Most Powerful and Locally Optimum Detection in additive Noise

Abstract

A fundamental problem in many types of digital communication receivers is that of discriminating between the presence or absence of a signal of known form mixed with additive noise. This paper considers the structure of two detectors designed for different optimality criteria respectively known as the most powerful (or Neyman-Pearson) detector and the locally optimum detector. It is shown that an affine relation between the nonlinearities of the two detectors is a necessary and sufficient condition for the two detectors to reduce to identical forms. This property is used to search for noise densities which have common detector structures.