Objective Evaluation Of Driver Fatigut By Spontaneous Pupillary Fluctuation

Objective:Accurate and objective assessment of driver fatigue have an important significance on road safety. Currently research is focused on active driving fatigue detection. Passive detection of driver fatigue is almost empty and this is the source of power for this article.Methods:This article introduces a fatigue detection system based on the pupil video which uses the spontaneous papillary fluctuation to detect physiological fatigue. The detection can be the following steps:(1) A 3 minute-long video of the pupil was captured by infrared cameras. Then the video is transmitted into computer through frame grabber.(2) Calculate the diameter of pupil by image processing software.(3) Plot the papillary waveform versus time and make spectral analysis.Results:(1) Significant differences can be found in the pupil-diameter waveforms between alert and fatigue people. In a drowsy situation, pupil diameter fluctuates with large amplitude; while in an alert situation pupil size remains stable for a long time and oscillates with amplitudes rarely exceeding±7% (about 0.5mm).(2) To further evaluate the differences between the two sets of data, this study proposed a parameter PDV to illustrate the variation of pupil diameter. The results show significant differences between two groups. The mean of PDV in sleepy group is five times larger than the one in alert group. Results also demonstrated that the difference between the groups were highly significant by using Mann-Whitney U-test (U=26, P<0.05).(3) The PERCLOS of the alert group is 10 % and drowsiness group is 40%, which is consistent with previous findings.(4) In power spectrum, the frequency of the pupil waveform mainly concentrated in the 0-0.050Hz. In this band, the energy in fatigue group and alert group is: 0.9283±0.0247;0.0257±0.0054. There is also statistically significant difference between the groups by using Mann-Whitney U-test (U=28, P<0.05).