A comparison of sEMG and MMG signal classification for automated muscle fatigue detection

This study compares the classification performance of both sEMG and MMG signal from fatiguing dynamic contraction of the biceps brachii. Commonly used statistical features are compared with a recently developed evolved pseudo-wavelet. Based on the literature, wavelet-based methods are a promising feature extraction technique for both types of signals (sEMG and MMG) during dynamic contractions. MMG results show that the evolved pseudo-wavelet improved the classification rate of muscle fatigue by 4.70 percentage points to 27.94 percentage points when compared to other standard wavelet functions, giving an average correct classification of 80.63%, with statistical significance (p < 0.05). For sEMG signals the evolved pseudo-wavelet improved the classification rate of muscle fatigue by 4.45 percentage points to 14.96 percentage points when compared to other standard wavelet functions (p < 0.05), giving an average correct classification of 87.90%. The comparison demonstrates that for both the sEMG and the MMG signal, the feature giving best classification results was the evolved pseudo-wavelet.