SAFE：视觉-语言-动作模型的多任务故障检测

SAFE: Multitask Failure Detection for Vision-Language-Action Models

June 11, 2025

作者: Qiao Gu, Yuanliang Ju, Shengxiang Sun, Igor Gilitschenski, Haruki Nishimura, Masha Itkina, Florian Shkurti

cs.AI

摘要

尽管视觉-语言-动作模型（VLAs）在多种操作任务中展现出了颇具前景的机器人行为，但在面对全新任务时，其开箱即用的成功率仍显有限。为确保这些策略能够安全地与环境互动，我们需要一种故障检测器，它能及时发出警报，使机器人能够停止、回退或寻求帮助。然而，现有的故障检测器仅针对一项或少数特定任务进行训练与测试，而VLAs则要求检测器能够泛化，并在未见过的任务和新环境中也能有效识别故障。本文首次提出了多任务故障检测问题，并推出了SAFE——一款专为通用型机器人策略（如VLAs）设计的故障检测器。我们深入分析了VLA的特征空间，发现VLAs对任务成功与失败具备足够的高层次认知，这种认知在不同任务间具有通用性。基于这一洞察，我们设计了SAFE，使其能够从VLA内部特征中学习，并预测一个指示任务失败可能性的单一标量。SAFE在成功与失败的执行轨迹上均接受训练，并在未见任务上进行评估。SAFE兼容多种策略架构，我们已在OpenVLA、pi_0及pi_0-FAST上，在仿真与真实环境中进行了广泛测试。通过与多种基线方法对比，我们展示了SAFE在故障检测性能上达到了业界领先水平，并利用保形预测实现了准确性与检测时间的最佳平衡。更多定性结果请访问https://vla-safe.github.io/。

English

While vision-language-action models (VLAs) have shown promising robotic behaviors across a diverse set of manipulation tasks, they achieve limited success rates when deployed on novel tasks out-of-the-box. To allow these policies to safely interact with their environments, we need a failure detector that gives a timely alert such that the robot can stop, backtrack, or ask for help. However, existing failure detectors are trained and tested only on one or a few specific tasks, while VLAs require the detector to generalize and detect failures also in unseen tasks and novel environments. In this paper, we introduce the multitask failure detection problem and propose SAFE, a failure detector for generalist robot policies such as VLAs. We analyze the VLA feature space and find that VLAs have sufficient high-level knowledge about task success and failure, which is generic across different tasks. Based on this insight, we design SAFE to learn from VLA internal features and predict a single scalar indicating the likelihood of task failure. SAFE is trained on both successful and failed rollouts, and is evaluated on unseen tasks. SAFE is compatible with different policy architectures. We test it on OpenVLA, pi_0, and pi_0-FAST in both simulated and real-world environments extensively. We compare SAFE with diverse baselines and show that SAFE achieves state-of-the-art failure detection performance and the best trade-off between accuracy and detection time using conformal prediction. More qualitative results can be found at https://vla-safe.github.io/.