Approximation Bayesian Reinforcement Learning based on Estimation of Plant Variation and its Application to Peg-in-Hole Task

• Senda Kei

  Graduate School of Engineering, Kyoto University

• Hishinuma Toru

  Graduate School of Engineering, Kyoto University

• Tani Yurika

  Graduate School of Engineering, Kyoto University

Bibliographic Information

Other Title

• プラント変動の推定に基づく近似ベイジアン強化学習とペグ・イン・ホール・タスクへの適用
• プラント ヘンドウ ノ スイテイ ニ モトズク キンジ ベイジアン キョウカ ガクシュウ ト ペグ ・ イン ・ ホール ・ タスク エ ノ テキヨウ

Abstract

In a general reinforcement learning problem, a plant, i.e. state transition probabilities, is estimated, and a learning policy for the estimated plant is applied to a real plant. If there is a difference between the estimated plant and the real plant, the obtained policy may not work well for the real plant. In this study, the real plant variation is parameterized by an interpolation of several estimated plants. This study proposes a reinforcement learning method based on estimation of parameter variation, and applies this method to 2-dimensional Peg-in-Hole Task. The effectiveness of the proposed method is demonstrated by numerical and experimental results.

Journal

• Transactions of the Institute of Systems, Control and Information Engineers

  Transactions of the Institute of Systems, Control and Information Engineers 29 (3), 122-129, 2016

  THE INSTITUTE OF SYSTEMS, CONTROL AND INFORMATION ENGINEERS (ISCIE)

Keywords

• reinforcement learning
• plant variation

Details

• CRID: 1390001205167170560
• NII Article ID: 130005157767
• NII Book ID: AN1013280X
• DOI: 10.5687/iscie.29.122
• ISSN: 2185811X
• ISSN: 13425668
• HANDLE: 2433/227087
• NDL BIB ID: 027191061
• Web Site: https://ndlsearch.ndl.go.jp/books/R000000004-I027191061
• Web Site: https://www.jstage.jst.go.jp/article/iscie/29/3/29_122/_pdf
• Text Lang: ja
• Data Source:
  • JaLC
  • IRDB
  • NDL
  • Crossref
  • CiNii Articles
  • KAKEN
• Abstract License Flag: Disallowed