1-108 Adaptive Actor-Critic Control with a Model-based Actor and Multi-Step Simulated Experiences :
-
- Syam,Rafiuddin
- Dept. of Advanced Systems Control Engineering, Saga University
-
- Watanabe,Keigo
- Dept. of Advanced Systems Control Engineering, Saga University
-
- Izumi,Kiyotaka
- Dept. of Advanced Systems Control Engineering, Saga University
-
- Kiguchi,Kazuo
- Dept. of Advanced Systems Control Engineering, Saga University
この論文をさがす
抄録
This paper describes a multi-step prediction of actor critic method, as a kind of temporal difference (TD) algorithm. This approach tries to control nonholonomic mobile robot, a tracking trajectory control problem for the position coordinates and azimuth. In our approach the TD-error is embedded from value functions and utility function. Value functions are constructed from the critics that are formulated by radial basis function neural network (RBFNN). The model based actor (MBA) or controller of this algorithm is embedded of a network, that we call it, resolved velocity control network (RVCN). The effectiveness of the method for adaptive control is tested on a nonholonomic mobile robot by some simulations.
収録刊行物
-
- インテリジェントシステム・シンポジウム講演論文集
-
インテリジェントシステム・シンポジウム講演論文集 12 33-38, 2002-11-14
日本機械学会
- Tweet
詳細情報 詳細情報について
-
- CRID
- 1540572720370814080
-
- NII論文ID
- 110002496477
-
- NII書誌ID
- AA1190206X
-
- 本文言語コード
- en
-
- データソース種別
-
- NDL-Digital
- CiNii Articles