具身开发者社区

外观

论文arXiv 2610.2060914 天前

SafeDriveVLA：用于冲突感知端到端自动驾驶的导航条件化世界模型 Dreaming ​

提出SafeDriveVLA，用导航条件化世界模型dreaming与合规模式预测，在Bench2Drive验证冲突感知端到端驾驶。

AI 整理
  • 感知与导航
  • 数据与采集
  • 具身模型

原文链接：SafeDriveVLA: Navigation-Conditioned World Model Dreaming for Conflict-Aware End-to-End Autonomous Driving（arXiv 2610.20609 · 发布日期：2026-10-14）

SafeDriveVLA: Navigation-Conditioned World Model Dreaming for Conflict-Aware End-to-End Autonomous Driving

提出SafeDriveVLA，用导航条件化世界模型dreaming与合规模式预测，在Bench2Drive验证冲突感知端到端驾驶。

核心要点 ​

  • 开放自然语言导航接口可能接收粗心或恶意的危险指令，带来新的安全风险。
  • 构建同时评估导航跟随、冲突感知与对抗鲁棒性的首个基准套件。
  • 方法显式预测三级导航合规模式，并用严格跟随导航信号的世界模型dreaming辅助判断。
  • 在Bench2Drive上达到88.16 DS和68.64% SR，并较最佳SimLingo变体降低不安全指令下碰撞。

验证范围与局限：验证范围为Bench2Drive闭环语言跟随与冲突/对抗场景，并提供项目页视频演示。

英文摘要 (Abstract) ​

Vision-Language-Action (VLA) models are gaining attention in autonomous driving, inheriting commonsense priors from large language models (LLMs) and widening the navigation interface from a fixed command set to free-form natural-language instructions. However, this open-ended interface exposes a new vulnerability surface: an inattentive driver may issue instructions that are unsafe given the current scene, and a malicious user may issue dangerous ones on purpose. We present the first benchmark suite that evaluates navigation following and conflict awareness together, along with adversarial robustness. Existing VLAs fall short on both: SimLingo completes no language-instructed lane change in closed loop, and all models degrade sharply under conflicting signals. We propose SafeDriveVLA, the first driving VLA that (1) explicitly predicts a three-level navigation-compliance mode to decouple conflict reasoning from action generation, and (2) uses navigation-conditioned world model dreaming to inform this prediction. Unlike prior world models that roll out expert or planned actions, ours dreams of strictly following the navigation signal, showing when following is safe. SafeDriveVLA reaches 88.16 Driving Score (DS) and 68.64% Success Rate (SR) on Bench2Drive, lifts closed-loop language following by 49.3% and reduces collisions under unsafe instructions from 60 to 38 compared to the best SimLingo variant. Video demonstrations are available on our project page.

作者：Shaoyuan Xie, Zihan Zhang, Jingxuan Wang, Jiashu Qu, Xiaoqing Liang, Lingdong Kong, Junchi Lu, Henrik I Christensen, Qi Alfred Chen

来源

本页事实对应的公开页面，以原文为准。

  1. arXiv 2610.20609：SafeDriveVLA: Navigation-Conditioned World Model Dreaming for Conflict-Aware End-to-End Autonomous Driving访问 2026-10-21