Information-Theoretic Information Fusion

Abstract : This Final Report summarizes research on information fusion based on finite-set statistics (FISST). FISST provides a fully unified, scientifically defensible, probabilistic foundation for the following aspects of multisource, multitarget, multiplatform data fusion: (1) multisource integration (detection, identification, and tracking) based on Bayesian filtering and estimation; (2) sensor management using control theory; (3) performance evaluation using information theory; (4) expert-systems theory (fuzzy logic, the Dempster-Shafer theory of evidence, rule-based inference); (5) distributed fusion; and (5) aspects of situation threat assessment. The core of FISST is a multisource-multitarget differential and integral calculus based on the fact that belief-mass functions are the multisensor-multitarget counterparts of probability-mass functions. One purpose of this calculus is to enable signal processing engineers to directly generalize conventional, engineering-friendly statistical reasoning to multisensor, multitarget, multi-evidence applications. A second purpose is to extend Bayesian (and other probabilistic) methodologies so that they are capable of dealing with (1) imperfectly characterized data and sensor models; and (2) two sensor models and true target models for multisource-multitarget problems. One consequence is that FlSST encompasses certain expert-system approaches that are often described as "heuristic"--fuzzy logic, the Dempster-Shafer theory of evidence, and rule-based inference-as special cases of a single probabilistic paradigm. Section A and Appendix 1 of the report summarize FISST and its basic consequences. Section B summarizes progress made during the course of the contract. Section 0 summarizes our progress in transitioning this USARO-funded basic research into practical applied-research funded by other DoD agencies.