A virtual force field-based cooperative balance control of an active variable stiffness pneumatic lower limb exoskeleton

Addressing the challenges posed by the limited flexibility and delayed response of conventional rigid-driven rehabilitation robots in complex human-computer interaction environments, a novel approach is proposed: a reconfigurable active variable stiffness principle based on the VSA model. This approach enables the adjustment and adaptation of joint stiffness by modifying the spring preload. It involves the design of a variable stiffness lower extremity exoskeleton for rehabilitation training, integrating pneumatic muscle and magnetorheological brake technologies. Theoretical stiffness models are developed, and stiffness characteristics are thoroughly simulated and analyzed in conjunction with the variable stiffness principle. To enhance the exoskeleton control system's tracking precision, a composite control strategy is formulated, combining virtual force field and impedance control methodologies. Subsequently, the stiffness characteristics and tracking performance of the joint are empirically validated. The research findings demonstrate the efficacy of the active variable stiffness model, augmented by virtual force field adaptive impedance control. Notably, this approach enhances parameter iteration speed, reduces maximum joint tracking errors by 16.634%, and extends the stiffness output range from 13.33 N.m/° to 26.15 N.m/°.