GAM: General affordance-based manipulation for contact-rich object disentangling tasks

GAM: General affordance-based manipulation for contact-rich object disentangling tasks
复制标题

GAM:针对接触丰富的对象解开任务的基于通用可供性的操作

DOI:
10.1016/j.neucom.2024.127386
复制
发表时间:
2024
期刊:
影响因子:
6
通讯作者:
Yang X
Yang X
中科院分区:
计算机科学2区
文献类型:
--
作者:
Yang X

文献摘要

参考文献

相似文献

由于纠缠物体具有丰富的接触动力学,拾取纠缠物体是一项困难的操作任务。大多数现有的解决方案未能产生把握构成,使可靠的操纵,由于对简化的假设的运动策略的依赖。由这些方法生成的抓取倾向于掉落物体或引起非抓取物体的不期望的移动。为了改善这种对象分解任务,我们建议扩展基于强化学习(RL)的启示的概念,以包括任意动作后果,并实现一个通用的基于启示的操纵(GAM)框架。在GAM中,我们训练了一个RL代理,该代理使用更细粒度的动作,并且比以前的方法表现更好,掉落物体和与非抓取钩子接触的机会更小。然后,一个操纵启示预测(MAP)模型进行训练,以估计RL代理的性能。最后,基于操作能力的抓取过滤器(MAGF)选择抓取姿势,提供所需的操作性能,在五个具有挑战性的钩解开任务的模拟显示出显着的改善。实验表明:(1)TAG生成器的局限性,(2)过滤TAG的有效性与预测的操作性能的基础上,一般的启示理论,和(3)的重要性,避免接触非把握对象的接触丰富的操作。
Picking up an entangled object is a difficult manipulation task due to its rich contact dynamics. Most existing solutions fail to produce grasp poses to enable reliable manipulation due to the dependence on simplified assumptions for the motion policies. Grasps generated by these methods tend to drop objects or cause undesired movements of non-grasped objects. To improve such object-disentangling tasks, we propose to extend the concept of reinforcement learning (RL)-based affordance to include arbitrary action consequences and implement a general affordance-based manipulation (GAM) framework. In the GAM, we train an RL agent that uses more fine-grained actions and outperforms previous methods with a smaller chance of dropping objects and making contact with non-grasped hooks. Then, a manipulation affordance prediction (MAP) model is trained to estimate the performances of the RL agent. Finally, the manipulation affordance-based grasp filter (MAGF) selects grasp poses that afford the desired manipulation performances, showing substantial improvements in five challenging hook disentangling tasks in simulation. The experiments show (1) the limitation of TAG generators, (2) the effectiveness of filtering TAGs with predicted manipulation performances based on the general affordance theory, and (3) the importance of avoiding contact with non-grasped objects in contact-rich manipulation.
DOI: 10.1080/01691864.2017.1394912
发表时间: 2017-10
期刊: Advanced Robotics
影响因子: 2
作者:
通讯作者: --
DOI: 10.1146/annurev-control-060117-104848
发表时间: 2018-01-01
期刊: ANNUAL REVIEW OF CONTROL, ROBOTICS, AND AUTONOMOUS SYSTEMS, VOL 1
影响因子: --
作者:
Mason, Matthew T.
通讯作者: Mason, Matthew T.
DOI: --
发表时间: 2020
影响因子: 5.2
作者:
Hongtao Wu;G. Chirikjian
通讯作者: G. Chirikjian
机器人操作和任务在成功衡量标准中的作用
DOI: 10.1038/s42256-019-0078-4
发表时间: 2019
影响因子: 23.8
作者:
V. Ortenzi;M. Controzzi;F. Cini;J. Leitner;M. Bianchi;M. Roa;Peter Corke
通讯作者: Peter Corke
使用深度强化学习在随机箱拣选中分离缠结的工件
DOI: --
发表时间: 2021
期刊: Procedia CIRP
影响因子: --
作者:
Marius Moosmann;Marco Kulig;Felix Spenrath;Manuel Mönnig;Simon Roggendorf;O. Petrovic;Richard Bormann;Marco F. Huber
通讯作者: Marco F. Huber