Newsletter Volume 14 - 2nd Quarter 2007
Study of human grasping leading to the development of robotic hands
CONTENTS
Case Study: Study of human grasping leading to the development of robotic hands.
Software: New 2007 versions of LifeMOD™, LifeMOD/NeckSIM™, and LifeMOD/KneeSIM™ are now shipping
BRG announces LifeMOD Employment Referral Network.
Publications: Journal Papers, Magazine Articles and Book Chapters
This issue of the newsletter presents a study of human grasping leading to the development of robotic hands. The model presented in the study exercises many new features in the new release of LifeMOD™ and is included as a self-training tutorial in the product.
BRG has been really been on the move this quarter. After many months of negotiations, BRG announces the acquisition of Tiger Software, Inc. To support this acquisition and recent professional staff increases, BRG has doubled the size of its corporate headquarters in San Clemente, Ca. The acquisition, increase in staff and facilities were necessary to support the growing LifeMOD community and to continue to drive the state-of-the-art in human modeling.
We would also like to announce a major release in our core product LifeMOD™ v.2007, a major release in the industry's premier total knee replacement solution, LifeMOD/KneeSIM™ v.2007, and a brand new product, LifeMOD/NeckSIM™ v.2007. The technology represented in these programs represent many man-years of development and reflect our commitment to responding to our active customer base as well as advancing the state-of-the-art in biomechanics simulation technology.
Also, our partnership with the Orthopaedic Research Laboratories (ORL), Cleveland Ohio, directed by A. Seth Greenwald continues to grow. The ORL has selected LifeMOD/KneeSIM as a basis for comparative studies between commercially available total knee replacement systems (TKR). They also have selected LifeMOD/NeckSIM as a basis of comparative studies for spinal devices.
In an effort to unite our ever-growing user community, we are introducing the new LifeMOD user forum. Please join today and see how creative our worldwide body of LifeMOD users can be.
To serve our growing customer base we have partnered with 11 new distributors around the world.
Finally, with the many employment inquiries from students and scientists, as well as many inquiries from our many industry clients we have decided to link up the two with the LifeMOD Employment Referral Network.
A STUDY OF HUMAN GRASPING LEADING
TO THE DEVELOPMENT OF ROBOTIC HANDS
Introduction
In normal and pathological conditions, the movement of the human hand is a very complex interaction of active and passive forces. Robotic hands are still a long way from matching the grasping, dexterity and manipulation capability of the human hands. One method to accelerate the development of robotic hands is to use computer simulation to learn more about the human grasping mechanism.
Figure 1. Shadow Dextrous Hand (left) and the LifeMOD hand (right).
In this study, we describe the detailed construction of a human hand and use it to perform a grasping exercise. The muscle-tendon forces are generated using an unique inverse-dynamics to direct-dynamics "training" method. The hand adapts to the shape of the grasped object through the modeling of the surface contact forces between the object and the complete metacarpal chains.
The complex relationship between the active muscle-tendon forces and the passive elasticities resident in the hand are studied. The model is validated by loading the tendons of cadaver hands, observing the finger behavior, and comparing this behavior to model predictions.
Features of this new model include:
- Twenty one bones modeled (carpal bones lumped as one)
- Passive tissues modeled to stabilize the joint
- Tendons guided through sheaths
- Forward dynamics - muscle driven
- Full contact between flesh and grasped object
- Anthropometrically scalable (using any of the 5 LifeMOD databases)
- Personalizable (using MRI data)
- Validated (limited)
- Quick solving (less than 1 minute clock time per simulation)
- Generalized model can be used for many gripping activities
Model Description
The human hand consists of fingers which in themselves, can be considered as a linkage system of intercalated bony systems balanced by muscle and tendon forces and joint constraints. An arm model was created by LifeMOD using parameters for a 50%-ile male from one of the internal anthropometric databases. The hand was further refined using the process outlined below.
Figure 2. The twenty one bones and the twenty joints of the hand model. The carpal bones are lumped into a single segment.
Bones
There are 27 bones within the wrist and hand. The wrist itself contains eight small bones,
called carpals. The carpals join with the two forearm bones, the radius and
ulna, forming the wrist joint. Further into the palm, the carpals connect to the
metacarpals. There are five metacarpals forming the palm of the hand. One metacarpal
connects to each finger and thumb. Small bone shafts called phalanges line up to form
each finger and thumb. Each bone was included in the model as a separate part except for the
carpal bones which were lumped into a single segment.
Joints
In the hand model, each articular joint is modeled with a kinematic spherical joint. Rotation
limitations are imposed by the ligaments described in the following section. The main knuckle
joints are formed by the connections of the phalanges to the metacarpals. These joints are
called the metacarpophalangeal joints (MCP joints). The three phalanges in each finger
are separated by two joints, called interphalangeal joints (IP joints). The one
closest to the MCP joint (knuckle) is called the proximal IP joint (PIP joint).
The joint near the end of the finger is called the distal IP joint (DIP joint).
The thumb only has one IP joint between the two thumb phalanges. Each metacarpal chain is
jointed to the lumped carpal bone segment with a joint representing the carpometacarpal
joints.
Ligaments
To restrict the range of motion of each joint in the hand, ligaments are created. Ligaments
are tough bands of tissue that connect bones together. Two important structures, called
collateral ligaments, are found on either side of each finger and thumb joint.
The function of the collateral ligaments is to prevent abnormal sideways bending of each
joint. In the PIP joint (the middle joint between the main knuckle and the DIP joint), the
strongest ligament is the volar plate. This ligament connects the proximal
phalanx to the middle phalanx on the palm side of the joint. The ligament tightens as the joint
is straightened and keeps the PIP joint from bending back too far (hyperextending). Finger
deformities can occur when the volar plate loosens from disease or injury. The
force-displacement behavior of the ligaments was modeled using data from [An, '83, Brand, '75].
Viscous damping effects for the ligaments were modeled using data derived from [Micks, '81].
Figure 3. The collateral and volar plate ligaments which restrict the joint articulation in the metacarpals.
Muscle-Tendons
The tendons that allow each finger joint to straighten are called the extensor tendons.
The extensor tendons of the fingers begin as muscles that arise from the backside of the
forearm bones. These muscles travel towards the hand, where they eventually connect to the
extensor tendons before crossing over the back of the wrist joint. As they travel into the
fingers, the extensor tendons become the extensor hood. The extensor hood flattens
out to cover the top of the finger and sends out branches on each side that connect to the
bones in the middle and end of the finger. Locations of tendon sheaths and force deflection
properties for the tendons were were obtained from [An, '88, Wells, '85, and Buchner, '83].
The muscle elements are created using LifeMOD's trainable muscle elements. When the muscles are first created they will record the contractile history given external manipulation of the hand. A subsequent forward dynamics simulation is performed with proportional-derivative (PD) controllers creating the muscle forces necessary to reproduce the motion. The PD controllers in the muscles represent a control scheme very appropriate for robotic control.
Figure 4. Dorsal view (right) and the ventral view (left) of the hand displaying tendon and muscle groups in the model.
Contact Forces
The hand is moved into position and a model of the
tennis ball is imported into LifeMOD and positioned in the palm of the hand. To enable the
physical reaction between the hand and the ball, a contact force, which includes normal and
frictional effects, is created by each ellipsoid on the hand and the ball. Data for force
deflection of the combined effects of the skin and the tennis ball were derived from
[Pawluk, '99, Serina, '98]. The friction between the ball and the skin was derived from
[Anitescu, '97].
Figure 5. Skin contact elements used to create contact and frictional forces between the hand and tennis ball.
Model Validation
For a first order validation of the model, the finger behavior of a cadaver will be compared to the finger behavior of the model under a load to the flexor tendons [Ranney, '87, Valero-Cuevas, '00]. The flexor digitorum profundus in both the cadaver and the model is loaded with a 10 N force and the resulting joint angles (DIP, PIP and MP) in the sagittal plane are reported when the system reaches equilibrium. Figure 6 displays the comparison, where it is noted that the model performs reasonably well in this simple experiment.
The final angle of model prediction are smaller than the cadaver tests. It is believed that the major reason for this is that the extension expansion was simplified as a single tendon. The lack of lateral bands requires greater tendon excursions of the single extensor tendon to achieve a give degree of flexion. The larger passive elasticities generated in the extensor muscles tends to restrict flexion of the joints.
Figure 6. Comparison of finger joint angles for a 10 N force applied to the flexor digitorum profundus of the model prediction and a cadaver.
Simulations
With a first order validation performed, the model will be used to simulate the grasping of the tennis ball. Data recorded from the simulations will include muscle forces, joint forces and finger contact forces.
Training the Muscles of the Hand
With the musculoskeletal model of the hand developed, the muscles must be trained to record
the contraction profiles necessary to grasp the ball. Motion agents are driving elements,
external to the hand which move based on an user-input motion trajectory. They are attached
to the fingers using spring elements so as to influence the motion of the hand, rather than
govern the motion. A simulation is then performed with the motion agents directing the motion
of the fingers. The fingers will conform to the tennis ball by virtue of the skin
contact elements created in the previous step.
Figure 7. Training the muscles of the hand involves moving the fingers to grasp the ball using external motion agents. The motion agents move the fingers based on user-specified trajectories. During this process muscle contractions are recorded.
Using the Trained Muscles in a Forward Dynamics Simulation
With the muscle contractile histories recorded, they may now be used in a forward dynamics
simulation to drive the hand model. LifeMOD/BodySIM uses an effective force approach to
muscle modeling. Using this approach, muscles produce the necessary forces in order to replicate
the desired motion of the body, while staying within each muscle's physiological limit. The
assumption is that if enough muscles are included, the calculated muscle forces will be
very close the the physiological muscle force values for the same activity.
Results
To provide insight to the physics of the human grasping mechanism, the forward dynamics grasping activity is investigated in terms of the flexion muscle forces, and the fingertip contact forces. Figure 8 below displays two frames of the simulation animation with the muscle graphics scaled based on activity level. It can be observed in the bottom picture that the flexor pollicis longus, flexor pollicis brevis and the adductor pollicis controlling the thumb have a high degree of activity as noted in the plot in Figure 9.
Figure 8. Two frames of the simulation animation with the initial condition (top) and the resulting grasp (bottom). The muscle graphics are scaled based on muscle activity levels.
Figure 9. Flexor muscle force for the grasping activity including data for flexor digitorum profundus (red curve), adductor pollicis (black curve) and the flexor pollicis longus (blue curve).
Figure 10. Finger pad contact forces for the index and the thumb.
Discussion
This study presents the development of a hand model with moderate complexity with the intention to provide insights to the development of robotic hands. In this simulation the ball is resting on the palm of the hand close enough to the thumb in order to generate the first contact with the digits at the thumb. The flexor muscles (figure 7) and the finger pad contact forces (figure 8) reflect this sequence. This model can provide some insights to the timing and force magnitudes necessary to create cable driven robotic hands with linear actuators. For torque driven actuators, the model could easily be reformulated with PD torque controllers at the joints to produce information on the timing and torque magnitudes necessary to create torque driven hands. Future models will include the addition of complex scheduling-type controllers to explore various grasping maneuvers.
References
- An, K.N. et al. (1983), J.Biomech. Vol. 16(6): 419-425.
- Brand, P.W. (1975), Clinical Biomech. of the Hand. C.V. Mosby.
- Micks and Reswick (1981), j. Hand Surg. 6(5): 462-267
- Wells, R.P.et al. (1985), Biomech. Current Interdisciplinary Research, Martinus Nijliof Publishers: 743-748.
- Buchner, H.J. et al (1988), J.Biomech. Vol.21 (6): 459-468.
- Seo, N.J. e al (2005), ASB 2005, Measure of Static Coefficient of Friction at Low Hand Contact Forces.
- Ranney, D. et al (1987) , J. Hand Surg. 12A(4): 566-575
- Pawluk, D. et al (1999) Dynamic lumped element response of the human fingerpad. J Biomech Engineering, 121(2):178-183, Apr 1999.
- Serina, ER, et al (1998) A structural model of the forced compression of the fingertip pulp. J. Biomech, 31(7):639-46, 1998.
- Anitescu, M, et al (1997) Formulating dynamic multi-rigid-body contact problems with friction as solvable linear complementarity problems. Nonlinear Dynamics, 14:231-247, 1997.
- Valero-Cuevas, FJ, et al (2000) Quantification of fingertip force reduction in the forefinger following simulated paralysis of extensor and intrinsic muscles. J Biomech, 33:1601-1609, 2000.
SOFTWARE
LifeMOD™
We would like to announce a powerful new release of
LifeMOD.
This release was developed as part of our long-term modeling strategy
and with close cooperation from commercial clients and our university
consortium. The user interface has been completely restructured and has
been reorganized with libraries for motion capture data sets, models,
postures, environments, contact conditions, etc. We have also expanded
our muscle modeling capability to include Hill type muscles in addition
to our LifeMOD trainable muscle elements. There is a muscle wrapping
tool which can model accurate lines of action as muscles wrap and
slide around structures. We have included ergonomic features such as
vision cones and body center-of-mass reporting. Expanded the
self-training tutorials include a tennis player, grasping hand
(featured in this newsletter) and detailed whiplash model. This
functionality is in addition to many convenience features that you will
see in the program.
LifeMOD/KneeSIM™
We would like to introduce a new version
LifeMOD/KneeSIM,
a complete virtual product development solution for Total Knee
Replacement Systems. The product, developed in collaboration with the
world's leading orthopedic companies, can be used for in vivo
simulation to ensure proper tibial rotation, validate femoral rollback,
minimize patella shear, maximize quadriceps efficiency and increase
durability and design robustness. LifeMOD/KneeSIM has proven itself in
the design of PCL/ACL sacrificing/retaining and fixed/mobile bearing
component systems. With simulation times of less than five minutes,
hundreds of design iterations may be characterized in a day. With this
powerful tool, engineers gain insight into cause and effect
relationships between design changes and response in the body.
See case study.
See press release.
LifeMOD/NeckSIM™
We would also like to introduce a new product, LifeMOD/NeckSIM, a
complete detailed cervical spine modeling tool. LifeMOD/NeckSIM is a
plug-in to the main LifeMOD program offering the incorporation
of detailed cervical spine models in a full body LifeMOD model.
LifeMOD/NeckSIM automates the vertebrae discretization, disk force
generation and soft tissue generation. The muscles may be either
Hill-type or trainable LifeMOD elements. Complete with the new tissue
wrapping feature, accurate lines of action may also be represented in
the spine model. This tool has great utility with applications in
injury, safety, orthopedics, surgical simulation, etc.
See case study.
Learn more by visiting our web page or by contacting us at newsletter@lifemodeler.com.
SERVICES
The Biomechanics Research Group, Inc. is a service-based organization with many major commercial successes using LifeMOD human simulation in the design process. We are expert in the development of the appropriate models, simulation cases and human response signature development to accelerate innovation of your product and greatly reduce the time to market. Contact us to see how we have done this in many industries.
OTHER NEWS
LifeMOD/KneeSIM
To replicate normal knee function, Smith & Nephew conducted in-depth analyses of natural
knee kinematics and the inherent limitations in knee replacement systems.
LifeMOD/KneeSIM played a key role in
this investigation. "We chose to partner with BRG because of their experience,
technology, and expertise in the area of applying virtual product development engineering to
human activity" states Brian McKinnon, one of the lead engineers on the Journey Knee
Program. This effort led to the innovative JOURNEY anatomical knee system designed to move
and feel like a normal knee, and addresses problems still found in conventional systems such
as instability and limited flexion. The new Journey Knee System was voted the
"Best in Show" at the 2006 AAOS meeting in Chicago last year.
See press release.
ORL
This month we are excited to announce the expansion of our partnership with the
Orthopaedic Research Laboratories, in Cleveland Ohio,
under the direction of A. Seth Greenwald, D.Phil. (Oxon). In addition to using
LifeMOD/KneeSIM as a basis
for comparative studies between commercially available total knee replacement systems,
they have chosen
LifeMOD/NeckSIM
as a basis of comparative studies for spinal devices. These capabilities will complement
their existing efforts to provide functional performance data to a significant number
of national and international manufacturers, the Food and Drug Administration (FDA) and
the orthopaedic surgeon community. All these efforts are directed toward an optimization of
surgical techniques, orthopaedic devices and surgeon education to improve patient outcome.
LifeMOD User Forum
In an effort to foster communication with the worldwide LifeMOD community at large, we are
introducing the LifeMOD User Forum.
In addition to LifeMOD users we have many experts in the general multibody dynamics community
enlisted to help answer general ADAMS-related questions. Please join today and take part in
this active knowledge base.
Partners
To respond to the ever-growing LifeMOD community we have developed partnerships with several
more full service resellers across the world including:
- China --- Zhongtian-Noah Sports Science Co. Ltd.
- France --- INCAT France
- Germany --- VI-Grade, LLC
- Greece --- O Vision 2000 S.A.
- India --- CSM Software Private Ltd.
- Italy --- Lista Studio, SRL
- Japan --- Inter-Reha CO., Ltd.
- Japan --- Nihon VI-Grade
- Korea --- AhTTi
- Netherlands --- SayField International
- Romania --- Magic Engineering, SRL
- South Africa --- MSC-Africa Simulating Reality (Pty), Ltd.
- Spain --- Analisis y Simulacion S.L.
- Taiwan --- Taiwan Auto-Design Co.
- Thailand --- Sigma Solutions Co., Ltd.
- Turkey --- Bias Muhendislik/Engineering
- UAE --- Nezmin Trading
- UK --- Marlbrook, Ltd.
- USA --- Diverse Solutions, Inc.
- USA --- ESA Corp.
Referral Network
Throughout the years, we have received many employment inquiries from student graduates,
and professionals seeking employment to utilize their skills with LifeMOD. Also, we have
been repeatedly asked for references by our top industry clients for LifeMOD-trained talent. We
have decided to combine both these needs in the LifeMOD Employment Referral Network. If you
are a LifeMOD trained investigator or an industry professional in need of a LifeMOD trained
investigator please contact us.
New Publications
New Publication: "The Biomechanical Revolution ", Watt Now, March 2007
New Publication:"A step towards biomechanical analysis of surgeon's gesture on Adams-LifeMod platform", F. Cavallo, et al., Computer Assisted Radiology and Surgery, Berlin, Germany, June 27-30, 2007
New Publication: "LifeMOD modelling of a complete human body: a walk with a right knee varus and valgus movement", G. Agnesina, et al., 5th World Congress of Biomechanics, Munich, Germany, July 29, 2006
New Publication: "Evaluation of Wheelchair Occupant Safety in Frontal & Side Impact of Wheelchair Loaded Vehicle by Computer Simulation Analysis Method (ADAMS+LifeMOD)", SM Kim, et al. , 5th World Congress of Biomechanics, Munich, Germany, July 29, 2006
New Publication: "Mass Customisation of Medical Devices and Implants", K.W. Dalgarno, et al., Virtual and Physical Prototyping, Vol 1, Sept 2006. See PDF
New Publication: "Design and research of biomechanical models of human with joint replacements", K. Kazlauskiene, Kaunas University of Technology Doctoral Dissertation, Lithuania. See PDF
New Publication: "Design of an active walking-aid for elderly people", P. Mederic, et al., Laboratoire de Robotique de Paris. See PDF
New Publication: "Investigation of a robot-human impact", S. Oberer, et al., SME Robot Special Session. See PDF
If you would like further information on our software and services, please give us a call.
Copyright© 2007 LifeModeler, Inc.