Distributed Detection with Waveform Observations: Correlated Observation Processes

We consider the problem of detection using distributed sensors and limited communication. In particular we examine the case of detecting a known signal in noise which is correlated between sensors, and that of detecting an unknown signal in white noise. As the optimal detection law cannot be determined, we optimize over a class of suboptimal detection laws based on local likelihood ratio tests. Numerical examples illustrate the performance of these laws in comparison to naive decentralized and to centralized decision laws.