Compliant Control of an Upper Limb Rehabilitation Robot Based on Impedance Design

Abstract

Robot-assisted therapy is an essential rehabilitation method to regain motor function in upper limb impairment after neurological events such as stroke. The most challenging and problematic issue in this field is to realize safe and natural human-robot interaction. Conventional robots, which are generally controlled by rigid position controllers, cannot be compliantly adjusted according to the patient's voluntary movement or spasm. The appearance of high interaction forces is also a potential cause of secondary damage. This is a big research gap that requires further research on advanced compliant control. This paper is devoted to addressing this problem by developing an adaptive impedance controller to realize compliant control for an upper limb rehabilitation robot. The dynamic model of the robot is established in the paper, and the adaptive impedance control law is formulated. Then the parameters of the impedance controller can be adjusted in real-time based on interaction forces. To validate the effectiveness of the proposed controller, several simulation analyses under different scenarios are conducted, such as tracking a predefined trajectory without any external interference and with interaction forces generated by simulated patients. The simulation results indicate that besides the controller can guaranteeing stable and accurate tracking, the controller also has a good compliance performance, i.e., the interaction forces decrease significantly when external disturbances from patients appear. The effectiveness and superiority of the adaptive impedance control design are verified by the simulation results. The physical rehabilitation system based on the adaptive impedance control design has a good theoretical reference and simulation basis.