Detection and Recognition of Motor Unit Action Potentials in the Multi-Channel Surface EMG Signals

The method about the detection and recognition of MUAPs was explored by using the multi-channel sEMG signals at the low contraction force. A combined detection method of continuous wavelet transform and hypothesis testing was applied for each channel sEMG signal. The comprehensive judgemental rule was implemented and utilized to analyze the motor unit firing. Then, the single channel signal which contained the candidate MUAP waveforms was derived from the average processing of multi-channel sEMG signals. With the averaged signal, the clustering algorithm was applied, and the unclassified waveforms were identified. The experimental results showed that some illusive shapes were rejected and the satisfying recognition performance could be acquired by using the averaged sEMG signals.