Research map / 3D vision / 3D generation
Vehicle Trajectory Prediction Works, but Not Everywhere
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Generate adversarial scenes with smooth turns, double turns, and ripple roads | Argoverse scenes evaluated with DATF · Argoverse generated scenes (All) | HOR (%): 2 -> 82 | Hurt |
| Generate adversarial scenes with smooth turns, double turns, and ripple roads | Argoverse scenes evaluated with WIMP · Argoverse generated scenes (All) | HOR (%): 1 -> 63 | Hurt |
| Generate adversarial scenes with smooth turns, double turns, and ripple roads | Argoverse scenes evaluated with LaneGCN · Argoverse generated scenes (All) | HOR (%): 1 -> 66 | Hurt |
| Add physical-feasibility constraints (ablation) | Generated high-speed scenarios evaluated with LaneGCN · High-speed generated scenarios | SOR (%): 47 -> 33 | Helped |
| Add physical-feasibility constraints (ablation) | Generated high-speed scenarios evaluated with MPC · High-speed generated scenarios | HOR (%): 3 -> 1 | Helped |
| Fine-tune LaneGCN on original data plus generated scenes | LaneGCN · Generated scenes, Pow=9 (Full) | SOR (%): 23 -> 14 | Helped |
| Fine-tune LaneGCN on original data plus generated scenes | LaneGCN · Generated scenes, Pow=9 (Full) | HOR (%): 66 -> 46 | Helped |
| Fine-tune LaneGCN on original data plus generated scenes | LaneGCN · Original Argoverse scenes | average / final displacement error: 1.35 / 2.98 m | No clear effect |
| Use brute-force search instead of Bayesian optimization (ablation) | Scene generation for LaneGCN · Generated scenes | SOR (%): 13 -> 23 | Helped |
| Use brute-force search instead of Bayesian optimization (ablation) | Scene generation for LaneGCN · Generated scenes | GPU hours: 17.5 -> 4.2 | Helped |
| Increase transformation power | LaneGCN · Generated scenes | HOR (%): 8 -> 66 | Hurt |
| Add lane merging as a scene transformation | Scene-generation framework · Generated scenes | HOR (%): 14% HOR | Hurt |
Closest papers by meaning
- Generating Multimodal Driving Scenes via Next-Scene Prediction (CVPR 2025 · similarity 0.87)
- SceneDiffuser++: City-Scale Traffic Simulation via a Generative World Model (CVPR 2025 · similarity 0.87)
- FreeVS: Generative View Synthesis on Free Driving Trajectory (ICLR 2025 · similarity 0.86)
- Future-Aware End-to-End Driving: Bidirectional Modeling of Trajectory Planning and Scene Evolution (NeurIPS 2025 · similarity 0.86)
- UniScene: Unified Occupancy-centric Driving Scene Generation (CVPR 2025 · similarity 0.85)
- Driving into the Future: Multiview Visual Forecasting and Planning with World Model for Autonomous Driving (CVPR 2024 · similarity 0.85)
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.