Distributed Hypothesis Testing in Distributed Sensor Networks

摘要

This report presents research results on distributed situation assessment in distributed sensor networks (DSNs). Information fusion algorithms for hypothesis formation and evaluation are presented. The algorithms are based on the concept of an information graph and issues that only consistent hypotheses are formed and information is not used redundantly. DSNs where the nodes have sensors observing different attributes are also considered. A theory for tracking groups of targets has been developed based on a Bayesian Theory for multitarget tracking. The general approach is used to develop algorithms for tracking with various types of sensors including acoustic. Some simulation results are presented. Keywords include: Distributed sensor networks; Distributed estimation, Distributed multitarget tracking; Information fusion; Group tracking; Distributed system testbed design; Acoustic tracking; and Attribute-based tracking.