Anomalous behavior detection of marine vessels based on Hidden Markov Model

In recent years, marine traffic has increased dramatically. Therefore, some vessels may show abnormal behaviors which may indicate different dangers or threats, and also decreases the marine security. These abnormal behaviors are called anomalies. Automatic detection of anomalies is one of the current issues in the maritime domain. Similar works have been proposed that often used parameters speed and location to detect anomalies, and also high false alarm rate is one of the main problems of those. In this paper, a novel approach to maritime vessels anomalous behavior detection based on Hidden Markov Model with the use of parameters speed, location and course have been proposed to increase the situational awareness. Implementation results of the proposed method represent low false alarm rate and more than 96% accuracy in the classification of normal behavior from anomalous behaviors.