Research map / Label-efficient and robust learning / Recognition and long-tail
SafeDrive: Fine-Grained Safety Reasoning for End-to-End Driving in a Sparse World
Jungho Kim, Jiyong Oh, Seunghoon Yu, Hongjae Shin, Donghyuk Kwak, Jun Won Choi
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 |
|---|---|---|---|
| replace DiffusionDrive head with ProposalNet (ablation) | SafeDrive · NAVSIM navtest | PDMS: 88.6 -> 89.1 | Improved |
| add scene-level safety evaluation (ablation) | ProposalNet · NAVSIM navtest | PDMS: 89.1 -> 90.1 | Improved |
| add SWNet (ablation) | ProposalNet + scene-level safety · NAVSIM navtest | PDMS: 90.1 -> 90.9 | Improved |
| add FRNet without SWNet (ablation) | ProposalNet + scene-level safety · NAVSIM navtest | PDMS: 90.1 -> 90.9 | Improved |
| add both SWNet and FRNet (ablation) | ProposalNet + scene-level safety · NAVSIM navtest | PDMS: 90.9 -> 91.6 | Improved |
| use sparse world instead of BEV world with fine-grained safety (ablation) | SafeDrive · NAVSIM navtest | PDMS: 90.9 -> 91.6 | Improved |
| BEV world model without safety evaluation vs no world model (ablation) | SafeDrive · NAVSIM navtest | PDMS: 90.1 -> 90.3 | No clear effect |
| add PwNC head (ablation) | FRNet · NAVSIM navtest | PDMS: 90.9 -> 91.5 | Improved |
| add TwDAC head (ablation) | FRNet · NAVSIM navtest | PDMS: 90.9 -> 91.4 | Improved |
| add both PwNC and TwDAC (ablation) | FRNet · NAVSIM navtest | PDMS: 90.9 -> 91.6 | Improved |
| SafeDrive vs prior best on Bench2Drive | SafeDrive · Bench2Drive | Driving Score: 63.46 -> 66.77 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- DriveTransformer — DriveTransformer: Unified Transformer for Scalable End-to-End Autonomous Driving (ICLR 2025) · better on 2 of 2 shared results
- DriveSuprim — DriveSuprim: Towards Precise Trajectory Selection for End-to-End Planning (AAAI 2026) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- SafeAuto: Knowledge-Enhanced Safe Autonomous Driving with Multimodal Foundation Models (ICML 2025 · similarity 0.88)
- AlignDrive: Aligned Lateral-Longitudinal Planning for End-to-End Autonomous Driving (NeurIPS 2026 · similarity 0.87)
- Hybrid-Driving: An Autonomous Driving Decision Framework Integrating Large Language Models, Knowledge Graphs and Driving Rules (AAAI 2025 · similarity 0.87)
- DriveCombo: Benchmarking Compositional Traffic Rule Reasoning in Autonomous Driving (CVPR 2026 · similarity 0.85)
- SGDrive: Scene-to-Goal Hierarchical World Cognition for Autonomous Driving (CVPR 2026 · similarity 0.85)
- OneVL: One-Step Latent Reasoning and Planning with Vision-Language Explanation (NeurIPS 2026 · similarity 0.84)
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.
