Myoelectric Pattern Discrimination using an HMM-based Neural Network

N. Bu, O. Fukuda, S. Noda, and T. Tsuji (Japan)

Keywords

Neural Networks, Pattern Discrimination, EMG, Recurrent Neural Network.

Abstract

A pattern discrimination method of electromyogram (EMG) signal for prosthetic control is presented in this paper. This method uses a novel recurrent neural network based on the hidden Markov model. This network includes recurrent connections to enable the capability of modeling time series such as the EMG signal. Weight coef´Čücients in the network can be learned using a well-known back propagation through time algorithm. The pattern discrimination experiments were conducted to demonstrate feasibility and performance of the proposed method. This method successfully discriminated forearm motions using the EMG signal, and achieved considerably high discrimi nation performance in comparison with the other discrimination methods.

Important Links:



Go Back