Relation between Probabilistic Giant Swing Behavior of a Robot and Its Reward Using Reinforcement Learning
-
- HIGASHIURA Takuya
- 横浜国立大学大学院工学府
-
- MATSUMOTO Satoru
- 横浜国立大学大学院工学府
-
- YABUTA Tetsuro
- Yokohama National Univ. Graduate School of Engineering
Bibliographic Information
- Other Title
-
- 確率的なゆらぎを有する強化学習を用いた大車輪ロボットの行動獲得と報酬の関係について
Abstract
We have succeeded in acquiring forward actions to various robot systems using Reinforcement Learning. We have also succeeded in acquiring a giant swing motion as dynamic task by devising its rewards. Although the giant swing robot has a continuous dynamic motion such as its angle and angler velocity, its state of the motion must be divided into discrete states in order to apply the reinforcement learning. Moreover, this giant swing robot system is not under Markov decision process by both control and defective sensation problem. For these reason, this robot shows probabilistic behavior. Then, this paper attempts to clarify the effect of probabilistic behavior of giant swing on the view point of various rewards, whose results are visualized using rotation rate. The results also show that features of the effect of probabilistic behavior are different for each reward.
Journal
-
- TRANSACTIONS OF THE JAPAN SOCIETY OF MECHANICAL ENGINEERS Series C
-
TRANSACTIONS OF THE JAPAN SOCIETY OF MECHANICAL ENGINEERS Series C 79 (807), 4335-4339, 2013
The Japan Society of Mechanical Engineers
- Tweet
Details 詳細情報について
-
- CRID
- 1390282681363710592
-
- NII Article ID
- 130003386490
-
- ISSN
- 18848354
- 03875024
-
- Text Lang
- ja
-
- Data Source
-
- JaLC
- Crossref
- CiNii Articles
-
- Abstract License Flag
- Disallowed