Controllable risk scenario generation from human crash data for autonomous vehicle testing
This addresses the problem of rigorous safety testing for autonomous vehicles by providing a method to simulate rare crash-like behaviors, though it is incremental as it builds on existing agent modeling approaches.
The paper tackles the challenge of generating realistic and rare safety-critical scenarios for autonomous vehicle testing by introducing the Controllable Risk Agent Generation (CRAG) framework, which improves diversity and enables controllable risk scenario generation for efficient AV evaluation.
Ensuring the safety of autonomous vehicles (AV) requires rigorous testing under both everyday driving and rare, safety-critical conditions. A key challenge lies in simulating environment agents, including background vehicles (BVs) and vulnerable road users (VRUs), that behave realistically in nominal traffic while also exhibiting risk-prone behaviors consistent with real-world accidents. We introduce Controllable Risk Agent Generation (CRAG), a framework designed to unify the modeling of dominant nominal behaviors and rare safety-critical behaviors. CRAG constructs a structured latent space that disentangles normal and risk-related behaviors, enabling efficient use of limited crash data. By combining risk-aware latent representations with optimization-based mode-transition mechanisms, the framework allows agents to shift smoothly and plausibly from safe to risk states over extended horizons, while maintaining high fidelity in both regimes. Extensive experiments show that CRAG improves diversity compared to existing baselines, while also enabling controllable generation of risk scenarios for targeted and efficient evaluation of AV robustness.