![Joint/Bone composition of human hand [2]](https://www.researchgate.net/profile/Alba-Perez-Gracia/publication/277064253/figure/fig1/AS:294574359629830@1447243408869/Joint-Bone-composition-of-human-hand-2_Q320.jpg)
Content uploaded by Alba Perez-Gracia
Author content
All content in this area was uploaded by Alba Perez-Gracia on Sep 03, 2015
Content may be subject to copyright.
Abstract—This paper presents the mechatronic design of a
robotic hand for prosthetic applications. The main
characteristic of this robotic hand is its biologically-inspired
parallel actuation system, which is based on the
behavior/strength space of the Flexor Digitorum Profundus
(FDP) and the Flexor Digitorum Superficialis (FDS) muscles.
The design separates the strength space of the FDS and FDP
muscles into a lighter strength region where finer manipulation
and general approach tasks are executed, and a higher strength
region where the more robust grasps are achieved. Two
parallel actuator types and kinematic structures are designed
to complement the requirements of both strength space regions.
This unique structure is intended to be driven by
electromyographical (EMG) signals captured at the surface of
the skin. The direct relation between signal and actuation
system lends itself well to interpreting the EMG signals from
the FDP and FDS muscles into effective task execution, with the
goal of helping the user to achieve a good approximation of the
full capabilities associated with the human hand, without
compromising strength, dexterity, appearance, or weight;
which are common issues associated with prosthetic hands.
The designed finger’s capability of having a strength space
similar to that of the FDS and FDP muscles is validated via
direct inputs from a power supply and then via a controller
using an actual EMG signal input from the human forearm.
The controller is a simple feed forward system at this point in
the research but provides the appropriate framework to
integrate more elaborate control schemes and EMG signal
conditioning as this portion of the research area matures.
Keywords – Prosthetics, Parallel Actuation Structure, Robotics,
Hand, FDS and FDP Muscles
I. INTRODUCTION
HERE Have been many different approaches taken in
the development of an effective prosthetic hand. These
varying strategies often find themselves focusing on one of
the following categories: implementing a new actuator type
[11-16], developing a more effective kinematic structure
[18,23,34], integrating effective compliance [18,23-25],
Manuscript received May 24, 2010. this research was performed under
an award/contract from Telemedicine Advanced Technology Research
Center (TATRC), of the U.S. Army Medical Research and Materiel
Command (USAMRMC) of the U.S. Department of Defense.
A. L. Crawford was a Ph.D. student at Idaho State University, Pocatello,
ID 83209 (e-mail: crawanth@isu.edu)
J. Molitor is a Masters student at Idaho State University, Pocatello, ID
83209 (e-mail: molijeff@isu.edu)
M. A. Perez-Gracia is an assistant professor at Idaho State University,
Pocatello, ID 83209 and is currently at Institut de Robotica i Informatica
Industrial (UPC/CSIC), Barcelona, Spain (e-mail: perealba@isu.edu)
generating effective control strategies [25-31], and
interpreting/conditioning input signals [25]. Advances in
these areas have resulted in robotic hands that perform many
tasks with a high similarity to that of the human hand, such
as the DLR hand [33], I-Limb hand [11], Shadow hand [12],
and Fluidhand [32] to name a few. However, a prosthetic
hand that is nimble, quick, strong, lightweight, quiet, and
efficient [1] has yet to be achieved.
The primary reason for the current state of prosthetic
hands has been the complexity associated with the human
hand as a result of its multiple bones and joints (Fig 1). This
is further compounded by the fact that the human hand as a
functioning unit does not just embody the palm and its digits
but also the wrist, forearm muscles, nervous system, and the
body’s energy generation system. As a result, the entire
prosthetic hand actuation structure (inputs, power, strength,
kinematics, etc.) must fit in a significantly reduced volume
compared to the human hand that it is replacing.
To address some of the challenges described above, this
research implements a unique perspective of the FDS and
FDP muscles’ strength space in the human forearm and
proposes a novel design and parallel actuation structure that
complements this perspective. The goal is to create a direct
relation between the forearm’s EMG signals and the
actuation system, in order to help the user achieve a good
approximation of the full capabilities associated with the
human hand in a compact design.
Fig 1: Joint/Bone composition of human hand [2]
Design of a Robotic Hand and Simple EMG Input Controller with a
Biologically-Inspired Parallel Actuation System for Prosthetic
Applications
Dr. Anthony L. Crawford, Member, IEEE, Jeffrey Molitor, Member, IEEE,
Dr. Alba Perez-Gracia Member, IEEE, Dr. Steve C. Chiu, Member, IEEE
T
Sections 2 and 3 of this paper describe the FDS and
FDP muscles’ strength space, how it relates to the human
hand’s capabilities, as well as the actuators and actuation
structures of current prosthetic/robotic hands. Sections 4
and 5 will provide a description of the mechanical design
and testing results that justify the design’s ability to execute
the strength space perspective developed in this paper. The
paper will finally present conclusions associated with the
testing results and an identification of future work. The
actuation structure’s mechanical design presented here is an
expansion of the content being published at the ASME
IDETC 2010 conference with the inclusion of friction in the
force calculations, implementation of the mechatronics, and
the EMG inputs being an expansion of that presented
research.
The FDS and FDP muscles are the primary flexor
muscles in the human finger and are primarily opposed by
the extensor digitorum (ED) muscle. As shown in Fig 2.
The FDP muscle is attached to the distal phalanx and is
capable of full hand closure; it is considered to be the more
active of both finger flexion muscles. The FDS muscle is
attached to the middle phalanx and its full capacity is
primarily achieved when activation of the DIP joint is not
required or when full hand closure tasks require additional
strength [3].
The strength space of the FDS and FDP muscles is
shown in Fig 3. The figure demonstrates the normalized
maximal force exertion of the FDS and FDP muscles (y-
axes) during maximum force execution of the hand (x-axes).
The FDP muscle is shown to reach its maximal force
execution (120N [5]) at approximately 35% of the total
flexural effort; however, the FDS muscle continues to exert
force until it reaches its maximal force execution (240 N [5])
at about 100% of the total flexural effort. The FDS and FDP
behavior can be attributed to the learned neurological
activation of these muscles [6] as well as the finger’s
associated kinematic structure.
Fig 2: Graphic of FDS and FDP muscle in finger [4]
Fig 3: Strength space of FDS and FDP muscles (x-axis: normalized postion
of hand from open (0) to closed (1), y-axis: normalized force)[7]
The FDS and FDP strength space comprises the strength
requirements for all the tasks that a hand must execute
[8][9]. In general, most everyday tasks don’t require
extensive force but do require a certain amount of dexterity.
For manipulation or approaching an object, the finger can
employ both the FDS and FDP muscles to nimbly
accommodate various shapes and execute both general and
complex movement paths. Fig 3 shows that the FDS and
FDP muscles are both active for activities below 35% of the
maximum force capacity of the hand and are most likely
employed during manipulation/object approaching
movements.
Gripping tasks generally require less dexterity and more
strength than manipulation and general object approach.
The size and weight of the object as well as the
characteristics of the grasp (e.g. friction between pads or
force closure) determines how much force is required from
the finger’s strength space shown in Fig 3. It is to be noted
that the actual act of the grip also generally requires much
less complex finger motion than that of manipulation.
Based on these observations, we divided the FDS and FDP
strength space into the regions shown in Fig 4. In Fig 4
region 1 is populated by the more frequent dexterous tasks
and region 2 is populated by the less frequent and more
strength-based tasks.
Fig 4: Divided Strength Space of FDS and FDP muscles (x-axis:
normalized hand position from open (0) to closed (1), y-axis: normalized
force)
Though the bones, joints, and muscles of the thumb are
somewhat different than that of the finger, the same FDS
and FDP strength space division philosophy is applied in the
mechanical design of the prosthetic thumb.
II. ACTUATION STRUCTURES
To the authors’ knowledge, all currently developed
prosthetic/robotic hands use a single actuator type to execute
all the tasks embodied in the FDS and FDP strength space.
This technique results in the shortcomings of the chosen
actuator being carried throughout the strength space, let it be
pneumatic, electromechanical, ultrasonic, or shape memory
alloy. This could include excessive size and weight, or
reduced time response and energy inefficiency to name a
few.
The actuator types used in current robotic/prosthetic
hands and considered in this design included
electromagnetic [10][11], pneumatic [12][13], hydraulic
[14], ultrasonic [15], and shape memory alloy [16].
Although all the listed actuator types have been employed in
multiple prosthetic hand designs based on their
advantageous features, one or more shortcoming (weight,
noise, size, efficiency, and speed) have been accepted as
well.
Prosthetic hands have employed the above actuator types
into two general types of kinematic structures. These two
structures are referred to as fully actuated and
underactuated. The underactuated structure often uses a
single input to actuate the multiple joints and essentially
wraps the phalanges of the finger around an object [17].
Typically this is achieved by a flexible tendon routed
through the finger structure which allows one of the finger’s
phalanges to be stopped by the object without preventing the
remaining phalanges from continuing to wrap around the
target
Some underactuated structures couple all three degrees
of freedom associated with finger flexure to one actuator.
However, there are structures where the underactuation
exists only between the PIP and DIP joints. This behavior is
typically executed by the human hand and will be referred to
herein as nearly fully actuated. One example of this type of
nearly fully actuated structure is that adopted by Yamano,
Takemura and Maeno [15]. Dollar and Howe present many
other types of couplings that have been employed in various
underactuated hands [18]. The primary disadvantage of
these structures is the lack of manipulation capabilities.
Fully actuated or nearly fully actuated structures do
allow the greater manipulability lacking in underactuated
structures. The consequence of this flexibility is the
increased number of motors required to actuate these
degrees of freedom, where each motor must also be of
adequate size to apply the required forces. This in turn
increases the size, weight, and control complexity of the
prosthetic hand.
Few prosthetic hands are actually fully actuated. This
could arise from the complexity that is introduced in the
design or from the fact that the tasks which are to be
performed are modeled after the underactuated human hand
for grasping actions only. One hand that does fully actuate
the finger’s degrees of freedom is the UB-3 hand [19]. In
this finger each phalange has a tendon attached to it and is
able to actuate all flexing degrees of freedom independently.
In the design presented in this paper, the actuators and
actuation structure were chosen to specifically complement
the task characteristics of each region shown in Fig 4. The
smaller, faster, and efficient electromagnetic motors are
incorporated into a nearly fully actuated kinematic structure
and chosen to perform the tasks associated with region 1.
The quiet, lightweight, strong shape memory actuators
implemented in a parallel underactuated structure were
selected to provide sufficient strength to the system when
required. The complementary function of both actuation
systems provides the prosthetic hand with a broad capability
for grasping and manipulating actions while trying to
optimize actuator size and performance.
III. PROSTHETIC HAND DESIGN
The developed design is shown in Fig 5. The design is
dimensionally consistent with that of an average male
human hand [2] and possesses the same degrees of freedom.
The anthropomorphic aspect of the hand is intended to
enhance the amputee’s acceptance and usability. The DIP
and PIP joints of the finger and the IP and MCP joints of the
thumb are coupled. This is achieved by connecting a single
actuator to both the PIP joint (bevel gears) and DIP joint
(pulley connection on metacarpal phalange). This coupling
technique is common among many prosthetic/robotic hands
as noted in the section above.
Fig 5: Prosthetic hand design
The developed parallel actuation structures discussed in
the previous section are shown in Fig 6 and Fig 7 for the
finger and Fig 8 and Fig 9 for the thumb. The movements
associated with region 1 in Fig 4 are achieved by two DC
motors. The DC motors actuating the coupled DIP/PIP
joints of the finger and IP/MCP joints of the thumb are
embedded in the proximal phalange of the finger and the
metacarpal phalange of the thumb. The DC motor in the
metacarpal phalange of the finger actuates the horizontal
degree of freedom of the MCP joint. The DC motor at the
base of the thumb actuates the CMC joint to obtain an
approximation of the abduction/adduction motion. The
second degree of freedom of the finger’s MCP joint
(abduction/adduction) is only subject to compliance without
actuation. The second degree of freedom in the thumb’s
CMC joint (flexion/extension) is actuated by the region 2
actuation structure.
Fig 6: Region 1 actuation scheme for the finger
Fig 7: Region 2 actuation scheme for the finger
Fig 8: Region 1 actuation scheme for the thumb
Fig 9: Region 2 actuation scheme for the thumb
The actuation structure corresponding to region 2 in Fig
4 for the finger includes a light cable that passes over two
restraining shafts in the MCP joint of the finger, coils in the
proximal phalange, and embeds in the middle phalange. The
string is kept in light tension by a tension unit at the back of
the hand while the region 1 actuation structure is active.
When region 2 actuation is required the shape memory alloy
actuates a spring loaded cam which in turn pinches the string
between itself and a roller beneath it. As the shape memory
alloy continues to actuate, the cam introduces the additional
force required for region 2 tasks. At task completion the
electric signal causing the shape memory alloy to heat up is
stopped and the DC motors and cam spring extend the shape
memory alloy back to its original state. The spring loaded
cam mechanism is more definitively shown in Fig 10 and
Fig 11.
Fig 10: Region 2 spring loaded cam mechanism prior to SMA actuation
(thinner line representing cable corresponds to lower tension applied by
tensioner)
Fig 11: Region 2 spring loaded cam mechanism during SMA actuation
(thicker line representing cable corresponds to additional tension applied by
SMA via cam mechanism)
The thumb’s region 2 actuation structure is similar to that
of the finger’s region 2 actuation structure. However, unlike
the finger, this structure actuates the degree of freedom at
the CMC joint that is not actuated by the DC motor. This is
based on the observation that this degree of freedom is more
apposing of the fingers during tasks that would require
additional force (power grasp, high force pinch grasp, lateral
grasp, etc. [8]).
The design shown in this section has been manufactured
using a rapid prototyping machine. The prototype can be
seen in the Results and Discussion section.
IV. CONTROLLER DETAILS
The control of the DC motors in this research was
accomplished through the use of pulse width modulation
(PWM) and direction control. PWM allowed the applied
voltage to be varied continuously which controlled the speed
and torque of the motor. Direction control was used to
determine the spin direction of the motor.
PWM and direction control functionality was provided
by a Pololu Qik 2s12v10. The Qik motor controller
provides two channels of speed and direction control for
brushed DC motors and is controlled by a serial interface. In
this research, the motor controller was connected to a PC
running LabVIEW to provide the serial commands. The
connection from the PC to the Qik was made through an
intermediate device, the Pololu Jrk 21v3, to provide the
conversion from USB to the required serial connection. The
Jrk also provides motor control functionality, but it was not
implemented in this research.
Additionally, EMG signal capturing capability was
added to control the DC motors. The raw signals from the
EMG sensor are shown in Fig 12 and Fig 13.
Fig 12: Raw EMG signal with finger in relaxed state
Fig 13: Raw EMG signal with finger in flexed state
The raw signals were processed by taking the maximum
absolute value over a 100 sample interval and generating a
new data set. In order to get the full flex and extend ability
of the finger from this new data set two threshold levels
where introduced. Above one threshold the motor would
spin in one direction (flex) and below another the motor
would spin in the opposite direction (extend). The deadzone
was implimented as a buffer area to transition between the
flexation and extention zones.
The EMG signal was acquired from the first author’s
digit III FDP muscle and supplied as an analog voltage from
a separate PC incorporating a Delsys Bagnoli EMG system.
The analog voltage was sampled through a National
Instruments data acquisition card installed in the motor
control PC. The analog signal was then filtered and scaled
using LabVIEW to a value representing the amplitude of the
EMG signal, which was then used to control the direction of
the motors. A sample of the filtered and scaled signal with
thresholds is shown in Fig 14.
Fig 14: EMG signal as finger was flexed and relaxed
The entire controller setup can be seen in the schematic
shown below in Fig 15.
Fig 15: Motor controller setup
V. EXPERIMENTAL SETUP
The experiment associated with this research was
performed to validate the designed actuation structure’s
ability to span a two-region strength space similar to the one
identified above for the human hand. The actuators used in
the prototype were Pololu 298:1 Micro Metal Gearmotor HP
DC motors [21] and the Electric Piston SMA actuator from
Raychem [22]. These actuators are relatively inexpensive
and the implementation of more expensive actuators could
further enhance the values reported below.
The experiment consisted of having the finger and the
thumb grip a FSR sensor fastened to a dense foam ball in the
large grip and then the close grip configurations as shown in
Fig 16 through Fig 19. The large grip setup simulated the
finger/thumb performing a more robust grasp on a larger
object and the small grip setup simulated the actuation
structure of the finger/thumb grasping smaller objects or
performing the more dexterous pinch or lateral grasp. The
values obtained experimentally are compared to the
expected calculated values using a simple friction inclusive
static calculation of the reaction force required at the FSR
sensor to resist the stall torque of the DC motors and the
measured tension provided by the SMA.
Fig 16: Finger in large grip configuration.
Fig 17: Finger in close grip configuration.
Fig 18: Thumb in close grip configuration.
Fig 19: Thumb in large grip configuration.
The experimental values were compared to the expected
ones using friction inclusive static calculations based on the
free body and geometric diagrams shown in Fig 20 and Fig
21. The “T” vectors in the free body diagrams represent the
tension in the cable. The tension values are calculated using
(1) where the Vi and Vj are the geometric vectors of the cable
on either side of a pivot point, ti and tj are the tension values
in these vectors, and μ is the friction coefficient of the pivot
point. Because tj is the only unknown in this equation it is
separated and solved for in order to determine the next
tension vector in the system.
(
)
VtVtVtVt j
i
i
j
j
i
i=+−
μ
(1)
Fig 20: Primary variables used in friction inclusive static calculations of
finger in both close and large grip configurations.
Fig 21: Primary variables used in friction inclusive static calculations of
thumb in both close and large grip configurations.
With the tension vectors determined, the simple static
equations (1) and (2) for the finger and thumb are used to
solve for the forces/torques of each joint from the distal
phalange to the knuckle in terms of the variable F value then
using the equations to solve for F. The stall torque applied
by the DC motors (90 oz-in) replaced the M1 and M2a/M2b
variables. As described in the design section M2a and M2b
are coupled and their relationship is shown in (3). The
gravitational forces generated by each phalange was applied
at the center of the link in the downward direction using the
mass values for the finger of (Mprox=21.6g, Mmid=9.45g,
Mdist=6.7g) and for the thumb of (Mmet=33.3g, Mprox=10.5g,
Mdist=8.5g).
∑=0M (2)
∑=0F (3)
inozbMaM −=+ 902
14
8
2 (4)
VI. RESULTS AND DISCUSSION
Each of the four configuration setups were performed
three times with each actuator type individually and then
with them combined. The results are shown in Table 1
where the columns correspond to the resulting forces from
the SMA, DC motors, the results of the SMA and DC
motors column summed together, the experienced force
when the two actuator types are physically applied at the
same time, and the value calculated using basic statics. The
table shows that when applied individually, the resulting
forces from the SMA are for the most part larger than the
applied DC motor forces for all four configurations. The
average combined forces for the finger are generally greater
than the experienced but the values for the thumb are
approximately the same.
Table 1: Finger & Thumb Testing Results
Finger M1 & M2 & SMA (Newtons)
Finger CG SMA DC Motors Combined
(Sum)
Combined
(Experimental)
Combined
§
(Calculated)
Combined
£
(Calculated)
Average 7.75 9.78 17.52 15.23 16.55 16.85
Std Dev 0.21 0.22 0.15 0.20 0.08 0.08
Finger LG SMA DC Motors Combined
(Sum)
Combined
(Experimental)
Combined
§
(Calculated)
Combined
£
(Calculated)
Average 8.17 8.67 16.84 15.08 15.23 15.46
Std Dev 1.26 1.14 1.27 0.80 0.30 0.30
Thumb M1 & SMA (Newtons)
Thumb CG SMA DC Motor Combined
(Sum)
Combined
(Experimental)
Combined
§
(Calculated)
Combined
£
(Calculated)
Average 4.70 1.2 4 5.94 5.60 5.14* 5.61*
Std Dev 0.90 0.35 0.86 1.14 0.05 0.06
Thumb LG SMA DC Motor Combined
(Sum)
Combined
(Experimental)
Combined
§
(Calculated)
Combined
£
(Calculated)
Average 4.56 3.4 5 8.01 8.18 7.70* 8.27*
Std Dev 1.85 0.45 2.45 0.39 0.49 0.58
* Values combined from calculated SMA values and measured DC Motor values
§ Values where calculated using friction coefficients of μ(nylon)=0.102 and μ(steel)=0.09
£ Values where calculated using friction coefficients of μ(nylon)=0.52 and μ(steel)=0.09
The finger behavior is expected because the SMA was
only partly isolated from the motors as the motors were still
employed during the SMA test in order to maintain the
structure of the finger against the ball prior to testing. This
resulted in the SMA having to press against the ball and
overcome the minor resistance presented by the motors. The
thumb behavior was expected as well due to the slightly
compliant units isolating the motor during the motor test
absorbing some of the applied motor force. The calculated
forces where also expectedly higher than the measured data
as friction from the system was not incorporated in their
development.
It is to be noted that the combined (calculated) values for
the thumb are only calculated for the SMA input and the DC
Motor experimental value has been added to the calculated
value. This was done because a part of the component in the
prototype thumb broke causing the gearing to not mesh
appropriately, giving significantly smaller force values
(approx. 3N) than those calculated (approx. 14N).
The controller was only applied to the DC motors of
finger and thumb in their close grip configurations. The
objective of these tests where to merely validate that the
controller platform had the power and functional capacity to
produce similar results to that of the power supply inputs
using an EMG signal input.
The EMG voltage and current outputs for both the finger
and thumb were consistent with that of the direct power
supply inputs. The EMG thumb test resulted in a DC motor
force of 0.95N which agreed with the 1.24N + 0.35N
generated by the direct power supply tests. It was also
visually verified that the finger squeezed the ball to a similar
degree with the EMG controller as the direct power supply
test system. However, a later identified disconnect in the
FSR sensor did not allow the generation of reliable force
data for the finger. Successive experiments for the finger
resulted in a break in the rapid prototyped proximal
phalange. This experiment is being repeated immediately
with a soon to be re-built proximal phalange
VII. CONCLUSIONS AND FUTURE WORK
This paper presents a novel design and actuation system
for a prosthetic hand. The actuation structure was shown to
effectively span a two-region strength space to execute grip
configurations similar to those found in the three primary
grips of the human hand (power, pinch, and lateral). The
design also showed the ability of an underactuated and
nearly fully actuated kinematic structure to exist in a single
actuation system using remarkably different types of
actuators, without compromising the required size and
weight of the prosthetic hand.
According to these results, the parallel actuation structure
is a good starting point for the design. The results also
validated that the design could be effectively driven by an
EMG signal. Complete testing of the parallel actuation
system’s performance will require the expansion of the
EMG inputs to perform different grasps and manipulations.
Future work includes further testing, development of the
controller, fabrication of a prototype that is more robust yet
of the same form as the rapid prototype components tested
here, and the use of the dynamical model of the hand for
manipulation tasks, in order to improve the design. More
precisely selected actuators will be used in the final design.
The stronger actuators will add the appropriate scale to the
developed strength space so as to mimic the force generation
capabilities of the system as well as the strength space form.
ACKNOWLEDGMENT
This research was performed under an award/contract
from Telemedicine Advanced Technology Research Center
(TATRC), of the U.S. Army Medical Research and Material
Command (USAMRMC) of the U.S. Department of
Defense.
REFERENCES
[1] Bicchi, A., 2000, “Hands for dexterous manipulation and robust
grasping: a difficult road toward simplicity”, IEEE Transactions on
Robotics and Automation, 16(6):652-662.
[2] Foegelle, M., Li, K., Pavacic, A., Moller, P., “Test & Measurement:
Developing a Standard Hand Phantom for Wireless Testing,” [Online].
Available:
http://www.wirelessdesignmag.com/ShowPR.aspx?PUBCODE=055&
ACCT=0029956&ISSUE=0811&RELTYPE=tm&PRODCODE=W02
60&PRODLETT=A&CommonCount=0,
[3] Norkin, C., Levangie, P., 1992, Joint Structure & Function: A
Comprehensive Analysis: Second Edition, F. A. Davis Publications,
262-299.
[4] Nelson, D. L. “Tendon Laceration Page,” [Online]. Available:
http://www.davidlnelson.md/Tendon_laceration.htm,
[5] Valero-Cuevas, F. J., Towles, J. D., Hentz, V. R., “Quantification of
fingertip force reduction in the forefinger following simulated
paralysis of extensor and intrinsic muscles,” Journal of Biomechanics
33 (2000) 1601-1609.
[6] Kevin G. Keenan, K. G., Veronica J. Santos, V. J., Venkadesan, M.
and Valero-Cuevas, F.J., Maximal Voluntary Fingertip Force
Production Is Not Limited by Movement Speed in Combined Motion
and Force Tasks, The Journal of Neuroscience, July 8, 2009 •
29(27):8784–8789.
[7] Kutch, J. J., Valero-Cuevas, F. J., “All muscles are redundant but
some are less redundant than others,” poster presentation for USC
Viterbi School of Engineering Brain-Body Dynamics Lab.
[8] Perez-Gracias, M. A., http://progeny.isu.edu/~alba/, Idaho State
University.
[9] Cutkosky, M.R., 1989, “On grasp choice, grasp models, and the
design of hands for manufacturing tasks”, IEEE Transactions on
Robotics and Automation, 5(3):269-279.
[10] Butterfass, J.; Fischer, M.; Grebenstein, M.; Haidacher, S. and
Hirzinger, G.., 2004, “Design and Experience with DLR Hand II”,
Proceedings of the World Automation Congress 2004, Seville, Spain,
2004.
[11] Touch Bionics Inc [Online]. Available: www.touchbionics.com
[12] Shadow Robot Company [Online].
www.shadowrobot.com/hand/papers.shtml
[13] Fite, K., Withrow, T., Shen, Z., Wait, K., Mitchell, J., Goldfarb, M.,
2008, “A Gas-Actuated Anthropomorphic Prosthesis for Transhumeral
Amputees,” IEEE Transactions on Robotics, 24(1):159:169.
[14] Pylatiuk, C., Schulz, S., Kargov, A., Bretthauer, G., 2004, “Two
Multiarticulated Hydraulic Hand Prostheses,” Artifical Organs,
28(11),980:986.
[15] I. Yamono, T. Maeno, 2005, “Five-fingered Robot Hand using
Ultrasonic Motors and Elastic Elements,” Proceedings of the 2005
IEEE International Conference on Robotics and Automation,
Barcelona, Spain, Aprill 2006.
[16] Andrianesis, K., Tzes, A., “Design of an Anthropomorphic Prosthetic
Hand Driven by Shape Memory Alloy Actuators,” Proceedings of the
2nd Biennial IEEE/RAS-EMBS International Conference on
Biomedical Robotics and Biomechatronics, Scottsdale, AZ, USA,
October 19-22, 08.
[17] Carrozza, M., Cappielo, G., Micera, S., Edin, B., Beccai, L., Cipriani,
C., 2006, “Design of a cybernetic hand for perception and action,”
Biol Cybern (2006) 95:629-644.
[18] Dollar, A.M., and Howe, R.D., 2006, “Joint coupling design of
underactuated grippers”, Proc. of the ASME 2006 International
Design Engineering Technical Conferences and Computers and
Information in Engineering Conference, September 10-13, 2006,
Philadelphia, PA, USA.
[19] Dollar Lotti, F., Tiezzi, P., Vassura, G., 2005, “Development of UB
Hand 3: Early Results,” Proceedings of the 2005 IEEE International
Conference on Robotics and Automation, Barcelona, Spain, April
2005.
[20] Hollerbach, J., “ME 6220: Introduction to Robotics”, Chapters 9 & 10
of Course Material, 2007.
[21] Pololu Robotics & Electronics, “298:1 Micro Metal Gearmotor HP,”
[Online] Available: http://www.pololu.com/catalog/product/994
[22] Electric Piston Instruction Manual, [Online]. Available:
www.robotstore.com/download/EPiston_Instr_V12.pdf
[23] Carrozza, M., Cappiello, G., Stellin, G., Zaccone, F., Vecchi,
F.,Micera, S., Dario, P., “A Cosmetic Prosthetic Hand with Tendon
Driven Under- Actuated Mechanism and Compliant Joints: Ongoing
Research and Preliminary Results,” Proc. of the 2005 IEEE
International Conference on Robotics and Automation, April 2005,
Barcelona, Spain.
[24] Vanderniepen, I., Van Ham, R., Van Damme, M., Lefeber, D.,
“Design of a powered elbow orthosis for orthopaedic rehabilitation
using compliant actuation,” Proc. of the 2nd Biennial IEEE/RAS-
EMBS International Conference on Bimedical Robotics and
Biomechatronics, October 19-22, 2008, Scottsdale, AZ, USA.
[25] Zollo, L., Roccella, S., Guglielmelli, E., Carrozza, M., Dario,
“Biomechatronic Design and Control of an Anthropomorphic
Artificial Hand for Prosthetic and Robotic Applications,”
IEEE/ASME Transactions on Mechatronics, Vol. 12, No. 4, August
20007.
[26] Engeberg, E. D., Meek, S., “Improved Grasp Force Sensitivity for
Prosthetic Hands Through Force-Derivative Feedback,” IEEE
Transactions on Biomedical Engineering, Vol. 55, No. 2, February
2008.
[27] Engeberg, E. D., Meek, S., “Model-Based Feedforward Derivative
Compensation for Prosthetic Hands,” Proc. of the 2nd Biennial
IEEE/RAS-EMBS International Conference on Bimedical Robotics
and Biomechatronics, October 19-22, 2008, Scottsdale, AZ, USA.
[28] Engeberg, E. D., Meek, S. G., Minor, M. A., “Hybrid Force-Velocity
Sliding Mode Control of a Prosthetic Hand,” IEEE Transactions on
Biomedical Engineering, Vol. 55, No. 5, February 2008.
[29] Cipriani, C., Zaccone, F., Micera, S., Carrozza, M., “On the Shared
Control of an EMG-Controlled Prosthetic Hand: Analysis of User-
Prosthesis Interaction,” IEEE Transactions on Robotics, Vol. 24, No.
1, February 2008.
[30] Blank, A., Okamura, A. M., Kuchenbecker, K. J., “Effects of
Proprioceptive Motion Feedback on Sighted and Non-Sighted Control
of a Virtual Hand Prosthesis,” Symposium on Haptic Interfaces for
Virtual Environments and Teleoperator Systems, March 13-14,
2008, Reno, Nevada, USA.
[31] Robinson, D. W., Pratt, G. A., “Force Controllable Hydro-Elastic
Actuator,” Proc. of the 2000 IEEE International Conference on
Robotics and Automation, April 200, San Fransico, CA, USA.
[32] Pylatiuk, C., Schulz, S., Kargov, A., Bretthauer, G., 2004, “Two
Multiarticulated Hydraulic Hand Prostheses,” Artifical Organs,
28(11),980:986.
[33] Institute of Robotics and Mechatronics [Online]:
http://www.dlr.de/rm/en/desktopdefault.aspx/tabid-3975/6161_read-
245/
[34] Winfrey, Rex C., 2008, “Prosthetic Hand Having a Conformable,
Compliant Grip, and Opposable, Functional Thumb,” Patent
7,361,197, April 22, 2008.
