Density mapping of ship traffic

An increasing number of satellites relay information from the Automatic Identification System (AIS) anti-collision system. This has resulted in a global coverage of ship position data, and several commercial service providers currently offer ship traffic analysis products based on such data. While these services are useful in work to maintain maritime situational awareness, military users may have to avoid them due to concerns about disclosing information about their areas of interest etc. The aim of the present work was to develop software for creating traffic analysis products, for military use, on in-house hardware. The Norwegian Defence Research Establishment (FFI) maintains a database of information, available to government users, from vessel reporting systems like the AIS, Long Range Identification and Tracking (LRIT), and Vessel Monitoring System (VMS). Data on ship positions and associated metadata, accumulated over time, may be used to create density maps of ship traffic. In this context, a density map is a geographical map where the graphics primitives (pixels or polygons) represent the expected number of ships in an area at a given time, or related quantities. Density values are plotted using color gradients that reveal patterns in ship traffic. In particular, density maps are useful for analysing patterns of life, i.e., habitual behavior. Providing user-friendly, on-demand density map services presents several challenges due to the large amount of data that must be processed: during a single month, the number of messages received from unclassified data sources approaches one billion. The report describes measures for reducing the amount of data and preparing data for density map products, including aggregation, quantization, and indexing. In particular, a subdivison of the Earth’s surface into cells of constant physical size (0:1 0:1 nm2) is defined. The grid provides an index scheme for position data and a way to quantize ship tracks, thereby reducing the amount of data. For example, a single ship track is represented as the set of grid cells it intersects and the length of time it visits each cell. Moreover, grid cells correspond to graphics primitives (polygons) in density maps, with associated metadata, that can be aggregated in large-scale maps, hence increasing performance and reducing file size. A range of prototype processing services have been implemented and made available for the Norwegian Defence Joint Head Quarters, the Royal Norwegian Coast Guard, and other users with similar interests. These services create and deliver density map products on demand based on input parameters in network requests from end users. Based on the experience gained from this work, some recommendations for the deployment of density mapping services on military networks are provided. A separate classified document, related to this report, describes military operational use of these types of products.