Biomechanical Interface System and Neural-like Cooperative Control for the Intelligent Prosthetic Arm

To address the problem of prosthetic limb control for patients with physical disability, an sEMG (surface electromyography) based intelligent prosthetic arm system is developed to achieve coordinated hand-elbow control for patients with a higher degree of arm disability. Firstly, the non-negative matrix factorization (NMF) method is applied to extracting muscle synergy based on the muscle synergy theory, and hand movement recognition and continuous motion estimation of the elbow joint are implemented. Secondly, a "feedforward-feedback" control framework is constructed based on the intention recognition results, and feedforward supervision and feedback detection are performed on the subjects to improve the comfort and robustness of the prosthetic system by adjusting the desired control input based on the feedforward-feedback results. Then, an adaptive grip force adjustment framework is constructed for hand movements to achieve adaptive grip of objects of different stiffness and shapes through alternating force and position information control; for elbow movements, an impedance control algorithm based on recognition results is designed to achieve stable human-machine interaction control of the hand-elbow integrated prosthesis. Finally, the above control strategy is experimentally verified by 6 healthy subjects and an arm handicapped subject in order to achieve more accurate intention recognition for the overall arm motion, and the result indicates that the proposed approach can complete stable elbow flexion and extension as well as hand grasping function to achieve coordinated control of the integrated hand-elbow. The system was realized in the Beijing 2022 Winter Paralympic Games for application demonstration.