Decentralized detection by a large number of sensors (Q1111520)

The author considers the decentralized detection problem, in which N identical sensors transmit a finite D-valued function of their observations to a fusion center which then decides which one of M hypotheses is true. It is assumed that the sensors are conditionally independent. In case the number N of sensors tends to infinity it is asymptotically optimal to divide the sensors in M(M-1)/2 groups, with all the sensors in each group using the same decision rule in deciding what to transmit. The optimal number of sensors in each group may be determined by solving a mathematical programming problem. In case \(M=2\) and \(D=2\) the solution simplifies considerably. The author also considers a Neyman-Pearson type of formulation of the problem. The main result of the paper is also interpreted in a coding context where one would like to transmit a, say binary message along a collection of noisy, memoryless, and independent channels, where one is allowed to transmit a message a total of N times.