By Paolo Franchi
Maritime traffic graphs: construction criteria and performance metrics. Coscia, Pasquale; Braca, Paolo; Millefiori, Leonardo M.; Palmieri, Francesco A. N.; Willett, Peter K.. CMRE-FR-2018-011. January 2019.
We propose an unsupervised procedure to automatically extract a graph-based model of commercial maritime traffic routes from historical Automatic Identification System (AIS) data. The network represents compactly the most significant elements of the traffic, such as waypoints and edges, which act as manoeuvre regions and sea-lanes, respectively. Vessel dynamics are defined by multiple Ornstein- Uhlenbeck (OU) processes with different long-run mean parameters estimated with a change detection procedure that discovers spatial points representative of velocity changes. A density-based clustering algorithm groups relevant points rejecting outliers. Two performance criteria, using real-world trajectory data of vessels sailing along the Iberian Coast and the English Channel, are tested to tune and validate the graph-based representation.