A Comparison of EMG and EEG Signals for Prostheses Control using Decision Tree

In spite of availability of various approaches, the control of prosthetic limb would be more effective if it is based on Electromyogram (EMG) signals from remnant muscles and Electroencephalogram (EEG). The analysis of these signals depends on various factors such as amplitude, time and frequency domain properties. EEG signals are obtained from the experiments conducted in Biomedical laboratory using 27 different subjects with four different hand movements viz., finger open (fopen), finger close (fclose), clock wise wrist rotation (cw) and counter clock wise wrist rotation (ccw). The EMG dataset for the same conditions were obtained from the NINAPRO DATABASE, a resource for bio robotics community of hand movements. The statistical features were extracted for both EMG and EEG signals and classified using decision tree (C4.5) algorithm. The comparison of classification accuracies for both EMG and EEG signals is presented.