No Cover Image

Journal article 794 views 512 downloads

Control Design of a Marine Vessel System Using Reinforcement Learning

Zhao Yin, Wei He, Chenguang Yang, Changyin Sun

Neurocomputing, Volume: 311, Pages: 353 - 362

Swansea University Author: Chenguang Yang

Check full text

DOI (Published version): 10.1016/j.neucom.2018.05.061

Abstract

In this paper, our main goal is to solve optimal control problem by using reinforcement learning (RL) algorithm for marine surface vessel system with known dynamic. And this algorithm is an optimal control algorithm based on policy iteration (PI), and it can obtain the suitable approximations of cos...

Full description

Published in: Neurocomputing
ISSN: 09252312
Published: 2018
Online Access: Check full text

URI: https://cronfa.swan.ac.uk/Record/cronfa40816
first_indexed 2018-06-26T19:27:54Z
last_indexed 2018-09-10T12:55:54Z
id cronfa40816
recordtype SURis
fullrecord <?xml version="1.0"?><rfc1807><datestamp>2018-09-10T11:07:24.6298508</datestamp><bib-version>v2</bib-version><id>40816</id><entry>2018-06-26</entry><title>Control Design of a Marine Vessel System Using Reinforcement Learning</title><swanseaauthors><author><sid>d2a5024448bfac00a9b3890a8404380b</sid><ORCID/><firstname>Chenguang</firstname><surname>Yang</surname><name>Chenguang Yang</name><active>true</active><ethesisStudent>false</ethesisStudent></author></swanseaauthors><date>2018-06-26</date><abstract>In this paper, our main goal is to solve optimal control problem by using reinforcement learning (RL) algorithm for marine surface vessel system with known dynamic. And this algorithm is an optimal control algorithm based on policy iteration (PI), and it can obtain the suitable approximations of cost function and the optimized control policy. There are two neural networks (NNs), where critic NN aims to estimate the cost-to-go and actor NN is utilized to design suitable input controller and minimize the tracking error. A novel tuning method is given for critic NN and actor NN. The stability and convergence are proven by Lyapunov&#x2019;s direct method. Finally, the numerical simulations are conducted to demonstrate the feasibility and superiority of presented algorithm.</abstract><type>Journal Article</type><journal>Neurocomputing</journal><volume>311</volume><paginationStart>353</paginationStart><paginationEnd>362</paginationEnd><publisher/><issnPrint>09252312</issnPrint><keywords>Reinforcement LearningCritic Neural NetworksActor neural networksLyapunov methodMarine Vessel</keywords><publishedDay>31</publishedDay><publishedMonth>12</publishedMonth><publishedYear>2018</publishedYear><publishedDate>2018-12-31</publishedDate><doi>10.1016/j.neucom.2018.05.061</doi><url/><notes/><college>COLLEGE NANME</college><CollegeCode>COLLEGE CODE</CollegeCode><institution>Swansea University</institution><apcterm/><lastEdited>2018-09-10T11:07:24.6298508</lastEdited><Created>2018-06-26T15:45:42.1419073</Created><path><level id="1">Faculty of Science and Engineering</level><level id="2">School of Engineering and Applied Sciences - Uncategorised</level></path><authors><author><firstname>Zhao</firstname><surname>Yin</surname><order>1</order></author><author><firstname>Wei</firstname><surname>He</surname><order>2</order></author><author><firstname>Chenguang</firstname><surname>Yang</surname><orcid/><order>3</order></author><author><firstname>Changyin</firstname><surname>Sun</surname><order>4</order></author></authors><documents><document><filename>0040816-29062018112737.pdf</filename><originalFilename>yin2018.pdf</originalFilename><uploaded>2018-06-29T11:27:37.3930000</uploaded><type>Output</type><contentLength>19636509</contentLength><contentType>application/pdf</contentType><version>Accepted Manuscript</version><cronfaStatus>true</cronfaStatus><embargoDate>2019-05-26T00:00:00.0000000</embargoDate><copyrightCorrect>true</copyrightCorrect><language>eng</language></document></documents><OutputDurs/></rfc1807>
spelling 2018-09-10T11:07:24.6298508 v2 40816 2018-06-26 Control Design of a Marine Vessel System Using Reinforcement Learning d2a5024448bfac00a9b3890a8404380b Chenguang Yang Chenguang Yang true false 2018-06-26 In this paper, our main goal is to solve optimal control problem by using reinforcement learning (RL) algorithm for marine surface vessel system with known dynamic. And this algorithm is an optimal control algorithm based on policy iteration (PI), and it can obtain the suitable approximations of cost function and the optimized control policy. There are two neural networks (NNs), where critic NN aims to estimate the cost-to-go and actor NN is utilized to design suitable input controller and minimize the tracking error. A novel tuning method is given for critic NN and actor NN. The stability and convergence are proven by Lyapunov’s direct method. Finally, the numerical simulations are conducted to demonstrate the feasibility and superiority of presented algorithm. Journal Article Neurocomputing 311 353 362 09252312 Reinforcement LearningCritic Neural NetworksActor neural networksLyapunov methodMarine Vessel 31 12 2018 2018-12-31 10.1016/j.neucom.2018.05.061 COLLEGE NANME COLLEGE CODE Swansea University 2018-09-10T11:07:24.6298508 2018-06-26T15:45:42.1419073 Faculty of Science and Engineering School of Engineering and Applied Sciences - Uncategorised Zhao Yin 1 Wei He 2 Chenguang Yang 3 Changyin Sun 4 0040816-29062018112737.pdf yin2018.pdf 2018-06-29T11:27:37.3930000 Output 19636509 application/pdf Accepted Manuscript true 2019-05-26T00:00:00.0000000 true eng
title Control Design of a Marine Vessel System Using Reinforcement Learning
spellingShingle Control Design of a Marine Vessel System Using Reinforcement Learning
Chenguang Yang
title_short Control Design of a Marine Vessel System Using Reinforcement Learning
title_full Control Design of a Marine Vessel System Using Reinforcement Learning
title_fullStr Control Design of a Marine Vessel System Using Reinforcement Learning
title_full_unstemmed Control Design of a Marine Vessel System Using Reinforcement Learning
title_sort Control Design of a Marine Vessel System Using Reinforcement Learning
author_id_str_mv d2a5024448bfac00a9b3890a8404380b
author_id_fullname_str_mv d2a5024448bfac00a9b3890a8404380b_***_Chenguang Yang
author Chenguang Yang
author2 Zhao Yin
Wei He
Chenguang Yang
Changyin Sun
format Journal article
container_title Neurocomputing
container_volume 311
container_start_page 353
publishDate 2018
institution Swansea University
issn 09252312
doi_str_mv 10.1016/j.neucom.2018.05.061
college_str Faculty of Science and Engineering
hierarchytype
hierarchy_top_id facultyofscienceandengineering
hierarchy_top_title Faculty of Science and Engineering
hierarchy_parent_id facultyofscienceandengineering
hierarchy_parent_title Faculty of Science and Engineering
department_str School of Engineering and Applied Sciences - Uncategorised{{{_:::_}}}Faculty of Science and Engineering{{{_:::_}}}School of Engineering and Applied Sciences - Uncategorised
document_store_str 1
active_str 0
description In this paper, our main goal is to solve optimal control problem by using reinforcement learning (RL) algorithm for marine surface vessel system with known dynamic. And this algorithm is an optimal control algorithm based on policy iteration (PI), and it can obtain the suitable approximations of cost function and the optimized control policy. There are two neural networks (NNs), where critic NN aims to estimate the cost-to-go and actor NN is utilized to design suitable input controller and minimize the tracking error. A novel tuning method is given for critic NN and actor NN. The stability and convergence are proven by Lyapunov’s direct method. Finally, the numerical simulations are conducted to demonstrate the feasibility and superiority of presented algorithm.
published_date 2018-12-31T19:26:28Z
_version_ 1821344203900190720
score 11.04748

Similar Items