Vis enkel innførsel

dc.contributor.authorMartinsen, Andreas Bell
dc.contributor.authorLekkas, Anastasios M.
dc.date.accessioned2019-05-07T10:20:47Z
dc.date.available2019-05-07T10:20:47Z
dc.date.created2019-01-14T08:37:53Z
dc.date.issued2018
dc.identifier.isbn978-1-5386-4814-8
dc.identifier.urihttp://hdl.handle.net/11250/2596760
dc.description.abstractThis paper proposes a methodology for solving the curved path following problem for underactuated vehicles under unknown ocean current influence using deep reinforcement learning. Three dynamic models of high complexity are employed to simulate the motions of a mariner vessel, a container vessel and a tanker. The policy search algorithm is tasked to find suitable steering policies, without any prior info about the vessels or their environment. First, we train the algorithm to find a policy for tackling the straight line following problem for each of the simulated vessels and then perform transfer learning to extend the policies to the curved-path case. This turns out to be a much faster process compared to training directly for curved paths, while achieving indistinguishable performance.nb_NO
dc.language.isoengnb_NO
dc.publisherInstitute of Electrical and Electronics Engineers (IEEE)nb_NO
dc.relation.ispartofOCEANS 2018 MTS/IEEE Charleston
dc.titleCurved Path Following with Deep Reinforcement Learning: Results from Three Vessel Modelsnb_NO
dc.typeChapternb_NO
dc.description.versionacceptedVersionnb_NO
dc.identifier.doi10.1109/OCEANS.2018.8604829
dc.identifier.cristin1655779
dc.description.localcode© 2018 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.nb_NO
cristin.unitcode194,63,25,0
cristin.unitnameInstitutt for teknisk kybernetikk
cristin.ispublishedtrue
cristin.fulltextoriginal
cristin.qualitycode1


Tilhørende fil(er)

Thumbnail

Denne innførselen finnes i følgende samling(er)

Vis enkel innførsel