Characteristic extraction of fatigue driver's EEG signals based on wavelet entropy

摘要

This study aims to develop a method to detect driver's fatigue using the EEG signals. Experiments have been designed to test the subjects under simulated driving and actual driving, and the fatigue driver's Electroencephalogram (EEG) signals were collected. Wavelet transform method was applied to de-noise the raw EEG data. The H, R (H=α/β; R= (α+θ)/β) wavelet entropy were calculated. The results show that the fatigue driver's H, R wavelet entropy decreased after rest (P<0.05). It is concluded that there are significant difference in brain function between fatigue states and recovered after rest. It is shown that H, R wavelet entropy is an effective eigenvalue to measure driver's fatigue.