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

Vision-Language-Action (VLA) models have emerged as a promising paradigm for end-to-end autonomous driving, yet their reliance on implicit parametric knowledge limits generalization in long-tail scenarios. While Retrieval-Augmented Generation (RAG) offers a solution by accessing external expert priors, standard visual retrieval suffers from high latency and semantic ambiguity. To address these challenges, we propose $\mathbf{\text{VLADriver-RAG}}$, a framework that grounds planning in explicit, structure-aware historical knowledge. Specifically, we abstract sensory inputs into spatiotemporal semantic graphs via a $\mathit{\text{Visual-to-Scenario}}$ mechanism, effectively filtering visual noise. To ensure retrieval relevance, we employ a $\mathit{\text{Scenario-Aligned}}\mathit{\text{Embedding}}\mathit{\text{Model}}$ that utilizes Graph-DTW metric alignment to prioritize intrinsic topological consistency over superficial visual similarity. These retrieved priors are then fused within a query-based VLA backbone to synthesize precise, disentangled trajectories. Extensive experiments on the Bench2Drive benchmark establish a new state-of-the-art, achieving a Driving Score of 89.12.