论文/arXiv 2610.14682·18 天前

成功但不安全：用于揭示视觉-语言导航中碰撞的分析基准

针对VLN成功指标掩盖碰撞的问题，提出VLN-Safe双评估机制，并在R2R-CE六个已发表智能体上验证安全差距与碰撞防护。

AI 整理

• 本体 SDK 与控制
• 感知与导航
• 具身模型

原文链接：Successful but Unsafe: An Analytical Benchmark for Revealing Collisions in Vision-and-Language Navigation（arXiv 2610.14682 · 发布日期：2026-10-10）

Successful but Unsafe: An Analytical Benchmark for Revealing Collisions in Vision-and-Language Navigation

针对VLN成功指标掩盖碰撞的问题，提出VLN-Safe双评估机制，并在R2R-CE六个已发表智能体上验证安全差距与碰撞防护。

核心要点

• 现有VLN评价关注到达与效率，滑动机制可能掩盖不安全动作影响。
• VLN-Safe用sliding与blocking两种机制互补分析碰撞安全。
• 在R2R-CE六个智能体上，滑动成功中38.5-56.8%含接触。
• 无训练碰撞防护在blocking下降低碰撞步62-98%，并提升成功率。

验证范围与局限：验证范围为R2R-CE上的六个已发表智能体，以及五类循环、拓扑和视觉-语言-动作智能体的防护评估。

英文摘要 (Abstract)

Vision-and-language navigation (VLN) agents are typically evaluated by whether they reach a target and how efficiently they follow an instructed route, yet these metrics provide little information about collision behavior. Moreover, continuous VLN evaluation commonly uses a sliding regime, where an agent can slide along obstacles after a collision, potentially masking the effect of unsafe actions on navigation performance. We introduce VLN-Safe, a benchmark for analyzing collision safety under two complementary evaluation regimes: sliding, reflecting common VLN evaluation practice, and blocking, where collision-inducing actions prevent forward motion and more directly expose their consequences. Across six published agents on R2R-CE, 38.5-56.8% of successful episodes contain contact under sliding, while switching to blocking decreases success rate by 11.7-36.1 percentage points. We further analyze signals for anticipating imminent collisions and find that depth geometry and recent movement history are more informative than policy-derived signals, while safe alternatives near the policy's preferred action are rarely available. Motivated by these findings, we introduce a training-free collision guard that vetoes risky forward actions, executes an escape maneuver, and returns control to the original policy. Under blocking, the guard reduces collision steps by 62-98% and improves success rate by 6.9-29.3 percentage points across five recurrent, topological, and vision-language-action agents. Under sliding, it leaves success rate unchanged while raising the share of collision-free successes by 4.1-12.4 points. These results reveal a substantial gap between task success and collision-safe navigation and show that lightweight collision-aware intervention can improve both safety and navigation performance.

作者：Hieu D. Pham, Quoc-An Luong, Dung D. Le

来源

本页事实对应的公开页面，以原文为准。

1. arXiv 2610.14682：Successful but Unsafe: An Analytical Benchmark for Revealing Collisions in Vision-and-Language Navigation访问 2026-10-21