EMG-based Assessment of Shank Muscle Fatigue During Dynamic Exercise

Muscle fatigue is a common occurrence in daily life. For stroke rehabilitation patients, secondary injuries caused by falls due to muscle fatigue during rehabilitation training are very common. Therefore, it is of great significance to monitor muscle state in real time to prevent muscle fatigue during rehabilitation training. Currently, there are few methods for assessing muscle fatigue. And some methods have high limitations, some can only target the state of muscles in a static state. This paper proposes a method that can monitor and evaluate muscle fatigue in real time during dynamic exercise. First, a Hammerstein model is established through EMG and joint motion angle data. According to the identified model parameters, combined with the characteristics of the parameters, an augmented matrix is constructed. Then we use hierarchical clustering to cluster fatigue levels. The classification of fatigue grades obtained by this method is proved by side experiments to conform to objective laws. The method is convenient, efficient, and cost-effective, and the required parameters can be obtained without redundant sensors. The obtained fatigue level classification is proved to conform to the obj ective law through side experiments, which has guiding significance for the estimation of human muscle state.