Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Automated and Complete Generation of Traffic Scenarios at Road Junctions Using a Multi-level Danger Definition

Aren A. Babikian, Attila Ficsor, Oszkár Semeráth, Gunter Mussbacher, Dániel Varró

TL;DR

The paper tackles the challenge of certifying autonomous vehicles by ensuring broad abstract coverage of dangerous traffic scenarios at road junctions. It introduces a multi-level approach that first generates a complete set of dangerous logical scenarios from a functional specification, then refines them into concrete, collision-inducing paths with speed profiles for simulation. The method provides theoretical guarantees of completeness at the logical level and demonstrates practicality by evaluating a state-of-the-art AV controller in CARLA across two junctions, revealing how danger scales with the number of external actors, maneuver type, and road geometry. The results underscore the importance of abstract scenario coverage for robust AV safety testing and offer a scalable framework for certification-oriented scenario generation and evaluation.

Abstract

To ensure their safe use, autonomous vehicles (AVs) must meet rigorous certification 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 situations. 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 maneuvers. As a step towards AV certification, we propose an approach to derive a complete set of (potentially dangerous) abstract scenarios at any given road junction, i.e. all permutations of overlapping abstract paths assigned to actors (including the AV) for a given set of possible abstract paths. From these abstract scenarios, we derive exact paths that actors must follow to guide simulation-based testing towards potential collisions. We conduct extensive experiments to evaluate the behavior of a state-of-the-art learning-based AV controller on scenarios generated over two realistic road junctions with increasing number of external actors. Results show that the AV-under-test is involved in increasing percentages of unsafe behaviors in simulation, which vary according to functional- and logical-level scenario properties.

Automated and Complete Generation of Traffic Scenarios at Road Junctions Using a Multi-level Danger Definition

TL;DR

The paper tackles the challenge of certifying autonomous vehicles by ensuring broad abstract coverage of dangerous traffic scenarios at road junctions. It introduces a multi-level approach that first generates a complete set of dangerous logical scenarios from a functional specification, then refines them into concrete, collision-inducing paths with speed profiles for simulation. The method provides theoretical guarantees of completeness at the logical level and demonstrates practicality by evaluating a state-of-the-art AV controller in CARLA across two junctions, revealing how danger scales with the number of external actors, maneuver type, and road geometry. The results underscore the importance of abstract scenario coverage for robust AV safety testing and offer a scalable framework for certification-oriented scenario generation and evaluation.

Abstract

To ensure their safe use, autonomous vehicles (AVs) must meet rigorous certification 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 situations. 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 maneuvers. As a step towards AV certification, we propose an approach to derive a complete set of (potentially dangerous) abstract scenarios at any given road junction, i.e. all permutations of overlapping abstract paths assigned to actors (including the AV) for a given set of possible abstract paths. From these abstract scenarios, we derive exact paths that actors must follow to guide simulation-based testing towards potential collisions. We conduct extensive experiments to evaluate the behavior of a state-of-the-art learning-based AV controller on scenarios generated over two realistic road junctions with increasing number of external actors. Results show that the AV-under-test is involved in increasing percentages of unsafe behaviors in simulation, which vary according to functional- and logical-level scenario properties.

Paper Structure

This paper contains 19 sections, 7 figures, 1 table, 2 algorithms.

Table of Contents

  1. Introduction
  2. Preliminaries
  3. Traffic scenes and scenarios
  4. Levels of abstraction in traffic scenarios
  5. Overview of our approach
  6. Logical-level path regions
  7. Dangerous concrete scenarios
  8. Deriving concrete paths
  9. Simulating concrete scenarios
  10. Evaluation
  11. Road junctions
  12. General measurement setup
  13. RQ1: Danger and preventive maneuvers
  14. RQ2: Increasing number of external actors
  15. RQ3: Functional-level maneuver analysis
  16. ...and 4 more sections

Figures (7)

  • Figure 1: A traffic scenario represented at various levels of abstraction
  • Figure 2: Overview of our scenario generation approach
  • Figure 3: Logical scenarios at a road junction. The designated path region of the ego and external actors (with arrows depicting their orientation) are depicted in red and in blue, respectively. Overlapping regions are depicted in gray.
  • Figure 4: Scenes depicting the moment of detection (left) and of collision (right) for avoidable and unavoidable collisions.
  • Figure 5: The road junctions under test
  • ...and 2 more figures