Driving fatigue affects a driver’s ability to control a vehicle and react to events. The goal of this research was to develop an online driving fatigue monitoring and early-warning system. Data reporting on electroencephalograph signal and driver reaction time were collected from twenty simulation experiments. Lane departure images were collected from five road experiments. A support vector machine was used to divide driver mental state into three levels. The Sobel edge operator was applied to search the edge of the lane line. The results show that the average driver reaction time significantly increases when there is serious driving fatigue. Moreover, driver ability sharply declines, causing the vehicle to sway from side to side at a low frequency. In consideration of these two factors, this study involved installing and testing an easily-mounted driving fatigue monitoring and early-warning system. Road tests demonstrated that the system could realize all the required functions, including being able to detect fatigue quickly and run steadily.

CICTP