Human Movements Classification Using Multi-channel Surface EMG Signals and Deep Learning Technique

Abstract

Electromyography (EMG) signals can be used for human movements classification. Nonetheless, due to their nonlinear and time-varying properties, it is difficult to classify the EMG signals and it is critical to use appropriate algorithms for EMG feature extraction and pattern classification. In literature various machine learning (ML) methods have been applied to the EMG signal classification problem in question. In this paper, we extracted four time-domain features of the EMG signals and use a generative graphical model, Deep Belief Network (DBN), to classify the EMG signals. A DBN is a fast, greedy deep learning algorithm that can rapidly find a set of optimal weights of a deep network with many hidden layers. To evaluate the DBN model, we acquired EMG signals, extracted their time-domain features, and then utilized the DBN model to classify human movements. The real data analysis results are presented to show the effectiveness of the proposed deep learning technique for both binary and 4-class recognition of human movements using the measured 8-channel EMG signals. The proposed DBN model may find applications in design of EMG-based user interfaces.