JUICER：用于机器人装配的数据高效模仿学习
JUICER: Data-Efficient Imitation Learning for Robotic Assembly
AI总结：
本文提出JUICER流程，通过结合表达性策略架构与数据集扩展及模拟增强技术，在少量人类演示预算下提升机器人长时程装配任务的模仿学习性能，并在模拟中实现多部件组装，优于基线方法。
摘要：
虽然从演示中学习是获取视觉运动策略的强大方法，但在缺乏大型演示数据集的情况下，对于需要精确、长时程操作的任务，高性能模仿仍然具有挑战性。本文提出了一种在少量人类演示预算下提升模仿学习性能的流程。我们将该方法应用于装配任务，这些任务需要在长时程和多个任务阶段中精确抓取、重新定向和插入多个部件。我们的流程结合了表达性策略架构以及多种数据集扩展和基于模拟的数据增强技术。这些技术有助于扩展数据集支持，并在需要高精度的瓶颈区域附近用局部纠正动作监督模型。我们在模拟环境中的四个家具装配任务上展示了该流程，使机械臂能够直接从RGB图像组装多达五个部件，持续近2500个时间步，优于模仿和数据增强基线。项目网站：https://imitation-juicer.github.io/。
英文摘要：
While learning from demonstrations is powerful for acquiring visuomotor policies, high-performance imitation without large demonstration datasets remains challenging for tasks requiring precise, long-horizon manipulation. This paper proposes a pipeline for improving imitation learning performance with a small human demonstration budget. We apply our approach to assembly tasks that require precisely grasping, reorienting, and inserting multiple parts over long horizons and multiple task phases. Our pipeline combines expressive policy architectures and various techniques for dataset expansion and simulation-based data augmentation. These help expand dataset support and supervise the model with locally corrective actions near bottleneck regions requiring high precision. We demonstrate our pipeline on four furniture assembly tasks in simulation, enabling a manipulator to assemble up to five parts over nearly 2500 time steps directly from RGB images, outperforming imitation and data augmentation baselines. Project website: https://imitation-juicer.github.io/.
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。
