Semg-based posture recognition of elbow flexion and extension

Surface electromyographic signal (sEMG) is used in some fields such as human machine interaction and measurement of human motor function, because it can reflect the activation of human muscle. Though the recognition of motion pattern of human limbs has been researched for many years, continuous recognition for human elbow motion without load is still difficult because of low signal noise ratio (SNR). In this paper, we proposed an improved weighted peaks method to process the filtered sEMG signals from the biceps muscle and adapted linear fitting method to obtain the elbow motion in sagittal plane. The experiments showed the proposed method can effectively process the sEMG signals and obtain the activation of biceps muscle. The experimental results show the similar data of elbow motion compared to the data derived from an inertia sensor.