AutoVLA: A Vision-Language-Action Model for End-to-End Autonomous Driving with Adaptive Reasoning and Reinforcement Fine-Tuning

TL;DR

AutoVLA is an innovative end-to-end autonomous driving model that integrates reasoning and action generation using a unified autoregressive approach, employing adaptive reasoning and reinforcement fine-tuning for improved performance.

Contribution

The paper introduces AutoVLA, a novel vision-language-action model that unifies reasoning and action generation, and incorporates reinforcement fine-tuning for adaptive, efficient autonomous driving.

Findings

AutoVLA achieves competitive results on nuPlan, nuScenes, Waymo, and CARLA datasets.

The model demonstrates effective adaptive reasoning in diverse driving scenarios.

Reinforcement fine-tuning reduces unnecessary reasoning in simple cases.

Abstract

Recent advancements in Vision-Language-Action (VLA) models have shown promise for end-to-end autonomous driving by leveraging world knowledge and reasoning capabilities. However, current VLA models often struggle with physically infeasible action outputs, complex model structures, or unnecessarily long reasoning. In this paper, we propose AutoVLA, a novel VLA model that unifies reasoning and action generation within a single autoregressive generation model for end-to-end autonomous driving. AutoVLA performs semantic reasoning and trajectory planning directly from raw visual inputs and language instructions. We tokenize continuous trajectories into discrete, feasible actions, enabling direct integration into the language model. For training, we employ supervised fine-tuning to equip the model with dual thinking modes: fast thinking (trajectory-only) and slow thinking (enhanced with chain-of-thought reasoning). To further enhance planning performance and efficiency, we introduce a reinforcement fine-tuning method based on Group Relative Policy Optimization (GRPO), reducing unnecessary reasoning in straightforward scenarios. Extensive experiments across real-world and simulated datasets and benchmarks, including nuPlan, nuScenes, Waymo, and CARLA, demonstrate the competitive performance of AutoVLA in both open-loop and closed-loop settings. Qualitative results showcase the adaptive reasoning and accurate planning capabilities of AutoVLA in diverse scenarios.