ArticlePDF Available

Mechanisms of the Anatomically Correct Testbed Hand

Authors:

Abstract and Figures

We have built an anatomically correct testbed (ACT) hand with the purpose of understanding the intrinsic biomechanical and control features in human hands that are critical for achieving robust, versatile, and dexterous movements, as well as rich object and world exploration. By mimicking the underlying mechanics and controls of the human hand in a hardware platform, our goal is to achieve previously unmatched grasping and manipulation skills. In this paper, the novel constituting mechanisms, unique muscle to joint relationships, and movement demonstrations of the thumb, index finger, middle finger, and wrist of the ACT Hand are presented. The grasping and manipulation abilities of the ACT Hand are also illustrated. The fully functional ACT Hand platform allows for the possibility to design and experiment with novel control algorithms leading to a deeper understanding of human dexterity.
Content may be subject to copyright.
1
Mechanisms of the Anatomically Correct Testbed
(ACT) Hand
Ashish D. Deshpande*, Zhe Xu*, Michael J. Vande Weghe*, Benjamin H. Brown, Jonathan Ko, Lillian Y. Chang,
David D. Wilkinson, Sean M. Bidic and Yoky Matsuoka
Abstract—We have built an Anatomically Correct Testbed
(ACT) hand with the purpose of understanding the intrinsic
biomechanical and control features in human hands that are
critical for achieving robust, versatile, and dexterous movements
as well as rich object and world exploration. By mimicking the
underlying mechanics and controls of the human hand in a
hardware platform our goal is to achieve previously unmatched
grasping and manipulation. In this paper the novel constituting
mechanisms, unique muscle to joint relationships, and movement
demonstrations of the thumb, index finger, middle finger, and
wrist of the ACT Hand are presented. The grasping and ma-
nipulation abilities of the ACT Hand are also illustrated. The
fully functional ACT Hand platform allows for the possibility to
design and experiment with novel control algorithms leading to
a deeper understanding of human dexterity.
*These authors are all first authors and the author list is in
reverse chronological order in contribution.
I. INT ROD UC TI ON
Human hands can perform many dexterous grasping and
manipulation tasks. Hand dexterity is the ability to precisely
control movements and forces using all of the hand’s degrees
of freedom to perform a variety of tasks. Examples include the
ability to play musical instruments, use chopsticks, gesture,
and perform daily tasks such as cooking and writing.
Researchers have been designing robotic hands for more
than four decades [51], [54], [35], and while many advance-
ments have been made in the newer robotic hands, the grasping
and manipulation abilities of the current robotic hands do
not match the versatile dexterity of the human hand. Most
of the existing robotic hands are designed to meet specific
task requirements, for example, prosthetic robot hand are de-
signed to achieve basic grasping with low-weight mechanisms
while hands for industrial applications are designed to handle
specific parts. Striving for human-like capabilities, current
robotic hands have been designed to be anthropomorphic,
having fingers and an opposable thumb with human-like
shapes and degrees of freedoms (DOF). However, the actual
mechanisms of actuation and controls, in most cases, have not
been anatomical in these prostheses. This is in part due to
lack of clear understanding of biomechanical features of the
human hand and also due to the difficulty in translating human
features in a machine form.
Versatile and robust dexterity are achieved in the human
hand through a combination of hand biomechanics, which can
be thought as the ‘hardware’ and neural controls, which can
be thought as the ‘software’ of the hand. To design a robotic
hand that achieves human-like dexterity, we have embarked
upon a distinct approach toward robotic hand design. We have
Fig. 1. The Anatomically Correct Testbed (ACT) Hand.
constructed the Anatomically Correct Testbed (ACT) Hand,
as shown in Figure 1 in which the mechanical elements are
designed to mimic the intricate features of hand biomechanics,
including bone structures and tendon arrangements, and the
software controls are based on the human hand neuromuscular
control system. With these attributes the ACT Hand open up
the hitherto unavailable possibility of addressing critical design
questions about human hand functionality.
Our long term goals are to achieve human-like dexterity in
the robotic form, and to enhance the understanding of human
hand functionality through experiments with the developed
robotic platform. Toward these goals we have accomplished a
critical step which is the development of a robotic hand with
mechanical parts that closely imitate the biomechanical fea-
tures of the human hand. Previous researchers have attempted
to develop an understanding of human hand biomechanics and
controls by either conducting experiments with cadaver hands,
or by developing computational models of hand biomechanics
and movements. However, since the most critical human
hand features are exposed only through dynamic physical
interactions with objects, and since the nonlinear interactions
between the muscles, tendons, bones and joints are extremely
challenging to recreate in computational models, a physical
realization – in the form of a detailed anatomical model – is
necessary to define and analyze the human hand features. None
of the existing hands are designed to possess the hardware and
software intricacies to tackle the important research questions
about dexterity of hands, while the ACT Hand has the potential
to achieve human-like muscle control.
The purpose of this paper is to present our novel ideas
for the constituting mechanisms, and summarize the devel-
opments and results in the ACT Hand design. We present
2
the constituting mechanisms. One key advantage of human-
like mechanisms is that the robotic hand possesses nonlinear,
nonconstant relationships between the actuators (muscles), and
joints. We develop mathematical models for these relations
based on the data collected from experiments with the ACT
Hand. We demonstrate that the moment arm variations match
with the known relations in the human hand. Moreover, our
model is more comprehensive than the existing models for the
human hand thus leading to a better understanding of human
hand biomechanics. The moment arm relations are key for
achieving reliable position control of the digits of the ACT
Hand. We present results from position tracking of the index
finger and thumb.
II. RE LATE D WO RK
Robotic hands have been developed for many decades for a
variety of applications ranging from industrial manufacturing
to prosthetics to humanoid robot research. Although numerous
grasping and manipulation results have been demonstrated by
existing anthropomorphic hands (e.g. fast manipulation by the
hand developed by the University of Tokyo [57]), the capa-
bilities of the current robotic hands do not match the abilities
of the human hand. Design of a versatile and robust robotic
hand that demonstrates human-like grasping and manipulation
capacity is a challenging task. The design decisions include
the number of fingers, number of joints, degrees of freedom
(DOFs), range of motion for the joints, speed of movements
and force generation capacity. The design choices have to
be made under the tight space and weight constraints. Focus
has been on increasing the number of DOFs, adding sensors,
and implementing novel controls, including machine learning
methods. Table I presents a list of representative robotic hands
with important characteristics of these hands. Note that in
many of cases the robotic hands are still under development
so the grasping and manipulation abilities are based on the
design specification not from actual demonstrations.
Many robotic hands are driven directly or through gears,
for example, the Stanford/JPL Hand [54], Barrett Hand [64],
the Southampton Hand [42], the Gifu Hand II [38] , the
NASA/JPL Robonaut Hand [11], the NAIST hand [67], and
the high speed hand from the University of Tokyo [57] (the
drive mechanisms for the recently developed DEKA arm-hand
prosthesis [24] have not been published). Examples of tendon-
driven hands include: the Okada Hand [51], the Utah/MIT
Hand [35], [29], [61], the DLR II Hand [43], the UB Hand
3 [45], the Vanderbilt Hand [22], the CyberHand [15], the
Karlsruhe Hand [36], [64], the smart motor and air muscle
Shadow Hands [58], and the Keio Hand [73].
The idea of designing smart mechanisms to simplify con-
trols has been explored in some of the existing hands. For
example, in the under-actuated grippers such as the SDM
Hand [27] and SPRING Hand [17], joint compliance is
implemented to be able to conform to different shapes for
grasping with relatively simple controls. Some researchers
have developed soft skin to accommodate against errors and
to embed sensors [56], [8]. However, the ability of these hands
to execute human-level manipulation tasks remains limited.
Despite the desire for dexterity in prosthetics, the most
commonly used prosthetic hand is a mechanically controlled
hook prosthesis [34] which was designed over a century ago.
Several researchers [22], [50], [16] and companies [63], [47],
[52] have designed robotic hands specifically for prosthetic
purposes with attention toward minimizing weight, simplifying
controls, and aesthetics. Some of the current commercial
prostheses are controlled by means of electromyographic
(EMG) signals recorded using surface electrodes, which detect
electrical activity related to the patient’s arm muscles [63],
[52]. Because of the difficulty in translating the user intent
into useful controls signals, current prosthetic hands have
only one or two DOFs; thus leading to limited functionality.
Some surveys reveal that 30 50% of the upper extremity
amputees do not use their prosthetic hand regularly because of
reduced functionality, poor cosmetic appearance, and limited
controllability [3], [59], [16]. Ideas for controlling of robotic
hands using neural signals have also been explored. Recent
studies enable monkeys to control the 3D movement of a
robotic arm to achieve self-feeding tasks [49], [69]. Over thirty
human arm/hand amputees have received nerve reinnervation
surgery to rewire the peripheral nerves that used to go into
the hand/arm to the chest muscle instead [40]. The signals
amplified by the natural muscle can then be tapped into with
surface EMG for prosthetic arm/hand control.
A number of researchers have developed computational
models of hand biomechanics [25], [33]. While some of the
recent models do include human-like muscles, tendons and
bones [33], [62], [65], these models still face difficulties
in simulating critical nonlinear relationships, for example,
tendons sliding over bones, and nonlinear joint movements.
For instance Sueda et al. present an automatic technique for
generating the motion of tendons and muscles [62], and Tsang
et al. present an anatomically accurate inverse dynamics of
the hand [65]. However, these methods do not represent the
nonlinear moment arm relationships for the hand muscle which
are critical in defining biomechanics of the hand. Considering
the challenges in modeling nonlinear relations and advan-
tages gained through experiments with physical prototypes,
we believe that it is important to continue developing both
computation and physical models of human hand.
To improve the performance and capabilities of the robotic
hands new ideas have to be introduced for both the robot
hardware and controls. Robotics researchers can greatly ben-
efit from a better understanding of the biomechanics and
neuromuscular controls of the human hand. Although many
robotic hands are designed to be anthropomorphic, the intrinsic
mechanisms of actuation and controls, in most cases, have not
been anatomical. In this context, the ACT Hand is designed
to be a tool to investigate human dexterity. By incorporating
the biomechanical features of the human hand the ACT Hand
allows for the identification of the critical factors that lead to
dexterity in the human hand.
1T, I, M, R and P denote thumb, index, middle, ring, and little finger,
respectively.
3
TABLE I
FEATU RES O F EX EMP LA RY ROBO TIC H AN DS
Robotic hands #
identical
fingers
# joints / DOF
(Total DOFs)1
Range of
motion
Speed of
motion
Activation / transmission
method
Types of grasps / ma-
nipulation
Hosmer hook [34] 2 split
hooks
1/1 (1 DOF) <human
hand
<human
hand
body-powered splitting hook pinch
Utah Arm / Liberating /
OttoBock [47], [44], [52]
2 T-1/1, I-1/0, M-1/0
(1 DOF)
<human
hand
<human
hand
EMG signal driven, DC
motor, cable
three-finger pinch
USC/Belgrade [7] 4 T-3/2, I-3/0.5,
M-3/0.5, R-3/0.5,
P-3/0.5 (4 DOFs)
<human
hand
<human
hand
DC motor, cable, linkage grasp: power & finger tip
Harvard SDM [28] 4 4×2/1 (1 DOF) >human
hand
<human
hand
DC motor, cable, elastic
joints
enveloping grasp
Gatech Dusty [72] 1 2/1 (1 DOF) <human
hand
DC motor, cable, spring
hinge joints
nonprehensile grasp
Barrett [64] 3 T(Right)-2/1.5,
T(Left)-2/1.5, I-2/1
(4 DOFs)
>human
hand
1.2×human
hand
DC motor, worm drives
integrated with cable drive
and breakaway clutch
grasp: power & finger tip
i-Limb [63] 4 T-3/1, I-2/1, M-2/1,
R-2/1, P-2/1 (5
DOFs)
<human
hand
human
hand
DC motors, belt transmission grip: key, hook, power &
precision;
grasp: spherical & pal-
mar
Southampton [42] 4 T-2/2, I-3/1, M-3/1,
R-3/1, P-3/1 (6
DOFs)
<human
hand
0.22×human
hand
DC motor, worm-wheel, lead
screw
power grasp, lateral
pinch
Cyber [15] 5 T-4/2, I-3/1, M-3/1,
R-3/1, P-3/1 (6
DOFs)
0.22×human
hand
0.38×human
hand
geared DC motor, lead
screw, cable, exensor spring
lateral pinch;
grasp: cylindrical, spher-
ical & tripod
Univ. of Tokyo Hand [57] 3 T(R)-3/3, I-2/2,
T(L)-3/3 (8 DOFs)
>human
hand
=15×human
hand
DC motor, harmonic and
bevel gear transmission
grasp: power & finger
tip;
dynamic manipulation
Stanford/JPL [54] 3 T-3/3, I-3/3, M-3/3
(9 DOFs)
>human
hand
DC motor, cable finger tip grasp
DARPA hand [23] 4 T-3/3, I-3/2, M-3/2,
R-3/2, P-3/2 (11
DOFs)
<human
hand
<human
hand
DC motor, cable, gear
transmission
grasp: hook & power
Robonaut [46] 4 T-5/3, I-4/3, M-4/3,
R-3/3, P-3/1 (11
DOFs)
human
hand
<human
hand
DC motor, flex shaft, lead
screw, cable
grasp: power & finger
tip;
lateral pinch
Naist [67] 4 T-4/3, I-4/3, M-4/3,
R-4/3 (12 DOFs)
human
hand
human
hand
geared DC motor, bevel gear power grasp
DLR II [14] 4 T-4/4, I-4/3, M-4/3,
R-4/3 (13 DOFs)
>human
hand
=3×human
hand
DC motor, belt, harmonic
drive, bevel gears
grasp: power & finger
tip;
lateral pinch
Utah/MIT [35] 4 T-4/4, I-4/4, M-4/4,
R-4/4 (16 DOFs)
<human
hand
1.82×human
hand
pneumatic actuator, cable finger tip grasp
/manipulation
Gifu III [48] 4 T-4/4, I-4/3, M-4/3,
R-4/3, P-4/3 (16
DOFs)
human
hand
1.35×human
hand
DC motor, gear transmission,
linkage mechanism
power grasp
UB III [45] 4 T-3/4, I-4/4, M-4/3,
R-4/2, P-4/3 (16
DOFs)
<human
hand
0.51×human
hand
DC motor, cable, helical
spring
grasp: power & finger tip
Shadow [58] 4 T-5/5, I-4/3, M-4/3,
R-4/3, P-4/3 (17
DOFs)
human
hand
0.5×human
hand
air muscle, cable and spring grasp: finger tip & power
Keio [73] 4 T-4/4, I-4/4, M-4/4,
R-4/4, P-4/4 (20
DOFs)
human
hand
2×human
hand
ultrasonic motors, elastic
elements, cable
grasp: power & finger
tip;
lateral pinch
4
III. ACT HAND MEC HA NI SM S
This section describes the mechanical design and fabri-
cation details of the ACT Hand. Overall, the focus is on
mimicking the intrinsic biomechanics, actuation and control
behavior to achieve human-like dynamic movements, rather
than minimizing the total weight of the hand. The following
subsections describe the mechanical components in the ACT
Hand, including its finger bones, joints, tendons and actuators.
A. ACT Hand Finger Bones
In an early version of the ACT Hand the fingers were
designed with cylindrical bones [71] with the assumption that
a biological tendon arrangement would lead to anatomical
performance even with engineering shapes for the bones.
However, experiments with this design revealed that the bi-
ological shapes of the human finger bones create moment
arms for the tendons that vary with joint angle, a behavior
critical for accurate hand function [1], [2], [68], [18]. The
variable moment arms are necessary for achieving human-like
joint-muscle movement relationships. The mass and inertia
properties of the bones also affect the dynamic behavior of the
fingers. To address these issues, we designed the finger bones
by accurately matching the size, shape and mass properties of
human bones. We used Stratasys Corporation’s existing laser-
scan model of human left hand bones supplied in STL format,
imported the tesselate facets into Pro/Engineer, and created
solid models for each bone by fitting new surfaces to the scan
geometry. The lengths of the phalanges are given in Table II.
The composition of the finger bones was designed with two
primary goals in mind: ease of manufacturing the complex
surface shapes, and high strength at the joints and tendon
attachment points. An earlier version of the ACT Hand utilized
cast aluminum and 7075 aluminum bones, which were difficult
to fabricate and offered only moderate strength, especially, to
hold the small threaded fasteners used at the tendon insertion
points. The current version uses innovative design with two
separate components. The core of the bones is comprised
of a set of steel beams, which offers superior strength and
durability; and, although not easy to machine, the beams are
much more straightforward to fabricate than the complex sur-
face shapes in the previous design. Attached to the beams are
outer plastic shells fabricated using stereolithography. Because
the shells are only used in compression in our application
the strength provided by the plastic material is sufficient.
The stereolithography manufacturing process makes it easy
to experiment with modified surface geometries and replace
the shells when they become worn or broken.
B. ACT Hand Finger Joints
The design of the finger joints plays a critical role in
matching ACT Hand kinematics with human kinematics. The
locations of the degrees of freedom (DOF) and axes of rota-
tions, and the ranges of motion (ROM) for all the finger joints
are debated in the biomechanics literature. In the ACT Hand
we have mimicked the most broadly accepted biomechanical
model of DOF and ROM for the fingers [13], [31], [32]. All
TABLE II
ACT H AND P HA LAN GE L ENG TH S
Finger Phalange Length (cm)
Index MCP to PIP 5.10
PIP to DIP 2.69
Distal phalange 1.55
Middle MCP to PIP 5.38
PIP to DIP 3.58
Distal phalange 1.80
Thumb CMC Flex to CMC Ab-Ad 2.31
CMC Ab-Ad to MCP 4.31
MCP to PIP 3.65
Distal phalange 2.00
fingers have four DOFs, while the opposable thumb has five
non-orthogonal, non-intersecting DOFs. The base of the ring
and little fingers in the palm has additional DOFs.
Thumb, index and middle fingers are actuated by anatom-
ically routed tendons and muscle-equivalent actuators. All
fingers can hyper-extend, as can human fingers. We chose
to defer the completion of actuation for the last two fingers
until we investigate the performance of multi-finger object
manipulations with three fingers. The wrist has two DOFs, and
all finger tendons are routed with moment arms preserved, i.e.,
wrist movements influence finger movements, as for humans.
In first version of the ACT Hand, joint design was based
on human joint biomechanics; ligaments connect bones and
create a joint capsule filled with cartilage and synovial fluid
to achieve a low friction joint [10]. Although mimicking
joint geometry yielded accurate motion vectors it resulted in
high friction and reduced range of motion. In the current
design, we implemented the joints as machined pin joints
and sought to align the joint axes to best approximate the
more complex motion of each human finger joint. In some
cases, we discovered that joints which at the first glance might
appear to be 3-DOF ball-and-socket joints were actually better
represented by two carefully aligned pin joints. Figures 2 and
3 show the CAD models for the index finger and thumb.
Fig. 2. ACT Hand index finger bones are made of two materials. The outer
shell, made of plastic, matches the human shape and size, while the inner
steel beam structure allows for anatomical joints.
5
1) Index and Middle Finger Joints: There are three joints in
the index and middle finger, namely, the metacarpophalangeal
(MCP), proximal interphalangeal (PIP) and distal interpha-
langeal (DIP). These are modeled by a novel design involving
pin joints in the ACT Hand. The PIP joint is located at the
distal end of the proximal phalangeal bone, and the DIP joint
is located at the distal end of the middle phalangeal bone. The
MCP joint has two DOFs: one to achieve flexion-extension and
another to realize abduction-adduction finger motion. These
two DOFs are realized by a gimbal mechanism at the distal
end of the metacarpophalangeal bone. To match the anatomical
joint properties of the human index finger, as described in [12],
the abduction-adduction joint axis is oriented at 60with
respect to the metacarpophalangeal bone as shown in Figure
2.
Fig. 3. Bones and joints in the thumb of the ACT thumb. The thumb has three
joints, namely, the carpometacarpal (CMC), metacarpophalangeal (MCP), and
interphalangeal (IP), and five DOFs.
2) Thumb Joints: The three thumb joints are the car-
pometacarpal (CMC), metacarpophalangeal (MCP), and in-
terphalangeal (IP) joints. The IP joint possesses one rotation
DOF in the flexion-extension direction. The DOF and joint
axes locations for the CMC and MCP joints are debated
in hand biomechanics research. Recent research shows that
defining two DOFs with two non-intersecting, non-orthogonal
axes [31], [32], [55], [60] instead of one universal joint at
each of these joints [21], [37] leads to more anatomically
correct thumb behavior. In the ACT Hand, the two DOFs at
the CMC joint are realized by two non-perpendicular, non-
intersecting pin joints, and the two DOFs at the MCP joint
are realized by a gimbal mechanism, as shown in Figure 3.
Supported by a pair of miniature ball bearings, the gimbal
piece rotates around the MCP abduction-adduction (AA) axis
fixed within the metacarpal bone. A small pin joint in the
gimbal piece represents the MCP flexion-extension (FE) axis,
which is fixed relative to the proximal phalange via a link
arm. The sweep of the joint cavity restricts the movement of
the gimbal assembly to the appropriate MCP joint range of
motion.
The CMC joint involves two pin joints at the ends of a single
link arm to realize the AA and FE degrees of freedom. Though
the CMC and MCP joints are conceptually similar in that they
both have FE and AA degrees of freedom, a gimbal design is
not suitable for the CMC joint because its two rotational axes
are located in separate bones. One pin joint coincides with
the CMC AA axis in the proximal end of the metacarpal,
while the other pin joint represents the CMC FE axis, which
intersects the trapezium carpal bone. Joint range of motion for
each of the two axes is constrained by narrow slot cuts in the
metacarpal and trapezium bones. The IP joint design consists
of single pin joint to represent the flexion-extension degree of
freedom between the two phalangeal bones. A link arm rigidly
attached to the distal phalange rotates about an axle coinciding
with the IP FE axis in the proximal phalange. The geometry of
the articulating bone ends was maintained except for a narrow
slot that allows the small diameter link arm to rotate around
the IP FE axis pin. The span of the cavity enforces the joint
range of motion.
3) Joint Range of Motion: Joint limits for the flexion joints
are imposed by creating internal beam features that interfere
with one another at the limits. The ranges of motion were
chosen to match those of the human finger, and are shown
in Table III. Because the gimbals used for the MCP joints
are free to rotate by ±180, we designed the bone shells to
limit the range of motion by contacting each other at the joint
limits. For example, Figure 2 shows the CAD model of the
MCP joint of the ACT Hand’s index finger with the bone shell
around the gimbal joint to ensure the correct range of motion.
TABLE III
ACT H AND FI NG ER JO IN T MOT ION L IM ITS
Finger Joint Minimum Maximum
Index MCP 30extension 90flexion
35abduction 35adduction
PIP 0extension 110flexion
DIP 0extension 70flexion
Middle MCP 30extension 90flexion
35abduction 35adduction
PIP 0extension 110flexion
DIP 0extension 70flexion
Thumb CMC 40extension 40flexion
40abduction 40adduction
MCP 60extension 60flexion
15abduction 15adduction
IP 20extension 80flexion
C. ACT Hand Tendons
Two types of muscles control hand movements: (1) those
located in the palm, called the intrinsic muscles, and (2)
those located in the forearm, called the extrinsic muscles.
The muscles are connected to the bones by long tendons that
6
pass over the joints, terminating at the insertion points on the
finger bones. Muscle contractions lead to hand movement and
force generation. In the ACT Hand, we used the brushless DC
motors (described later in the paper) as muscles and we fabri-
cated our tendons with 0.46 mm Spectra(R) strings. The string
was chosen because of its strength (200 N breaking strength),
stiffness (4800 N/strain), and ability to slide smoothly over
the bones. The tendons in the human hand have elastic
properties [74] which play a critical role in the hand dexterity
and controls. The choice of a string with high stiffness allows
for reliable and quantifiable relationships between the motors
and joints. Mimicking the stiffness properties of the human
tendons, either through hardware or software controls is part
of our future work.
1) Tendon Routing over the Wrist Joint: The human wrist
position affects finger joint properties (for example, joint
stiffness) because the tendons connecting the extrinsic muscles
to the fingers travel past the wrist joint, and the joint position
changes their path lengths. While designing the wrist for the
ACT Hand, we faced a tradeoff of minimizing tendon friction
versus routing tendons to be accurately affected by wrist
position. We chose to reduce tendon friction and explicitly
designed a routing to minimize the effect of wrist position on
tendon length. As shown in Figure 4, each tendon crosses the
wrist joint via a pair of sheaves: a large central sheave near
the flexion/extension axis and a smaller outer sheave. The size
and placement of the tendon sheaves have been chosen so that
during wrist flexion/extension, the tendons unwind from one
sheave as they are being wound onto the other, minimizing
any change in path length. Wrist abduction/adduction has
a greater effect on path length due to the line of central
sheaves rotating out of the plane of the outer sheaves, but
the effect is small, particularly for tendons routed near the
abduction/adduction pivot point. Since the tension in each
tendon is under independent software control, it is possible
to undo the abduction/adduction effects and apply the desired
flexion/extension effects as a function of wrist position.
Fig. 4. Wrist joint mechanism with two degrees of freedom
2) Extensor Tendon Hood: On the dorsal side of the fingers,
the tendons are connected from the bone insertion points to
the actuators via an elaborate extensor tendon web. These
networks of tendons play a central role in defining the biome-
chanics and control of the digits [1], [12]. To closely match
the moment arm variations in the human hand, we developed
a web of tendons that mimics the human structure with the
spectra strings. The current version of the extensor mechanism
is fabricated by crocheting nylon composite to emulate the
geometry and functionality of the human counterpart as closely
as possible as shown in Figure 5. We prioritized our tendon
structure design to achieve any posture that a typical human
finger can achieve, mimic the overall geometry based on [30],
produce a smooth surface to facilitate a sliding motion, and
match the stiffness of a real tendon. The critical features of
the extensor tendon web are the insertion points, the lateral
bands and the hood, as shown in Figure 5.
(a) Schematic (b) Schematic
(c) Actual (d) Actual
Fig. 5. Tendon hood structure for the index finger and thumb.
We have mimicked the exact locations of the bone insertion
points in our tendon arrangements. The lateral bands serve
the following functions: (1) assist lateral movements via the
interosseous muscles when these muscles are activated as
antagonists, (2) allow the extension of the distal phalanx and
the finger as a whole when the interosseous muscles are
activated as agonists in concert with the EDC muscle, and
(3) coordinate the two distal joints (DIP and PIP) in flexion
and extension [71]. The hood structure enables the flexion of
the MCP joint independent of other joints and rotation at the
MCP joint [71].
3) Flexor Side: The tendons on the flexion side are con-
nected directly from the bone insertion points to the actuators.
For example, notice the tendon routing in on the flexion side
of the thumb in Figure 3. The flexion tendons pass through
guiding ‘rockers’ that allow for smooth travel of the tendon
while holding it close to the bone forms in order to achieve
accurate moment arm lengths about the finger joints.
D. ACT Hand Actuators
The ACT Hand possesses the same number of muscles as
the human hand: the index and middle fingers each have six
muscles, namely, EI, RI, PI, LUM, FDS and FDP, and the
thumb has eight muscles, namely, EPL, EPB, APB, APL, OP,
ADP, FPB, and FPL. The wrist is actuated by four muscles.
All muscles are realized by brushless DC motors located in
the forearm. The tendon strings are wound on the threaded
motor shaft, and the loose end is secured by a pin and
knot arrangement, as shown in Figure 6. The string-motor
arrangement leads to matching the musculotendon property
of one-dimensional actuation, that is, the muscles can only
7
actuate by contraction. We investigated the possibility of using
a direct linear muscle-like actuator, such as a Shape Memory
Alloy [41], McKibben muscles [70], or other artificial mus-
cles [66]. However, none of these actuators can provide the
fast response time (<200ms), total excursion lengths (>4cm)
and possibility of variable stiffness that a human muscle can
produce. At this stage of the ACT Hand’s design our focus is
on mimicking the static and dynamic force and movement
generation capacity, rather than the shape, size, or weight
of the real muscle. Hence, the DC motors were chosen as
actuators. To overcome the torque ripple in the DC motors, we
custom designed the DC motors 2based on the Kollmorgen
motor [39] as described below.
Fig. 6. ACT Hand actuators arranged in the forearm.
1) Magnet Skewing: One of the known issues with the
permanent magnet DC motors is the motor cogging, which
shows up as torque ripple at very low speeds [9]. The cogging
effects can lead to errors in finger positions and jerky finger
motions, especially during low speed operations. To address
this, we designed a rotor with permanent magnets cut into
three pieces, longitudinally, which are staggered at 10axial
rotation with respect to each other. This method, called step
skewing of the magnet, is a known, low cost way to reduce
the cogging torques [9]. The magnet skewing reduced the
cogging torque to less than 30% of the original value, while
the peak torque reduced by only 5%.
2) Motor Controller: Each motor is connected to a brush-
less servo controller (called the Puck [6]). Each controller has
an embedded photo-sensor and an encoder wheel (with 114
ticks/deg) that allows for a high precision position sensing of
the motor rotation to mimic muscle spindles (see Figure 6).
The controllers are connected to an RTAI Linux [53] machine,
which provides motor position readings at high frequency
(>500 Hz).
3) Motor Housing in the Forearm: All motors are located in
the forearm, as shown in Figure 6. The motors are arranged in
clusters of six units, and six clusters are connected end-to-end.
The modular design of the clusters and motor housings allows
2The brushless DC motor properties are: length 3cm, outside radius 2.2
cm, peak torque 40 mN-m, stall torque 40 mN-m, torque constant 0.13 mN-
m/A and rise time 0.1s.
for any motor to be easily replaced and entire clusters to be
added or removed to match the number of actuators required
for a particular setup. Additionally, the clusters incorporate
guiding sheaves to route the tendon strings from their radial
orientation after leaving the motor shaft to an axial orientation
to reach the wrist.
E. Mounting the Hand on to an Arm
Since our focus is on understanding the movements of
the hand and fingers, we mimicked the anatomy of the
fingers and palm. Our wrist design is not anatomical in that
it does not match the bone shapes and joint axes of the
human wrist. However, we mimicked its flexion-extension and
abduction/adduction motions.
The wrist has two DoFs: a “yaw" joint (±15 degrees)
attached to the end of the arm, and a “pitch" joint (±30
degrees) connected to the hand. Three banks of pulleys are
mounted on the gimbal structure joining the two DoFs and
guide the tendons from the arm-mounted actuators to the finger
joints. These pulleys minimize friction in the tendon paths and
route the tendons near the pitch axis to mimic the kinematic
coupling present in a human hand. As explained above, all
ACT Hand actuators are located in the forearm. The forearm
and hand are mounted on a Whole Arm Manipulator (WAM)
developed by Barrett Technologies, Inc. [6]. The WAM is a
4-DOF, cable-driven back-drivable manipulator. Its DC motor
control allows position and force control modes. The ACT
Hand forearm was connected to the WAM at the elbow joint
using the same physical connection that connects the WAM
forearm. Figure 6 shows the entire assembly.
IV. MUS CL E- JO IN T MAP PI NG S: M OM EN T ARM S
An important characteristic of the human hand is the me-
chanical advantage, called the moment arm, that each muscle-
tendon combination has on each joint. The muscle moment
arms in the human hand are configuration dependent and
play a critical role in hand movement control; however, the
exact properties of the moment arm variations are not known.
Because the ACT Hand mimics hand biomechanics, through
bone shapes and tendon hood structure, its muscle moment
arms are also configuration dependent, and determining the
exact moment arm relationships is critical for ACT hand
controls. Also determination of the moment arm relationships
in the ACT Hand can potentially lead to a better understanding
of human hand biomechanics.
We developed a method to acquire the moment arm relation-
ships for the ACT Hand that is based on an analysis of motion
capture data for finger and muscle-motor movements. What
follows are sample results from the moment arm analysis for
the index finger of the ACT Hand. To determine the moment
arms, we moved the index finger through its joint range of
motion and recorded the joint angles and changes in muscle
lengths. We then found a functional mapping, fi, among all
four joint angles and each muscle excursion using a Gaussian
process (GP) based regression model.
li=fi(θ)i= 1, ..., 6,(1)
8
where liis the excursion length for the muscle iwhich
is a member of the vector of muscle excursions (l=
[l1, l2, l3, l4, l5, l6]T) and θis a vector of finger joint angles
(θ= [θ1, θ2, θ3, θ4]T). We define θ1- MCP Ab-Ad, θ2- MCP
Flex, θ3- PIP Flex and θ4- DIP Flex, with the abduction and
flexion as positive directions. The moment arm relationships
were determined by taking the partial derivatives of the muscle
excursion functional mappings with respect to the joint angles.
In the case of the ACT Hand index finger, the moment arm is
defined by a matrix Rof dimension 6×4.
˙
l=R(θ)˙
θ(2)
where,
Rij (θ) = li
∂θj
=∂fi
∂θj
i= 1, ..., 6and j= 1, ..., 4.(3)
Figure 7(a) shows the variations in the excursions of EI
muscles as two finger angles vary. The dots in the figure are the
data points, and the surfaces show the fitted mapping functions.
Figure 7(b) shows the variations in the moment arm of the EI
muscle wrt to the MCP flex angle, and Figure 7(c) shows the
variations in the moment arm of the EI muscle wrt to the PIP
angle.
PIP flex (deg)
0
10
20
EI
60
MCP flex (deg)
90
0
30
0
30
60
(a) EI excursions
MCP Flex (deg)
0
0
PIP Flex (deg)
30
60
90
30
60
90
10.5
11.5 EI
(mm)
(b) Variations in the EI moment arm
wrt. MCP flex
(c) Variations in the EI moment arm
wrt. PIP flex
Fig. 7. Muscle excursions and moment arm variations for the EI muscle
with respect to the MCP and PIP flexion-extension, as MCP flexion and PIP
angles change.
The mean error across all data points with the GP-based
mapping is 0.65 (sd 0.33) mm. This error is low when
compared with the total excursions which are in the range
of 30 40 mm. Our results show that the excursion lengths
and moment arms for all muscles of the index finger depend
significantly on all the joint positions of the finger.
A. Validation
We conducted experiments to test the quality of the angle to
muscle length mapping determined by our method. Figure 8
28 29 30 31 32 33 34 35 36 37 38
−2
0
2
4
6
8
10
12
14
16
18
time (sec)
EI excursion (mm)
Actual
GP
Fig. 8. Actual vs estimated length changes in the EI muscle during a typical
finger motion.
shows the mapped muscle length when it was projected to EI
muscle length estimation. Table IV shows the mean absolute
error for all muscle length excursions when tested on a data
set of over 200,000 angles and muscle lengths combinations.
The data covers the physiological ranges of motion for all
the finger joints. The actual excursion length was recorded
directly from the encoder values, and the other estimations
were from the joint angle information from a motion capture
system translated to muscle lengths. We used an optical motion
tracking system (Vicon 360 with six M2 cameras) to record
motions of the finger involving all four joint angles. Thirteen
markers, each 3mm in diameter were placed on the ACT
finger and the distribution of the markers was as follows: five
on the MCP bone, three on the proximal and middle phalange,
and two on the distal phalange. The XYZ positions of the
markers were recorded at 120 Hz, and finger joint angles were
determined by using an angle determination algorithm built
into the motion analysis software (Vicon iQ 2.5).
TABLE IV
MEA N ABS OL UTE E RRO RS IN M US CLE T O JOI NT M APP IN GS (MM )
EI PI FDP LUM FDS RI
error 0.1838 0.4193 0.1350 0.3192 0.1408 0.4567
B. Comparison with Cadaver Data
We compare our results with index finger cadaver data
from [1] which is one of the most comprehensive data sets
available. To make the comparison variable moment arm data
from [1], which is from one female specimen, we generated
slice plots from our data by keeping the other joint angles
constant as it was done for the cadaver data. The moment
arm values are higher in our case by, on average 125% due
to the fact that we modeled the ACT Hand size after a male
subject. A comparison between the ACT Hand moment arms
and scaled human moment arms (by 125%) gives an overall
mean error 3±1.5mm. The sign of moment arms, indicat-
ing contribution to flexion/extension or abduction/adduction,
for all muscles, except for LUM in parts of finger flexion,
match with the cadaver plots. Table V gives the correlation
coefficients between individual muscle plots for MCP Ab-ad
and flexion variations. The first row gives the moment arms
9
wrt MCP flexion angle and the second row gives the moment
arm wrt MCP abduction-adduction angle. The differences in
variations in LUM and RI with flexion angle and PI with
adduction might have arisen due to the differences in the
structure of LUM in the cadaver hands and the ACT Hand.
The ACT lumbrical tendon and associated motor is attached
to the equivalent of a skeletal anchor point. In contrast, the
lumbrical tendon and associated muscle in human hands are
attached to another sliding tendon.
TABLE V
COR REL ATIO N COE FFIC IEN TS B ETW EE N MOM EN T ARM C URVE S FRO M TH E
CA DAVER AN D ACT H AND .
EI PI FDP LUM FDS RI
MCP FE 0.94 0.47 0.83 -0.71 0.85 -0.95
MCP AA 0.99 -0.66 0.99 0.67 0.93 0.99
C. Physical Interpretation of Moment Arm Variations
The extrinsic muscle excursions show significant depen-
dency on flexion but little variation with abduction-adduction.
In contrast, the intrinsic muscle excursions show significant
dependency on both flexion and abduction. This means that
moment arms for all muscles of the index finger depend
significantly on all the joint positions of the finger and, for
some of the muscles, the moment arms change sign. A negative
value for the moment arm of a muscle with respect to a joint
indicates that the muscle is contracting for positive change in
joint angle, meaning that the muscle actively contributes to
joint movement. Conversely, a positive value for the moment
arm means that the muscle is stretched for positive change in
joint angle. The moment arms for the muscle with respect to
the primary angles show large variations and maintain the sign
of the moment arms.
−20
0
20
40
60
MCP Flexion angle (deg)
MCP Ad-Ab angle (deg)
RI muscle velo (mm/sec)
−40
−20
0
20
40
−20
−15
−10
−5
0
5
10
constant
moment arm
variable
moment arm
Fig. 9. Variations in muscle excursion velocity of RI as for constant joint
angle velocities with a variable moment arm and with a constant moment arm
Our analysis leads to a mathematical model of the variations
in the index finger moment arms when multiple joint move
simultaneously. Since the ACT Hand structures imitate human
anatomy the moment arm variations provide insight into the
human moment arm properties. Previous studies have analyzed
moment arm variations with respect to motions of only one
joint at a time. Also, our study analyzes moment arm variation
with respect all four joints of the index finger which is missing
in the previous studies which focused only on the two MCP
joints of the index finger. Thus using the ACT Hand we have
determined previously unknown moment arm relationship in
the human index finger.
Variable moment arms play a significant role in the move-
ment control of the finger. For example, Figure 9 shows a
plot of variations in muscle excursion velocity of RI for the
same angular velocities of MCP Ab-Ad (0.5rad/sec), MCP
Flex(1.0rad/sec), PIP Flex (1.0rad/sec) and DIP Flex (1.0
rad/sec). As the MCP Ab-Ad and PIP Flex angles vary, while
MCP Flex and DIP Flex are kept constant, the RI velocity
changes from positive to negative. This means that the muscle
switched from being active and pulling to being passive and
stretching. The figure also shows that a constant moment
arm model, assumed in all the previous hand biomechanics
studies, does not capture the variations in muscle velocities
as functions of the finger configuration. Thus to control the
finger joint velocities muscle velocities are generated based
on the finger configuration using the variable moment arms.
A constant moment arms assumption will lead to erroneous
model of neuromuscular controls.
V. GR AS PI NG A ND HA ND MOV EM EN T ILL US TR ATIO N
Taking advantage of the DOFs and ranges of motion in
the ACT Hand, which are copied from the human hand, the
ACT Hand is able to grasp and manipulate a number of
objects, which are part of daily activities. Figure 10 shows
nine examples of object grasping using the ACT Hand. These
pictures demonstrate the DOFs and range of motion of the
finger joints, and overall anthropomorphic grasping abilities.
These nine objects were chose to demonstrate the wide variety
of grasps that can be achieved with the ACT Hand, for
example, power grasp in case of the water bottle and pinch
grasp in case of the spoon. Grasping performance is greatly
affected by the object interactions through the skin contact.
Currently we are developing tactile skin for the ACT Hand.
These grasps were achieved through a direct muscle control
scheme [26]. For each grasp we started from a neutral position
of thumb, index finger and middle finger, such that all fingers
are open, and pre-calculated the joint angles for the specific
grasp. The joint angles for the grasps were calculated by
manually moving the fingers and recording the joint angles
using the optical motion tracking system (Vicon 360 with six
M2 cameras). We then calculated the muscle excursions, i.e.
contraction and stretch, necessary to achieve the desired joint
angle and implemented position-integration control on the DC
motor driving the muscle. Figure 11 shows the changes in
lengths for the thumb and index finger muscles during the key
grasp maneuver shown in Figure 10. The top figure shows
the muscle excursions for the six index finger muscles and
the bottom figure shows the muscle excursions for the eight
thumb muscles. For the starting neutral position the muscle
excursions start at close to zero. Negative excursion means the
muscle is contracting and positive excursions means that the
muscle is stretching. As the finger joints flex, the flexors are
10
contracted and extensors are stretched for both the thumb and
index finger. Notice that the muscle lengths change smoothly
over the maneuver leading to smooth motions of the fingers.
The mean tracking error over all muscles is 1.67 (sd 0.34)
mm.
Considering that thumb movements account for more than
50% of hand function [20], we also implemented a thumb
movement of rubbing against the index finger, which is part
of many daily hand movements such as counting money
and opening a plastic bag. Because of the mimicking of
human thumb biomechanics in the ACT Hand, including axes
locations, number of DOF, and muscle arrangements, we were
able to generate an abduction-adduction motion at the MCP
joint in the ACT Hand thumb. Figure 12 shows the snapshots
of the thumb motion and corresponding length changes of the
eight muscles during this motion.
00.5 1 1.5 2 2.5 3 3.5 4
−1.5
−1
−0.5
0
0.5
1
1.5
2Index Finger
Excursion (cm)
PI
FDP
FDS
LUM
EI
RI
0
−1
−0.5
0
0.5
1
Thumb
Time (sec)
Excursion (cm)
FPB
FPL
OP
APB
ADP
EPL
APL
EPB
0.5 1 1.5 2 2.5 3 3.5 4
neutral position
neutral position
key grasp
key grasp
Fig. 11. Length changes in thumb and index finger muscles during key grasp
maneuver starting from neutral position (fingers open)
FPB
FPL
OP
APB
ADP
EPL
APL
EPB
Fig. 12. With the anatomical DOFs the thumb in the ACT Hand can be
moved to rub the side of the index finger which is a common and useful
human motion. The top row shows a sequence of photos taken at different
times (dotted lines) during the movement execution. And the bottom row
shows the plots of length changes (in mm) in the eight muscles of the thumb
during this movement.
VI. CO NC LU SI ON S AN D FUT UR E WOR K
This paper presents the novel constituting mechanisms,
unique muscle to joint relations, and movement demonstra-
tions of the thumb, index finger, middle finger, and wrist
of the ACT Hand. The ACT Hand is designed to further
our understanding of human hand mechanisms and control
and to provide guidelines for building versatile prosthetic
and dexterous hands. We reviewed the state-of-the-art robotic
hands and demonstrated that the unique design goals and
features distinguish the ACT Hand when compared to the
existing robotic hands. In Section III we have presented a
number of novel mechanical elements which are designed
and built to mimic human hand biomechanics. These include
the bone structures, joints, tendon arrangements and actuators
in the fingers and thumb, and also the tendon routing and
actuation of the wrist.
With the anatomically correct mechanical elements the ACT
Hand uniquely mimics the human hand biomechanics. We
have carried out a data driven kinematic analysis of the
relationships between the muscles and joints, defined by the
moment arms. Our analysis shows that the values for the
moment arms for the muscles in the ACT Hand vary with
the configuration of the hand, and these variations match with
the available data on human hand biomechanics. Our model
for moment arm variations is more comprehensive than the
existing models for the human hand, thus leading to a better
understanding of human hand biomechanics. Development of
tactile skin and passive joint properties in the ACT Hand are
part of ongoing research.
Completion of the ACT Hand mechanisms and software
platform allows us to conduct many experiments, which were
hitherto impossible, to study human hand properties and trans-
lation of important properties into robotic forms. For example,
now we can implement novel control algorithms to develop
a deeper understanding of human dexterity. Our group has
investigated the existence and importance of muscle synergies
during hand movement and force control [5], [4] and we are
implementing these mechanisms in the ACT Hand. We are also
investigating control strategies for achieving neuromuscular
control of ACT Hand muscles [26]. We are planning to
develop replicas of the ACT Hand and make them available
for other researchers. This will allow for the simultaneous
experimentation and growth in understanding of human hand
complexities.
REF ER EN CE S
[1] K. N. An, Y. Ueba, E. Y. Chao, W. P. Cooney, and R. I. Linscheid,
“Tendon excursion and moment arm of index finger muscles,” Journal
of Biomechanics, vol. 16, pp. 419–425, 1983.
[2] K.-N. An, “Tendon excursion and gliding: Clinical impacts from humble
concepts,” Journal of Biomechanics, vol. 40, pp. 713–718, 2007.
[3] D. J. Atkins, D. C. Heard, and W. Donovan, “Epidemiologic overview of
individuals with upper limb loss and their reported research priorities,”
Internal Journal of Prosthetics and Orthotics, vol. 8, 1996.
[4] R. Balasubramanian and Y. Matsuoka, “Biological stiffness control
strategies for the anatomically correct testbed (ACT) hand,” in Pro-
ceedings of the 2008 IEEE International Conference on Robotics and
Automation., 2008.
[5] ——, “The role of small redundant actuators in precise manipulation,”
in Proceedings of the 2009 IEEE international conference on Robotics
and Automation., 2009.
[6] Barrett Technology, Inc., “WAM specifications,” September 2009,
http://www.barrett.com/robot/products-arm-specifications.htm.
[7] G. Bekey, R. R. Tomovic, and I. Zeljkovic, “Control architecture for the
belgrade/USC hand,” Dextrous Robot Hands., pp. 136–149, 1990.
11
Fig. 10. The ACT Hand can grasp everyday objects with human-like finger postures. Figures show grasping of nine objects, namely, a coffee plate, key,
credit card, hand towel, glasses, water bottle, cordless phone, medicine bottle, and spoon, which are identified as important for ALS patients [19].
[8] L. Biagiotti, P. Tiezzi, Vassura, and C. G. Melchiorri, Modelling and
controlling the compliance of a robotic hand with soft finger-pads,
F. Barbagli, D. Prattichizzo, and K. Salisbury, Eds. Springer Tracts
in Advanced Robotics, 2005.
[9] N. Bianchi and S. Bolognani, “Design techniques for reducing the
cogging torque in surface-mounted PM motors,” Industry Applications,
IEEE Transactions on, vol. 38, no. 5, pp. 1259–1265, Sep/Oct 2002.
[10] S. Bidic, J. Imbriglia, and Y. Matsuoka, “An anatomical hand for
instruction and simulation,” in American Society for Surgery of the Hand
Meeting, Chicago, 2003.
[11] W. Bluethmann, R. Ambrose, M. Diftler, S. Askew, E. Huber, M. Goza,
F. Rehnmark, C. Lovchik, and D. Magruder, “Robonaut: A robot
designed to work with humans in space,” Autonomous Robots, vol. 14,
pp. 179–197, 2003.
[12] P. W. Brand and A. M. Hollister, Clinical Mechanics of the Hand, 3rd ed.
Mosby, 1999.
[13] P. W. Brand and M. H. Anne, Clinical Mechanics of the Hand. St.
Louis: Mosby-Year Book, Inc., 1993.
[14] J. Butterfass, M. Fischer, M. Grebenstein, S. Haidacher, and
G. Hirzinger, “Design and experiences with DLR hand II,” in Automation
Congress, 2004. Proceedings. World, vol. 15, 28 2004-July 1 2004, pp.
105–110.
[15] M. C. Carrozza, G. Cappiello, S. Micera, B. B. Edin, L. Beccai, and
C. Cipriani, “Design of a cybernetic hand for perception and action,”
Biological Cybenetics, vol. 95, no. 6, pp. 629–644, 2006.
[16] M. C. Carrozza, B. Massa, S. Micera, R. Lazzarini, M. Zecca, and
P. Dario, “The development of a novel prosthetic hand- ongoing research
and preliminary results,” IEEE/ASME Transactions on Mechatronics,
vol. 7, pp. 108–114, 2002.
[17] M. C. Carrozza, C. Suppo, F. Sebastiani, B. Massa, F. Vecchi, R. Laz-
zarini, M. R. Cutkosky, and P. Dario, “The spring hand: Development
of a self-adaptive prosthesis for restoring natural grasping,Autonomous
Robots, vol. 16, no. 2, pp. 125–141, 2004.
[18] L. Y. Chang and Y. Matsuoka, “A kinematic thumb model for the ACT
hand,” in Proceedings of the 2006 IEEE International Conference on
Robotics and Automation., 2006.
[19] Y. S. Choi, T. Deyle, and C. C. Kemp, “Benchmarking assistive mobile
manipulation: A list of household objects for robotic retrieval prioritized
by people with ALS,” in International Conference on Rehabilitation
Robotics, 2009.
[20] J. C. Colditz, “Anatomic considerations for splinting the thumb,Reha-
bilitation of the hand: surgery and therapy, 1990, m. E. J. Hunter J. M.,
Callahan A. D., Ed. Philadelphia: C. V. Mosby Company.
[21] W. P. Cooney, M. J. E. Lucca, Y. Chao, and R. L. Linscheid, “The
kinesiology of the thumb trapeziometacarpal joint,” Journal of Bone
Joint Surg Am, vol. 63, pp. 1371–1381, 1981.
[22] S. A. Dalley, T. E. Wiste, T. J. Withrow, and M. Goldfarb, “Design
of a multifunctional anthropomorphic prosthetic hand with extrinsic
actuation,” IEEE-ASME Transactions on Mechatronics, vol. 14, pp. 699–
706, 2009.
[23] DEKA, 2008, http://www.dekaresearch.com.
[24] DEKA-HAND, “Dean Kamen Luke arm prosthesis,”
http://www.dekaresearch.com/index.shtml, 2010.
[25] S. Delp and J. Loan, “A graphics-based software system to develop and
analyze models of musculoskeletal structures,” Computers in biology
and medicine, vol. 25, no. 1, pp. 21–34, 1995.
[26] A. Deshpande, J. Ko, D. Fox, and Y. Matsuoka, “Anatomically correct
testbed hand control: Muscle and joint control strategies,” in Proceedings
of the 2009 IEEE International Conference on Robotics and Automation,
2009, pp. 4416–4422.
[27] A. M. Dollar and R. D. Howe, “A robust compliant grasper via shape
deposition manufacturing,” IEEE/ASME Transactions on Mechatronics,
vol. 11, pp. 154–161, 2006.
[28] A. Dollar and R. Howe, “Simple, robust autonomous grasping in
unstructured environments,” in 2007 IEEE International Conference on
Robotics and Automation, April 2007, pp. 4693–4700.
[29] K. A. Farry, I. D. Walker, and R. G. Baraniuk, “Myoelectric teleoper-
ation of a complex robotic hand,” IEEE Transactions on Robotics and
Automation, vol. 12, pp. 775–788, 1996.
[30] M. Garciaelias, K. AN, L. Berglund, R. Linscheid, W. Cooney, and
E. Chao, “Extensor mechanism of the fingers. I: A quantitative geometric
study,Journal of Hand Surgery-American Volume, vol. 16A, no. 6, pp.
1130–1136, NOV 1991.
[31] A. Hollister, W. L. Buford, L. M. Myers, D. J. Giurintano, and
A. Novick, “The axes of rotation of the thumb carpometacarpal joint.
Journal of Orthopaedic Research, vol. 10, no. 3, pp. 454–460, May
1992.
12
[32] A. Hollister, D. J. Giurintano, W. L. Buford, L. M. Myers, and
A. Novick, “The axes of rotation of the thumb interphalangeal and
metacarpophalangeal joints.” Clin Orthop Relat Res, no. 320, pp. 188–
193, November 1995.
[33] K. Holzbaur, W. Murray, and S. Delp, “A model of the upper extremity
for simulating musculoskeletal surgery and analyzing neuromuscular
control,” Annals of biomedical engineering, vol. 33, no. 6, pp. 829–840,
2005.
[34] Hosmer Dorrance Corporation, “Body-powered prosthetic hand,”
September 2009, http://www.hosmer.com/products/hooks/index.html.
[35] S. Jacobsen, E. Iversen, D. Knutti, R. Johnson, and K. Biggers, “Design
of the Utah/M.I.T. dextrous hand,” in Proceedings. 1986 IEEE Interna-
tional Conference on Robotics and Automation., vol. 3, Apr 1986, pp.
1520–1532.
[36] A. Kargov, C. Pylatiuk, R. Oberle, H. Klosek, T. Werner, W. Roessler,
and S. Schulz, “Development of a multifunctional cosmetic prothetic
hand,” in IEEE International Conference on Rehabilitation Robotics,
2007.
[37] J. A. Katarincic, “Thumb kinematics and their relevance to function,
Hand Clinics, vol. 17, pp. 169–174, 2001.
[38] H. Kawasaki, T. Komatsu, and K. Uchiyama, “Dexterous anthropo-
morphic robot hand with distributed tactile sensor: Gifu Hand II,”
Mechatronics, IEEE/ASME Transactions on, vol. 7, no. 3, pp. 296–303,
Sep 2002.
[39] Kollmorgen, Inc., “Kollmorgen motors,” 2010,
http://www.kollmorgen.com/.
[40] T. Kuiken, G. Dumanian, R. Lipschutz, L. Miller, and K. Stubblefield,
“The use of targeted muscle reinnervation for improved myoelectric
prosthesis control in a bilateral shoulder disarticulation amputee,” Pros-
thetics and Orthotics International, vol. 28, no. 3, pp. 245–253, DEC
2004.
[41] M. Kumon, I. Mizumoto, Z. Iwai, and A. Indou, “Shape memory
alloy actuator with simple adaptive control,” in Second International
Conference on Innovative Computing, Informationand Control., 2007.
[42] P. J. Kyberd, C. Light, P. H. Chappell, J. M. Nightingale, D. Whatley,
and M. Evans, “The design of anthropomorphic prosthetic hands: A
study of the southampton hand,” Robotica, vol. 19, no. 6, pp. 593–600,
2001.
[43] T. Lan, Y. W. Liu, M. H. Jin, S. W. Fan, H. G. Fang, J. J. Xia,
and H. Liu, “Dsp & fpga-based joint impedance controller for dlr/hit
ii dexterous robot hand,” in IEEE/ASME International Conference on
Advanced Intelligent Mechatronics, 2009.
[44] Liberating Technologies, Inc., “Upper extremity prosthetics„” September
2009, http://www.liberatingtech.com.
[45] F. Lotti, P. Tiezzi, G. Vassura, L. Biagiotti, G. Palli, and C. Melchiorri,
“Development of UB Hand 3: Early results,” in Proceedings of the 2005
IEEE International Conference on Robotics and Automation, April 2005,
pp. 4488–4493.
[46] C. Lovchik and M. Diftler, “The Robonaut hand: a dexterous robot hand
for space,” in Proceedings of the 1999 IEEE International Conference
on Robotics and Automation., vol. 2, 1999, pp. 907–912.
[47] Motion Control, Inc., “The motion control ETD„” September 2009,
http://www.utaharm.com/.
[48] T. Mouri, H. Kawasaki, Y. Keisuke, J. Takai, and S. Ito, “Anthropo-
morphic robot hand: Gifu hand III,” Proc. Int. Conf. ICCAS, p. 1288,
2002.
[49] M. Nicolelis, “Actions from thoughts,” NATURE, vol. 409, no. 6818,
pp. 403–407, JAN 18 2001.
[50] D. Nishikawa, Y. Ishikawa, W. Yu, M. Maruishi, I. Watanabe, H. Yokoi,
Y. Mano, and Y. Kakazu, “On-line learning based emg prosthetic hand,
Electrophysiology and Kinesiology, 2000.
[51] T. Okada, “Object handling system for manual industry,” IEEE Trans-
actions on Systems, Man and Cybernetics, vol. 9, pp. 79–89, 1979.
[52] OttoBock HealthCare, Inc., “Cable-controlled arm prostheses„” Septem-
ber 2009, http://www.ottobock.com.
[53] RTAI, “http://www.rtai.org,” 2005.
[54] J. K. Salisbury and J. J. Craig, “Articulated hands: Force control and
kinematic issues,” International Journal of Robotics Research, vol. 1,
1982.
[55] V. J. Santos and F. J. Valero-Cuevas, “Anatomical variability naturally
leads to multimodal distributions of denavit-hartenberg parameters for
the human thumb,” Engineering in Medicine and Biology Society, vol. 2,
pp. 1823– 1826, 2003.
[56] K. Sasaki, Y. Fujikake, H. Takahashi, and M. R. Cutkosky, “Manipu-
lation of an object using slip sensing tactile sensor,” in Proceedings of
the ICMA’98 - Advanced Mechatronics: First-Time-Right, 1998.
[57] T. Senoo, Y. Yamakawa, S. Mizusawa, A. Namiki, M. Ishikawa, and
M. Shimojo, “Skillful manipulation based on high-speed sensory-motor
fusion,” in Proceedings of the 2009 IEEE International Conference on
Robotics and Automation., May 2009, pp. 1611–1612.
[58] ShadowHand, “Shadow robot company,
http://www.shadowrobot.com/hand/, 2010.
[59] D. H. Silcox, M. D. Rooks, R. R. Vogel, and L. L. Fleming, “Myoelectric
prostheses - a long-term follow-up and a study of the use of alternate
prostheses,” Jounral of Bone and Joint Surgey - Americal Volume, vol.
75A, pp. 1781–1789, 1993.
[60] W. P. Smutz, A. Kongsayreepong, R. E. Hughes, G. Niebur, W. P.
Cooney, and K. N. An, “Mechanical advantage of the thumb muscles,”
Jounral of Biomechanics, vol. 31, pp. 565–570, 1998.
[61] T. H. Speeter, “Primitive based control of the UTAH/MIT hand,” in
International Conference on Robotics and Automation, 1991.
[62] S. Sueda, A. Kaufman, and D. Pai, “Musculotendon simulation for hand
animation,” in ACM SIGGRAPH 2008 papers. ACM, 2008, pp. 1–8.
[63] Touch Bionics, Inc., “Fitting and service manual„” May 2009,
http://www.touchbionics.com/.
[64] W. Townsend, “The Barrett hand grasper - programmably flexible part
handling and assemble,” The International Journal of Industrial Robot,
vol. 27, no. 3, pp. 181 – 188, 2000.
[65] W. Tsang, K. Singh, and E. Fiume, “Helping hand: an anatomically
accurate inverse dynamics solution for unconstrained hand motion,” in
Proceedings of the 2005 ACM SIGGRAPH/Eurographics symposium on
Computer animation. ACM, 2005, pp. 319–328.
[66] T. Tsuji, S. Miyata, T. Hashimoto, and H. Kobayashi, “Controller
design for robot with pneumatic artificial muscles,” in International Joint
Conference SICE-ICASE, 2006.
[67] J. Ueda, Y. Ishida, M. Kondo, and T. Ogasawara, “Development of the
NAIST-Hand with vision-based tactile fingertip sensor,” in Proceedings
of the 2005 IEEE International Conference on Robotics and Automa-
tion., April 2005, pp. 2332–2337.
[68] M. Vande Weghe, M. Rogers, M. Weissert, and Y. Matsuoka, “The ACT
hand: Design of the skeletal structure,” in Proceedings of the 2004 IEEE
International Conference on Robotics and Automation., 2004.
[69] M. Velliste, S. Perel, M. C. Spalding, A. S. Whitford, and A. B.
Schwartz, “Cortical control of a prosthetic arm for self-feeding,” Nature,
vol. 453, no. 7198, pp. 1098–1101, June 2008.
[70] S. Wakimoto, K. Suzumori, and T. Kanda, “Development of intelligent
mckibben actuator,” in IEEE/RSJ International Conference on Intelligent
Robots and Systems., 2005.
[71] D. Wilkinson, M. Vande Weghe, and Y. Matsuoka, “An extensor mecha-
nism for an anatomical robotic hand,” in Proceedings of the 2003 IEEE
International Conference on Robotics and Automation., Sept. 2003, pp.
238–243.
[72] Z. Xu, T. Deyle, and C. Kemp, “1000 Trials: An empirically validated
end effector that robustlygrasps objects from the floor,” in Proceedings of
the 2009 IEEE International Conference on Robotics and Automation.,
May 2009, pp. 2160–2167.
[73] I. Yamano and T. Maeno, “Five-fingered robot hand using ultrasonic mo-
tors and elastic elements,” in Proceedings of the 2005IEEE International
Conference on Robotics and Automation., April 2005, pp. 2673–2678.
[74] F. E. Zajac, “Muscle coordination of movement: A perspective,Journal
of Biomechanics, vol. 26, no. Supplement 1, pp. 109 – 124, 1993.
... An open-source musculoskeletal handwrist model is proposed in [1] incorporating the properties of moment arms and passive joint torques, and is validated within multiple scenarios, such as lateral pinch and passive grasp/release. Meanwhile, anthropomorphic robotic tendondriven hands have been developed as the counterpart of the human hand, aiming to achieve the dexterity of human hands in daily tasks [6]- [8] and to provide a platform for exploring the biomechanical properties of human hands by mimicking the geometry and tendon structure of human hands [9] [10]. ...
... For example, the tendon excursion method was proposed and applied to identify the moment arm curves of cadaver index fingers in [11] and was extended to all fingers in study [12]. The study [16] designed an experiment using a regression method to determine the moment arm curves of tendons using the ACT test-bed [9]. The passive viscoelasticity property of the human index finger has been investigated in [17][15] [18] through in vivo experiments, in which the study [15] indicated that the joint viscoelasticity dominates the finger dynamics over rigid body dynamics during movement. ...
... An open-source musculoskeletal handwrist model is proposed in [1] incorporating the properties of moment arms and passive joint torques, and is validated within multiple scenarios, such as lateral pinch and passive grasp/release. Meanwhile, anthropomorphic robotic tendondriven hands have been developed as the counterpart of the human hand, aiming to achieve the dexterity of human hands in daily tasks [6]- [8] and to provide a platform for exploring the biomechanical properties of human hands by mimicking the geometry and tendon structure of human hands [9] [10]. ...
... For example, the tendon excursion method was proposed and applied to identify the moment arm curves of cadaver index fingers in [11] and was extended to all fingers in study [12]. The study [16] designed an experiment using a regression method to determine the moment arm curves of tendons using the ACT test-bed [9]. The passive viscoelasticity property of the human index finger has been investigated in [17][15] [18] through in vivo experiments, in which the study [15] indicated that the joint viscoelasticity dominates the finger dynamics over rigid body dynamics during movement. ...
Preprint
Full-text available
This study addresses the absence of an identification framework to quantify a comprehensive dynamic model of human and anthropomorphic tendon-driven fingers, which is necessary to investigate the physiological properties of human fingers and improve the control of robotic hands. First, a generalized dynamic model was formulated, which takes into account the inherent properties of such a mechanical system. This includes rigid-body dynamics, coupling matrix, joint viscoelasticity, and tendon friction. Then, we propose a methodology comprising a series of experiments, for step-wise identification and validation of this dynamic model. Moreover, an experimental setup was designed and constructed that features actuation modules and peripheral sensors to facilitate the identification process. To verify the proposed methodology, a 3D-printed robotic finger based on the index finger design of the Dexmart hand was developed, and the proposed experiments were executed to identify and validate its dynamic model. This study could be extended to explore the identification of cadaver hands, aiming for a consistent dataset from a single cadaver specimen to improve the development of musculoskeletal hand models.
... Fully actuated hands have the same number of motors as the number of degrees of freedom. Deshpande et al. (2013) constructed a 21-degree-of-freedom anatomically correct test bed (ACT) hand with a human-like tendon structure, and Yamano and Maeno (2005) designed a 20-degree-of-freedom Keio hand actuated by ultrasonic motors with elastic elements. Grebenstein et al. (2011) designed a 19-degree-offreedom Deutsches Zentrum für Luft-und Raumfahrt (DLR) hand actuated by tendons. ...
Article
Full-text available
The traditional underactuated grippers can only passively adapt to the contour of the object, and the passive contact process may lead to the object slipping, affecting the stability of the grasping process. In this paper, an anthropomorphic modular gripper finger actuated by antagonistic wire and shape-memory alloy (SMA) springs, which can actively control the grasping morphology according to the characteristics of the objects to be grasped, is proposed. The wire drive simulates the flexor muscle, and the SMA and reset springs simulate the extensor muscles of the finger, which antagonistically control the grasping morphology of the finger. It is more in line with the grasping characteristics of the human hand. According to the moment equilibrium principle of the finger joints, the deformation model of the gripper is established, the influence of the wire tension and the equivalent stiffness of the finger joints on the grasping morphology is analyzed, and the theoretical joint angle results are verified by the Adams simulation; finally, the experimental system of the gripper is constructed, and the verification of the deformation morphology of the single finger and the gripper's enveloping–grasping experiments is completed. The results show that according to the contour size of the object, by actively controlling the wire force of the gripper and the equivalent stiffness of the interphalangeal joints, the enveloping–grasping action of different objects can be completed and the stable grasping of objects of different shapes and sizes can be realized.
... Fully actuated grippers have the same number of actuators as the number of degrees of freedom. Deshpande et al. [1] constructed a 21-degree-of-freedom ACT hand with a human-like tendon structure, and Yamano et al. [2] designed a 20-degree-offreedom Keio hand actuated by ultrasonic motors with elastic elements. Grebenstein et al. [3] designed a 19-degree-of-freedom DLR hand actuated by tendons. ...
Article
Full-text available
The traditional underactuated grippers can only passively adapt to the shape contour of the object to complete the envelope or grasping action. In this paper, a wire and shape memory alloy (SMA) spring-based differential drive gripper finger are proposed, which can actively control the grasping morphology according to the objects to be grasped, and are more in line with the grasping characteristics of the human hand. The wire simulates the flexor muscle, and the SMA and reset springs simulate the extensor muscle of the finger, which together control the finger morphology. According to the principle of moment equilibrium, the static mechanical model of the gripper is established, and the influence of the wire driving force and the equivalent stiffness of the finger joints on the grasping morphology are analysed, and the theoretical grasping force is verified by ADAMS simulation. A comparative analysis of the grasping force between the non-active and active morphology control is presented. Finally, the experimental system of the gripper is constructed, and the verification of the deformation morphology of the single finger and the gripper’s grasping experiments is completed. The results show that according to the contour size of the object, by actively controlling the wire force of the gripper and the equivalent stiffness of the joints, the enveloping or grasping action for different objects can be completed. In addition, the grasping force of the finger can be improved by actively controlling the grasping morphology.
... The intention was to make the system more dexterous [12][13][14][15]. In recent years, advances in materials science, sensor technology, and artificial intelligence have continued to drive the development of more advanced and sophisticated robotic hands [16][17][18][19][20][21]. Today, there are a wide variety of robotic hands and prosthetic devices that are capable of mimicking the complex movements and dexterity of the human hand, and these devices are being used in a variety of applications, from manufacturing and logistics to healthcare and rehabilitation [22]. ...
Article
Full-text available
The development of robotic hands that can replicate the complex movements and dexterity of the human hand has been a longstanding challenge for scientists and engineers. A human hand is capable of not only delicate operation but also crushing with power. For performing tasks alongside and in place of humans, an anthropomorphic manipulator design is considered the most advanced implementation, because it is able to follow humans’ examples and use tools designed for people. In this article, we explore the journey from human hands to robot hands, tracing the historical advancements and current state-of-the-art in hand manipulator development. We begin by investigating the anatomy and function of the human hand, highlighting the bone-tendon-muscle structure, skin properties, and motion mechanisms. We then delve into the field of robotic hand development, focusing on highly anthropomorphic designs. Finally, we identify the requirements and directions for achieving the next level of robotic hand technology.
Article
Robots often interact with unstructured and cluttered environments, requiring a dexterous robotic hand that can adapt to unpredictability. In this paper, we propose a 16 degrees of freedom (DOFs) Thumb-Inspired Tactile four-finger hand (TIT Hand) for high-precision dexterous fingertip manipulation, which can realize 6-DOFs manipulation of the entire hand and 4-DOFs control of a single finger in Cartesian space. Specifically, we design a novel 3-DOFs spherical six-bar mechanism to imitate the human thumb’s carpometacarpal joint, enabling finger flexion/extension, abduction/adduction and pronation/supination motions. To realize active compliance interaction with objects, we equip the fingertip with a high-resolution visuotactile sensor and develop a compliance controller based on tactile feedback. Experiments show that the controller can adjust the interaction force within 0.6 s. Finally, the proposed hand is evaluated by extensive experiments, such as strength&stiffness tests, which show that the hand has a fingertip repeatability of 0.0302 mm and a maximum grasp force of 36 N. The experimental results verify the effectiveness of this novel hand design and the robustness of the tactile-driven compliance controller.
Conference Paper
Full-text available
This paper proposes a small Shape Memory Alloy actuator unit and its control algorithm. Since the unit contains a part to sense its displacement and stiffness, no external sensors are needed. SMA deformation is driven by thermal dynamics which can be significantly effected by surrounding environment. In order to attenuate this effect, a robust and simple control algorithm, Simple Adaptive Control is adopted. The proposed method was evaluated by the developed actuator unit. Experimental results showed that the stiffness was well controlled and the position could be controlled to converge the desired displacement.
Chapter
This chapter describes design and control features of a five-fingered anthropomorphic end-effector designed primarily for grasping tasks. Advantages and limitations of the design are discussed, and special emphasis is placed on its suitability for autonomous, non-numerical or reflex control of grasp. Following a discussion of its mechanical design, we present the controller and sensor features incorporated into the current finger model. A knowledge-based control of hand preshape (prior to grasping) is then outlined, and the hand’s suitability as a testbed for the study of human and robot hand motion control is discussed. The final section of this chapter describes future directions.
Article
Commercially available prosthetic hands are simple grippers with one or two degrees of freedom; these pinch type devices have two rigid fingers in opposition to a rigid thumb. This paper focuses on an innovative approach for the design of a myoelectric prosthetic hand. The new prosthesis features underactuated mechanisms in order to achieve a natural grasping behavior and a good distribution of pinching forces. In this paper it is shown that underactuation allows reproducing most of the grasping behaviors of the human hand, without augmenting the mechanical and control complexity.
Conference Paper
This paper presents an anthropomorphic prototype hand prosthesis that is intended for use with a multiple channel myoelectric interface. The hand contains 16 joints, which are differentially driven by a set of five independent actuators. The hand prototype was designed with the minimum number of independent actuators required to provide a set of eight canonical hand postures. This paper describes the design of the prosthesis prototype, demonstrates the hand in the desired eight canonical postures, and experimentally characterizes the force and speed capability of the device. A video is included in the supplemental material that also illustrates the functionality and performance of the hand.
Article
A realistic biomechanical model of the thumb would enhance our understanding of the functional consequences of orthopedic and neurological diseases, and of treatments. Our work has shown that a kinematic description with five orthogonal and intersecting axes of rotation cannot predict realistic thumbtip forces. An alternative kinematic description proposes five axes of rotation that need not be orthogonal or intersecting. In order to make this description amenable for roboticist-use, we described the model in Denavit-Hartenberg (DH) notation. We explored the effects of reported anatomical variability on the DH parameters using Monte Carlo simulations. We report the DH parameters as statistical distributions that can be used for robotics-based models of the hand and stochastic analyses. We found three characteristic sets of kinematic descriptions. In 65.2% of the 3,140 simulations, the metacarpophalangeal flexion-extension axis was distal to the metacarpophalangeal adduction-abduction axis. We pose the question: Are multiple types of kinematic descriptions necessary to account for the natural anatomical variability of the thumb? This question is important for the biomechanical modeling of the hand, as the debate continues of whether patient-specific models are needed to simulate hand function for clinical applications, or if a single common model suffices.
Article
Kinematic and control issues are discussed in the context of an articulated, multifinger mechanical hand. Hand designs with particular mobility properties are illustrated, and a definition of accuracy points within manipulator workspace is given. Optimization of tlte physical dimensions of the Stanford-JPL hand is described. Several architectures for position and force control of this multiloop mechanism are described, including a way of dealing with the internal forces inherent in such systems. Preliminary results are shown for the joint torque subsystem used in the hand controller.