Non-conventional Control and implementation of an electric wheelchair designed to climb up stairs, controlled via electromyography and supported by artificial neural network processing

Abstract

This work is part of a project which main objective is to design and implement an electric wheelchair capable of climbing up stairs. The wheelchair is operated thru silent speech non-conventional control presented on this paper using electromyographic (EMG) signals which are fed to a multiple artificial neural network (ANN) system for pattern recognition. Mechanical design of the wheelchair, powering and sensing aspects are presented on additional papers. Electromyographic signals are captured from the patient's anterior triangle of the neck muscle area using Surface Electromyography (SEMG) with electrodes in bipolar configuration. Silent speech commands from patient's gestural movements of three, four and five phonetically different words are used to interact with a Graphic User Interface (GUI) for wheelchair navigation. The system was tested on five patients, achieving an overall recognition accuracy of 95% with 0.6s maximum response time. © 2013 Springer-Verlag Berlin Heidelberg.