基于驾驶员紧急制动行为特征的危险估计算法

Threat Assessment Algorithm Based on Characteristics of Driver Emergency Braking Behavior

摘要:

基于真实交通工况下驾驶员的紧急制动行为特征，建立了一种新的危险估计算法用于汽车避撞系统的控制策略开发.通过在车辆上安装车辆行驶记录仪采集了目标区域内驾驶员的真实交通场景；对采集到的交通场景进行人工筛选，并按照美国高速公路安全管理局（NHTSA）的分类方法进行了分类，得到了6种典型的危险工况；通过视频图像处理等方法对典型危险工况下驾驶员的紧急制动行为进行了分析，得到了驾驶员紧急制动起始点的碰撞时间tTTC值以及车辆在紧急制动过程中的平均制动减速度，并利用这些数据建立了基于碰撞时间倒数tiTTC和期望减速度areq的危险估计算法.这种危险估计算法能够同时考虑两车快速靠近以及稳定跟车工况，并与驾驶员在真实交通环境中的紧急制动行为相对应.

Abstract:

Based on characteristics of driver emergency braking behavior in real traffic scenarios, a new threat assessment algorithm was established for control strategy development of automotive collision avoidance system. Real driving scenarios of drivers in the target area were collected by video drive record (VDR). After manual selection and classification according to the scenario typology defined by National Highway Traffic Safety Administration (NHTSA), six typical types of risk scenarios were obtained. Emergency braking behavior of drivers under these typical risk scenarios were analyzed by using video image processing method and then the tTTC(time to collision) values at the starting point of emergency braking as well as the average values of emergency braking decelerations were obtained. These data were then used to establish the threat assessment algorithm based on 1/tiTTC (tiTTC·tTTC=1) and the required deceleration areq. This threat assessment algorithm can take both the fast closing condition and stable following condition into consideration and correspond with driver emergency braking behavior in real traffic scenarios.