Dynamic simulation of the hand

Abstract

Robots and simulations provide complementary approaches for exploring different aspects of hand modeling. Robotic systems have the advantage in that all physical interactions, such as contact between tendons and bones, are automatically and correctly taken into account. Conversely, software simulations can more easily incorporate different material models, muscle mechanics, or even pathologies. Previously, no software for dynamic simulation could efficiently handle the complex routing and contact constraints of the hand. We address these challenges with a new simulation framework well suited for modeling the hand. We use the spline basis as the system’s dynamic degrees of freedom, and place them where they are most needed, such as at the pulleys of the fingers. Previous biomechanical simulation approaches, based on either line-of-force or solid mechanics models, are not wellsuited for the hand, due to the complex routing of tendons around various biomechanical constraints such as sheaths and pulleys. In line-of-force models, wrapping surfaces are used to approximate the curved paths of tendons and muscles near and around joints, but these surfaces affect only the kinematics, and not the dynamics, of musculotendons. In solid mechanics models, the fiber-like properties of muscles are not directly represented and must be added on as auxiliary functions. Moreover, contact constraints between bones, tendons, and muscles must be detected and resolved with a general purpose collision scheme. Neither of these approaches efficiently handles both the dynamics of the musculotendons and the complex routing constraints, while our approach resolves these issues.