发表机构
Nanjing University; The Hong Kong University of Science and Technology (Guangzhou); Beijing University of Technology（南京大学; 香港科技大学（广州）; 北京工业大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
SafeLoop提出一种非侵入式外部包装器，通过风险预测与回滚恢复机制增强VLA操控安全性，在LIBERO和真实机器人任务中将危险情况减少约70%。
AI 中文摘要
近期视觉-语言-动作（VLA）模型在通用操控任务中展现出巨大潜力，但长时程执行仍然脆弱。微小的状态估计或控制误差可能导致不可逆的失败（如碰撞和物体掉落）。避免这些风险需要一种能够预见危险的主动安全机制。在本文中，我们提出SafeLoop，一种非侵入式外部包装器，在不改变VLA模型参数的情况下，为其添加危险预测和基于回滚的恢复功能。SafeLoop利用视觉和本体感觉训练一个风险预测器，输出四个值：身体碰撞和物体故障的概率及发生时间。随后，一个轻量级控制器根据预测风险从三个动作中选择其一：继续执行（无操作）、保存安全检查点（记录）或在关节空间内回退（回滚）。回滚操作将机器人移回最近的安全路径点，并重新查询基础策略，从而可能获得替代的继续执行方案。在24个LIBERO任务（每个任务16个随机种子）和三个真实机器人任务（每个任务25次试验）中，SafeLoop相比其他方法实现了更强的整体安全-成功权衡，在保持任务成功率和基础策略控制率的同时，将危险情况减少了约70%。项目代码可在以下网址获取：https URL。
英文摘要
Recent vision-language-action (VLA) models are promising for general-purpose manipulation, but long-horizon execution remains fragile. Small state-estimation or control errors can lead to irreversible failures (e.g., collisions and object drops). Avoiding these risks requires a proactive safety mechanism capable of anticipating hazards. In this paper, we introduce SafeLoop, a non-invasive external wrapper that adds hazard prediction and rollback-based recovery to a VLA model without changing its parameters. SafeLoop trains a risk predictor from vision and proprioception to output four values: the probability and time-to-hazard for body collisions and for object failures. A lightweight controller then chooses one of three actions based on the predicted risk: continue execution (noop), save a safety checkpoint (record), or retreat in joint space (rollback). Rollback moves the robot back to a recent safe waypoint and queries the base policy again, which may yield an alternative continuation. Across 24 LIBERO tasks (16 random seeds each) and three real-robot tasks (25 rollouts each), SafeLoop achieves a stronger overall safety-success trade-off than alternative methods, reducing hazard cases by roughly 70% while preserving task success and the base-policy control rate. Project code is available at https://github.com/Loule0-0/SafeLoop/tree/release/safeloop.
Comments8 pages, 7 figures. Accepted to the 2026 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2026)
