Control brain machine interface for a power wheelchair

Abstract

Controlling a power wheelchair using a brain machine interface (BMI) requires sufficient subject training. A neural network based BMI design using motor imagery of four states is used to control the navigation of a power wheelchair. The online experiment results are presented for two indoor navigation protocols. Twotrained subjects participated in the study. Performance of the real-time experiments is assessed based on the targets reached, time taken to reach the targets and on completion of a given navigation protocol. A performance rate of 85.7% is achievable by both subjects for the real-time experiments.