Research map / Probabilistic and scientific modeling / Molecules and proteins
LTP: Lane-Based Trajectory Prediction for Autonomous Driving
Jingke Wang, Tengju Ye, Ziqing Gu, Junbo Chen
Check 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 lane segments as fine-grained proposals instead of entire lanes (ablation) | LTP · Argoverse validation | Top-1 FDE: 3.71 -> 2.81 | Improved |
| use 5m lane segments instead of 10m or 2m segments (ablation) | LTP · Argoverse validation | Top-1 FDE: 10m: 3.03; 5m: 2.81; 2m: 2.91 | Mixed |
| add attention modules (ablation) | LTP with 5m lane segments · Argoverse validation | Top-1 FDE: 3.66 -> 2.81 | Improved |
| increase cross-attention depth from one layer to four (ablation) | LTP with 5m lane segments · Argoverse validation | Top-1 FDE: 3.07 -> 2.81 | Improved |
| add adaptive variance-based NMS instead of no NMS (ablation) | LTP with 5m lane segments and C-4 attention · Argoverse validation | Top-6 FDE: 1.23 -> 1.07 | Improved |
| use adaptive variance-based NMS instead of fixed-threshold NMS (ablation) | LTP with 5m lane segments and C-4 attention · Argoverse validation | Top-6 FDE: 1.14 -> 1.07 | Improved |
| add cost-volume optimization using multi-modal predictions (ablation) | LTP closed-loop simulation · Internal driving logs | Frame collision rate: 5.2 -> 0.7 | Improved |
| add cost-volume optimization using multi-modal predictions (ablation) | LTP closed-loop simulation · Internal driving logs | Trip collision rate: 65.0 -> 17.2 | Improved |
| use lane-segment proposals and variance-based NMS | LTP · Argoverse test | b-minFDE6: 1.86 -> 1.856 | Improved |
Closest papers by meaning
- Divide-and-Conquer for Lane-Aware Diverse Trajectory Prediction (CVPR 2021 · similarity 0.90)
- LaneSegNet: Map Learning with Lane Segment Perception for Autonomous Driving (ICLR 2024 · similarity 0.86)
- AlignDrive: Aligned Lateral-Longitudinal Planning for End-to-End Autonomous Driving (NeurIPS 2026 · similarity 0.86)
- LaKD: Length-agnostic Knowledge Distillation for Trajectory Prediction with Any Length Observations (NeurIPS 2024 · similarity 0.86)
- HPNet: Dynamic Trajectory Forecasting with Historical Prediction Attention (CVPR 2024 · similarity 0.85)
- GoIRL: Graph-Oriented Inverse Reinforcement Learning for Multimodal Trajectory Prediction (ICML 2025 · 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.
