Research map / Label-efficient and robust learning / Recognition and long-tail
Perceive, Predict, and Plan: Safe Motion Planning Through Interpretable Semantic Representations
Abbas Sadat, Sergio Casas, Mengye Ren, Xinyu Wu, Pranaab Dhawan, Raquel Urtasun
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 end-to-end training of perception and planning (trajectory prediction) (ablation) | P3 variants · Open-loop planning | Collision rate @5s (%): 2.94 -> 2.21 | Improved |
| add end-to-end training (occupancy) (ablation) | P3 variants · Open-loop planning | Collision rate @5s (%): 1.96 -> 1.78 | Improved |
| use semantic occupancy instead of trajectory prediction (separate training) (ablation) | P3 variants · Open-loop planning | Collision rate @5s (%): 2.94 -> 1.96 | Improved |
| use semantic occupancy instead of trajectory prediction (end-to-end) (ablation) | P3 variants · Open-loop planning | Collision rate @5s (%): 2.21 -> 1.78 | Improved |
| rasterize predicted trajectories into occupancy instead of using them directly (ablation) | P3 variants · Open-loop planning | Collision rate @5s (%): 2.94 -> 2.36 | Mixed |
| use native occupancy instead of rasterized trajectories (ablation) | P3 variants · Open-loop planning | Collision rate @5s (%): 2.36 -> 1.96 | Improved |
| add end-to-end training (occupancy) (ablation) | P3 variants · Open-loop planning | Jerk: 1.23 -> 1.27 | Hurt |
| use P3 instead of PLT | motion planner · Open-loop planning | Collision rate @5s (%): 2.94 -> 1.78 | Improved |
| use P3 instead of NMP | motion planner · Open-loop planning | Collision rate @5s (%): 5.22 -> 1.78 | Improved |
| use P3 instead of NMP | motion planner · Open-loop planning | Progress: 31.7 -> 27.6 | Hurt |
| use occupancy end-to-end (M5) instead of trajectory-based (M2) (ablation) | autonomy system · Closed-loop simulation | Collision rate (% of runs): 18.5 -> 9.8 | Improved |
| use occupancy end-to-end (M5) instead of trajectory-based (M2) (ablation) | autonomy system · Closed-loop simulation | Jerk: 4.08 -> 1.85 | Improved |
Closest papers by meaning
- Prioritizing Perception-Guided Self-Supervision: A New Paradigm for Causal Modeling in End-to-End Autonomous Driving (NeurIPS 2025 · similarity 0.88)
- Drive-R1: Bridging Reasoning and Planning in VLMs for Autonomous Driving with Reinforcement Learning (AAAI 2026 · similarity 0.88)
- $AutoDrive\text{-}P^3$: Unified Chain of Perception–Prediction–Planning Thought via Reinforcement Fine-Tuning (ICLR 2026 · similarity 0.88)
- On the Road to Portability: Compressing End-to-End Motion Planner for Autonomous Driving (CVPR 2024 · similarity 0.88)
- Driving in the Occupancy World: Vision-Centric 4D Occupancy Forecasting and Planning via World Models for Autonomous Driving (AAAI 2025 · similarity 0.88)
- SGDrive: Scene-to-Goal Hierarchical World Cognition for Autonomous Driving (CVPR 2026 · similarity 0.88)
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.
