Spectral characteristics assessment in recognition of drivers' drowsiness using statistical tests

One of the main causes of traffic accidents is drowsiness while driving. Since brain signal (EEG) can report the brain state and its activity momentarily and simultaneously, many researches have been focused on EEG signal for detecting the driver's alertness state. In this study, multichannel EEG signal was recorded from 10 volunteers when each person played a driving game in a virtual environment passing barriers in the game. These subjects should stay awake for at least 20 hours before the test. Process of recording signal for each subject during the driving game lasted about 45 minutes. After the preprocessing and manual labeling of gathered EEG data and extracting spectral features, using paired sample t-test we showed which of these frequency characteristics can create a significant difference up to 95% between alertness and drowsiness signals, and in the future studies can be used as indexes for drowsiness in order to increase the drivers' safety.