paper-with-me

홈 › Papers

DynVLA: Learning World Dynamics for Action Reasoning in Autonomous Driving

2026-03-11 · Shuyao Shang, Bing Zhan, Yunfei Yan, Yuqi Wang, Yingyan Li, Yasong An, Xiaoman Wang, Jierui Liu, Lu Hou, Lue Fan, Zhaoxiang Zhang, Tieniu Tan arxiv

We propose DynVLA, a driving VLA model that introduces a new CoT paradigm termed Dynamics CoT. DynVLA forecasts compact world dynamics before action generation, enabling more informed and physically grounded decision-making. To obtain compact dynamics representations, DynVLA introduces a Dynamics Tokenizer that compresses future evolution into a small set of dynamics tokens. Considering the rich environment dynamics in interaction-intensive driving scenarios, DynVLA decouples ego-centric and environment-centric dynamics, yielding more accurate world dynamics modeling. We then train DynVLA to generate dynamics tokens before actions through SFT and RFT, improving decision quality while maintaining latency-efficient inference. Compared to Textual CoT, which lacks fine-grained spatiotemporal understanding, and Visual CoT, which introduces substantial redundancy due to dense image prediction, Dynamics CoT captures the evolution of the world in a compact, interpretable, and efficient form. Extensive experiments on NAVSIM, Bench2Drive, and a large-scale in-house dataset demonstrate that DynVLA consistently outperforms Textual CoT and Visual CoT methods, validating the effectiveness and practical value of Dynamics CoT. Project Page: https://yaoyao-jpg.github.io/dynvla.

📄 PDF Abstract BibTeX arXiv:2603.11041

Code (0)

등록된 구현이 없습니다.

Tasks

Autonomous Driving

Similar Papers 제목 키워드 기반

BrainWAM: Action-Space Coordination of Semantic Priors and Predictive Dynamics for Autonomous Driving

2026-08-13 · Bing Zhan, Shuyao Shang, Jiahao Gu, Shuo Lu 외 arxiv

Autonomous driving requires planning under both semantic constraints and predictive dynamics. Existing end-to-end driving approaches, however, typically emphasize only one side of this requirement: Vision-Language-Action…

Autonomous Driving

RADAR: Benchmarking Vision-Language-Action Generalization via Real-World Dynamics, Spatial-Physical Intelligence, and Autonomous Evaluation

2026-02-11 · Yuhao Chen, Zhihao Zhan, Xiaoxin Lin, Zijian Song 외 arxiv

VLA models have achieved remarkable progress in embodied intelligence; however, their evaluation remains largely confined to simulations or highly constrained real-world settings. This mismatch creates a substantial real…

Spatial Reasoning

Improving Adversarial Transferability in MLLMs via Dynamic Vision-Language Alignment Attack

2025-02-27 · Chenhe Gu, Jindong Gu, Andong Hua, Yao Qin

Multimodal Large Language Models (MLLMs), built upon LLMs, have recently gained attention for their capabilities in image recognition and understanding. However, while MLLMs are vulnerable to adversarial attacks, the tra…

WCog-VLA: A Dual-Level World-Cognitive Vision-Language-Action Model for End-to-End Autonomous Driving

2026-07-09 · Xuerun Yan, Zhexi Lian, Nuoheng Zhang, Shiyu Fang 외 arxiv

Vision-Language-Action (VLA) models have advanced end-to-end autonomous driving. However, existing methods either lack comprehensive world cognition or suffer from fragmented world foresight, inherently confining these m…

Autonomous Driving

CoWorld-VLA: Thinking in a Multi-Expert World Model for Autonomous Driving

2026-05-11 · Minqing Huang, Yujiao Xiang, Zihan Liang, Jiajie Huang 외 arxiv

Vision-Language-Action (VLA) models have emerged as a promising paradigm for end-to-end autonomous driving. However, existing reasoning mechanisms still struggle to provide planning-oriented intermediate representations:…

Collision AvoidanceAutonomous DrivingScene Generation