Intention-Driven Variable Impedance Control for Physical Human-Robot Interaction

Abstract

The technology of physical human-robot interaction (pHRI) aims to combine the advantages of both humans and robots, and it is now playing a significant role in many industrial and medical applications. However, the open issues of safety and efficiency have not been systematically addressed for pHRI. In this paper, a new intention-driven variable impedance controller is proposed for compliantly-driven robots, where the robot is controlled to track the estimated human motion intention under a variable desired impedance model. The human motion intention is represented with the desired angle of the human limb and approximated with adaptive neural networks (NNs). Then, the impedance parameters are adjusted online according to the offset between the desired angle and the current joint configuration of the robot. Such a formulation allows the robot to predict human behavior for better collaboration and hence improve the safety and the efficiency of pHRI. The stability of the closed-loop system is rigorously proved with Lyapunov methods, and experiments on a lower-limb exoskeleton are conducted to evaluate the performance of the proposed controller.