Data augmentation methods of dynamic model identification for harbor maneuvers using feedforward neural network

Wakita, Kouki; Miyauchi, Yoshiki; Akimoto, Youhei; Maki, Atsuo

doi:https://doi.org/10.1007/s00773-024-01036-w

Data augmentation methods of dynamic model identification for harbor maneuvers using feedforward neural network

Wakita, Kouki; Miyauchi, Yoshiki; Akimoto, Youhei et al.

Journal of Marine Science and Technology, 2024

アクセス数:312件（2025-07-26 22:23 集計）

固定URL: https://hdl.handle.net/11094/100126

閲覧可能ファイル

ファイル	フォーマット	利用条件	サイズ	閲覧回数	利用開始日	説明	information
JMarineSciTechnol_s00773-024-01036-w	pdf	なし	30.8 MB	22	2025.01.28

Creative Commons : 表示

論文情報

ファイル出力 EndNote Basic出力 Mendeley出力

タイトル	Data augmentation methods of dynamic model identification for harbor maneuvers using feedforward neural network
著者	ORCID 0000-0003-0363-1396 0000-0003-0363-1396 著者名 Wakita, Kouki
	著者名 Miyauchi, Yoshiki
	NRID 1000020709654 1000020709654 著者名 Akimoto, Youhei
	NRID 1000050556496 1000050556496 ORCID 0000-0002-2819-1297 0000-0002-2819-1297 researchmap permalink atsuo_maki atsuo_maki Scopus Author ID 24503751100 24503751100 著者名 Maki, Atsuo
キーワード等	Jittering
	Maneuvering model
	Neural network
	Slicing
内容	The version of record of this article, first published in Journal of Marine Science and Technology, is available online at Publisher’s website: https://doi.org/10.1007/s00773-024-01036-w.
抄録	A dynamic model for an automatic berthing and unberthing controller has to estimate harbor maneuvers, which include berthing, unberthing, approaching maneuvers to berths, and entering and leaving the port. When the dynamic model is estimated by the system identification using feedforward neural networks, a large number of tests or trials are required to measure the various motions of harbor maneuvers. However, the amount of data that can be obtained is limited due to the high costs and time-consuming nature of full-scale ship trials. Therefore, this paper introduces data augmentation to improve the generalization performance of dynamic models identified from a limited dataset. This study used slicing and jittering as data augmentation methods and confirmed their effectiveness by numerical experiments using the free-running model tests. Results of numerical experiments demonstrated that slicing and jittering are effective data augmentation methods but could not improve generalization performance for extrapolation states of the original dataset.
出版者	Springer
収録物名	Journal of Marine Science and Technology
刊行年月	2024-11-20
言語	英語
ハンドルURL	https://hdl.handle.net/11094/100126
PISSN	09484280
関連情報 (isIdenticalTo)	DOI (出版社版) https://doi.org/10.1007/s00773-024-01036-w
アクセス権	open access
権利情報	This article is licensed under a Creative Commons Attribution 4.0 International License.
出版タイプ	Version of Record
カテゴリ	学術雑誌論文 Journal Article

資源タイプ	学術雑誌論文
ローカル資源タイプ	学術雑誌論文
DCTERMS.bibliographicCitation	Journal of Marine Science and Technology. 2024
DC.title	Data augmentation methods of dynamic model identification for harbor maneuvers using feedforward neural network
DC.creator	Wakita, Kouki
	Miyauchi, Yoshiki
	Akimoto, Youhei
	Maki, Atsuo
DC.publisher	Springer
DC.language' scheme='DCTERMS.RFC1766	英語
DCTERMS.issued' scheme='DCTERMS.W3CDTF	2024-11-20
DC.identifier	https://doi.org/10.1007/s00773-024-01036-w
DC.subject	Jittering
	Maneuvering model
	Neural network
	Slicing
DC.description	The version of record of this article, first published in Journal of Marine Science and Technology, is available online at Publisher’s website: https://doi.org/10.1007/s00773-024-01036-w.
DCTERMS.abstract	A dynamic model for an automatic berthing and unberthing controller has to estimate harbor maneuvers, which include berthing, unberthing, approaching maneuvers to berths, and entering and leaving the port. When the dynamic model is estimated by the system identification using feedforward neural networks, a large number of tests or trials are required to measure the various motions of harbor maneuvers. However, the amount of data that can be obtained is limited due to the high costs and time-consuming nature of full-scale ship trials. Therefore, this paper introduces data augmentation to improve the generalization performance of dynamic models identified from a limited dataset. This study used slicing and jittering as data augmentation methods and confirmed their effectiveness by numerical experiments using the free-running model tests. Results of numerical experiments demonstrated that slicing and jittering are effective data augmentation methods but could not improve generalization performance for extrapolation states of the original dataset.
DC.rights	This article is licensed under a Creative Commons Attribution 4.0 International License.
DC.format	100126
citation_title	Data augmentation methods of dynamic model identification for harbor maneuvers using feedforward neural network
citation_author	Wakita, Kouki
	Miyauchi, Yoshiki
	Akimoto, Youhei
	Maki, Atsuo
citation_publisher	Springer
citation_language	英語
citation_date	2024-11-20
citation_journal_title	Journal of Marine Science and Technology
citation_doi	https://doi.org/10.1007/s00773-024-01036-w
citation_public_url	https://hdl.handle.net/11094/100126

Data augmentation methods of dynamic model identification for harbor maneuvers using feedforward neural network

閲覧可能ファイル

論文情報

論文詳細を表示