OSMA-Bench：家庭机器人导航中障碍物条件化安全的基准测试
Benchmarking Obstacle-Conditioned Safety in Household Robot Navigation
AI 导读
研究者推出 OSMA-Bench（障碍物条件化安全运动适应基准），并提出语义敏感指数 SSI，衡量机器人留隙、速度与平滑度是否随障碍物安全等级调整。评测显示 VLM 与 VLA 策略均无法将安全语义差异持续反映到运动中，即使加入显式安全提示也无效。
正文 · 原文
Abstract:Safe household navigation requires robots to regulate clearance and speed according to what an obstacle is and how close it is, without unnecessary conservatism. Yet existing evaluations do not test whether semantic safety distinctions are reflected in continuous motion. To that end, we introduce OSMA-Bench, an Obstacle-Conditioned Safe Motion Adaptation Benchmark built on spatially matched scenario sets that differ only in obstacle type. We also propose the \emph{Semantic Sensitivity Index} (SSI), which measures whether clearance and distance-dependent speed and smoothness follow the caution ordering of obstacles. Evaluating VLM-based and vision-language-action (VLA) policies on the benchmark, we find that neither consistently adapts its motion to obstacle safety, even with explicit safety prompts. Our analysis shows that the policies can correctly distinguish obstacles by safety relevance but fail to reflect this distinction in the motion parameters or actions they produce. For VLM-based policies, which recognize the obstacle and then generate a navigation program, adding an explicit parameter-inference step between these two stages improves adaptation. These findings expose a gap in \emph{semantic-to-motion grounding}, highlighting the need to translate semantic caution into appropriately safe motion.
| Subjects: | Robotics (cs.RO); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV) |
| Cite as: | arXiv:2610.34674 [cs.RO] |
| (or arXiv:2610.34674v1 [cs.RO] for this version) | |
| https://doi.org/10.48550/arXiv.2610.34674 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Hyunwoong Bae [view email]
[v1]
Mon, 26 Oct 2026 01:06:22 UTC (11,524 KB)
来源：arXiv cs.AI · arxiv.org