基于事故数据挖掘的乘用车追尾卡车 AEB测试规程研究

基于未来出行交通事故场景研究数据库中的乘用车追尾卡车事故，分析并提出了以卡车为目标物的乘用车自动紧急制动系统的典型测试工况。采用K-means聚类算法得出可代表实际卡车的目标物颜色，基于事故数据分析提取卡车目标物尾部特征参数，设计并制作了一种新型的具有与真实车辆反射特性和机器视觉识别特性接近的重型厢式卡车目标物。卡车目标物静止，测试车辆分别以45、50、55、60 km/h的速度进行100%重叠自动紧急制动系统测试，验证了该目标物的可行性和有效性。可为车辆主动安全相关标准法规研究提供数据支撑，推动车辆主动安全测试技术的发展。

Study on AEB Testing for Passenger Cars Rear-Ending Trucks Based on Accident Data Mining and Analysis

The high casualty rate among passengers in cars rear-ending trucks has aroused a great deal of attention. However, neither domestic nor international standards and regulations have published AEB tests specific to rear-end collisions between passenger cars and trucks. The paper proposes a new typical AEB test scenario for passenger cars rear-ending trucks, based on real traffic accident data in the future mobile traffic accident scenario study(FASS) database. The K-means clustering algorithm is used to identify the test target color that represents actual trucks, and the characteristic parameters of the truck''s rear are extracted based on accident data analysis. A new AEB test target is designed and produced, resembling a red, heavy and box-type truck, with reflective characteristics and machine vision features similar to those of real trucks. Finally,the feasibility and effectiveness of the test target were verified by real vehicle tests, where the truck target remained stationary while the test vehicle speeds were set at 45 , 50 , 55 , 60 km/h, with a 100% overlap. This study provides data support for the development of relevant standards and regulations of vehicle active safety and promotes the improvement of vehicle active safety testing technology.