EMG signal based control of an intelligent wheelchair

This paper presents a novel artificial neural network approach to control an intelligent wheelchair using myoelectric signals. The work is divided into six stages out of which feature extraction and classification are the main stages for this research. The type of classification technique used is Multi-Layer Perceptron. The EMG data is collected by placing the electrodes on the forearm muscles. This data is segmented for every 200 milliseconds after which the feature extraction is performed using mean absolute value. The signals are fed to the artificial neural networks and processed to attain parameters that are related to the muscles temporal hand activities. The resulting commands are sent to drive the wheelchair according to the user's intention. The software was tested on the intelligent wheelchair in real-time, which confirm that the system is robust for different gender and environments.