Muscle fatigue detection method based on feature extraction and deep learning

This paper investigates muscle fatigue recognition method by using the deep learning technique. We designed a F_GRU model to classify the surface electromyography (sEMG) collected on the back muscle group of the subjects. A set of nine time-domain and frequency-domain features were first extracted from each sEMG segment along with the sliding time window, and then recombine feature time sequences, which can characterize the original sEMG in both time domain and frequency domain. Next, the trained F_GRU model was used to classify all the sEMG samples to recognize fatigue state. The accuracy of muscle fatigue detection is higher than 98%, The F_GRU model has the advantages of less time-consuming and less memory consumption than the other machine learning methods.