构建首个面向人机协作的机器人操作数据集，让机器人学会与人协同工作。
HABIT: Human-Aware Behavior and Interaction Training Dataset for Robot Manipulation
- 采集60项任务、超10,000个回合，覆盖协作、共事、监督三类人机交互模式
- 在协作任务中实现时空同步，在监督任务中理解手势指令
- 适合研究人机协作、具身智能的科研人员和工程师
大规模示范数据集推动了通用机器人策略的发展。然而，现有数据集均在无人环境下采集，仅基于此类数据训练的策略虽能独立完成任务，却缺乏对人类行为的感知。为此，我们提出HABIT，一个面向人机共存环境的大规模机器人示范数据集。将任务分为三类角色：协作者（共同完成任务）、同事（共享空间中各自执行任务）和监督者（指导机器人）。数据集包含超过10,000个回合、超过160小时的视频，涵盖60项任务。实验表明，基于人存在数据训练可催生机器人仅用无人数据无法实现的人类意识行为：协作任务中的时空同步、共事任务中的主动让位，以及监督任务中的手势理解。此外，使用HABIT训练的模型可快速适应新的人机交互任务。通过引入人类参与作为数据多样性的新维度，HABIT使机器人策略拓展至与人类共享的复杂环境。
原文摘要 · Abstract (English)
Large-scale demonstration datasets have been central to recent progress in general-purpose robot policies. However, existing datasets are collected in human-absent settings, and policies trained on such data may perform tasks competently in isolation but fail to exhibit human-aware behaviors. To address this gap, we introduce HABIT, a large-scale robot demonstration dataset for human-present environments. We organize tasks into three roles capturing distinct modes of human-robot interaction: Collaborator, where human and robot jointly accomplish a task; Coworker, where they pursue separate tasks in a shared space; and Supervisor, where the human directs the robot. The dataset comprises over 10K episodes and over 160 hours across 60 tasks. Our experiments show that training on human-present data elicits human-aware behaviors that robot-only data fails to produce: spatiotemporal synchronization in Collaborator tasks, yielding in Coworker tasks, and gesture grounding in Supervisor tasks. Moreover, training on HABIT enables rapid adaptation to new human-robot interaction tasks. By introducing human presence as a new axis of dataset diversity, HABIT extends robot policies to environments shared with humans.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。