Somatosensory-Enhanced Decoupled Motion Control Strategy for a Novel Redundant Actuation Rehabilitation Robot Featuring Dual Parallel Washout

Abstract Background With the increasing demand for balance disorder treatment, rehabilitation training robots are continuously evolving and advancing. Lower-limb end-effector rehabilitation robots can perform human-like gait movements and provide assistance to help patients complete rehabilitation training. The degree of realism in the somatosensory experience and the level of movement reproduction provided to patients directly determine the therapeutic effect. This study focuses on a rehabilitation training robot designed with a modular structural approach, featuring serial-parallel hybrid redundant actuation, and conducts theoretical and experimental research on its somatosensory enhancement technology and motion control strategies. Methods The washout algorithm(WA) is commonly used for dynamic simulation and motion control, which can enhance patients' training experience and treatment effectiveness in the field of rehabilitation. This study uses multi-objective genetic algorithm to optimize the filter parameters of WA, with the aim of reducing the motion perception error caused by the robot to patients. And different optimization objectives are used for different structural modules of the robot to better match the modules' motion with human perception thresholds and the dynamic response characteristics of the modules. A motion control strategy with enhanced motion perception was proposed by applying the optimized WA to the control of the robot. Collect walking gait motion data from healthy volunteers and use it as a predefined work trajectory for the robot. Through simulation analysis, the improvement in somatosensory simulation performance brought by the optimization algorithm has been verified. The feasibility of the motion control strategy was tested through experiments. Results Considering the structural characteristics and motion advantages of different modules, distinct optimization objectives were established for the optimization of algorithms in each module, avoiding conflicts between multiple objectives within the optimization model or a tendency to optimize towards a particular objective. Through simulation analysis, the maximum error in somatosensory acceleration after optimization decreased by approximately 30%, and there was also an improvement in the error of somatosensory angular velocity. The results from human-robot coupling gait motion experiments demonstrated that the two structural modules worked together in coordination, and under the influence of the WA, the robot was able to reproduce continuous human-like walking gait movements within a limited workspace, proving the feasibility of the proposed motion control strategy. Conclusions The motion decoupling control strategy of dual parallel WA with enhanced somatosensory proposed for the rehabilitation robot can improve the realism of the robot's somatosensory simulation and accelerate the rehabilitation process; The control algorithm is perfectly integrated with the two module devices, allowing them to fully utilize their respective motion advantages and work together stably.