EMG motion pattern classification through design and optimization of Neural Network

Abstract

This paper illustrates the classification of EMG signals through design and optimization of Artificial Neural Network (ANN). Different types of ANN models are basically structured with many interconnected network elements which can develop pattern classification strategies based on a set of input/training data. The ANN models work in parallel thus providing higher computational performance than traditional classifiers which function sequentially. The EMG signals obtained for different kinds of hand motions, which further denoised and processed to extract the features. Extracted time and time-frequency based feature sets are used to train the neural network. A back-propagation neural network with Levenberg-Marquardt training algorithm has been utilized for the classification of EMG signals. The results show that the designed network is optimized for 10 hidden neurons with 7 input features and able to efficiently classify single channel EMG signals with an average success rate of 88.4%.