objectives

The project has addressed the maritime traffic modelling by probabilistic and unsupervised machine learning approaches and exploits the maritime traffic models for ship abnormal behaviour detection and classification.

Besides, the main objective is achieved by developing tools for probabilistic characterisation of the main routes and the operational profiles of the maritime traffic off the continental coast of Portugal, using historical ship trajectory data and expert knowledge of the local traffic characteristics.

This characterisation of the traffic patterns and ship routes provides a means for developing advanced unsupervised tools for automatic detection of abnormal traffic behaviour of ships, which provide alerts to support monitoring, control and surveillance tasks.