Performance analysis of adaptive interaction control laws for rehabilitation robotics

Abstract

In this work a comparative analysis is carried out of adaptive compliance control laws developed by the authors for biomedical applications of human-robot interaction. They are the self-regulated compliance control in the joint space in [1], the coactivation-based compliance control in the joint space in [2] and the torque-dependent compliance control in the joint space in [3]. The control laws try to overcome limitations of the traditional interaction control by making the control gains adaptable to the interacting environment. To this purpose, inspiration from biological motor control is taken, with particular attention to the mechanisms of visco-elastic regulation of the human arm. A specific performance metrics is developed which allows measuring control performance especially in the interaction with the environment through the analysis of impact dynamics. Experimental trials of motion in the free space and impact against an obstacle are arranged with the MIT-Manus robot available in our lab, and a comparison with state-of-the-art traditional compliance controls is carried out.