A Novel Synthesis of Computational Approaches Enables Optimization of Grasp Quality of Tendon-Driven Hands

Exploring the high-dimensional structure of muscle redundancy via subject-specific and generic musculoskeletal models

Subject-specific and generic musculoskeletal models are the computational instantiation of hypotheses, and stochastic techniques help explore their validity. We present two such examples to explore the hypothesis of muscle redundancy. The first addresses the effect of anatomical variability on static force capabilities for three individual cat hindlimbs, each with seven kinematic degrees of freedom (DoFs) and 31 muscles. We present novel methods to characterize the structure of the 31-dimensional set of feasible muscle activations for static force production in every 3-D direction. We find that task requirements strongly define the set of feasible muscle activations and limb forces, with few differences comparing individual vs. species-average results. Moreover, muscle activity is not smoothly distributed across 3-D directions. The second example explores parameter uncertainty during a flying disc throwing motion by using a generic human arm with five DoFs and 17 muscles to predict muscle fiber velocities. We show that the measured joint kinematics fully constrain the eccentric and concentric fiber velocities of all muscles via their moment arms. Thus muscle activation for limb movements is likely not redundant: there is little, if any, latitude in synchronizing alpha-gamma motoneuron excitation-inhibition for muscles to adhere to the time-critical fiber velocities dictated by joint kinematics. Importantly, several muscles inevitably exhibit fiber velocities higher than thought tenable, even for conservative throwing speeds. These techniques and results, respectively, enable and compel us to continue to revise the classical notion of muscle redundancy for increasingly more realistic models and tasks.

Toward isometric force capabilities evaluation by using a musculoskeletal model: Comparison with direct force measurement

Developing formalisms to determine force capabilities of human limbs by using musculoskeletal models could be useful for biomechanical and ergonomic applications. In this framework, the purpose of this study was to compare measured maximal isometric force capabilities at the hand in a set of Cartesian directions with forces computed from a musculoskeletal model of the upper-limb. The results were represented under the form of a measured force polytope (MFP) and a musculoskeletal force polytope (MSFP). Both of them were obtained from the convex hull of measured and simulated force vectors endpoints. Nine subjects participated to the experiment. For one posture recorded with an optoelectronic system, maximum isometric forces exerted at the hand were recorded in twenty six directions of the Cartesian space with a triaxial force sensor. Results showed significant differences between the polytopes global shapes. The MSFP was more elongated than the MFP. Concerning the polytopes volumes, no significant difference was found. Mean maximal isometric forces provided by MFP and MSFP were 509.6 (118.4)N and 627.9 (73.3)N respectively. Moreover, the angle between the main axes of the two polytopes was 5.5 (2.3)° on average. Finally, RMS error values between MFP and MSFP were lower than 100N in 88% of the considered directions. The proposed MSFP based on a musculoskeletal model gave interesting information on optimal force orientation parameters. The possible applications in the frame of ergonomics, rehabilitation and biomechanics are proposed and discussed.

Anthropomorphic tendon-driven robotic hands can exceed human grasping capabilities following optimization

How functional versatility emerges in vertebrate limbs in spite of their anatomical complexity is a longstanding question. In particular, fingers are actuated by numerous muscles pulling on tendons following intricate paths. In contrast, the tendon-driven robotic hands with intuitive tendon routings preferred by roboticists for their ease of analysis and control do not perform at the level of their biological counterparts. Thus there is much debate on whether and how the anatomy of the human hand contributes to grasp capabilities. These parallel questions in biology and robotics arise partly because it is unclear how the number and routing of tendons offer functional benefits. We use a novel computational approach that analyzes tendon-driven systems and quantifies grasp quality to compare the precision grasp capabilities of thousands of robotic index finger and thumb designs vs. the capabilities measured in human hands. Our exhaustive search finds that neither the symmetrical designs sometimes preferred by roboticists nor randomly generated designs approach the grasp capabilities of the human hand (they are on average 73% weaker). However, optimizing for anatomically plausible asymmetry in joint centers, tendon routings, and maximal tendon tensions produces designs that can exceed the human hand by 13–45%, and outperform the preferred robotic designs by up to 435%. Thus, the grasp capabilities of prosthetic or anthropomorphic hands can be greatly improved by judiciously altering design parameters, at times in counter-intuitive ways. Moreover we conclude that, in addition to its other capabilities, the human hand’s anatomy is very advantageous for precision grasp as it greatly outperforms numerous alternative robotic designs.