Muscle fatigue is required to be assessed in real-time to maintain the best physical condition,especially for sports and rehabilitation areas. In recent years, numerous studies proposed muscle fatigueestimation methods with non-invasive surface electromyography (sEMG). However, the previous approacheswere limited to discerning whether muscle fatigue occurs and were unable to quantify the fatigue level dueto individual differences in muscle characteristics. In this study, we propose a novel method for quantitativemuscle fatigue estimation that is applicable for various people without individual calibration. Because musclemass is closely related to muscular endurance, it is utilized as a standard parameter in our assessment process.We introduce a new concept of muscle fatigue score (MFS), based on the cosine similarity between musclemass and representative fatigue indicators. The MFS exhibits a high correlation coefficient (|R|=0.7398)with key muscle characteristics compared to previous representative muscle fatigue indicators calculatedfrom sEMG: mean frequency (|R| =0.2848), median frequency (|R| =0.1972), and low-frequency ratio(|R|=0.0346).
© Copyright 2020 IEEE Access. Alle Rechte vorbehalten.