The BesMan Learning Platform for Automated Robot Skill Learning

被引：10

作者：

Gutzeit, Lisa ^{[1
]}

Fabisch, Alexander ^{[2
]}

Otto, Marc ^{[1
]}

Metzen, Jan Hendrik ^{[3
]}

Hansen, Jonas ^{[1
]}

Kirchner, Frank ^{[1
,2
]}

Kirchner, Elsa Andrea ^{[1
,2
]}

机构：

[1] Univ Bremen, Robot Res Grp, Bremen, Germany

[2] DFKI GmbH, Robot Innovat Ctr, Bremen, Germany

[3] Robert Bosch GmbH, Bosch Ctr Artificial Intelligence, Renningen, Germany

来源：

FRONTIERS IN ROBOTICS AND AI | 2018年 / 5卷

关键词：

robotics; imitiation learning; reinforcement learning; manipulation; behavior segmentation;

D O I：

10.3389/frobt.2018.00043

中图分类号：

TP24 [机器人技术];

学科分类号：

080202 ; 1405 ;

摘要：

We describe the BesMan learning platform which allows learning robotic manipulation behavior. It is a stand-alone solution which can be combined with different robotic systems and applications. Behavior that is adaptive to task changes and different target platforms can be learned to solve unforeseen challenges and tasks, which can occur during deployment of a robot. The learning platform is composed of components that deal with preprocessing of human demonstrations, segmenting the demonstrated behavior into basic building blocks, imitation, refinement by means of reinforcement learning, and generalization to related tasks. The core components are evaluated in an empirical study with 10 participants with respect to automation level and time requirements. We show that most of the required steps for transferring skills from humans to robots can be automated and all steps can be performed in reasonable time allowing to apply the learning platform on demand.

引用

页数：14

共 57 条

[1] A Shift in Task Routines During the Learning of a Motor Skill: Group-Averaged Data May Mask Critical Phases in the Individuals' Acquisition of Skilled Performance [J].