Human Muscle Fatigue Research And Detection System Design Based On Multi-modal Data Fusion

The detection of human muscle fatigue has great research and application value in the fields of sports competition and rehabilitation training to prevent damage caused by muscle overtraining.Inspired by this,this thesis will study three physiological signals?surface electromyography signal,mechanonmyography signal,and muscle oxygen saturation?that can characterize muscle fatigue in a method of data fusion to achieve more accurate detection for the muscle fatigue.The research content and results of this thesis are as follows:?1?This thesis build a multi-source physiological signal acquisition system.It mainly includes the selection of mechanonmyography signal sensor,muscle oxygen saturation sensor,A/D converter,communication module and main control chip,the design and performance test of surface electromyography sensor,lower computer program and upper computer program design.Design a muscle fatigue experiment and collect multi-source physiological signals of biceps brachii muscle.At the same time,the level of fatigue is independently calibrated according to the fatigue level evaluation table.?2?Extracte and analyze the characteristics of surface electromyography signal,mechanonmyography signal,and muscle oxygen saturation.Conduct regression and significance analysis on surface EMG signal IEMG and RMS parameters in time domain,MPF and MF parameters in frequency domain and C₀ complexity and Ap En feature quantity of nonlinear dynamic angle,and finally select the optimal feature parameters from them.Linear regression analysis and significance analysis are performed on the MPF and MF parameters of the mechanonmyography signal in the frequency domain,as well as the LZC,FD,MLE,and Ap En feature quantities of the non-linear dynamics angle,and the optimal parameters are selected.Muscle oxygen saturation index is used as a parameter to characterize muscle state,and muscle fatigue correlation analysis is performed on this index.?3?This thesis Builds a multi-modal data fusion muscle fatigue rating evaluation model.This thesis will band together optimal feature parameters extracted from surface electromyographic signal,muscle sound signal and muscle oxygen signal,and will establish a muscle fatigue level evaluation model based on SVM?Support Vector Machines?.This thesis designs a comparative experiment of the equivalent evaluation of muscle fatigue with the fusion of single-modal data and multi-modal data.The experimental results show that under the same conditions,the recognition rate of multi-modal data fusion for muscle fatigue is 89.2%.While when muscle and tone signals are used as the basis for judging muscle fatigue,its accuracy rates are 84.2%and 80.8%respectively.The experimental results also show that the recognition rate of multi-modal data fusion has been improved by 5%and 8.4%respectively compared with the recognition rates of two single-mode.Furthermore,the accuracy and detection performance of evaluating muscle fatigue has been improved and stabilized.