Fatigue Analysis Of Lower Extremity Muscle Based On Surface Electromyography And Gait Signal

Muscle fatigue is defined as a reduction in the maximum force produced by a muscle contraction,usually caused by strenuous exercise or prolonged loading.Muscle fatigue affects the synergistic muscle activation patterns and alters the body’s motor function.Selecting appropriate physiological signals to study muscle fatigue can provide patients or athletes with a reasonable exercise plan and reduce the risk of injury,which is of great significance in the fields of medical rehabilitation and sports competition.Surface EMG signals have the unique advantages of being non-invasive,and gait signals can reflect human motor function well,so they are widely used for muscle fatigability analysis.However,there are some problems with the current research: 1)Muscle fatigue experiments mostly involve subjects doing active work to achieve muscle fatigue,while studies of fatigue due to passive exercise under fixed load conditions are relatively lacking.2)Most of the fatigue muscle synergy studies analyze the difference of synergistic patterns at the analysis level,and lack of reasonable explanation from the aspect of neural control strategy.3)Muscle fatigue analysis based on surface EMG unimodal signals is difficult to comprehensively interpret the motor function changes caused by fatigue,and the accuracy of the fatigue state detection system thus constructed needs to be improved.To address the above issues,this study first designed an experimental protocol for subjects to reach fatigue by passive exercise on a treadmill,recruited 18 subjects to participate,and simultaneously collected surface EMG and gait signals from five muscles of the lower extremities,and specifically carried out the following two aspects of research work:1.In response to the current situation that the neural control strategy of fatigue muscle synergy is not yet clear,this study extracted muscle synergy based on the surface EMG signal using non-negative matrix decomposition algorithm(Nonnegative Matrix Factorization,NMF)and determined the number of muscle synergy according to the parameter VAF(Variance Acounted For),and it was analyzed that: 1)Four muscle synergies were extracted before and after muscle fatigue,which did not change and had a high similarity in synergistic structure.2)There were significant differences in muscle activation weights and activation factors before and after fatigue,as evidenced by significant increases or decreases in muscle activation weights,increases in peak of muscle activation factors and shifts in peak occurrence time back and forth during the gait cycle.In this study,we suggest that the above changes are a strategy for the nervous system to compensate for muscle strength under fatigue conditions to maintain the original motor state and avoid injury.2.In order to be able to more comprehensively analyze the differences in motor function before and after muscle fatigue,this study proposed a joint analysis method that can use EMG and gait signals,and designs a muscle fatigue state classification model using machine learning methods as well as a gait prediction model using deep learning methods,and obtains the following conclusions: 1)In the state of muscle fatigue,the muscles need a greater degree of activation and longer activation time in order to maintain the original movement state? in addition,muscle fatigue reduces the stability of gait and increases the degree of left-right sway of the body,which leads to a non-fluent human movement.2)Based on the fused feature set of gait and EMG data,the classification accuracy of fatigue state can reach up to 98.93%,which is 4.87% more accurate than the classification accuracy using surface EMG unimodal signal features.3)Based on the correlation between surface EMG and gait timing signals,the lowest mean absolute deviation(Mean Absolute Pencentage Error,MAPE)of gait prediction index before and after muscle fatigue could reach 19.19% and 22.09%,respectively.Although the difficulty of predicting gait signals after muscle fatigue has increased,it can still do a good job of prediction.Finally,in order to make the above part of the study have practical application value,a human sports muscle fatigue analysis system was developed based on Matlab GUI tool,which realized four functions of data loading,data pre-processing,feature extraction and feature analysis,which can effectively analyze the changes of motor function caused by muscle fatigue and more targeted rehabilitation training for patients with muscle fatigue.