Unified Bayesian Cluster Target Tracking and Discrimination

"Most single- and multi-target tracking algorithms are designed to track point targets.However, in many real-world applications the targets of interest are not point targets butEXTENDED TARGETS and GROUP TARGETS, e.g. missile re-entry clusters.Under this effort we propose a systematic, fully probabilistic, and theoretically rigorousapproach to cluster target tracking and discrimination. Our approach is a direct generalizationof BAYES-OPTIMAL RECURSIVE NONLINEAR FILTERING THEORY to the multisource-multitarget realm.The overall objective is to develop innovative, speculative, high-risk technologiesfor enhancing either the Theater Missile Defense (TMD) or National Missile Defense (NMD) capabilities.Phase I showed the feasibility assessment for multitarget tracking/discrimination and ``bulk'' clustertracking by utilizing a novel approximation of multi-target non-linear filtering based on the spectralcompression (SPECC) non-linear filter (NLF) implementation of Stein-Winter probability hypothesis densities (PHDs).The Phase II objectives are to further develop, analyze, and refine the PHD filtering approach by extending the resultsof Phase I. Specific Phase II tasks are: (1) Implement multipeak extraction algorithms, (2) Implement alternativeNLF implementations of PHD, (3) Extend approach to Joint Track & Non-Cooperative Target Identification (NCTI), (4) Developdiscrimination-based information about clusters to determine presence or non-pr