Scholarlink

<p>To ensure their safe use, autonomous driving systems (ADSs) must meet rigorous safety assurance criteria that involve executing maneuvers safely within arbitrary scenarios where other actors perform their intended maneuvers. For that purpose, existing scenario generation approaches optimize search to derive scenarios with high probability of dangerous interactions. In this paper, we hypothesize that at road junctions, potential danger predominantly arises from overlapping paths of individual actors carrying out their designated high-level (abstract) maneuvers. As a step toward ADS safety assurance, we propose an approach to derive an exhaustive set of potentially dangerous logical scenarios at any given road junction, i.e., all permutations of overlapping maneuvers assigned to actors, including the ADS, for a given set of possible maneuvers. From these logical scenarios, we derive concrete-level exact paths that actors must follow to guide simulation-based testing toward potential collisions. We conduct extensive experiments over two realistic road junctions with increasing number of external actors to (1) compare our scenario generation approach to the state-of-the-art <span>Scenic</span> tool and to (2) evaluate the behavior of a state-of-the-art learning-based ADS controller. Results show that (1) our approach outperforms <span>Scenic</span> both in terms of achieved coverage and runtime, and (2) that the ADS-under-test is involved in increasing percentages of unsafe behaviors in simulation, which vary according to abstract scenario properties.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Automated and logically exhaustive generation of traffic scenarios at road junctions using a multi-level danger definition

  • 期刊论文

  • 发表日期：

    2026年5月21日

查看全文
  • Aren A. Babikian,
  • Attila Ficsor,
  • Oszkár Semeráth,
  • Gunter Mussbacher,
  • Dániel Varró

摘要

To ensure their safe use, autonomous driving systems (ADSs) must meet rigorous safety assurance criteria that involve executing maneuvers safely within arbitrary scenarios where other actors perform their intended maneuvers. For that purpose, existing scenario generation approaches optimize search to derive scenarios with high probability of dangerous interactions. In this paper, we hypothesize that at road junctions, potential danger predominantly arises from overlapping paths of individual actors carrying out their designated high-level (abstract) maneuvers. As a step toward ADS safety assurance, we propose an approach to derive an exhaustive set of potentially dangerous logical scenarios at any given road junction, i.e., all permutations of overlapping maneuvers assigned to actors, including the ADS, for a given set of possible maneuvers. From these logical scenarios, we derive concrete-level exact paths that actors must follow to guide simulation-based testing toward potential collisions. We conduct extensive experiments over two realistic road junctions with increasing number of external actors to (1) compare our scenario generation approach to the state-of-the-art Scenic tool and to (2) evaluate the behavior of a state-of-the-art learning-based ADS controller. Results show that (1) our approach outperforms Scenic both in terms of achieved coverage and runtime, and (2) that the ADS-under-test is involved in increasing percentages of unsafe behaviors in simulation, which vary according to abstract scenario properties.