Evaluation Of Lower Extremity Rehabilitation Based On SEMG

Due to the aging of population and other social problems,the number of disabled and elderly people in China has been high.This phenomenon leads to a large number of lower extremity patients and a large demand for lower extremity rehabilitation,so it is necessary and practical to carry out the evaluation of lower extremity rehabilitation.The acquisition system selected in this paper is composed of E.M.G.System TB0810 EMG instrument and JANALYSIS E100 C software.According to the characteristics of sEMG and the research of pattern recognition,this paper designs the evaluation scheme for the rehabilitation status of lower limbs.In order to realize the comparative study of the two groups,we collected,denoised,extracted features and identified patterns of the surface EMG signals of the lower extremities.According to the comparison of the characteristic values of integral electromyography between the two groups,the recovery of four muscles(gastrocnemius muscle,peroneal longus muscle,soleus muscle and tibial anterior muscle)in the lower limbs of the patient group was obtained,and the rehabilitation training effect of the four muscles was reported to the doctor in time.In the denoising part of the paper,there are three parts: moving end detection,eighth order bandpass Vos filter and wavelet filter.The moving end detects the beginning and end of the lower limb movement.The eighth order bandpass Vos filter filters the invalid signal according to the surface muscle electrical signal characteristics,and the wavelet filter filters the burr of the waveform.The three parts realize the premise of keeping the effective information Next,noise reduction of the signal is carried out.In feature extraction,the size ratio of time-domain eigenvalue iEMG is used to find that the tibialis anterior muscle plays the most important role in the process of walking,the muscle strength is poor,and the walking stability is good.The time-frequency domain eigenvalues are selected as the final parameters of feature extraction for the next step of pattern recognition.In pattern recognition,radial basis function of support vector machine is selected and combined with one-to-one classification method to build a classifier model.Particle swarm optimization(PSO)is selected to optimize the kernel parameters,which makes the classifier more effective.Through the final designed classifier to recognize the four lower limb movements(left,right,up,down)of patients with lower limbs,from the recognition results,it can be concluded that the rehabilitation status of men’s lower limbs is far better than that of women,and the rehabilitation training program before evaluation is more suitable for men.In the rehabilitation training,we should pay attention to the muscle strength training of gastrocnemius and peroneal longus,so as to achieve better results in the rehabilitation of lower limbs.