Posture recognition of elbow flexion and extension using sEMG signal based on multi-scale entropy

Recognition for human elbow motion with surface electromyographic signal (sEMG) research receives more and more attention, especially in some fields like human machine interaction and measurement of human motor function due to the reason the EMG can reflect the activation of human muscle. However, continuous recognition for human elbow motion without load is still difficult due to the low signal noise ratio (SNR). In this paper, we utilized the multi-scale entropy and moving-window method to reveal the elbow motion information hidden in the filtered sEMG signals from the biceps muscle with good performance compared to the angel record derived from an inertia sensor.