BusyBot: Learning to Interact, Reason, and Plan in a BusyBoard Environment
BusyBot: Learning to Interact, Reason, and Plan in a BusyBoard Environment
复制标题
DOI：
10.48550/arxiv.2207.08192
复制
发表时间：
2022-07
期刊：
影响因子：
--
通讯作者：
Zeyi Liu;Zhenjia Xu;Shuran Song
中科院分区：
文献类型：
--
作者：
Zeyi Liu;Zhenjia Xu;Shuran Song
我们介绍BusyBoard，一个玩具启发的机器人学习环境，利用一组不同的铰接对象和对象间的功能关系，为机器人交互提供丰富的视觉反馈。基于这种环境，我们引入了一个学习框架，BusyBot，它允许代理以集成和自我监督的方式联合获得三个基本功能（交互，推理和规划）。利用BusyBoard提供的丰富的感知反馈，BusyBot首先学习与环境有效交互的策略;然后利用策略收集的数据，BusyBot通过因果发现网络推理对象间的功能关系;最后通过结合学习到的交互策略和关系推理技能，智能体能够执行目标条件的操作任务。我们评估BusyBot在模拟和现实世界的环境中，并验证其泛化到看不见的对象和关系。视频可在https://youtu.be/EJ98xBJZ9ek上获得。
We introduce BusyBoard, a toy-inspired robot learning environment that leverages a diverse set of articulated objects and inter-object functional relations to provide rich visual feedback for robot interactions. Based on this environment, we introduce a learning framework, BusyBot, which allows an agent to jointly acquire three fundamental capabilities (interaction, reasoning, and planning) in an integrated and self-supervised manner. With the rich sensory feedback provided by BusyBoard, BusyBot first learns a policy to efficiently interact with the environment; then with data collected using the policy, BusyBot reasons the inter-object functional relations through a causal discovery network; and finally by combining the learned interaction policy and relation reasoning skill, the agent is able to perform goal-conditioned manipulation tasks. We evaluate BusyBot in both simulated and real-world environments, and validate its generalizability to unseen objects and relations. Video is available at https://youtu.be/EJ98xBJZ9ek.