Passivity-Based Learning Control for Torque and Cadence Tracking in Functional Electrical Stimulation (FES) Induced Cycling

Abstract

This paper examines torque tracking accomplished by the activation of lower-limb muscles via Functional Electrical Stimulation (FES) and cadence regulation by an electric motor. Challenges arise from the fact that skeletal muscles evoke torque via FES in a time-varying, nonlinear, and delayed manner. A desired torque trajectory is constructed based on the crank position and determined by the knee joint torque transfer ratio (i.e., kinematic efficiency of the knee), which varies as a periodic function of the crank angle. To cope with this periodicity, a repetitive learning controller is developed to track the desired periodic torque trajectory by stimulating the muscle groups. Concurrently, a sliding-mode controller is designed for the electric motor to maintain cadence tracking throughout the entire crank cycle. A passivity-based analysis is developed to ensure stability of the torque and cadence closed-loop systems.