sEMG Based Classification of Hand Movements Using Artificial Neural Network
A human machine interface (HMI) is an interface that permits the interaction between a human being and a machine. Surface Electromyogram (sEMG) is an electro-physiological signal that provides inherent information about the activities of the human skeleton muscle. This paper propose sEMG pattern recognition system to control the myoelectric hand system using sEMG for HMI with the help of neural networks. Six parametric feature extraction algorithms are used to extract the features from sEMG signals such as AR (Autoregressive) Burg, AR Yule Walker, AR Covariance, AR Modified Covariance, Levinson Durbin Recursion and Linear Prediction Coefficient. The sEMG signals are modeled using Radial Basis Function Neural Network (RBFNN), Probabilistic Neural Network (PNN)and General Regression Neural Network (GRNN).The performance of the HMI system has obtained average mean classification accuracy of92.02% for AR Burg features using RBFNN. From the results, it is observed that AR Burg using Radial basis neural network outperformed the other networks in classifying twelve different finger movements.