Deep Reinforcement Learning Algorithms for Steering a Underactuated Ship

Le Pham Tuyen; Layek Abu; Ngo Anh Vien; TaeChoong Chung

doi:10.1109/MFI.2017.8170388

Deep Reinforcement Learning Algorithms for Steering a Underactuated Ship

Le Pham Tuyen, Layek Abu, Ngo Anh Vien, TaeChoong Chung

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

19 Citations (Scopus)

1352 Downloads (Pure)

Abstract

Based on state-of-the-art deep reinforcement learning (Deep RL) algorithms, two controllers are proposed to pass a ship through a specified gate. Deep RL is a powerful approach to learn a complex controller which is expected to adapt to different situations of systems. This paper explains how to apply these algorithms to ship steering problem. The simulation results show advantages of these algorithms in reproducing reliable and stable controllers

Original language	English
Title of host publication	2017 IEEE International Conference on Multisensor Fusion and Integration for Intelligent Systems (MFI 2017)
Number of pages	6
ISBN (Electronic)	978-1-5090-6064-1, 978-1-5090-6063-4
DOIs	https://doi.org/10.1109/MFI.2017.8170388
Publication status	Published - 11 Dec 2017
Event	2017 IEEE International Conference on Multisensor Fusion and Integration for Intelligent Systems (MFI 2017) - , Korea, Republic of Duration: 16 Nov 2017 → 18 Nov 2017

Conference

Conference	2017 IEEE International Conference on Multisensor Fusion and Integration for Intelligent Systems (MFI 2017)
Country/Territory	Korea, Republic of
Period	16/11/2017 → 18/11/2017

Access to Document

10.1109/MFI.2017.8170388Licence: Unspecified

Deep Reinforcement Learning Algorithms for Steering an Underactuated Ship
© 2017 IEEE. This work is made available online in accordance with the publisher’s policies. Please refer to any applicable terms of use of the publisher.
Accepted author manuscript, 1.19 MBLicence: Unspecified

Cite this

@inproceedings{737f8b6ff22449b69ab4617e734cc670,

title = "Deep Reinforcement Learning Algorithms for Steering a Underactuated Ship",

abstract = "Based on state-of-the-art deep reinforcement learning (Deep RL) algorithms, two controllers are proposed to pass a ship through a specified gate. Deep RL is a powerful approach to learn a complex controller which is expected to adapt to different situations of systems. This paper explains how to apply these algorithms to ship steering problem. The simulation results show advantages of these algorithms in reproducing reliable and stable controllers",

author = "{Pham Tuyen}, Le and Layek Abu and Vien, {Ngo Anh} and TaeChoong Chung",

year = "2017",

month = dec,

day = "11",

doi = "10.1109/MFI.2017.8170388",

language = "English",

isbn = "978-1-5090-6065-8",

booktitle = "2017 IEEE International Conference on Multisensor Fusion and Integration for Intelligent Systems (MFI 2017)",

note = "2017 IEEE International Conference on Multisensor Fusion and Integration for Intelligent Systems (MFI 2017) ; Conference date: 16-11-2017 Through 18-11-2017",

}

Pham Tuyen, L, Abu, L, Vien, NA & Chung, T 2017, Deep Reinforcement Learning Algorithms for Steering a Underactuated Ship. in 2017 IEEE International Conference on Multisensor Fusion and Integration for Intelligent Systems (MFI 2017). 2017 IEEE International Conference on Multisensor Fusion and Integration for Intelligent Systems (MFI 2017), Korea, Republic of, 16/11/2017. https://doi.org/10.1109/MFI.2017.8170388

TY - GEN

T1 - Deep Reinforcement Learning Algorithms for Steering a Underactuated Ship

AU - Pham Tuyen, Le

AU - Abu, Layek

AU - Vien, Ngo Anh

AU - Chung, TaeChoong

PY - 2017/12/11

Y1 - 2017/12/11

N2 - Based on state-of-the-art deep reinforcement learning (Deep RL) algorithms, two controllers are proposed to pass a ship through a specified gate. Deep RL is a powerful approach to learn a complex controller which is expected to adapt to different situations of systems. This paper explains how to apply these algorithms to ship steering problem. The simulation results show advantages of these algorithms in reproducing reliable and stable controllers

AB - Based on state-of-the-art deep reinforcement learning (Deep RL) algorithms, two controllers are proposed to pass a ship through a specified gate. Deep RL is a powerful approach to learn a complex controller which is expected to adapt to different situations of systems. This paper explains how to apply these algorithms to ship steering problem. The simulation results show advantages of these algorithms in reproducing reliable and stable controllers

U2 - 10.1109/MFI.2017.8170388

DO - 10.1109/MFI.2017.8170388

M3 - Conference contribution

SN - 978-1-5090-6065-8

BT - 2017 IEEE International Conference on Multisensor Fusion and Integration for Intelligent Systems (MFI 2017)

T2 - 2017 IEEE International Conference on Multisensor Fusion and Integration for Intelligent Systems (MFI 2017)

Y2 - 16 November 2017 through 18 November 2017

ER -

Deep Reinforcement Learning Algorithms for Steering a Underactuated Ship

Abstract

Conference

Access to Document

Fingerprint

Cite this