Distributed detection with censoring sensors under dependent observations

Distributed detection in censoring sensor networks, where each sensor transmits “informative” observations to the Fusion Center (FC), and censors those deemed “uninformative”, has been investigated by many researchers, but under the assumption of conditionally independent observations. In this paper, we consider a more realistic situation in a censoring sensor network where observations may not be independent. We derive optimal fusion rules at the FC under both Neyman-Perason (NP) and Bayesian frameworks, assuming that each sensor sends complete observations to the FC only when its observation falls out of a certain no-send region. Simulation results are provided to demonstrate the superior performance of our fusion rule compared with several other fusion rules derived in earlier work.