Performance improvement in decentralized fusion problems via additive dependent noises

This paper investigates effects of adding dependent random signals (additive noises) to observations of local sensors in decentralized fusion problems, where the decision of each local sensor is directly transmitted to the fusion center for the final decision according to Neyman-Pearson criterion. Threshold of the fusion center and random signals added to observations of local sensors are modeled as random variables dependent on each other but independent from the observations at the local sensors. Simulation results are presented to investigate performance of the proposed approach.