Adaptive Admittance Control Based on Linear Quadratic Regulation Optimization Technique for a Lower Limb Rehabilitation Robot

Abstract

Compliant, natural and safe physical human-robot interaction is of practical significance for rehabilitation robots. In our recently developed lower limb rehabilitation robot (LLRR), an adaptive admittance control based on linear quadratic regulation (LQR) optimization technique was designed to regulate parameters synchronously with the variable impedance property of human-robot interactive system. Firstly, a computed torque PD control was designed to guarantee the accuracy and stability of trajectory tracking. Secondly, an observer was designed to estimate human-robot interaction torque (HRIT) during cooperative task. Finally, a LQR optimization technique was employed to optimize admittance model parameters and minimize tracking errors and human efforts. Simulation studies were conducted on the LLRR and the results show that the HRIT can be estimated by the observer correctly and the desired trajectory was deformed smoothly and rightly with the interaction torque.