Human-Centred sEMG-Based Collaborative Control for Voluntary Participation in Rehabilitation

Abstract

This systematic review examined how surface electromyography-based (sEMG-based) human-machine collaborative control supports voluntary participation in rehabilitation for individuals with motor impairment. Following PRISMA guidelines, five databases were searched, and 91 studies were included in the main review, with 7 supplementary studies. Across the reviewed studies, sEMG mainly entered the rehabilitation control loop through intention-driven control, contribution regulation, and parameter regulation. The roles supported variable impedance control, admittance or compliant control, assist-as-needed control, shared control, and sliding mode control, and were further reflected in function-oriented active training, ability-adaptive training modes, and state and feedback regulation. Overall, the findings suggest that sEMG is increasingly used not only to detect movement intention but also to organize assistance, training progression, and feedback updates.