This paper considers discrete-time distributed detection of a constant, weak signal in dependent noise. A number of sensors transmit binary decisions to a central processor, which assimilates them to produce a final verdict. Asymptotic relative efficiency (ARE) performance measure is used to compare the decentralized and centralized detection schemes. The precise definitions of ARE-robustness and asymptotic ARE-robustness are introduced, and utilized to understand better the trade-offs involved in multiple-sensor detection in the presence of dependent noise.
展开▼