A Comprehensive Study of Clustering-Based Techniques for Detecting Abnormal Vessel Behavior

Abnormal behavior detection is currently receiving much attention because of the availability marine equipment and data allowing maritime agents to track vessels. One most popular tools for developing an efficient anomaly system Automatic Identification System (AIS). The aim this paper explore performance existing well-known clustering methods detecting two dangerous abnormal behaviors based on AIS. include K-means, Density-Based Spatial Clustering Applications with Noise (DBSCAN), Affinity Propagation (AP), Gaussian Mixtures Model (GMM). In order evaluate methods, we also used AIS vessels, which were collected through Finnish transport agency from whole Baltic Sea three months. Although studies focus ocean route recognition, deviations regulated routes, or irregular speed, focused dark ships those sets vessels that turn off perform illegal activities spiral vessel movements. experimental results demonstrate K-means method can effectively detect movements, are threatening events safety.