Research map / Image generation / Autoregressive and latent generation
Safe-Sim: Safety-Critical Closed-Loop Traffic Simulation with Diffusion-Controllable Adversaries
Wei-Jer Chang, Francesco Pittaluga, Masayoshi Tomizuka, Wei Zhan, Manmohan Chandraker
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use Safe-Sim instead of STRIVE | rule-based planner on nuScenes · nuScenes | Collision rate (%) / Realism: 36.4 -> 43.2 collision; realism 0.85 -> 0.38 | Improved |
| use Safe-Sim instead of DiffScene | rule-based planner · nuPlan | Collision rate (%): 56.7 -> 80 | Improved |
| Safe-Sim vs DiffScene adversary off-road and relative speed on nuPlan | Safe-Sim · nuPlan | Adv offroad (%): 5.0 -> 11.7 | Hurt |
| add partial diffusion with varied trajectory proposals (ablation) | Safe-Sim · nuScenes | Collision point variance (m): 1.62 -> 5.44 | Improved |
| add regularization term J_reg (ablation) | Safe-Sim · nuScenes | Adv offroad (%): 23.1 -> 14.6 | Improved |
| add regularization term J_reg (ablation) | Safe-Sim · nuScenes | Ego-adversary collision rate (%): 53.5 -> 23.9 | Hurt |
| increase TTC cost weight from 0.0 to 2.0 (ablation) | Safe-Sim · nuScenes | Collision rate (%): 48.2 -> 60.7 | Improved |
| increase TTC cost weight from 0.0 to 2.0 (ablation) | Safe-Sim · nuScenes | Realism: 0.76 -> 0.81 | Hurt |
| add J_col term to regularization (ablation) | Safe-Sim · nuScenes | Collision rate (%): 25.0 -> 43.2 | Improved |
| add J_route term to regularization (ablation) | Safe-Sim · nuScenes | Other offroad (%): 5.6 -> 1.9 | Improved |
| relative speed control target from -2.0 to 2.0 m/s (ablation) | Safe-Sim · nuScenes | Ego-Adv relative speed (m/s): 0.90 -> 1.94 | Improved |
Closest papers by meaning
- DiffScene: Diffusion-Based Safety-Critical Scenario Generation for Autonomous Vehicles (AAAI 2025 · similarity 0.91)
- Causal Composition Diffusion Model for Closed-loop Traffic Generation (CVPR 2025 · similarity 0.89)
- World Models as Adversaries: Multi-Agent Self-Play Fine-Tuning for Robust Motion Planning (NeurIPS 2026 · similarity 0.89)
- SceneDiffuser: Efficient and Controllable Driving Simulation Initialization and Rollout (NeurIPS 2024 · similarity 0.87)
- Diffusion-Based Planning for Autonomous Driving with Flexible Guidance (ICLR 2025 · similarity 0.87)
- BehaviorGPT: Smart Agent Simulation for Autonomous Driving with Next-Patch Prediction (NeurIPS 2024 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
