Integrating Robust Gait Tracking Control and Variable Admittance Model for Pediatric Exoskeleton: An ‘Assist-as-Needed’ Framework

Abstract

Traditional passive-assist rehabilitation, using a fixed trajectory, often falls short in engaging patients with residual muscle strength, leading to substandard outcomes. This work proposes a novel ‘assist-as-needed’ control strategy for a pediatric lower-limb exoskeleton system. It employs variable admittance control (VAC) based on the neuro-fuzzy inference technique, encouraging realistic subject-exoskeleton interactions and participation. Thereafter, a robust adaptive backstepping sliding mode (ABSM) control with a rapid reaching law (RRL) for gait tracking is introduced. The stability of the proposed position control is established using carefully selected Lyapunov candidate functions. Through numerical simulations, we compare our approach (ABSM-VAC) with an adaptive backstepping-fixed admittance control (AB-FAC), demonstrating improved tracking, compliance, and safety in active-assist gait training. This innovative method can significantly enhance rehabilitation outcomes, particularly for those with partial muscle strength, providing a more engaging and effective therapy experience.