Isokinetic Muscle Strength Training Based on Optimal Speed Planning

Abstract

This work proposes a novel speed control frame-work for isokinetic muscle strength training systems with step-change and unknown dynamic characteristics, which consists of a dynamic speed planning strategy and a speed control strategy. The dynamic speed planning strategy can achieve optimal configuration for smooth transition of speed step signals, effectively suppressing system oscillations caused by abrupt speed changes and ensuring human-robot interaction friendliness. The speed control strategy includes a super-twisting non-singular fast terminal sliding mode control (SNFTSMC) and a nonlinear disturbance observer (NLDO). The NLDO is used to address the adverse effects of system modeling uncertainties and nonlinear coupling disturbances on speed control, achieving lumped disturbance estimation and compensation; And combining this with the proposed SNFTSMC forms a composite controller with a feedforward compensation term and state-feedback control, which significantly improves the system’s speed control accuracy and anti-interference performance. The method is verified to have good speed trajectory tracking performance based on both simulation and speed trajectory tracking experiments on a single-joint muscle strength training robot (SJMSTR), with speed tracking error in isokinetic training maintained below 3%.