Previous studies have used the anaerobic threshold (AT) to non-invasively predict muscle fatigue. This study proposes a novel method for the automatic classification of muscle fatigue based on surface electromyography (sEMG). The sEMG data were acquired from 20 participants during an incremental test on a cycle ergometer using sEMG sensors placed on the vastus rectus femoris (RF), vastus lateralis (VL), vastus medialis (VM), and gastrocnemius (GA) muscles of the left leg. The ventilation volume (VE), oxygen uptake (VO₂), and carbon dioxide production (VCO₂) data of each participant were collected during the test. Then, we extracted the time-domain and frequency-domain features of the sEMG signal denoised by the improved wavelet packet threshold denoising algorithm. In this study, we propose a new muscle fatigue recognition model based on the long short-term memory (LSTM) network. The LSTM network was trained to classify muscle fatigue using sEMG signal features. The results showed that the improved wavelet packet threshold function has better performance in denoising sEMG signals than hard threshold and soft threshold functions. The classification performance of the muscle fatigue recognition model proposed in this paper is better than that of CNN (convolutional neural network), SVM (support vector machine), and the classification models proposed by other scholars. The best performance of the LSTM network was achieved with 70% training, 10% validation, and 20% testing rates. Generally, the proposed model can be used to monitor muscle fatigue.

This study proposed a muscle fatigue classification method based on surface electromyography (sEMG) signals to achieve accurate muscle fatigue detection and classification. A total of 20 healthy young participants (14 men and 6 women) were recruited for fatigue testing on a cycle ergometer, and sEMG signals and oxygen uptake were recorded during the test. First, the measured sEMG signals were denoised with an improved wavelet threshold method. Second, the V-slope method was used to identify the ventilation threshold (VT) to reflect the muscle fatigue state. The time- and frequency-domain features of the sEMG signals were extracted, including root mean square, integrated electromyography, median frequency, mean power frequency, and band spectral entropy. Third, the time- and frequency-domain features of the sEMG signals were labeled either “normal” or “fatigued” based on the VT. Finally, the statistical features of 16 participants were selected as the training data set of the Convolutional Neural Network–Support Vector Machine (CNN-SVM), Support Vector Machine, Convolutional Neural Network, and Particle Swarm Optimization–Support Vector Machine algorithms. In addition, the statistical features of the four remaining participants were used as the test data set to analyze the classification accuracy of the four aforementioned algorithms. Experimental results indicated that the denoising effect of the improved wavelet threshold algorithm proposed in this study was satisfactory. The CNN-SVM algorithm achieved accurate muscle fatigue classification and 80.33%–86.69% classification accuracy.