Presentation 2013-01-24
Control of the falling cat motion by using path-integral reinforcement learning
Daichi NAKANO, Shin-ichi MAEDA, Shin ISHII,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) The falling-cat motion is a motion for controlling the cat's posture under no existence of external force. To obtain a controller that emits such a motion is of practice, because of its applicability to controlling problems of various systems including space robots. It is not an easy task, however, because the cat's posture control is a nonholonomic system to which general state feedback control is hard to be applied. In this study, we propose an approach to use path integral reinforcement learning for identifying a control policy that is dependent on time but independent of system state. We will show that the proposed method can acquire such a control policy that achieves a goal state (a goal posture) of a dynamical system which models the cat's posture control, in an efficient manner.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Nonholonomic system / falling-cat motion / path integral reinforcement learning
Paper # NLP2012-107,NC2012-97
Date of Issue

Conference Information
Committee NLP
Conference Date 2013/1/17(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Nonlinear Problems (NLP)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Control of the falling cat motion by using path-integral reinforcement learning
Sub Title (in English)
Keyword(1) Nonholonomic system
Keyword(2) falling-cat motion
Keyword(3) path integral reinforcement learning
1st Author's Name Daichi NAKANO
1st Author's Affiliation Department of System Science, Graduate School of Informatics, Kyoto University()
2nd Author's Name Shin-ichi MAEDA
2nd Author's Affiliation Department of System Science, Graduate School of Informatics, Kyoto University
3rd Author's Name Shin ISHII
3rd Author's Affiliation Department of System Science, Graduate School of Informatics, Kyoto University
Date 2013-01-24
Paper # NLP2012-107,NC2012-97
Volume (vol) vol.112
Number (no) 389
Page pp.pp.-
#Pages 6
Date of Issue