A method for analyzing muscle fatigue characteristics based on sEMG signal multifractal

Abstract

Aiming at the problem that surface EMG signals (sEMG) are inaccurate in estimating muscle fatigue due to their non-stationary, nonlinear, self-similarity and other complex characteristics, a method for analyzing muscle fatigue characteristics based on sEMG signal multifractal downtrend moving average method (MFDMA) was proposed. Firstly, the MFDMA method was used to analyze the nonlinear dynamics of the collected sEMG signal, shuffle signal and Gaussian white noise signal; secondly, MFDMA method was used to calculate the multifractal spectrum width, Hurst exponent variation difference, probability measure value and peak singularity exponent of sEMG signal; finally, the significant difference in multifractal characteristics between muscle fatigue and non-fatigue state was analyzed by t-test. The results show that MFDMA method can describe the multifractal behavior of sEMG signal, and the multifractal characteristics such as spectral width have significant differences between muscle fatigue and non-fatigue state. The proposed method can reliably characterize exercise-induced muscle fatigue, and provide some feature reference for muscle fatigue recognition model and rehabilitation medicine research.

Keywords

• rehabilitation engineering; surface electromyography signal; multifractal; muscle fatigue; nonlinear characteristics