Research map / Vision-language and multimodal / Robot action and navigation
ORION: A Holistic End-to-End Autonomous Driving Framework by Vision-Language Instructed Action Generation
Haoyu Fu, Diankun Zhang, Zongchuang Zhao, Jianfeng Cui, Dingkang Liang, Chong Zhang, Dingyuan Zhang, Hongwei Xie, Bing Wang, Xiang Bai
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 |
|---|---|---|---|
| replace diffusion planner with VAE generative planner (ablation) | ORION · Bench2Drive closed-loop | Driving Score: 71.97 -> 77.74 | Improved |
| add generative planner output instead of plain text trajectory output (ablation) | ORION with QT-Former · Bench2Drive closed-loop | Driving Score: 42.23 -> 77.74 | Improved |
| use generative planner instead of MLP decoder for the special planning token (ablation) | ORION · Bench2Drive closed-loop | Driving Score: +7.01 DS, +9.5% SR, +6.28% mean ability | Improved |
| add explicit traffic state supervision to QT-Former (ablation) | ORION · Bench2Drive closed-loop | Driving Score: 56.33 -> 74.65 | Improved |
| add motion prediction to QT-Former perception head (ablation) | ORION · Bench2Drive closed-loop | Driving Score / Success Rate: DS 74.65 -> 74.07, SR 49.31 -> 49.77 | Mixed |
| add long-term memory bank with history queries to QT-Former (ablation) | ORION · Bench2Drive closed-loop | Driving Score: 74.07 -> 77.74 | Improved |
| add QT-Former designs to plain text output model (ablation) | plain text VLM planner · Bench2Drive closed-loop | Driving Score: 25.45 -> 42.23 | Improved |
| increase history queries from 0 to 16 (ablation) | ORION (planning-only training) · Bench2Drive closed-loop | Driving Score: 65.10 -> 74.10 | Improved |
| increase history queries from 16 to 32 (ablation) | ORION (planning-only training) · Bench2Drive closed-loop | Driving Score: 74.10 -> 62.46 | Hurt |
| add joint VQA and planning training instead of planning-only training (ablation) | ORION · Bench2Drive closed-loop | Driving Score: 74.10 -> 77.74 | Improved |
| add VQA fine-tuning stage (V to L) before planning training (ablation) | ORION · Bench2Drive closed-loop | Driving Score: 57.96 -> 65.10 | Improved |
| add third-stage end-to-end joint fine-tuning (V to L to A) (ablation) | ORION · Bench2Drive closed-loop | Driving Score: 65.10 -> 74.65 | Improved |
Later papers that compared with this method
13 later accepted papers report ORION as a baseline in their results tables; 12 of them report a better number on most of the results they share with it.
- UniDrive-WM: Unified Understanding, Planning and Generation World Model For Autonomous Driving (ECCV 2026 · better on 8 of 8 shared results)
- Generative Scenario Rollouts for End-to-End Autonomous Driving (NeurIPS 2026 · better on 5 of 5 shared results)
- Scene-Adaptive VLA: Efficient Autonomous Driving via Dynamic Layer Routing (NeurIPS 2026 · better on 4 of 4 shared results)
- SpaceDrive: Infusing Spatial Awareness into VLM-based Autonomous Driving (CVPR 2026 · better on 4 of 4 shared results)
- DeepSight: Long-Horizon World Modeling via Latent States Prediction for End-to-End Autonomous Driving (ICML 2026 · better on 3 of 3 shared results)
- Unifying Language-Action Understanding and Generation for Autonomous Driving (CVPR 2026 · better on 3 of 3 shared results)
- MindDrive: A Vision-Language-Action Model for Autonomous Driving via Online Reinforcement Learning (ECCV 2026 · better on 3 of 3 shared results)
- Explicit Geometric Chain-of-Thought for Vision-Language-Action in Autonomous Driving (NeurIPS 2026 · better on 2 of 2 shared results)
5 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
- SOLVE: Synergy of Language-Vision and End-to-End Networks for Autonomous Driving (CVPR 2025 · similarity 0.92)
- AutoVLA: A Vision-Language-Action Model for End-to-End Autonomous Driving with Adaptive Reasoning and Reinforcement Fine-Tuning (NeurIPS 2025 · similarity 0.92)
- Drive-R1: Bridging Reasoning and Planning in VLMs for Autonomous Driving with Reinforcement Learning (AAAI 2026 · similarity 0.91)
- More Than Meets the Eye? Uncovering the Reasoning-Planning Disconnect in Training Vision-Language Driving Models (NeurIPS 2026 · similarity 0.90)
- OneVL: One-Step Latent Reasoning and Planning with Vision-Language Explanation (NeurIPS 2026 · similarity 0.90)
- Generative Scenario Rollouts for End-to-End Autonomous Driving (NeurIPS 2026 · similarity 0.90)
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.
