Parallel Multiple-Hypothesis Tracker (PMHT) for GMTI Data Fusion

"Current radar systems generate enormous volumes of ground moving target indicator (GMTI) data that simply cannot be managed by existing automated systems or human operators. Our Phase II SBIR proposal features an innovative new approach to the GMTI fusionproblem that integrates precision tracking and group tracking algorithms into a parallel multiple-hypothesis fusion architecture. This approach is unique in that it enables precise tracking of individual ground vehicles in conjunction with more generalsurveillance level tracking that identifies groups of vehicles exhibiting common location and movement over time. Plug-and-play array processor cards are utilized within a COTS VME-based architecture to provide an extendible processing capability. Withinthis architecture, major tracking functions are encapsulated as separate software processes that are replicated and assigned to dedicated compute nodes. By adding processor cards, throughput can be scaled up to meet increasing requirements for additionaltargets, sensors, and revisit rates without modifications to the software system. A dedicated fusion management process automatically distributes the processing load evenly across the configured processing nodes. Data association is carried out under astrict multiple-hypothesis framework. The resource management of the array processor nodes is dynamic within the limitations of the number of physical processor nodes configured within