Enhancing Autonomous Driving Safety with Collision Scenario Integration

Autonomous vehicle safety is crucial for the successful deployment of self-driving cars. However, most existing planning methods rely heavily on imitation learning, which limits their ability to leverage collision data effectively. Moreover, collecting collision or near-collision data is inherently challenging, as it involves risks and raises ethical and practical concerns. In this paper, we propose SafeFusion, a training framework to learn from collision data. Instead of over-relying on imitation learning, Safefusion integrates safety-oriented metrics during training to enable collision avoidance learning. In addition, to address the scarcity of collision data, we propose CollisionGen, a scalable data generation pipeline to generate diverse, high-quality scenarios using natural language prompts, generative models, and rule-based filtering. Experimental results show that our approach improves planning performance in collision-prone scenarios by 56% over previous state-of-the-art planners while maintaining effectiveness in regular driving situations. Our work provides a scalable and effective solution for advancing the safety of autonomous driving systems.

SafeFusion

SafeFusion aims to improve the safety of neural planners by integrating collision data into training, using a planning vocabulary and Multi-Target Knowledge-Distillation to address the challenge of unavailable collision-avoidance trajectories in collision data. Our framework can effectively handle collision data, whether collected from real-world environments or generated synthetically. In this work, due to the lack of access to real-world collision datasets, we utilize the synthetic data CollisionGen generated for both training and testing.

CollisionGen

The pipeline begins by taking text descriptions of collision scenarios as input. A generator with a language interpreter and a generative transformer is then applied, followed by the use of predefined rules and a PDM simulator to filter out qualified collision scenarios.

---

Results

Please refer to the paper for detailed quantitative results and analysis. Below, we present visualizations showcasing performance. The red cars represent vehicles controlled by the planner, while the blue cars depict other vehicles exhibiting malicious behaviors.

---

PDM-Closed	Hydra-GT	SafeFusion (Ours)

Here are more visualizations.