Experiment Results for Automatic Ship Berthing using Artificial Neural Network Based Controller

Abstract

Experiment results are often worthy and would be a great topic of discussion. Since many years, research on automatic ship berthing had already been going on to make it actually possible so that one can minimize the annual cost of dock damage as well to ensure safe operation for such sophisticated multi input-multi output phenomenon. But, success up to the extent to execute it in real cases not yet possible. Recently upon developing the unique concept named 'virtual window', the authors of this paper had already challenged the automatic ship berthing for Esso Osaka 3-m model even considering wind disturbances up to 1.5 m/s which is 15 m/s for full scale ship and considered as dead end for most ports in Japan. After getting success, for the first time using this concept, automatic ship berthing is planned and executed for free running experiment. This paper is aimed to publish some of valuable experiment results in categorised way depending on behavior of Artificial Neural Network which is used as a controller during the berthing manoeuvre.