Rapidly Learning Generalizable and Robot-Agnostic Tool-Use Skills for a Wide Range of Tasks.

Rapidly Learning Generalizable and Robot-Agnostic Tool-Use Skills for a Wide Range of Tasks.
复制标题

DOI:
10.3389/frobt.2021.726463
复制
发表时间:
2021
影响因子:
3.4
通讯作者:
Scassellati B
Scassellati B
中科院分区:
其他
文献类型:
--
作者:
Qin M;Brawer J;Scassellati B

文献摘要

参考文献

被引文献

相似文献

许多现实世界的应用需要机器人使用工具。然而,机器人缺乏学习和执行许多基本工具使用任务所需的技能。为此,我们提出了TRansferrIng熟练的工具使用获得快速(TRI-STAR)框架的任务一般的机器人工具的使用。TRI-STAR有三个主要组成部分:1)从最少数量的培训演示中学习并将工具使用技能应用于各种任务的能力,2)将学到的技能推广到其他工具和操纵对象的能力,以及3)将学到的技能转移到其他机器人的能力。这些功能是由TRI-STAR的面向任务的方法实现的,该方法通过使用我们基于目标的任务分类法来识别和利用结构化任务知识。我们展示了这个框架与七个任务,施加不同的要求的工具的使用,其中六个分别在三个物理机器人与不同的运动学配置。我们的研究结果表明,TRI-STAR可以学习有效的工具使用技能,只有20个培训示范。此外,我们的框架将工具使用技能推广到形态上不同的对象,并将它们转移到新的平台,性能略有下降。
Many real-world applications require robots to use tools. However, robots lack the skills necessary to learn and perform many essential tool-use tasks. To this end, we present the TRansferrIng Skilled Tool Use Acquired Rapidly (TRI-STAR) framework for task-general robot tool use. TRI-STAR has three primary components: 1) the ability to learn and apply tool-use skills to a wide variety of tasks from a minimal number of training demonstrations, 2) the ability to generalize learned skills to other tools and manipulated objects, and 3) the ability to transfer learned skills to other robots. These capabilities are enabled by TRI-STAR’s task-oriented approach, which identifies and leverages structural task knowledge through the use of our goal-based task taxonomy. We demonstrate this framework with seven tasks that impose distinct requirements on the usages of the tools, six of which were each performed on three physical robots with varying kinematic configurations. Our results demonstrate that TRI-STAR can learn effective tool-use skills from only 20 training demonstrations. In addition, our framework generalizes tool-use skills to morphologically distinct objects and transfers them to new platforms, with minor performance degradation.
DOI: 10.1177/0278364919872545
发表时间: 2019-08-29
影响因子: 9.2
作者:
Kuan Fang;Zhu, Yuke;Savarese, Silvio
通讯作者: Savarese, Silvio
DOI: 10.1016/j.robot.2014.02.001
发表时间: 2014-06-01
影响因子: 4.3
作者:
Hoffmann, Heiko;Chen, Zhichao;Sinapov, Jivko
通讯作者: Sinapov, Jivko
DOI: 10.1038/s41586-020-2442-2
发表时间: 2020-07-09
期刊: NATURE
影响因子: 64.8
作者:
Burger, Benjamin;Maffettone, Phillip M.;Cooper, Andrew I.
通讯作者: Cooper, Andrew I.
DOI: 10.1109/lra.2021.3062315
发表时间: 2021-04-01
影响因子: 5.2
作者:
Gao, Wei;Tedrake, Russ
通讯作者: Tedrake, Russ
DOI: 10.1177/0278364917713116
发表时间: 2017-07-01
影响因子: 9.2
作者:
Lioutikov, Rudolf;Neumann, Gerhard;Peters, Jan
通讯作者: Peters, Jan