Overview of machine learning methods in maritime traffic monitoring based on Automatic Identification System

Abstract

The Automatic Identification System (AIS) was introduced to increase maritime safety. Its messages contain information on the identification, position, speed and course of ships. AIS enables the tracking of ships, the establishment of corridors, the detection of unusual manoeuvres (piracy, terrorism, problems with the ship or crew), the optimisation of shipping channels, strategic planning and the increase of shipping efficiency, as well as the reduction of pollution. AIS data is processed using various machine learning techniques for the purpose of extracting vessel trajectories, assessing collision risk, detecting maritime anomalies and analysing ship emissions to the environment. The development of machine learning techniques has a large and significant impact because they can quickly process a large amount of data and thus increase the safety and reliability of maritime transport. The paper gives an overview of the technical features and benefits of using AIS as well as the analysis of AIS data, algorithms for classifying and clustering AIS data and machine learning methods for detecting anomalies in maritime traffic.