Brain Computer Interface (BCI) offers a solution to independent mobility for people with moving difficulties. This paper proposes a BCI to smart control of a wheelchair. The paper describes the experience of developing a complete BCI system consisting of hardware and software parts to instruct a wheelchair by human intention to move to different directions, left, right, backward, and forward using noninvasive EEG brain waves. Three different machine-learning algorithms are used to classify human intention to move to the desired direction, K-nearest neighbour algorithm, support vector machine algorithm, and artificial neural network. The highest accuracy achieved is 79.2% for support vector machine algorithm. The results prove that the system can be used for medical purposes successfully and the concept can be extended for other applications. © 2018 Institution of Engineering and Technology. All rights reserved.

View