Research map / Reinforcement learning / Visual and representation RL
Autonomous Vehicle Path Planning by Searching with Differentiable Simulation
Asen Nachkov, Jan-Nico Zaech, Danda Pani Paudel, Xi Wang, Luc Van Gool
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 |
|---|---|---|---|
| add gradient step on actions (no search) (ablation) | DSS reactive policy · Waymax tracking | ADE: 2.9792 -> 0.4644 | Improved |
| increase accel gradient step from 1000 to 5000 (ablation) | DSS reactive · Waymax tracking | ADE: 0.4644 -> 0.7678 | Hurt |
| add Monte Carlo search with K=8 rollouts (no gradient) (ablation) | DSS · Waymax tracking | ADE: 2.9792 -> 1.0710 | Improved |
| add gradients on top of K=8 search (ablation) | DSS · Waymax tracking | ADE: 1.0710 -> 0.1766 | Improved |
| increase rollouts K from 1 to 8 with gradients (ablation) | DSS · Waymax tracking | ADE: 0.4644 -> 0.1766 | Improved |
| add search K=4 vs single rollout for collision/offroad planning (ablation) | DSS · Waymax path planning | ADE: 2.9817 -> 1.3861 | Improved |
| add gradients with K=4,T=10 (ablation) | DSS · Waymax path planning | ADE: 1.6016 -> 1.3861 | Improved |
| increase horizon T from 10 to 20 with K=4 (ablation) | DSS · Waymax path planning | offroad rate: 0.0638 -> 0.0782 | Hurt |
| add test-time planning to policy not trained on collisions (ablation) | DSS · Waymax path planning | overlap rate: 0.1380 -> 0.0442 | Improved |
| DSS vs PiDT/EasyChauffeur on collision and offroad | DSS · Waymax path planning | offroad rate: 0.0298 (PiDT) vs 0.0762 | Hurt |
Closest papers by meaning
- Diffusion-ES: Gradient-free Planning with Diffusion for Autonomous and Instruction-guided Driving (CVPR 2024 · similarity 0.88)
- Diffusion-Based Planning for Autonomous Driving with Flexible Guidance (ICLR 2025 · similarity 0.86)
- World Models as Adversaries: Multi-Agent Self-Play Fine-Tuning for Robust Motion Planning (NeurIPS 2026 · similarity 0.86)
- Future-Aware End-to-End Driving: Bidirectional Modeling of Trajectory Planning and Scene Evolution (NeurIPS 2025 · similarity 0.86)
- Versatile Navigation Under Partial Observability via Value-guided Diffusion Policy (CVPR 2024 · similarity 0.86)
- AlignDrive: Aligned Lateral-Longitudinal Planning for End-to-End Autonomous Driving (NeurIPS 2026 · similarity 0.86)
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.
