Comprehensive analytics for predicting traffic in maritime navigation

Maritime traffic has long been an area of interest for trade, tourism and governments alike. In recent years, the increase in capabilities to generate novel insights from large amounts of data has led to a larger interest in analysis in the maritime sector. One widely used data source is the Automatic Identification System (AIS), which requires extensive status reports to be sent by most ships in regular intervals. These reports are collected by data providers. In order to utilize this data, we introduce CAPTaiN, a real time analytics and prediction approach for maritime navigation. Our approach provides continuous monitoring of maritime traffic solely based on AIS reports, meaning the ports and port calls are automatically identified using AIS messages as well as the traffic network. Historic AIS data is used to predict the behavior of ships. This includes the prediction of destination ports, the taken routes and automatic identification of behavioral anomalies. We propose a novel iterative probability update which predicts the destination correctly at 44.9 % of a voyage when predicting the destination out of 854 possible ports. The median distance between the actual position and the closest of the 25 positions predicted with the highest probabilities, out of over 45,000, is 49 km for predictions 48 steps ahead.