EEG Controlled Semi-Autonomous Mobile Vehicle Design and Implementation

Abstract

The main goal of this paper is to control a mobile autonomous vehicle by studying EEG signals and their properties. EEG signals are analyzed to classify an example dataset provided by Internet resources. Different feature extraction methods were applied. Since EEG signals have high noise and information is embedded in the noise spectrum, frequency analysis was used to eliminate noise. Wavelet Transform was discussed in detail. Feature vectors were selected by Fisher Discriminate Analysis, that is an improved version of Linear Discriminant Analysis. With these data, mobile vehicle was controlled by differential driving technique. An attempt was made to avoid obstacles and to move with commands extracted from EEG signals.