Posture prediction and force/torque analysis for human hands

Human hands are the bridge between humans and the objects to be manipulated or grasped both in the real and virtual world. Hands are used to grasp or manipulate objects and one of the most important functionalities is to position the fingers, i.e., given the position of the fingertip and to determine the joint angles. Last year we presented a 25-degree of freedom (DOF) hand model that has palm arch functionality. In this paper we preset an optimization-based inverse kinematics approach to position this 25 DOF hand locally with respect to the wrist instead of the traditional Moore-Penrose pseudo-inverse and experiment methods. The hypothesis is that human performance measures govern the configuration and motion of the hand. We also propose contact force and joint torque prediction. The exposition addresses (1) the determination whether a point is reachable (i.e., is it within the reach envelope), (2) the prediction of a finger posture for a given target point, (3) given the finger contact force analyzing the joint torque, and (4) given joint torque analyzing finger contact force. We illustrate the methodology through examples.