Bl-Nonlinear Model Predictive Control Design: A Case Study on Wrist Rehabilitation Robot

To maximize the benefits of robot rehabilitation and increase patient involvement in their treatment, control algorithms must be improved. While carrying out the movement, the model predictive controller can effectively address this need. This study suggests the nonlinear model predictive control on the basis of Bayesian learning (BL-NMPC) method to increase wrist robot rehabilitation. Contrary to the NMPC control method, this method estimates the nonlinear term (regarded as uncertainty) with $\mathcal{B}\mathcal{L}$ algorithm and controls signal decomposition without the requirement to transform the nonlinear system to a semi-linear system or utilize an observer. Lastly, the simulation results of the rehabilitation robot are examined and contrasted with the outcomes of the nonlinear MPC approach to see how well the proposed method performs.