中文摘要：

为进一步完善汽车主动防碰撞预警系统,建立了一种基于模糊推理的驾驶员反应时间修正方法。该方法以驾驶员的实时反应时间为研究对象,通过实时采集驾驶员的加油频率、深加油比例、制动频率、深制动比例等数据,利用所建立的模糊推理规则判断出一个适应于当前驾驶员动态驾驶倾向的实时反应时间,然后对Berkeley算法进行修正,计算出车辆行驶所需要的安全车距,并在驾驶模拟器上进行了对比分析。经模拟实验验证,本方法所确定的15名驾驶员的理论平均反应时间为1.0000 s,方差为0.0417,而实际平均反应时间为1.0077 s,方差为0.0469,并且根据理论反应时间所确定的预警距离与根据实际反应时间所确定的预警距离比较接近。可见本方法能在一定程度上提高安全预警算法的准确性,能够充分考虑驾驶员的实时驾驶特性,降低预警系统的虚警率。

英文摘要：

In order to further improve the automobile active collision warning system, a driver reac-tion time correction method based on fuzzy inference was established. This method takes the drivers’ real-time response time as the research object. By real-time collecting the drivers’ accelerating fre-quency, deep accelerated ratio, braking frequency and deep braking ratio, the established fuzzy in-ference rules were used to determine a real-time response time adapted to the drivers’ current dynam-ic driving tendency, and then to modify the Berkeley algorithm to calculate the vehicle’s safety dis-tance. Finally, comparative analysis was carried on the driving simulator. It has been verified by simulation experiments that the average responding time of 15 drivers was 1. 000 0 s, and the vari-ance was 0. 041 7, while the average actual responding time was 1. 007 7 s, and the variance was 0. 046 9 . The warning distance predicted by the theory reaction time and actual reaction time were very close. Obviously, to some extent, this method could improve the accuracy of the warning algo-rithm, could give full consideration to the driver’s real-time driving characteristics, and could reduce the false alarm rate of the warning system.