Towards an automatic drowsiness detection system by evaluating the alpha band of EEG signals

Abstract

In this paper, we present the important role of alpha band for the evaluation of the drowsiness degree. By filtering the alpha band and by using the power spectral density of that same band, our data were analyzed using the percentiles as measures of dispersion. A threshold discriminating the two states was found, which helped to highlight the area of the brain responsible for the state of drowsiness for driver. So, in this work, we look to develop a drowsiness monitoring system in the goal to participate in reducing of the big number of road accidents to estimate the drowsiness level by analysis of EEG (electroencephalography) signals records. Finally, the algorithm developed in this work has been tested on twelve samples from the Physionet sleep-EDF database.