Functional Electrical Stimulation Modulation Strategy Based On Electromyographic Feedback

Functional electrical stimulation,as an emerging rehabilitation treatment method,has been applied to the clinical treatment of limb motor function impairment with remarkable results and has largely alleviated the medical resource problem of shortage of rehabilitation therapists.However,the open-loop control method in clinical practice suffers from poor individual adaptability,unclear criteria for optimization of stimulation current parameters and low patient initiative.To address these problems,the following research work was conducted in this paper:First,an experimental platform for functional electrical stimulation was built,and a wireless sensing network was established to collect EMG signals and joint angles simultaneously,and the stimulation parameters were adjusted based on the acquired EMG signals and angle information.In addition,to address the problems of spike artifact removal and loss of effective information in the existing methods of functional electrical stimulation,we proposed an adaptive artifact removal method for EMG signals under the conditions of functional electrical stimulation with variable parameters,combining the adaptive template method,variable modal decomposition and long and short-term memory network to effectively remove the stimulation current artifacts.After experimental validation,this method is more noise suppressive than traditional methods while retaining effective information.Second,a closed-loop regulation strategy of functional electrical stimulation based on Hill-Type muscle model is proposed.Through the established Hill-Type muscle model,the acquired bilateral EMG signals and joint angles are converted into muscle activity and muscle force magnitude.The BP-PID controller was used to regulate the current parameters of the functional electrical stimulator based on the difference of bilateral physiological indexes,so that the bilateral physiological indexes converge to achieve the closed-loop regulation of functional electrical stimulation.After experimental verification,the output results of the established Hill-type muscle model are consistent with the results of opensim simulation,and the tracking performance of the controller is tested.Third,a reinforcement learning-based optimization strategy for functional electrical stimulation parameters is proposed.By analyzing the muscle activity under different parameter combinations,the mapping relationship between controller output and stimulation parameters is established.And the reinforcement learning method was introduced to optimize the established stimulation parameter mapping to achieve efficient rehabilitation training while improving individual adaptability and delaying muscle fatigue and reducing patient discomfort,and finally,the effectiveness of the method was experimentally verified.