arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2302.08463cs.RO
收藏

基于学习型元控制器的动态抓取

Dynamic Grasping with a Learned Meta-Controller

Columbia University(哥伦比亚大学);University of Pennsylvania(宾夕法尼亚大学)

Yinsen Jia, Jingxi Xu, Dinesh Jayaraman, Shuran Song

AI总结:

提出一种基于强化学习的元控制器动态调整动态抓取流水线的前瞻时间和运动规划时间预算,显著提升了抓取成功率与效率,并展现出对未见障碍物场景的强泛化能力。

摘要:

抓取移动的物体是一项具有挑战性的任务,需要多个子模块,例如物体位姿预测器、机械臂运动规划器等。每个子模块都在其自身的元参数集下运行。例如,位姿预测器应该预测未来多远(即前瞻时间),以及运动规划器规划运动所花费的最大时间(即时间预算)。许多先前的工作为这些参数分配固定值;然而,在单次动态抓取过程中的不同时刻,最优值应根据当前场景而变化。在这项工作中,我们提出了一种带有元控制器的动态抓取流水线,该元控制器动态控制前瞻时间和时间预算。我们通过强化学习以稀疏奖励来学习元控制器。我们的实验表明,与最强基线相比,元控制器提高了抓取成功率(在最杂乱的环境中高达28%),并减少了抓取时间。我们的元控制器学会了推理可达工作空间,并将预测位姿保持在可达区域内。此外,它为运动规划器分配了少量但充足的时间预算。我们的方法可以处理不同的物体、轨迹和障碍物。尽管仅在3-6个随机长方体障碍物下进行训练,我们的元控制器能很好地泛化到7-9个障碍物,以及具有未见障碍物形状的更真实的域外家庭场景设置中。

英文摘要:

Grasping moving objects is a challenging task that requires multiple submodules such as object pose predictor, arm motion planner, etc. Each submodule operates under its own set of meta-parameters. For example, how far the pose predictor should look into the future (i.e., look-ahead time) and the maximum amount of time the motion planner can spend planning a motion (i.e., time budget). Many previous works assign fixed values to these parameters; however, at different moments within a single episode of dynamic grasping, the optimal values should vary depending on the current scene. In this work, we propose a dynamic grasping pipeline with a meta-controller that controls the look-ahead time and time budget dynamically. We learn the meta-controller through reinforcement learning with a sparse reward. Our experiments show the meta-controller improves the grasping success rate (up to 28% in the most cluttered environment) and reduces grasping time, compared to the strongest baseline. Our meta-controller learns to reason about the reachable workspace and maintain the predicted pose within the reachable region. In addition, it assigns a small but sufficient time budget for the motion planner. Our method can handle different objects, trajectories, and obstacles. Despite being trained only with 3-6 random cuboidal obstacles, our meta-controller generalizes well to 7-9 obstacles and more realistic out-of-domain household setups with unseen obstacle shapes.

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。

↑