跳到正文
arXiv cs.AI· Chen Chen, Omid Bagherieh, Naveen Nikamanth Appiah Balaji, Jesse T. Gonzalez, Negin Heravi, Fahad Islam, Xueting Li, Jean-Baptiste Regli, Kejia Ren, Petr \v{S}vec, David J. Yoon, Matthias M\"uller, Vladlen Koltun·· 2 小时前精选AI 评分71

arXiv 论文提出 AMP1 智能体操控策略,RoboLab-120 零样本成功率 85.6%

On Agentic Manipulation Policies

AI 导读

Chen Chen、Matthias Müller、Vladlen Koltun 等人在 arXiv 发表论文,提出让通用 AI 模型直接指挥机械臂的智能体操控策略 AMP1,在 RoboLab-120 上零样本达到 85.6% 成功率,高于领先 VLA 的 49.2% 和 WAM 的 42.9%。

推荐理由

论文提出通用智能体直接操控机械臂的 AMP1,零样本成绩与真实机器人试验数据齐全,读者可据此比较其与 VLA、WAM 路线的差异。

正文 · 原文

Authors:Chen Chen, Omid Bagherieh, Naveen Nikamanth Appiah Balaji, Jesse T. Gonzalez, Negin Heravi, Fahad Islam, Xueting Li, Jean-Baptiste Regli, Kejia Ren, Petr Švec, David J. Yoon, Matthias Müller, Vladlen Koltun

View PDF HTML (experimental)

Abstract:Recent work on robot manipulation has focused on specialist policies trained on demonstrations. We show that generalist AI models can control manipulators more reliably than specialist policies such as vision-language-action models (VLAs) and world-action models (WAMs). We systematically analyze agentic manipulation policies (AMPs), in which a generalist AI model commands the manipulator directly, adapting its actions as the task unfolds. We present AMP1, a reference harness that runs unchanged across single-arm, bimanual, and mobile manipulation. AMP1 achieves state-of-the-art results on the RoboLab-120 and CaP-Gym benchmarks, with no demonstrations or task-specific training. On RoboLab-120, AMP1 reaches an 85.6% success rate zero-shot, compared to 49.2% and 42.9% for leading VLA and WAM policies. The agent succeeds zero-shot in 104 of 108 real-world trials spanning twelve tasks and three robot instances, reducing failure rates by 20x and 15x relative to strong VLA and WAM baselines. The basic harness supplies no grasping, pushing, or placing skills; the agent develops its own strategies, such as searching for hidden objects, using objects as tools, and calibrating its cameras. With the addition of a persistent memory store, the agent can also learn from experience: an augmented harness with a memory representation that is carried across episodes lowers the failure rate on RoboDojo bimanual manipulation tasks by 4.6x relative to zero-shot deployment. Our analysis suggests that AMPs are substantially more general and robust than specialist manipulation policies, providing a promising foundation for robot manipulation. See this https URL for code and videos.
Comments: Code and videos at: this https URL
Subjects: Robotics (cs.RO); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
Cite as: arXiv:2611.09892 [cs.RO]
  (or arXiv:2611.09892v3 [cs.RO] for this version)
  https://doi.org/10.48550/arXiv.2611.09892

arXiv-issued DOI via DataCite

Submission history

From: Matthias Müller [view email]
[v1] Fri, 6 Nov 2026 18:59:34 UTC (24,323 KB)
[v2] Wed, 18 Nov 2026 17:12:53 UTC (24,188 KB)
[v3] Thu, 19 Nov 2026 17:48:05 UTC (24,188 KB)

来源:arXiv cs.AI · arxiv.org