Robust MPC with Integral Super-Twisting for Trajectory Tracking of an Exoskeleton Robot Arm

This work aims to enhance the robustness and reduce the control effort in the presence of input and state constraints while performing rehabilitation exercises with uncertainty in the model. To that end, a model predictive control (MPC) approach with an integral super-twisting algorithm for trajectory tracking is presented. The proposed method considers the uncertainties and disturbances due to the wearer-exoskeleton interactions while optimally exploiting the robot’s physical capabilities. Simulation and real-time results show that the proposed method outperforms traditional MPC and STA control methods regarding tracking accuracy and control effort reduction. As a result, the proposed approach can be applied to other wearable exoskeletons, providing a safe control solution for improving human motor skills in rehabilitation.