Upper Limb Electrical Stimulation Using Input-Output Linearization and Iterative Learning Control

Abstract

A control scheme is developed for multijoint upper limb reference tracking using functional electrical stimulation (FES). In accordance with the needs of stroke rehabilitation, FES is applied to a reduced set of muscles in the arm and shoulder, with support against gravity provided by a passive exoskeletal mechanism. The approach fuses input-output linearization with iterative learning control, one of the few techniques to have been applied in clinical treatment trials with patients. This powerful hybrid control structure hence extends performance and scope of clinically proven technology for widespread application in rehabilitation robotic and FES domains. In addition to simplifying tracking and convergence properties of the stimulated joints, the framework enables conditions for the stability of unstimulated joints to be derived for the first time. Experimental results confirm tracking performance of the stimulated joints, together with unstimulated joint stability.