VLADriver-RAG: Retrieval-Augmented Vision-Language-Action Models for Autonomous Driving

VLADriver-RAG is a retrieval-augmented Vision-Language-Action framework for autonomous driving, designed to enhance planning robustness through structure-aware historical scenario retrieval.

Project Page • arXiv •

News

• [2026/05/07] 🌐 Project page is live: Project.
• [2026/05/12] 👉 We released our paper on arXiv.

---

Visualization

In challenging corner-case driving scenarios, the baseline often produces unstable or unsafe planning results, whereas VLADriver-RAG (b) is able to generate a safer and more reliable trajectory. These qualitative results demonstrate that retrieved historical knowledge effectively improves planning robustness and decision stability under uncertain environments.

---

Quick start

Coming soon

---

Citation

@misc{zhao2026vladriverragretrievalaugmentedvisionlanguageactionmodels,
      title={VLADriver-RAG: Retrieval-Augmented Vision-Language-Action Models for Autonomous Driving}, 
      author={Rui Zhao and Haofeng Hu and Zhenhai Gao and Jiaqiao Liu and Gao Fei},
      year={2026},
      eprint={2605.08133},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2605.08133}, 
}