Hand Motion Recognition And Intelligent Prosthetic Hand Control Based On Surface Electromyography

Due to diseases,accidents and other reasons,the number of patients with physical disabilities is growing,but intelligent prosthetic hand devices have not yet fully met the needs of patients.With the advancement of rehabilitation medicine engineering technology,intelligent prosthetic hand control systems based on surface electromyography signal(sEMG)have received widespread attention.This system could help patients with forearm disabilities restore their partial limb functions and improved their quality of life of patients.Therefore,the research on intelligent prosthetic hand systems had large significance.This thesis constructed an online gesture recognition and intelligent prosthetic hand control system based on sEMG.The main research contents are as follows:(1)Collection and preprocessing of sEMG.In this paper,Myo armband was used to collect sEMG,which was easy to operate and had high safety.In order to enhance the characterization ability of sEMG in hand gesture recognition,a signal recombination and fusion algorithm(RCF)was proposed.The algorithm virtually amplified the signal channel by reconstructing the sample structure of the signal,which could provide richer input information for gesture recognition.(2)Feature extraction and feature optimization of sEMG.In the feature extraction stage,this paper constructed a feature set constructed from 8 time domain and 2 frequency domain features.Also,the principal component analysis method was used to reduce the data dimensionality for the requirement of high-dimensional feature data at this stage,aiming to reduce the amount of data analysis calculations.(3)Hand movement recognition based on sEMG.In this paper,three machine learning algorithms: K-nearest neighbor,linear discriminant and support vector machine were used to perform hand movement classification experiments,and the comparison of the results proved the effectiveness of the RCF algorithm.In order to improve the accuracy of hand movement recognition,the temporal convolutional network(TCN)was improved based on the preprocessing of the RCF data,and finally a classification accuracy of up to 93.69% for ten gestures was reached.(4)Real-time intelligent prosthetic hand control system based on sEMG.The intelligent prosthetic hand in the system used 3D printing technology,which had light weight and is easy to carry.The system could carry on online recognition of ten gestures based on the improved TCN,and used the recognition results to control the knuckle angle changes of the prosthetic hand in real time.Finally,the control of prosthetic hand based on sEMG was achieved.