危险驾驶场景要素分析  

Analysis of Hazardous Driving Scenario Elements

作　　者：陈启蒙 吴艳[1,2] 王丽芳 吴昊[1,2] Chen Qimeng;Wu Yan;Wang Lifang;Wu Hao(Institute of Electrical Engineering Chinese Academy of Sciences,Beijing 100190;University of Chinese Academy of Sciences,Beijing 101408)

机构地区：[1]中国科学院电工研究所,北京100190 [2]中国科学院大学,北京101408

出　　处：《中国汽车》2023年第7期31-38,共8页China Auto

基　　金：国家重点研发计划(2022YFB2503401)。

摘　　要：各国学者都致力于开发更加先进的自动驾驶系统来缓解日益频发的交通事故对于人们生命财产的威胁。在目前的自动驾驶系统测试中,仍然以需要大量资源投入的传统道路测试为主,而更加高效安全的虚拟场景测试又缺少危险且复杂的测试场景。本文分析了常见的道路事故特征,在此基础上对危险驾驶场景要素进行了分类和分析,将危险驾驶场景关键要素分为5个大类和20个要素维度,并详细分析了每个要素维度的具体属性。该项工作为后续的危险驾驶场景生成、场景库构建等研究工作提供技术支撑。Worldwide scholars are devoted to developing more advanced self-driving system to alleviate the threat to people's lives and property brought by more frequent traffic accidents.Currently,the traditional road-based test that requires substantial resources is still the most prevailing self-driving system testing approach,while the virtual-scene-based test which is more efficient and safer has insufficient hazardous and complicated test scenarios.Based on the analysis regard to the characteristics of common road accidents,this article classifies and analyzes the elements that constitute hazardous driving scenarios.The key elements of hazardous driving scenarios are divided into 5 categories and 20 elementary dimensions and the specific attribute of each dimension are analyzed in detail.This work lays the foundation for the subsequent research on the generation of hazardous driving scenarios and the construction of scenario libraries.

关 键 词：交通事故 自动驾驶系统测试 虚拟场景测试 危险驾驶场景 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]