Research map / Label-efficient and robust learning / Recognition and long-tail
TransFuser: Imitation With Transformer-Based Sensor Fusion for Autonomous Driving
Kashyap Chitta, Aditya Prakash, Bernhard Jaeger, Zehao Yu, Katrin Renz, Andreas Geiger
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 transformer-based global fusion instead of geometry-based fusion | TransFuser · Longest6 | Driving Score: 27.32 -> 47.30 | Improved |
| add transformer-based fusion instead of late fusion | TransFuser · Longest6 | Driving Score: 22.47 -> 47.30 | Improved |
| add LiDAR fusion via transformers on top of image-only Latent TransFuser | TransFuser · CARLA leaderboard | Driving Score: 45.20 -> 61.18 | Improved |
| add all four auxiliary losses (ablation) | TransFuser · Longest6 | Route Completion: 78.17 -> 92.28 | Improved |
| add all four auxiliary losses (ablation) | TransFuser · Longest6 | Driving Score: 44.29 -> 56.68 | Improved |
| add depth loss alone (ablation) | TransFuser · Longest6 | Driving Score: 56.23 vs 56.68 best seed | No clear effect |
| add global safety heuristic (also during regular driving) | Latent TransFuser · Longest6 | Driving Score: 42.19 -> 50.00 | Improved |
| add global safety heuristic (also during regular driving) | TransFuser · CARLA leaderboard | Driving Score: 50.57 -> 41.93 | Hurt |
| add global safety heuristic (also during regular driving) | TransFuser · Longest6 | Driving Score: 42.51 -> 49.49 | Improved |
| ensemble 3 models | TransFuser · CARLA leaderboard | Driving Score: 50.57 -> 61.18 | Improved |
| ensemble 3 models | Latent TransFuser · Longest6 | Driving Score: 42.19 -> 36.18 | Hurt |
| use 4 fusion scales instead of only deepest 1, 2 or 3 (ablation) | TransFuser · Longest6 | Driving Score: 49.35, 53.52, 48.77 for 1, 2, 3 scales vs 56.68 | Improved |
Later papers that compared with this method
16 later accepted papers report TransFuser as a baseline in their results tables; 15 of them report a better number on most of the results they share with it.
- StyleDrive: Towards Driving-Style Aware Benchmarking of End-To-End Autonomous Driving (AAAI 2026 · better on 4 of 4 shared results)
- VADv2: End-to-End Autonomous Driving via Probabilistic Planning (ICLR 2026 · better on 1 of 2 shared results)
- ResAD: Normalized Residual Trajectory Modeling for End-to-End Autonomous Driving (CVPR 2026 · better on 2 of 2 shared results)
- MOJITO: Modal Joint Learning for Unified End-to-End Autonomous Driving (ECCV 2026 · better on 2 of 2 shared results)
- DIVER: Reinforced Diffusion Breaks Imitation Bottlenecks in End-to-End Autonomous Driving (TPAMI 2026 · better on 2 of 2 shared results)
- Trajectory Evaluation via Rollout-Free World Model for End-to-end Autonomous Driving (NeurIPS 2026 · better on 1 of 1 shared result)
- DriveSuprim: Towards Precise Trajectory Selection for End-to-End Planning (AAAI 2026 · better on 1 of 1 shared result)
- SparseDriveV2: Scoring is All You Need for End-to-End Autonomous Driving (ECCV 2026 · better on 1 of 1 shared result)
8 more not listed.
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
- Multi-Modal Fusion Transformer for End-to-End Autonomous Driving (CVPR 2021 · similarity 0.96)
- Knowledge Distillation from Single-Task Teachers to Multi-Task Student for End-to-End Autonomous Driving (AAAI 2024 · similarity 0.87)
- GaussianFusion: Gaussian-Based Multi-Sensor Fusion for End-to-End Autonomous Driving (NeurIPS 2025 · similarity 0.87)
- DriveTransformer: Unified Transformer for Scalable End-to-End Autonomous Driving (ICLR 2025 · similarity 0.87)
- Prioritizing Perception-Guided Self-Supervision: A New Paradigm for Causal Modeling in End-to-End Autonomous Driving (NeurIPS 2025 · similarity 0.85)
- DrivingGaussian: Composite Gaussian Splatting for Surrounding Dynamic Autonomous Driving Scenes (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.
