Study Of Muscle Fatigue Identification And Prediction Based On EMG And Fabric Sensors

With the rapid development of modern medicine and modern technology,the study of human muscle fatigue has been intensified and muscle fatigue detection has gradually become a hot research topic in the fields of rehabilitation medicine and sports science.However,even the most mature EMG sensors,which have been developed as portable and wireless devices,are not suitable for use in strenuous exercise due to the ease of contamination of action potentials.In this study,we use a newly developed flexible fabric strain sensor to monitor changes in arm muscle circumference to study muscle fatigue in skeletal muscles.Compared to traditional sensors such as EMG,fabric sensors are more compatible with human skin,more resistant to interference and more flexible,and have great promise in the field of wearable device research.This study combined a portable surface EMG sensor and a wearable smart arm band made from a fabric strain sensor to design a signal acquisition platform capable of detecting surface EMG signals and muscle thickness changes in the biceps muscle in real time.A dynamic muscle fatigue experimental protocol using dumbbells for bending exercises was designed and experimental data were collected from 32 subjects,forming a database containing 580 samples of complete bending exercise rounds.After filtering and noise reduction of the collected sensor signals,the fatigue-related characteristics of muscle thickness were proposed based on muscle physiology and combined with the characteristics of the surface EMG signal to demonstrate the potential of this fabric strain sensor as an auxiliary and complementary method for the field of skeletal muscle fatigue research.And based on this,the fatigue cycle was successfully identified using a simple k-mean clustering technique.For potential applications,fatigue prediction models based on supervised learning methods were further proposed and compared,where the SVMbased fatigue prediction model had an overall accuracy of 83.3%,effective recall of 90%,F1-SCORE of 95% and AUC of 98.7%,showing that muscle thickness measured by a flexible sensor helped to predict muscle fatigue successfully.This work exploits deeper potentialities of the fabric strain sensors in muscle fatigue monitoring,the result and methodologies will inspire wider horizon of human-centered applications using novel flexible sensors.