Research On Acquisition And Reconstruction Of Semg Based On Compressed Sensing

Surface Electromyogram(sEMG)record the electrical signals of muscles.It is an important method for non-invasive detection of muscle activity on the body surface.It is often used in the fields of intelligent prosthesis control,rehabilitation training and evaluation,and human-computer interaction.Nowadays,sEMG acquisition systems are limited by processing time,storage capacity and high power consumption.The research of sEMG acquisition and reconstruction based on compressed sensing is of great significance to reduce the power consumption of sEMG acquisition system,solve the bandwidth limitation of data transmission,save data storage memory,and improve the accuracy and practicability of sEMG detection.Therefore,this paper studies the acquisition and reconstruction of sEMG based on compressed sensing,aiming at the problem of large power consumption in the transmission process and large information quantity in the storage of sEMG.The main contents of the paper are as follows:1.Design of sEMG reconstruction algorithm based on compressed sensingThe sparsity of sEMG in each orthogonal basis is analyzed and the digital model of measurement matrix is designed.The reconstruction of single channel sEMG is studied,and the influence of sparsity on the success rate of reconstruction is analyzed.Meanwhile,the joint reconstruction of multi-channel sEMG and multiple compressed data is studied to further reduce the number of observations and improve the reconstruction accuracy.2.Reconstruction method of active segment of sEMG based on modified KSVDIn the compressed sensing measurement domain,the active segment of the sEMG is detected directly.According to the characteristics of the compressed sensing observation sequence,the active segment is determined by the sample entropy algorithm and then the active segment of the sEMG is reconstructed.Design and improve the KSVD adaptive learning dictionary,the trained dictionary can reflect the overall characteristics of sEMG,so as to further improve the reconstruction speed and accuracy of multi-channel sEMG.3.Design of semi-physical simulation system for sEMG compression and reconstructionThe sEMG sensor rand data acquisition card,Simulink is used to build a semi-physical simulation system for real-time compression sampling and reconstruction of sEMG.The effect of data compression on gait recognition rate is analyzed.The high performance GK-SVM of particle swarm optimization is used to identify the three modes of ground walking,up and down stairs of the original and reconstruct sEMG respectively.