A4 Refereed article in a conference publication
Wide-Area Ship Movement Prediction Using Random Forests
Authors: Vähämäki,Tanja; Farahnakian, Farshad; Nevalainen, Paavo; Heikkonen, Jukka
Editors: Razminia, Abolhassan; Nguyen, Dinh Hoa
Conference name: International Symposium on Intelligent Technology for Future Transportation
Publisher: Springer Nature Switzerland
Publication year: 2025
Journal: Communications in Computer and Information Science
Book title : Intelligent Technology for Future Transportation: First International Symposium, ITFT 2024, Helsinki, Finland, October 19–21, 2024, Proceedings
Journal name in source: Communications in Computer and Information Science
Volume: 2378
First page : 220
Last page: 245
ISBN: 978-3-031-84147-7
eISBN: 978-3-031-84148-4
ISSN: 1865-0929
eISSN: 1865-0937
DOI: https://doi.org/10.1007/978-3-031-84148-4_18
Web address : https://doi.org/10.1007/978-3-031-84148-4_18
Maritime situational awareness requires real-time traffic prediction over a large area based on the Automatic Identification System (AIS). The second requirement is allowing input from all the traffic. We propose Random Forests (RF) for ship movement prediction and demonstrate how it can be adapted to varying zone shapes and anomaly detection tasks. We also apply it to the clustering of vessels to regularly and irregularly moving ships. Our research area is the Baltic Sea and the recording period of data is 26 July 2022... 12 August 2022. Results from the class of regularly behaving ships (499 ships out of 634) show 0.2... 2.1 km mean absolute error (MAE) over 15 min... 2 h which reaches the same accuracy as many published cases with more expensive computational models. The prediction for all supported time intervals can be updated every 10 min, which makes the implementation practical for large-scale situational awareness systems.
