Detection and Estimation

Abstract

In dealing with noise-like signals presented to the observer in a noisy environment, one is faced with two interrelated problems: First one must establish whether there is a signal present at all (detection) and second, if there is a signal, one must extract useful information from it. A paper presented at the most recent previous NATO Advanced Study Institute on Signal Processing attempted to summarize the theoretical background of the detection problem [1]. The present paper shifts the emphasis towards questions of information extraction, meaning primarily the estimation of source parameters such as bearing, range, and velocity. This is a choice dictated not only by space and time limitations, but also by the development of the field and the structure of this conference. Recent efforts in detection theory have been concentrated heavily in the area of adaptive processing, a subject treated in separate papers of this series. My comments concerning detection will therefore be limited to a few basic observations necessary to make this presentation self-contained and to those features of the detection problem which interface directly with parameter estimation.