Output Feedback Repetitive Learning Control of an Electrohydraulic Actuator of a Lower Limb Rehabilitation Exoskeleton

Abstract

This paper addresses the output feedback repetitive learning control of an electrohydraulic actuator of a lower limb rehabilitation exoskeleton. A repetitive learning extended state observer (RLESO) is proposed to estimate the unmeasurable system states, mismatched modeling uncertainties, and repetitive unknowns. Based on the backstepping technique and RLESO, an output feedback repetitive controller (OFRC) is developed to improve the tracking performance. Both the convergence of the RLESO and the stability of the closed-loop system are proved in a Lyapunov way. The efficiencies of the proposed RLESO and OFRC are verified via simulation. It is worthwhile to highlight that the repetitive learning scheme is integrated into both the observer and controller design parts. In this way, the remarkable performances of the output tracking via OFRC and the estimations of unmeasurable states via RLESO can be ensured.