Feature Analysis Of SEMG And Research On Muscle Fatigue Classification

Muscle fatigue is a common life phenomenon,which usually occurs after strenuous exercise or continuous exercise,and manifests as muscle aches and fatigue.When muscle fatigue reaches a certain level,it will cause human body damage.In order to avoid this situation,you can choose a more reasonable physiological signal to judge the degree of human body fatigue.The EMG signal has attracted much attention in evaluating muscle fatigue because it is easy to observe and has high real-time performance.The electromyography signal is a kind of bioelectric signal that is superimposed on time and space by the potential generated by the muscle motor unit and transmitted from the muscle neuron to the surface of the skin.The EMG signal can be collected non-destructively,long-term,and real-time,and the change of the signal is related to the activity level and functional state of the muscle.The EMG signal acquisition methods include needle EMG collection and surface EMG collection.Surface EMG signals are widely used because of their non-invasive and convenient detection advantages.In order to establish a muscle fatigue classification model and improve the correct recognition rate of the model,firstly,based on the commonly used time-domain and frequency-domain feature analysis methods,the time-frequency domain,nonlinearity and parametric model feature analysis methods are introduced;secondly,3 channels are designed.Surface EMG signal acquisition system,each channel can extract 16-dimensional feature parameters,and a total of 48-dimensional feature parameters are extracted;then a fatigue induction experiment is designed to extract the time domain,frequency domain,time-frequency domain,nonlinearity and parameter model features and Analyze the law of change,adopt three feature dimensionality reduction methods of mutual information measurement,principal component analysis and kernel principal component analysis to reduce the dimensionality of the feature set to reduce the redundancy between features;finally,the new features after dimensionality reduction Set and Fisher linear discriminant analysis,K-nearest neighbor and support vector machine three classifiers are combined with each other to establish nine fatigue classification models to classify muscles into three states: relaxed state,fatigue transition state,and fatigue state.The results show that the core The model combining component analysis and support vector machine has the highest average recognition rate for fatigue classification,reaching 91.5%,which is higher than other fatigue classification modelsIn summary,the surface EMG signal feature analysis and muscle fatigue classification methods proposed in this thesis can achieve better fatigue classification results,and have important research significance for the judgment of muscle fatigue.