Lane Departure Warning Model And Evaluation Algorithm

For the characteristics of driving behavior during lane departure, this paper set up a lane departure warning model based on lasting time and frequency. According to the thought of this method, a integrated evaluating model of the driving behavior in the lane departure was presented to evaluate the security status of the driver.In order to reduce the probability of false warning, the paper analyzed the subjective and objective factors of the lane departure when driving the vehicle. Fully considering the impact to lane departure such as the lane line, vehicle characteristics and driving behavior, a "people-oriented" lane departure warning algorithm was established. Road is divided into different levels of security areas in this model. Through real-time collection of traffic and road information, the coordinates of the vehicle, road alignment data, speed, the status of the turning lights and other information, the current vehicle status and location of the road which the vehicle located is determined, the duration and frequency of the vehicles driving into the various regions is accumulated, which is then compared with the set duration and frequency threshold, so as to be made as the basis of a trigger based automotive lane departure warning.At the same time, the concept of risk degree was introduced based on the father research of lane departure warning model. Then we established the risk degree model of the lasting time and frequency as to evaluate the security status of driver. A comprehensive evaluation algorithm of the driving behavior was set up during the lane departure. The results of the evaluation algorithm can be implemented on the driver's expedition, providing a viable means for the driver's training, education and management.Finally, the DSR-1000TS split three-screen driving simulator was conducted to simulate an actual road driving condition. Using the method of the comparative test, the better lane keeping and serious lane departure experiment was done as a comparison group. By output, analysis and processing of the data, comparative analysis and evaluation of the situation on the driver's lane departure is done to verify the algorithm is reasonable and practical to some degree.