# RefineDrive: Reliable Failure-Guided Learning for Vision-Language-Action Driving

arXiv: https://arxiv.org/abs/2609.35078 · HTML: https://mlsanity.com/paper/2609.35078/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in VLA for autonomous driving, one of 13 directions in Vision-language-action models.
- Builds on 28 of its 29 references that are in the map.
- First posted on arXiv 28 Sep 2026.

## Placed in

- Vision-language-action models / [VLA for autonomous driving](https://mlsanity.com/vla/vla-for-autonomous-driving/)
- End-to-end autonomous driving / [Reasoning VLA and VLM driving policies](https://mlsanity.com/driving/reasoning-vla-and-vlm-driving-policies/)

## Builds on (most cited first, 28 of 28)

- [Planning-oriented Autonomous Driving](https://mlsanity.com/paper/2212.10156/) (2022)
- [DriveLM: Driving with Graph Visual Question Answering](https://mlsanity.com/paper/2312.14150/) (2023)
- [DriveVLM: The Convergence of Autonomous Driving and Large Vision-Language Models](https://mlsanity.com/paper/2402.12289/) (2024)
- [TransFuser: Imitation with Transformer-Based Sensor Fusion for Autonomous Driving](https://mlsanity.com/paper/2205.15997/) (2022)
- [DriveGPT4: Interpretable End-to-end Autonomous Driving via Large Language Model](https://mlsanity.com/paper/2310.01412/) (2023)
- [ChauffeurNet: Learning to Drive by Imitating the Best and Synthesizing the Worst](https://mlsanity.com/paper/1812.03079/) (2018)
- [NAVSIM: Data-Driven Non-Reactive Autonomous Vehicle Simulation and Benchmarking](https://mlsanity.com/paper/2406.15349/) (2024)
- [GPT-Driver: Learning to Drive with GPT](https://mlsanity.com/paper/2310.01415/) (2023)
- [DiffusionDrive: Truncated Diffusion Model for End-to-End Autonomous Driving](https://mlsanity.com/paper/2411.15139/) (2024)
- [LMDrive: Closed-Loop End-to-End Driving with Large Language Models](https://mlsanity.com/paper/2312.07488/) (2023)
- [AutoVLA: A Vision-Language-Action Model for End-to-End Autonomous Driving with Adaptive Reasoning and Reinforcement Fine-Tuning](https://mlsanity.com/paper/2506.13757/) (2025)
- [ReCogDrive: A Reinforced Cognitive Framework for End-to-End Autonomous Driving](https://mlsanity.com/paper/2506.08052/) (2025)
- [OpenDriveVLA: Towards End-to-end Autonomous Driving with Large Vision Language Action Model](https://mlsanity.com/paper/2503.23463/) (2025)
- [Pseudo-Simulation for Autonomous Driving](https://mlsanity.com/paper/2506.04218/) (2025)
- [DriveVLA-W0: World Models Amplify Data Scaling Law in Autonomous Driving](https://mlsanity.com/paper/2510.12796/) (2025)
- [Reason2Drive: Towards Interpretable and Chain-based Reasoning for Autonomous Driving](https://mlsanity.com/paper/2312.03661/) (2023)
- [DriveSuprim: Towards Precise Trajectory Selection for End-to-End Planning](https://mlsanity.com/paper/2506.06659/) (2025)
- [DiffusionDriveV2: Reinforcement Learning-Constrained Truncated Diffusion Modeling in End-to-End Autonomous Driving](https://mlsanity.com/paper/2512.07745/) (2025)
- [Discrete Diffusion for Reflective Vision-Language-Action Models in Autonomous Driving](https://mlsanity.com/paper/2509.20109/) (2025)
- [DriveFine: Refining-Augmented Masked Diffusion VLA for Precise and Robust Driving](https://mlsanity.com/paper/2602.14577/) (2026)
- [Unleashing VLA Potentials in Autonomous Driving via Explicit Learning from Failures](https://mlsanity.com/paper/2603.01063/) (2026)
- [ExploreVLA: Dense World Modeling and Exploration for End-to-End Autonomous Driving](https://mlsanity.com/paper/2604.02714/) (2026)
- [Qwen-Drive-1.0: An Initial Step towards a Vision-Language Foundation Model for Autonomous Driving](https://mlsanity.com/paper/2609.00111/) (2026)
- [SafeAlign-VLA: A Negative-Enhanced Safe Alignment Framework for Risk-Aware Autonomous Driving](https://mlsanity.com/paper/2605.19524/) (2026)
- [FIRE-VLA: Failure-Informed Self-Evolution for Vision-Language-Action Models in Autonomous Driving](https://mlsanity.com/paper/2608.13395/) (2026)
- [DriveMA: Rethinking Language Interfaces in Driving VLAs with One-Step Meta-Actions](https://mlsanity.com/paper/2605.21273/) (2026)
- [Learning from Mistakes: Rollout-Retrieval Lifelong Policy Learning for Autonomous Driving](https://mlsanity.com/paper/2606.30537/) (2026)
- [Teaching Vision-Language-Action Models What to See and Where to Look](https://mlsanity.com/paper/2607.01658/) (2026)

## Abstract (arXiv)

Vision-Language-Action (VLA) models for autonomous driving rely heavily on successful expert demonstrations, leaving model-specific failures underexploited. Learning from these failures is hindered by unreliable diagnoses, poorly matched correction targets, and coarse rewards. We propose RefineDrive, a failure-guided post-training framework that learns from self-generated failures through targeted supervision and safety-aware reinforcement learning. Reliable Diagnosis derives structured, verifiable feedback on collisions and drivable-area violations directly from simulator states. Minimum-Correction Target Retrieval searches a clustered human trajectory bank for nearby corrections that satisfy hard-safety constraints in the current scene, prioritizing preservation of the failed prediction's motion pattern. Conditioned on the driving context and failed trajectory, Correction SFT learns to generate the diagnosis followed by the retrieved correction as a training-only auxiliary task. We then apply GRPO with a Safety-Layered Reward that strictly prioritizes hard-safe trajectories, retains continuous safety feedback for both unsafe and hard-safe trajectories, and rewards driving progress only after hard safety is satisfied. At inference, the policy directly predicts trajectories from the driving context without an explicit diagnosis or repair stage. On NAVSIM v1, RefineDrive improves the 4B base SFT policy from 87.7 to 91.7 PDMS. Using the same checkpoint without additional training, RefineDrive achieves 89.4 EPDMS on the original NAVTEST scenes evaluated with NAVSIM v2 extended metrics. Controlled ablations support the benefits of structured diagnosis supervision, retrieved corrections, and safety-layered optimization for direct planning.