BUMBLE：基于视觉-语言模型统一推理与行动的整栋建筑级移动操作
BUMBLE: Unifying Reasoning and Acting with Vision-Language Models for Building-wide Mobile Manipulation
AI总结：
针对整栋建筑级长时程移动操作任务，提出集成开放世界感知、多粒度技能与双层记忆的VLM统一框架BUMBLE，在多场景试验中成功率优于基线，用户满意度更高，展现了基础模型的提升潜力。
摘要：
为实现整栋建筑尺度的作业，服务机器人必须执行超长时程的移动操作任务，包括导航至不同房间、前往不同楼层，以及与种类繁多且未见过的日常物品交互。我们将这类任务定义为整栋建筑级移动操作（Building-wide Mobile Manipulation）。为应对这些本质上长时程的任务，我们提出BUMBLE——一个基于视觉-语言模型（VLM）的统一框架，集成了开放世界RGBD感知、从粗到细的广泛运动技能以及双层记忆机制。我们开展了累计90余小时的大量评估，结果显示，在需要按序执行多达12个真值技能、单次试验时长15分钟的长时程整栋建筑任务中，BUMBLE的表现优于多个基线方法。在不同建筑、任务、场景布局下，从不同起始房间和楼层开展的70次试验中，BUMBLE的平均成功率达47.1%。用户研究表明，用户对我们方法的满意度比当前最优的移动操作方法高22%。最后，我们证明了利用能力不断增强的基础模型可进一步提升性能。更多信息请访问https://robin-lab.cs.utexas.edu/BUMBLE/
英文摘要：
To operate at a building scale, service robots must perform very long-horizon mobile manipulation tasks by navigating to different rooms, accessing different floors, and interacting with a wide and unseen range of everyday objects. We refer to these tasks as Building-wide Mobile Manipulation. To tackle these inherently long-horizon tasks, we introduce BUMBLE, a unified Vision-Language Model (VLM)-based framework integrating open-world RGBD perception, a wide spectrum of gross-to-fine motor skills, and dual-layered memory. Our extensive evaluation (90+ hours) indicates that BUMBLE outperforms multiple baselines in long-horizon building-wide tasks that require sequencing up to 12 ground truth skills spanning 15 minutes per trial. BUMBLE achieves 47.1% success rate averaged over 70 trials in different buildings, tasks, and scene layouts from different starting rooms and floors. Our user study demonstrates 22% higher satisfaction with our method than state-of-the-art mobile manipulation methods. Finally, we demonstrate the potential of using increasingly-capable foundation models to push performance further. For more information, see https://robin-lab.cs.utexas.edu/BUMBLE/
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
