Research map / Vision-language and multimodal / Robot action and navigation
ReCogDrive: A Reinforced Cognitive Framework for End-to-End Autonomous Driving
Yongkang Li, Kaixin Xiong, Xiangyu Guo, Fang Li, Sixu Yan, Gangwei Xu, Lijun Zhou, Long Chen, Haiyang Sun, BING WANG, Kun Ma, Guang Chen and 3 more
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 driving VQA pre-training of the VLM (ablation) | InternVL3 trajectory model · NAVSIM navtest | PDMS: 82.4 -> 84.1 | Improved |
| add cognition-guided diffusion planner (ablation) | ReCogDrive · NAVSIM navtest | PDMS: 84.1 -> 86.5 | Improved |
| add DiffGRPO reinforcement learning (ablation) | ReCogDrive · NAVSIM navtest | PDMS: 86.5 -> 90.8 | Improved |
| replace text trajectory output with diffusion planner (ablation) | ReCogDrive · NAVSIM navtest | PDMS and latency: 84.1 -> 86.5 PDMS; 0.5839 -> 0.0750 s | Improved |
| use MLP head instead of diffusion planner (ablation) | VLM trajectory decoder · NAVSIM navtest | PDMS: 75.5 vs 85.4 (vanilla diffusion) | Hurt |
| add SwiGLU FFN, RMS Norm, RoPE and QK Norm to the diffusion planner (ablation) | VLM + Vanilla Diffusion · NAVSIM navtest | PDMS: 85.4 -> 86.5 | Improved |
| add chain-of-thought reasoning guidance (ablation) | ReCogDrive · NAVSIM navtest | PDMS: 90.8 -> 90.7 | No clear effect |
| add high-level command guidance (ablation) | ReCogDrive · NAVSIM navtest | PDMS: 90.8 -> 90.8 | No clear effect |
| increase behavior cloning loss weight from 0.001 to 0.1 (ablation) | DiffGRPO · NAVSIM navtest | PDMS: 90.8 -> 89.3 | Hurt |
| use discount factor 0.6 instead of 1.0 over diffusion steps (ablation) | DiffGRPO · NAVSIM navtest | PDMS: 90.3 -> 90.8 | Improved |
| set minimum diffusion noise sigma_min to 0.02 instead of 0.01 (ablation) | DiffGRPO · NAVSIM navtest | PDMS: 89.5 -> 90.8 | Improved |
| set minimum diffusion noise sigma_min to 0.04 instead of 0.02 (ablation) | DiffGRPO · NAVSIM navtest | PDMS: 90.8 -> 90.5 | Hurt |
Later papers that compared with this method
17 later accepted papers report ReCogDrive-8B as a baseline in their results tables; 17 of them report a better number on most of the results they share with it.
- SGDrive: Scene-to-Goal Hierarchical World Cognition for Autonomous Driving (CVPR 2026 · better on 5 of 5 shared results)
- DriveMind: Mind-Evolving Belief Tracking for Closed-Loop Autonomous Driving (NeurIPS 2026 · better on 3 of 4 shared results)
- DriveFine: Refining-Augmented Masked Diffusion VLA for Accurate and Robust Driving (ECCV 2026 · better on 4 of 4 shared results)
- AutoMoT: A Unified Vision-Language-Action Model with Asynchronous Mixture -of-Transformers for End-to-End Autonomous Driving (ICML 2026 · better on 2 of 3 shared results)
- MindDrive: A Vision-Language-Action Model for Autonomous Driving via Online Reinforcement Learning (ECCV 2026 · better on 2 of 2 shared results)
- DynVLA: Learning World Dynamics for Action Reasoning in Autonomous Driving (ICML 2026 · better on 1 of 1 shared result)
- DriveWorld-VLA: Unified Latent-Space World Modeling with Vision–Language–Action for Autonomous Driving (ICML 2026 · better on 1 of 1 shared result)
- DriveVLA-W0: World Models Amplify Data Scaling Law in Autonomous Driving (ICLR 2026 · better on 1 of 1 shared result)
9 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
- Drive-R1: Bridging Reasoning and Planning in VLMs for Autonomous Driving with Reinforcement Learning (AAAI 2026 · 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)
- SGDrive: Scene-to-Goal Hierarchical World Cognition for Autonomous Driving (CVPR 2026 · similarity 0.91)
- Distill to Think, Foresee to Act: Cognitive-Physical Reinforcement Learning for Autonomous Driving (NeurIPS 2026 · similarity 0.91)
- $AutoDrive\text{-}P^3$: Unified Chain of Perception–Prediction–Planning Thought via Reinforcement Fine-Tuning (ICLR 2026 · similarity 0.91)
- DriveCombo: Benchmarking Compositional Traffic Rule Reasoning in Autonomous Driving (CVPR 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.
