arXiv：cs.LG· Stabak Das, Priyesh Ranjan, Xiangfang Li, Lijun Qian·· 3 小时前AI 评分40
RoboGuard 安全监控存在精化缺口：高层机器人计划通过检查仍可能产生不安全执行
Mind the Refinement Gap: When Safe High-Level Robot Plans Produce Unsafe Executions
AI 导读
研究审计了 RoboGuard 的轨迹完整性假设，发现其基于表面计划的 LTL 安全判定与图精化轨迹的判定存在系统性偏差。在 28 个受控案例中，12 个目标抽象案例全部出现预期的表面与精化差异，16 个对照案例表现正常；另有 14 个端到端案例由 SPINE 从自然语言指令生成计划。作者提出基于图的轨迹精化作为轻量级缓解手段和物理 AI 安全监控的诊断工具。
正文
Abstract:Language-enabled robot systems increasingly combine semantic-graph planning with temporal-logic safety monitors. We investigate a trace-completeness assumption in these systems: whether the high-level action sequence checked by a monitor represents the navigation and implicit action effects induced during execution. We audit this assumption in RoboGuard by comparing its verdict on a surface plan with its verdict on a graph-refined trace under the same Linear Temporal Logic (LTL) specification. Our evaluation comprises 28 controlled cases spanning five action-abstraction families and 14 end-to-end cases in which SPINE [1] generates plans from natural-language instructions while RoboGuard generates scene-grounded safety specifications. In the controlled evaluation, all 12 targeted abstraction cases exhibit the predicted surface-versus-refined discrepancy while all 16 controls behave as expected, motivating graph-based trace refinement as a lightweight mitigation and a diagnostic tool for physical-AI safety monitors.
| Comments: | 5 pages, 1 figure |
| Subjects: | Machine Learning (cs.LG); Robotics (cs.RO) |
| Cite as: | arXiv:2610.02662 [cs.LG] |
| (or arXiv:2610.02662v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.02662 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Stabak Das [view email]
[v1]
Fri, 2 Oct 2026 01:31:57 UTC (18 KB)
来源：arXiv：cs.LG · arxiv.org