Maritime Traffic Model Aided Tracking

Toyon Research Corporation will perform research and development of algorithms and software to predict future behaviors of surface vessels in littoral domains. Of particular interest are small boats since these boats are challenging to identify and persistently track. An algorithm will be developed which learns maritime traffic patterns from recorded maritime surveillance data. This data-driven solution will employ statistical modelling of maritime motion and parameter estimation to learn motion patterns from data for regions and vessels of interest. Two software prototypes will be developed during the program as a basis for commercializing the technology. The MTP-Learn application will support the processing of recorded maritime surveillance data to understand and learn traffic pattern models. The MTP-Predict library will employ traffic pattern models and offer prediction-as-a-service to client applications that require high-fidelity prediction of future vessel motion. Raytheon Company, Space and Airborne Systems, will support the effort by supplying data and data collection services for the NextGen APY-10 radar system. The radar data will be used extensively to validate and analyze the traffic pattern learning and prediction technologies as solutions to improve the Navys ability to identify and persistently track small surface vessels in littorals using airborne radar.