一句话：提出 HazSim 城市驾驶模拟器，用于在降雨、镜头水滴等自然危险下测试自动驾驶感知系统的鲁棒性。
感知系统是自动驾驶安全的关键，但在暴雨、镜头水滴等自然危险下其鲁棒性仍待充分评估。该论文提出 HazSim——一个面向城市驾驶的模拟器，旨在为自动驾驶系统（ADS）的感知系统提供针对危险场景的测试环境。其动机来自许多国家监管机构对传统驾驶员提出的严格测试要求，因此自动驾驶感知系统也需要类似甚至更严格的危险条件测试。摘要未展开具体仿真技术、场景生成机制和实验数据；从描述看，评估方式是在模拟的自然危险条件下检验感知系统，以支持面向鲁棒性的测试。贡献在于将城市驾驶模拟与危险因素注入结合，为感知系统安全测试提供工具支撑。
Perception systems play a crucial role in ensuring the safety of autonomous vehicles. In intricate scenarios of self-driving technology, these vehicles need to effectively handle a range of natural hazards, including heavy rainfall or the presence of raindrops on camera lenses. Consequently, it is imperative to thoroughly evaluate the robustness of perception systems in autonomous vehicles against hazards, aligned with the stringent testing requirements set by regulatory bodies in many nations for conventional drivers.
DOI 原文 ·
@inproceedings{paperbot9176,
title = {HazSim: An Urban Driving Simulator for Testing Perception Systems of ADSes Against Hazards},
author = {Xiaodong Zhang and Jie Bao and Yulong Shen and Qin Xia and Zijiang Yang},
booktitle = {FSE},
year = {2025},
doi = {10.1145/3696630.3728589}
}