In order to provide some measure of the uncertainty inherent in the sorts of charting data that are provided to the end-user, we have previously proposed risk models that measure the magnitude of the uncertainty for a ship operating in a particular area. Calibration of these models is essential, but the complexity of the models means that we require detailed information on the sorts of ships, traffic patterns and density within the model area to make a reliable assessment. In theory, the ais system should provide this information for a suitably instrumented area. We consider the problem of converting, filtering and analysing the raw ais traffic to provide statistical characterizations of the traffic in a particular area, and illustrate the method with data from 2008-10-01 through 2008-11-30 around Norfolk, VA. We show that it is possible to automatically construct aggregate statistical characteristics of the port, resulting in distributions of transit location, termination and duration by vessel category, as well as type of traffic, physical dimensions, and intensity of activity. We also observe that although 60 days give us suffi- cient data for our immediate purposes, a large proportion of it—up to 52% by message volume—must be considered dubious due to difficulties in configuration, maintenance and operation of ais transceivers.
Journal or Conference Title
U.S. Hydrographic Conference
May 11 - May 14, 2009
Norfolk, VA, USA
International Hydrographic Organization
Calder, Brian R. and Kurt, Schwehr, "Traffic Analysis for the Calibration of Risk Assessment Methods" (2009). U.S. Hydrographic Conference. 451.