Art

J-GLOBAL ID：200902245028276324   Reference number：04A0184447

Learning a dynamic policy by using policy gradient: Application to biped walking

方策勾配法を用いた動的行動則の獲得:2足歩行運動への適用

Author (5)： MATSUBARA TAKAMITSU (奈良先端科学技術大学院大) ,  MORIMOTO JUN (ATR脳情報研) ,  NAKANISHI JUN (ATR脳情報研) ,  SATO MASAAKI (ATR脳情報研) ,  DOYA KENJI (ATR脳情報研)
Material： 電子情報通信学会技術研究報告  (IEICE Technical Report (Institute of Electronics, Information and Communication Engineers))
Volume： 103  Issue： 602(NC2003 119-134)  Page： 53-58  Publication year： Jan. 27, 2004 
JST Material Number： S0532B  ISSN： 0913-5685  Document type： Proceedings
Article type： 原著論文  Country of issue： Japan (JPN)  Language： JAPANESE (JA)

Thesaurus term： *bipedal locomotion ,...
Semi thesaurus term： 方策勾配法 ,...

   To see more with JDream III (charged).   {{ this.onShowAbsJLink("http://jdream3.com/lp/jglobal/index.html?docNo=04A0184447&from=J-GLOBAL&jstjournalNo=S0532B") }}

JST classification (2)： Motion and control of robots  (IC04012J) ,  Artificial intelligence  (JE08000Z)

Reference (21)：

• BENBRAHIM, H. Biped dynamic walking using reinforcement learning. 1997, 22, 283-302
• DOYA, K. Reinforcement learning in continuous time and space. Neural Computation. 2000, 12, 219-245
• HIRAI, K. The development of honda humanoid robot. IEEE International Conference on Robotics and Automation, 1998. 1998, 1321-1326
• KAGAMI, S. AutoBalancer : An Online Dynamic Balance Compensation Scheme for Humanoid Robots. Algorithmic and Computational Robotics : New Directions. 2001, 329-340
• KIMURA, H. An analysis of actor/critic algorithms using eligibility traces : Reinforcement learning with imperfect value functions. Proceedings of the 15th Int. Conf. on Machine Learning, 1998. 1998, 284-292

Terms in the title (5)： 方策勾配法 ,  行動 ,  獲得 ,  2足歩行 ,  運動