Study On Lower Limb Rehabilitation Training System Based On SEMG And Virtual Reality And Fatigue Detection

With the gradual increase in the number of stroke patients,rehabilitation training robots are gradually coming into people’s view.For stroke patients,the enthusiasm and safety of lower limb rehabilitation training are very important indicators.However,most of the current lower limb rehabilitation training robots can only carry out boring and boring training,and lack the function of detecting the fatigue state of stroke patients.To solve this problem,this paper will design a lower limb rehabilitation training system based on EMG signals and virtual reality,which includes three parts: passive rehabilitation training,active rehabilitation training game and fatigue detection,to assist stroke patients to carry out safe and effective rehabilitation training.Firstly,the passive rehabilitation training method and the virtual reality game of active rehabilitation training were designed,and the fatigue evaluation standard was set according to the perceived exertion scale(RPE).Then,sEMG signals of human lower limb muscles were collected,and the sEMG signals were preprocessed with Butterworth bandpass filter.Then,the sliding window method was used to extract the characteristics of sEMG signals in time and frequency domain under different training actions and different fatigue states.The CNN-LSTM model will be established,and the parameters of the CNN-LSTM model are continuously improved through several comparative tests.Finally,the recognition rate of CNN-LSTM model for EMG signals generated by different actions is up to 97%,and the recognition rate of EMG signals under different fatigue states is up to 95%.In the experiment part,the lower limb rehabilitation robot system experiment platform was built,including the upper computer,STM32 main control chip,electromyographic signal acquisition module,wireless communication module and rehabilitation robot.Through the experiment,it is verified that the control method of using EMG signal to control the lower limb rehabilitation robot has good accuracy,and the accuracy of judging the fatigue state of patients by EMG signal.The rehabilitation training system designed in this paper focuses on improving the human-computer interaction ability of the robot,enhancing the rehabilitation initiative and enthusiasm of patients,and ensuring the safety of training by judging the fatigue state of users during training,which provides a theoretical basis for the subsequent development of lower limb rehabilitation training.