The Research Of Analysis Methods About Fatigue Based On Eeg And Eog

Driving fatigue is the researching objection of this paper. The driving fatigue phenomena is occurred when driver spends too much time on driving which is very easy to cause traffic accidents because the physical and mental functions of driver decays gradually and his reaction and controlling speed becomes more slowly. In recent years, the traffic accidents caused by driving fatigue is becoming more frequent. Therefore it is very necessary to research and prevent driving fatigue.By extracting EEG and EOG data of driver while he is under energetic state and fatigue state respectively, and analyzing and comparing these data so that we can conclude those data which are relates closely with the driving fatigue, which is the method we used to research the driving fatigue and can be describe as follow:Firstly, we design a experiment which is using the traffic icon to stimulate drivers enter a simulating driving environment, then we pick up the EEG and EOG data of 8 drivers who under energetic state and fatigue state respectively by using the lab equipment Neuroscan.Then, we analysis and process the physical data which we have picked up respectively. We extract electrode 3C located in the middle area as the EEG data which have several waveforms with different frequencies, such as: wave α(7 ~ 14Hz), wave β(14 ~ 30Hz), waveθ(0.5 ~ 4Hz) etc. These waveforms and fatigue are highly related. The brain is active when the driver is under energetic state and the main EEG frequency is wave β. Wave β decreases gradually and wave α and waveθ increase gradually while the driver is under fatigue state. The average rate ratios of(α +θ) / β 、θ/ β 、α/ β are acquired by processing waveforms of different frequency bands, and the differences between the effect of the average rate ratio of α/ β to the energetic state and the fatigue state is the most obvious one. As for the EOG data, we extract horizontal eye electro-oculogram(HEO) and vertical eye electro-oculogram(VEO) these two electrical signal data. The average rate ratio of low and high frequency of HEO is extracted. Based on the researches have been studied by others and our own experiment results, weconclude that the average rate ratio is very sensible to the fatigue state and un-fatigue state when it is(0 ~ 2.5Hz) /(2.5 ~ 30Hz). As for the VEO, we mainly extract the features of blinking time, blinking speed and blinking frequency, etc.Finally, normalizing the electrical characteristics of EEG and EOG, we acquire a new variable curve by principal component analysis. Making a linear correlation analysis between these features and the new variable curve so that we can eventually figure out the features which relate to the fatigue relatively high. In this paper, the average rate ratio of α/ β of EEG, the average rate ratio of(0 ~ 2.5Hz) /(2.5 ~ 30Hz) of HEO, and blinking speed, all of them can reflect fatigue state very well. Blinking time and blinking frequency also can reflect the fatigue state to some extent, but they are susceptible to the interference of emergency. So the further research is needed if we want to estimate the fatigue of driver by using blinking time and blinking frequency.