CrashAgent: Crash Scenario Generation via Multi-modal Reasoning
This addresses the long-tail distribution issue in autonomous driving data, enabling safer algorithm development by providing diverse crash scenarios, though it is an incremental improvement in scenario generation.
The paper tackles the problem of limited safety-critical scenarios in autonomous driving datasets by generating crash scenarios from real-world crash reports using a multi-agent framework, resulting in a high-quality, large-scale dataset for training and evaluation.
Training and evaluating autonomous driving algorithms requires a diverse range of scenarios. However, most available datasets predominantly consist of normal driving behaviors demonstrated by human drivers, resulting in a limited number of safety-critical cases. This imbalance, often referred to as a long-tail distribution, restricts the ability of driving algorithms to learn from crucial scenarios involving risk or failure, scenarios that are essential for humans to develop driving skills efficiently. To generate such scenarios, we utilize Multi-modal Large Language Models to convert crash reports of accidents into a structured scenario format, which can be directly executed within simulations. Specifically, we introduce CrashAgent, a multi-agent framework designed to interpret multi-modal real-world traffic crash reports for the generation of both road layouts and the behaviors of the ego vehicle and surrounding traffic participants. We comprehensively evaluate the generated crash scenarios from multiple perspectives, including the accuracy of layout reconstruction, collision rate, and diversity. The resulting high-quality and large-scale crash dataset will be publicly available to support the development of safe driving algorithms in handling safety-critical situations.