论文arXiv 2611.02871今天

ReGATE控制器实现Unitree G1视觉中断下85%抓取成功率

ReGATE全身控制器通过循环门控融合本体感知与深度数据，无需外部物体位姿估计。仿真静态/动态抓取成功率达85.2%/85.9%，在Unitree G1实机验证了视觉遮挡下的鲁棒操作能力。

AI 整理

本体 SDK 与控制
感知与导航
具身模型

本文由 AI 根据公开内容整理，可能有误，请以原文为准：ReGATE: Recurrent Gated Visual Fusion for Reactive Humanoid-Object Interaction（arXiv 2611.02871）

核心内容

针对人形机器人移动操作中视觉观测不完整难题，研究提出ReGATE全身控制器。该方法基于Teacher-Student框架，利用循环适配器预测物体几何与接触状态，摆脱对外部物体位姿估计器的依赖。仿真测试中，ReGATE在330次静态物体获取试验中成功率达85.2%，显著优于VisualMimic的48.5%；在440次动态物体试验中成功率达85.9%，确立了其在非结构化环境下的性能优势。

系统架构采用Head-wise与Global Gates调节视觉更新至本体感知通路，结合视觉接地与数据损坏课程训练适配器，并通过相位感知初始化提升短暂接触转换覆盖率。该方案已在Unitree G1人形机器人上完成部署，成功执行指令化抓取-运输-释放序列及侧向位移箱体获取。硬件实验进一步证实，即使在初始观测后、接触指令前相机被完全遮挡的极端工况下，机器人仍能依靠记忆与本体反馈完成稳定抓取。

关键看点

核心指标：仿真静态/动态物体获取成功率分别达85.2%和85.9%，较VisualMimic基线提升约37个百分点
技术机制：采用循环门控视觉融合架构，通过Head-wise与Global Gates自适应调节视觉与本体感知权重
鲁棒验证：Unitree G1实机验证在相机遮挡中断视觉输入后，仍可依赖内部状态估计完成箱体抓取任务
训练策略：引入视觉接地与数据损坏课程学习，配合相位感知初始化以覆盖短暂接触转换的关键窗口期

工程与产业观察

ReGATE证明了在人形机器人操作中，通过循环门控机制深度融合本体感知与视觉历史，可有效替代高带宽实时视觉流。这为解决具身智能在遮挡、光照剧变等长尾场景下的操作失效提供了工程范式，降低了对外部高精度感知系统的刚性依赖，对提升人形机器人非结构化环境作业鲁棒性具有重要参考价值。

英文摘要 (Abstract)

Humanoid loco-manipulation requires coordinating object acquisition, transport, and release from onboard sensing while responding to user commands and incomplete visual observations. We present ReGATE, a whole-body controller that combines egocentric depth, proprioception, and online locomotion and contact commands. Built on a teacher-student framework, ReGATE integrates interaction-state estimation through a recurrent adapter that predicts object geometry and physical contact without an external object-pose estimator. Head-wise and global gates regulate visual updates to a proprioceptive pathway. A visual-grounding and corruption curriculum trains the adapter, while phase-aware initialization increases coverage of brief contact transitions. In simulation, ReGATE achieves 85.2% success across 330 static-object acquisition trials, compared with 48.5% for VisualMimic under the same protocol, and 85.9% across 440 moving-object acquisition trials. Deployment on a Unitree G1 demonstrates commanded grasp-transport-release sequences and acquisition of laterally displaced boxes. A qualitative hardware experiment further demonstrates successful box acquisition when the camera is occluded after an initial observation and before the contact command. To the best of our knowledge, this is the first work to extend recurrent gated fusion of onboard depth and proprioception from perceptive locomotion to humanoid loco-manipulation under interrupted visual sensing.

作者：Hao Zhu, Jorge Pe~na-Queralta