{"updated":"2025-01-19T09:40:25.674742+00:00","links":{},"metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00234910","sets":["1164:2735:11468:11669"]},"path":["11669"],"owner":"44499","recid":"234910","title":["世界モデル学習とそれを用いたモデル予測制御の実時間・実環境応用に向けた検討"],"pubdate":{"attribute_name":"公開日","attribute_value":"2024-06-13"},"_buckets":{"deposit":"7e4347d0-170e-4f13-b292-42f13fc5756a"},"_deposit":{"id":"234910","pid":{"type":"depid","value":"234910","revision_id":0},"owners":[44499],"status":"published","created_by":44499},"item_title":"世界モデル学習とそれを用いたモデル予測制御の実時間・実環境応用に向けた検討","author_link":["641042","641035","641036","641037","641034","641038","641040","641043","641039","641041"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"世界モデル学習とそれを用いたモデル予測制御の実時間・実環境応用に向けた検討"},{"subitem_title":"World Model Learning and Model Predictive Control Towards Realtime and Real World Robot Control","subitem_title_language":"en"}]},"item_keyword":{"attribute_name":"キーワード","attribute_value_mlt":[{"subitem_subject":"情報論的学習理論と機械学習3","subitem_subject_scheme":"Other"}]},"item_type_id":"4","publish_date":"2024-06-13","item_4_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"京都大学大学院情報学研究科"},{"subitem_text_value":"ATR脳情報研究所"},{"subitem_text_value":"京都大学大学院情報学研究科"},{"subitem_text_value":"京都大学大学院情報学研究科京都大学大学院情報学研究科 / ATR脳情報研究所"}]},"item_4_text_4":{"attribute_name":"著者所属(英)","attribute_value_mlt":[{"subitem_text_value":"Graduate School of Informatics, Kyoto University","subitem_text_language":"en"},{"subitem_text_value":"ATR Computational Neuroscience Labs","subitem_text_language":"en"},{"subitem_text_value":"Graduate School of Informatics, Kyoto University","subitem_text_language":"en"},{"subitem_text_value":"Graduate School of Informatics, Kyoto University","subitem_text_language":"en"},{"subitem_text_value":"Graduate School of Informatics, Kyoto University / ATR Computational Neuroscience Labs","subitem_text_language":"en"}]},"item_language":{"attribute_name":"言語","attribute_value_mlt":[{"subitem_language":"jpn"}]},"item_publisher":{"attribute_name":"出版者","attribute_value_mlt":[{"subitem_publisher":"情報処理学会","subitem_publisher_language":"ja"}]},"publish_status":"0","weko_shared_id":-1,"item_file_price":{"attribute_name":"Billing file","attribute_type":"file","attribute_value_mlt":[{"url":{"url":"https://ipsj.ixsq.nii.ac.jp/record/234910/files/IPSJ-MPS24148023.pdf","label":"IPSJ-MPS24148023.pdf"},"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-MPS24148023.pdf","filesize":[{"value":"3.0 MB"}],"mimetype":"application/pdf","priceinfo":[{"tax":["include_tax"],"price":"0","billingrole":"17"},{"tax":["include_tax"],"price":"0","billingrole":"44"}],"accessrole":"open_login","version_id":"5d6f4abc-6402-4fa7-8846-e891e0391afd","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 2024 by the Institute of Electronics, Information and Communication Engineers This SIG report is only available to those in membership of the SIG."}]},"item_4_creator_5":{"attribute_name":"著者名","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"竹中, 郁弥"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"山森, 聡"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"森田, 光紀"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"八木, 聡明"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"森本, 淳"}],"nameIdentifiers":[{}]}]},"item_4_creator_6":{"attribute_name":"著者名(英)","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"Fumiya, Takenaka","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Satoshi, Yamamori","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Mitsuki, Morita","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Satoshi, Yagi","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Jun, Morimoto","creatorNameLang":"en"}],"nameIdentifiers":[{}]}]},"item_4_source_id_9":{"attribute_name":"書誌レコードID","attribute_value_mlt":[{"subitem_source_identifier":"AN10505667","subitem_source_identifier_type":"NCID"}]},"item_4_textarea_12":{"attribute_name":"Notice","attribute_value_mlt":[{"subitem_textarea_value":"SIG Technical Reports are nonrefereed and hence may later appear in any journals, conferences, symposia, etc."}]},"item_resource_type":{"attribute_name":"資源タイプ","attribute_value_mlt":[{"resourceuri":"http://purl.org/coar/resource_type/c_18gh","resourcetype":"technical report"}]},"item_4_source_id_11":{"attribute_name":"ISSN","attribute_value_mlt":[{"subitem_source_identifier":"2188-8833","subitem_source_identifier_type":"ISSN"}]},"item_4_description_7":{"attribute_name":"論文抄録","attribute_value_mlt":[{"subitem_description":"近年，ヒューマノイドロボットをはじめとする多自由度ロボットの開発が注目を集めている．その動的な動作生成のため，すべての自由度を考慮した上でモデル予測制御を用いるアプローチが検討されているが，制御器の導出に必要となる計算時間や実環境とのモデル誤差のためにその実システムへの適用は容易ではない．そこで本研究では，強化学習分野で研究が進められてきたタスクに特化した世界モデルの学習および環境乱択化の手法をモデル予測制御に適用することにより，実環境において実時間での動作生成を可能とする枠組みの開発に向けた方法を検討した．具体的には，二脚倒立振子ロボットを対象に，動的なバランス制御をタスクとして検討したアプローチの評価をおこなった．その結果，世界モデルを用いることで計算時間が削減され，環境乱択化を行うことでモデル誤差に頑健な方策が得られることが確認された．","subitem_description_type":"Other"}]},"item_4_description_8":{"attribute_name":"論文抄録(英)","attribute_value_mlt":[{"subitem_description":"In recent years, the development of high-DoF robots, such as humanoid robots, has attracted much attention. In order to generate their dynamic behavior, approaches utilising model predictive control, which takes into account all degrees of freedom, have been considered. However, the application of these techniques to real systems is not straightforward due to the significant computational time required to derive controllers and model errors. In this study, we explore a method to develop a framework that enables real-time movement generation in real environments. Concretely, we propose to combine task-specific world model learning and domain randomization for model predictive control. We evaluated the proposed approach on a bipedal inverted pendulum robot, where dynamic balance control was considered as a target task. The results demonstrated that the utilization of a world model resulted in a reduction in computational time and that the acquired strategy was robust against modeling errors due to the domain randomization.","subitem_description_type":"Other"}]},"item_4_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"4","bibliographic_titles":[{"bibliographic_title":"研究報告数理モデル化と問題解決（MPS）"}],"bibliographicPageStart":"1","bibliographicIssueDates":{"bibliographicIssueDate":"2024-06-13","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"23","bibliographicVolumeNumber":"2024-MPS-148"}]},"relation_version_is_last":true,"weko_creator_id":"44499"},"id":234910,"created":"2025-01-19T01:36:48.156857+00:00"}