Vigilance Monitoring System Using Brain EEG Signal Processing

For many human machine interaction systems, the operators need to retain high and constant level of vigilance to prevent accidents. Comparing with other techniques that used for vigilance monitoring such as face recognition; a new technique was emerged that based on using the electroencephalogram (EEG) signals from the brain to reflect the vigilance level much sooner and more accurately. However, many difficulties exist in this field such as; how to label the EEG data, how to remove the noise from the EEG data, what are the most effective features for this type of signal and then what is the optimum classification technique. This paper introduced a new experiment for vigilance monitoring using Brain EEG signal processing. EEG data are analyzed using Fast Fourier Transform to extract features corresponding to two distinct vigilance levels: awake and falling asleep. Unsupervised learning method using multi layer Neural Network trained by a standard back propagation algorithm is used to classify the two classes of vigilance levels. Our preliminary results for estimating different vigilance levels with EEG signals are quite promising. We reached to 96.4% classification rate for the two considered vigilance levels, the result that surpassed the results from other researches in the same application. This research can give a direction for the vigilance labeling and features selection for the real time vigilance monitoring system in future. Key-Words: EEG signal processing, Vigilance monitoring system, Safety systems.