Study Of The Method Of Extracting Fatigue Information From EMG Signal

In recent ten years, along with the rapidly developing of nonlinear theory, the study and application of nonlinear time series analysis has been spread broadly into many study fields. Inspired by this, the author try to apply some new method of signal processing and analysis in the extracting fatigue information from EMG signals. These methods not only fit in the analysis of nonlinear signals but also fit in the analysis of linear signals. The main contents of this thesis are as follows:l)Test the nonlinearality in the surface EMG signal. 2)Study the way of application of the method of symbolic series analysis in the analysis of time series; Study the characteristic of symbol strain histogram of different kind of signals (deterministic, stochastic, linear, nonlinear, chaotic etc.). Then the correlation between the shannon entropy of the symbol strain histogram and the correspond muscular fatigue level was discovered, including the location of spike. Like the median frequency, the variation of shannon entropy of the symbol strain histogram can be the index of muscle fatigue level.3)Converte the one dimension time series into two dimension spot figure, then study the distribution of the spot around the 360 degree directions, compute the quadrant information entropy. The strong correlation between quadrant information entropy and the robust of muscle was discovered. But the . quadrant information entropy is not fit for being index of muscle fatigue. The distances numbered in 20% and 80% among the total spots to the (0,0) spot correlate with .the muscular fatigue level.4)Based on the results in this thesis, a method called short time isotonic contraction was suggested to detecting the fatigue level of muscle.