Study of Surface Electromyography Signal Based on Wavelet Transform and Radial Basis Function Neural Network

This paper makes use of the method of wavelet transform to analysis and extract wavelet coefficient as the eigenvalue from original two-channel surface electromyography signal (SEMG) in the Physics Labs. It is input to a radial base function (RBF) neural network as training sample to train. This network is used to pattern Qubi or Shenbi classification for the surface EMG of forearm. The experiment shows that RBF neural network classification´s accuracy rate is higher than BP neural network on the base function of the RBF. It can effectively identify the muscle-motion mode, and has more robustness and adaptability. So it proved to be a potential way in the field of muscle pattern recognition.