AIS data-based probabilistic ship route prediction

Abstract

Maritime transportation suffers from uncertainties caused by pilots’ decisions and environmental factors. The changing environment and surroundings of ships would pose high risks to safe ship navigation. The Automatic Identification System (AIS) was developed to monitor ship behaviours in order to avoid ship collision accidents. AIS can provide information-rich ship status data such as longitude, latitude, course over ground, and speed over ground. Therefore, they are widely used for ship collision avoidance, trajectory clustering, and prediction. This study integrates Douglas–Peucker (DP) algorithm, dynamic time warping (DTW), and Hierarchical Density-Based Spatial Clustering of Applications with Noise (HDBSCAN) to cluster ship trajectories using one-year AIS data of container ships navigating in a regional area. The result shows that the proposed method can identify routes correctly. Based on the clustered trajectories, DP and DTW are used to estimate the possibility of a ship tracking certain routes. Four ships are used as case studies to show how to apply the proposed method to carry out online probabilistic route prediction. The analysis result could be used to support decision-makers for collision avoidance.