In collaborative human-robot manipulation, a robot must predict human intents and adapt its actions accordingly to smoothly execute tasks. However, the human's intent in turn depends on actions the robot takes, creating a chicken-or-egg problem. Prior methods ignore such inter-dependency and instead train marginal intent prediction models independent of robot actions. This is because training conditional models is hard given a lack of paired human-robot interaction datasets. Can we instead leverage large-scale human-human interaction data that is more easily accessible? Our key insight is to exploit a correspondence between human and robot actions that enables transfer learning from human-human to human-robot data. We propose a novel architecture, InteRACT, that pre-trains a conditional intent prediction model on large human-human datasets and fine-tunes on a small human-robot dataset. We evaluate on a set of real-world collaborative human-robot manipulation tasks and show that our conditional model improves over various marginal baselines. We also introduce new techniques to tele-operate a 7-DoF robot arm and collect a diverse range of human-robot collaborative manipulation data, which we open-source.

翻译：在人机协同操作中，机器人必须预测人类意图并据此调整自身动作以流畅执行任务。然而，人类的意图又依赖于机器人采取的动作，形成了"先有鸡还是先有蛋"的相互依赖问题。现有方法忽视了这种相互依赖关系，转而训练独立于机器人动作的边际意图预测模型。这是因为在缺乏配对的人机交互数据集情况下，训练条件模型十分困难。我们能否转而利用更易获取的大规模人人交互数据？我们的关键洞察在于发掘人类与机器人动作之间的对应关系，从而实现从人人数据到人机数据的迁移学习。我们提出名为InteRACT的新型架构，该架构在大型人人数据集上预训练条件意图预测模型，并在小型人机数据集上进行微调。通过在实际人机协同操作任务中的评估，我们证明条件模型相较于各种边际基准模型具有显著优势。我们还引入了遥操作7自由度机器人手臂的新技术，并采集了多样化的协同操作数据，这些数据将以开源形式发布。