Abstract:Autonomous driving systems (ADSs) have undergone remarkable development and are increasingly employed in safety-critical applications. However, recently reported data on fatal accidents involving ADSs suggests that the desired level of safety has not yet been fully achieved. Consequently, there is a growing need for more comprehensive and targeted testing approaches to ensure safe driving. Scenarios from real-world accident reports provide valuable resources for ADS testing, including critical scenarios and high-quality seeds. However, existing scenario reconstruction methods from accident reports often exhibit limited accuracy in information extraction. Moreover, due to the diversity and complexity of road environments, matching current accident information with the simulation map data for reconstruction poses significant challenges. In this paper, we design and implement SoVAR, a tool for automatically generating road-generalizable scenarios from accident reports. SoVAR utilizes well-designed prompts with linguistic patterns to guide the large language model in extracting accident information from textual data. Subsequently, it formulates and solves accident-related constraints in conjunction with the extracted accident information to generate accident trajectories. Finally, SoVAR reconstructs accident scenarios on various map structures and converts them into test scenarios to evaluate its capability to detect defects in industrial ADSs. We experiment with SoVAR, using accident reports from the National Highway Traffic Safety Administration's database to generate test scenarios for the industrial-grade ADS Apollo. The experimental findings demonstrate that SoVAR can effectively generate generalized accident scenarios across different road structures. Furthermore, the results confirm that SoVAR identified 5 distinct safety violation types that contributed to the crash of Baidu Apollo.

Multi-Agent Vulnerability Discovery for Autonomous Driving Policy by Finding AV-Responsible Scenarios

Multi-Agent Vulnerability Discovery for Autonomous Driving with Hazard Arbitration Reward

DAnoScenE: a driving anomaly scenario extraction framework for autonomous vehicles in urban streets

SafeDrive: Knowledge- and Data-Driven Risk-Sensitive Decision-Making for Autonomous Vehicles with Large Language Models

Building Safer Autonomous Agents by Leveraging Risky Driving Behavior Knowledge

A Data-Driven Control-Policy-Based Driving Safety Analysis System for Autonomous Vehicles

SoVAR: Building Generalizable Scenarios from Accident Reports for Autonomous Driving Testing

Quantitative Representation of Scenario Difficulty for Autonomous Driving Based on Adversarial Policy Search

Risk-anticipatory autonomous driving strategies considering vehicles' weights, based on hierarchical deep reinforcement learning

Adversarial Driving Behavior Generation Incorporating Human Risk Cognition for Autonomous Vehicle Evaluation

Interactive Critical Scenario Generation for Autonomous Vehicles Testing Based on In-depth Crash Data Using Reinforcement Learning

Stackelberg Driver Model for Continual Policy Improvement in Scenario-Based Closed-Loop Autonomous Driving

Generating Critical Scenarios for Testing Automated Driving Systems

Driving-Policy Adaptive Safeguard for Autonomous Vehicles Using Reinforcement Learning

Adversarial Safety-Critical Scenario Generation using Naturalistic Human Driving Priors

A Hazardous Lane-Changing Scenario Search Approach Based on Adaptive Random Testing for Automated Vehicles

Adversarial Generation of Safety-Critical Lane-Change Scenarios for Autonomous Vehicles

Adversarial Stress Test for Autonomous Vehicle Via Series Reinforcement Learning Tasks with Reward Shaping

Acceleration Method for Hazardous Scenario Generation Based on Disengagement Cases

Graph Structure-Based Implicit Risk Reasoning for Long-Tail Scenarios of Automated Driving

Towards Cooperative Driving among Heterogeneous CAVs: A Safe Multi-Agent Reinforcement Learning Approach