Affiliation:
1. Department of Naval Architecture and Ocean Engineering, Research Institute of Marine Systems Engineering, Seoul National University , Seoul 08826, Republic of Korea
2. Maritime Safety and Environement Research Division, Korea Research Institute of Ships and Ocean Engineering , Daejeon 34103, Republic of Korea
Abstract
Abstract
Maritime Autonomous Surface Ships are in the development stage and they play an important role in the upcoming future. Present generation ships are semi-autonomous and controlled by the ship crew. The performance of the ship is predicted using the data collected from the ship with the help of machine learning and deep learning methods. Path planning for an autonomous ship is necessary for estimating the best possible route with minimum travel time and it depends on the weather. However, even during the navigation, there will be changes in weather and it should be predicted in order to reroute the ship. The weather information such as wave height, wave period, seawater temperature, humidity, atmospheric pressure, etc., is collected by ship external sensors, weather stations, buoys, and satellites. This paper investigates the ensemble machine learning approaches and seasonality approach for wave data prediction. The historical meteorological data are collected from six stations near Puerto Rico offshore and Hawaii offshore. We explore ensemble machine learning techniques on the data collected. The collected data are divided into training and testing data and apply machine learning models to predict the test data. The hyperparameter optimization is performed to find the best parameters before fitting on train data, this is essential to find the best results. Multivariate analysis is performed with all the methods and errors are computed to find the best models.
Funder
National Research Foundation of Korea
Ministry of Science and ICT, South Korea
Publisher
Oxford University Press (OUP)
Subject
Computational Mathematics,Computer Graphics and Computer-Aided Design,Human-Computer Interaction,Engineering (miscellaneous),Modeling and Simulation,Computational Mechanics
Cited by
16 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献