. 12, no. 3, pp. 253–264, 2007.
[59] L. Mesin, R. Merletti, and A. Rainoldi, “Surface EMG: The issue
of electrode location,” Journal of Electromyography and Kinesiology,
vol. 19, no. 5, pp. 719–726, 2009.
10
[60] J. Feinberg, “EMG: Myths and Facts,” HSS Journal, vol. 2, no. 1, pp.
19–21, 2006.
[61] X. Zhu, H. Ding, and M. Wang, “A numerical test for the closure prop-
erties of 3-d grasps,” IEEE Transactions on Robotics and Automation,
vol. 20, no. 3, pp. 543–549, 2004.
[62] X. Zhu and J. Wang, “Synthesis of force-closure grasps on 3-d objects
based on the q distance,” IEEE Transactions on Robotics and Automa-
tion, vol. 19, no. 4, pp. 669–679, 2003.
[63] L. Zollo, S. Roccella, E. Guglielmelli, M. Carrozza, and P. Dario,
“Biomechatronic design and control of an anthropomorphic artificial
hand for prosthetic and robotic applications,” IEEE/ASME Transactions
on Mechatronics, vol. 12, no. 4, pp. 418–429, 2007.
[64] A. Bicchi, “Hands for dexterous manipulation and robust grasping: A
difficult road toward simplicity,” IEEE Transactions on Robotics and
Automation, vol. 16, no. 6, pp. 652–662, 2002.
[65] H. Liu, P. Meusel, N. Seitz, B. Willberg, G. Hirzinger, M. Jin, Y. Liu,
R. Wei, and Z. Xie, “The modular multisensory DLR-HIT-Hand,”
Mechanism and Machine Theory, vol. 42, no. 5, pp. 612–625, 2007.
[66] H. Kawasaki, T. Komatsu, and K. Uchiyama, “Dexterous anthropo-
morphic robot hand with distributed tactile sensor: Gifu hand II,”
IEEE/ASME Transactions on Mechatronics, vol. 7, no. 3, pp. 296–303,
2002.
[67] S. Sun, C. Rosales, and R. Suarez, “Study of coordinated motions of
the human hand for robotic applications,” Proc. IEEE International
Conference on Information and Automation, pp. 776–781, 2010.
[68] T. Laliberte and C. Gosselin, “Simulation and design of underactuated
mechanical hands,” Mechanism and Machine Theory, vol. 33, no. 1-2,
pp. 39–57, 1998.
[69] C. Connolly, “Prosthetic hands from Touch Bionics,” Industrial Robot:
An International Journal, vol. 35, no. 4, pp. 290–293, 2008.
[70] Bebionic, “Bebionic hand,” http://www.bebionic.com/, 2011.
[71] P. Kyberd, C. Wartenberg, L. Sandsoj, S. Jonsson, D. Gow, J. Frid,
C. Almstrom, and L. Sperling, “Survey of upper-extremity prosthesis
users in Sweden and the United Kingdom,” Journal of Prosthetics and
Orthotics, vol. 19, no. 2, pp. 55–62, 2007.
[72] C. Cipriani, F. Zaccone, S. Micera, and M. Carrozza, “On the shared con-
trol of an EMG-controlled prosthetic hand: analysis of user–prosthesis
interaction,” IEEE Transactions on Robotics, vol. 24, no. 1, pp. 170–184,
2008.
[73] C. Pylatiuk, A. Kargov, and S. Schulz, “Design and evaluation of a low-
cost force feedback system for myoelectric prosthetic hands,” Journal
of Prosthetics and Orthotics, vol. 18, no. 2, pp. 57–61, 2006.
[74] I. Gaiser, C. Pylatiuk, S. Schulz, A. Kargov, R. Oberle, and T. Werner,
“The FLUIDHAND III: A Multifunctional Prosthetic Hand,” Journal of
Prosthetics and Orthotics, vol. 21, no. 2, pp. 91–97, 2009.
[75] T. Rhee, U. Neumann, and J. Lewis, “Human Hand Modeling from
Surface Anatomy,” Proc. Symposium on Interactive 3D Graphics and
Games, pp. 1–6, 2006.
[76] N. Motoi, M. Ikebe, and K. Ohnishi, “Real-time gait planning for
pushing motion of humanoid robot,” IEEE Transactions on Industrial
Informatics, vol. 3, no. 2, pp. 154–163, 2007.
Honghai Liu (M’02-SM’06) received his Ph.D de-
gree in robotics from King’s college London, UK,
in 2003. He joined the University of Portsmouth,
UK in September 2005. He previously held re-
search appointments at the Universities of London
and Aberdeen, and project leader appointments in
large-scale industrial control and system integration
industry.
Dr Liu has published over 250 peer-reviewed
international journals and conference papers includ-
ing four best paper awards. He is interested in
approximate computation, pattern recognition, intelligent video analytics and
cognitive robotics and their practical applications with an emphasis on
approaches which could make contribution to the intelligent connection of
perception to action using contextual information. He is Associate Editor of
IEEE Transactions on Industrial Informatics, IEEE Transactions on Systems,
Man and Cybernetics, Part C and International Journal of Fuzzy Systems.