Automatic Ship Berthing using Artificial Neural Network Based on Virtual Window Concept in Wind Condition

Abstract

Automatic berthing control is one of the unresolved problems in the field of ship control. Many methods and theories were developed to achieve the goal of performing automatic berthing of ship. Among them Artificial Neural Network (ANN) has found to be the most successful one by many researchers. But still questions arise about the consistency of teaching data which are used to train ANN and also about the capability of ANN to cope with wind disturbances. To investigate such phenomena, consistent teaching data based on four virtual windows (for restricted rudder angle ± 10°, ± 15°, ± 20° and ± 25°) are created using optimal steering with the help of nonlinear programming language (NPL) method. Then trained ANN is used to validate its workability in case of no wind condition first. In case of gust wind, separate ANN is trained with reconstructed teaching data and to deal with any abrupt condition, ANN-PD controller is introduced whose effectiveness is well verified not only for teaching data but also in case of non-teaching data even in case of severe wind condition.