Control Design for Human-like Reaching Movements using Redundancy in Robot Arm-Trunk Systems

Abstract

This paper develops a control algorithm to show the human-like reaching movements in humanoid redundant systems involving the trunk. This algorithm neither requires the computation of pseudo-inverse of Jacobian nor does it need the optimization of any artificial performance index. The control law accommodates the time-varying temporal properties of the muscle stiffness and damping as well as low-pass filter characteristics of human muscles. It uses a time-varying damping shaping matrix and a bijective joint muscle mapping function to describe the spatial characteristics of human reaching motion like quasi-straight line trajectory of the end-effector and symmetric bell shaped velocity profile as well as the temporal characteristics like the occurrence of the peak velocity of the trunk motion after the peak velocity of the arm motion. The aspect of self-motion is also analyzed using the null-space motion of the manipulator Jacobian. The effects of the control parameters on the motion pattern are analyzed in detail and some basic guidelines have been provided to select their proper values. Simulation results show the efficacy of the newly developed algorithm in describing human-motion characteristics.