Composit Tracking and Discrimination Module

The project objective is to develop a set of mathematically rigorous Composite Tracking and Discrimination Modules (CTDMs) for accurately fusing both kinematic and non-kinematic sensor information to contribute to a consistent Single Integrated Air Picture (SIAP) containing both Tactical Ballistic Missiles (TBMs) and Air Breathing Targets (ABTs). A distributed data fusion architecture is assumed. Local (sensor level) tracks are formed based on measurement-to-track fusion. Multi-sensor system tracks are formed based on track-to-track fusion. The target state vector includes both kinematic and feature states (such as radar cross section and color temperature). Non-linear state estimation methods employed include extended Kalman filtering, a Gaussian sum representation for the target state distribution, and the modified Euler method for approximating the solution to the target state SDE. Data association is formulated as a classical assignment problem. Data association hypotheses are generated using the Munkres algorithm. A graph-theoretic algorithm is used to form cluster tracks partitioning the data association problem into independent subproblems. The bandwidth required to communicate tracking data across the distributed network is reduced by sending pseudo-measurements that capture the information from multiple physical measurements. A Bayesian inference engine performs the target discrimination and classification function. Multi-sensor registration is performed using non-Gaussian methods.