Human-Robot Cooperative Adaptive Reinforcement Constraint Control for a Lower Limb Rehabilitation Exoskeleton Based on User’s Movement Intention

Abstract

Exoskeletons used for rehabilitation must operate together with the patient to adapt to the biomechanical-inspired movements of the regular human gait cycle rather than operate by following a predefined trajectory without considering the human-robot interaction effects. This work presents the assessment of a lower limb exoskeleton whose motion is performed according to a collaborative approach given the movements of the human user and the relative force concerning the exoskeleton structure. The Opensim™ software serves to define the force and position reference trajectories to follow during the gait cycle, which serves as a reference for the hybrid control. These forces and movements are compared to the results of a virtual model that considers the interaction between the user and the exoskeleton in two possible scenarios. The first scenario contemplates the implementation of a position controller to generate interaction-independent movement of the exoskeleton. The second scenario considers the force exerted by the exoskeleton on the patient to trigger a force-based controller after trespassing a predefined value. This leads to a hybrid control scheme, which considers the position restrictions in the closed-loop feedback control strategy. Using this approach, the exoskeleton can collaborate actively with the user and provide motion as required, responding to position-controlled motion if the user is not opposed to the exoskeleton motion. This novel strategy permits the evaluation of a hybrid position-force controller for wearing the active orthosis. Numeric simulations show the performance of the proposed system. These outcomes confirm the supposed advantages of the proposed controller.