SDH TS 00044 a study on an automatic ship berthing based on artificial neural network controller using head up coordinate system

Doctoral Dissertation A StudyonAn AutomaticShip BerthingBased on Artificial Neural Network Controller Using Head-up Coordinate System Supervisor Prof.. NamkyunIm August 2016 Mokpo Na

Doctoral Dissertation

A StudyonAn AutomaticShip BerthingBased on Artificial Neural Network Controller Using

Head-up Coordinate System

Supervisor Prof NamkyunIm

August 2016

Mokpo National Maritime University, Graduate School

Department of Maritime Transportation System Engineering

Van-Suong Nguyen

A StudyonAn AutomaticShip Berthing Based on Artificial Neural Network Controller Using

Head-up Coordinate System

Supervisor

Prof NamkyunIm

By

Van-Suong Nguyen

A Dissertation Submitted in Partial Fulfillment of the Requirements for the

Degree of

Doctor of Philosophy

In Department of Maritime Transportation SystemEngineering, Graduate School

of Mokpo National Maritime University

August 2016

A StudyonAn AutomaticShip Berthing Based on Artificial Neural Network Controller Using

Head-up Coordinate System

A Dissertation by Van-Suong Nguyen

Approved as to style and content by

Chairman Prof Gyei-Kark Park

Member Prof Chang-Hyun Jung

MemberProf Sang-Min Lee

MemberProf Yu-Chang Seong

MemberProf NamkyunIm

August 2016

Dedication

This dissertation is dedicated to the author’s family, Van-Xuyen Nguyen, Thi-Tuyen Do, Van-Vui Nguyen and ThiHien-Phuong Ha

I, the undersigned, hereby declare that I am the sole author of this thesis I have prepared my thesis specifically for the degree of Doctor of Philosophy, while under supervision at Mokpo National Maritime University, Republic of Korea, in accordance with academic rules and ethical conduct I also declare that, as required be these rules and conduct, I have fully cited and acknowledged all the ideas and work of others, whether published or unpublished, which are not original to my thesis

Van-Suong Nguyen

i

Acknowledgements

Many thoughts went through my mind as I compiled the work of the past threeyears to write this dissertation Most of all thoughts about the many people whoenabled me to perform this research work

Firstly, I am extremely grateful to my supervisor, Prof NamkyunIm, for hisguidance andfinancial support throughout the course of this research Hisenthusiasmand encouragement have been invaluable source of inspirationand motivation for me during the last three years

Secondly, I also would like to express my gratitude to all my Lab members, Ms Soo-Jin Hwang andMs Bora Choice for their help during the last three years I would like to thank all Korean friends,who helped and made my stay in Mokpo particularly joyful

Thirdly, I would like to thank the committee members: Prof Gyei-Kark Park, Prof Sang-Min Lee, Prof Chang-Hyun Jung, Prof Yu-Chang Seongand Prof Nam-KyunImfor all invaluable comments and suggestions

Lastly, I would like to thank all members in my family:Father Mr Nguyen Van Xuyen, Mother Mrs Do ThiTuyen, Brother Mr Nguyen Van Vui and My Wife Mrs.Ha Thi Hien Phuong, withall my love Thanks for sharing with me every joyful moments as well as difficultiesand disappointments.Their encouragement and support was immeasurable as well

as motivation in order that I can complete this course

Mokpo National Maritime University, Korea

Van-Suong Nguyen

August 2016

ii

A Study on An Automatic Ship Berthing Based on Artificial Neural

Network Controller Using Head-up Coordinate System

Van-Suong Nguyen

Department of Maritime Transportation System Engineering, Graduate School

Mokpo National Maritime University, August 2016

Abstract

Artificial Neural Network (ANN) has been known as one of most effective theories for automatic ship berthing because it has the learning ability and mimics action of human’s brain

in performing stages of ship berthing maneuver So far, there have been many researches which usedthe ANNstructure as main controllerto automatic ship berthing However, these controllers can only control ship into berth automatically and safely in a unique port where teaching data werecreated It means thatwhen ship arrives to new ports,these ANN controllers must be retrainedby new teaching data or the ship has to install multi-controllerfor different ports respectively It causes time-consuming, expensive computation and complicated structure in the control system Furthermore,theseANN controllers can only controlship into berthon fixed-berth-approachingdirection whichwastrained by teaching data in original port Therefore,when applying these controllers to different approaching directions,the structure ofmulti-controller must be equipped to ship.It is not feasible for economical characteristic To overcome limitations of previous researches on automatic ship berthing using artificial neural network, some new ideas are proposed in this research

Firstly, head-up coordinate system is suggested to consider two new inputs for neural controller, a novel ANN controller is proposed to automatic ship berthing.By usinghead-up

iii

coordinate systemwhichincludes relative bearing and distance from ship to berth, the proposed ANN controller can be used for different ports as well as different terminals adaptively withoutretraining the controller as well asconsidering the geographical feature of different ports Teaching data is created in original port to train the proposed ANN controller, after then the ship can be automatically controlled into berth in original port as well as other ports where inputs in head-up coordinate system are same as or similar to ones in teaching data To verify the effectiveness of the proposed ANN controller to automatic ship berthing, numerical simulations are carried out for two different imaginary ports and a real port The results show that the proposed controller has good performance for automatic ship berthing

Secondly, the symmetrical coordinate system is described to show two different berth-approaching directions in port Toemploy the proposed ANN controller to new berth-approaching direction which is symmetrical with one in teaching data according to symmetrical line, The modified ANN controller is suggestedbased on the proposed ANN controller andthe characteristic of ship velocity change in turning circle.The modified ANN controller is obtained by the combination of threenegative switchers and the proposed ANN controller By using the modified ANN controller, the ship can be applied to different approaching directions without training the controller again Numerical simulations are also performed in two imaginary ports and real port to verify the effectiveness ofthisproposal The results show that the modified ANN controllerhas good performance for automatic ship berthing control

iv

List of Contents

Acknowledgements……… i

Abstract………ii

List of Contents……… iv

List of Figures……….viii

List of Tables……… x

Nomenclatures……….xii

Chapter 1 Introduction 1.1 Introduction……… ……….……… 1

1.2 Purpose of This Research……… 3

1.3 Structure of This Research……… 4

Chapter 2 Mathematical Model for Ship Maneuvering 2.1 Introduction……… 6

2.2 Kinematics Equation of Ship……….………7

2.3 Ship Mathematical Model……….……… 9

Chapter 3 Artificial Neural Network Controller Using Head-Up Coordinate System 3.1 Introduction……….….….14

3.2 Head-up coordinate system……… 15

3.3 Formula for Data Converter……… ….16