Vessel route prediction for maritime surveillance

The paper investigates the integration of context information in maritime traffic monitoring systems. Enhanced monitoring performance has been achieved in critical operational conditions (lack of sensor measurements) and scenarios with dense vessel traffic. Specifically, the accuracy in the compilation of the traffic picture largely benefits from route prediction algorithms that exploit the available context information such as coastline and sea lanes. The idea of route prediction is to obtain an estimate of the vessel position after a long time interval under the influence of intrinsic (e.g., vessel motion characteristics) and external parameters (e.g., perturbing factors due to the environment). This can be used to support the compilation of the traffic picture and detect anomalous kinds of behaviors of the vessels. In this paper, an overview of the Route Propagation Module (RPM), developed for the EU-FP7-funded project NEREIDS, is presented. The module addresses the problem of tracking maneuvering vessels for long time intervals, in which no new information (reports) on the vessel state is received. For track maintenance purposes, the RPM algorithms exploit a priori information such as the knowledge of maritime traffic patterns, sea lanes, and the bathymetry of the monitored area of interest. The capabilities of this context-aided technique are assessed for realistic scenarios that include typical vessel maneuvers. The results on real data show that the use of the a priori information yields improvements in the accuracy of the predicted vessel position.