General Detection and Estimation Theory in an Adaptive Context

Abstract

This paper has two aspects: one is tutorial in nature and its objective is to present in a concise way the fundamental ideas of detection and estimation theory which are necessary to easily undestand the matter presented in the following part; the second is more a presentation of new material in the field of adaptive detection, and particularly of signal detection in noise with fluctuating power. Thus this second part appears as the subsequent part of a sequence of papers already delivered in previous NATO ASI (1) (2) (3). In (1) was discussed the concept of optimality for an adaptive detection system and particularly its application to the detection of a deterministic signal in spherically invariant noise. In (2) the concept of Noise Alone Reference (NAE) already used in spatial signal processing was introduced in order to present a geometrical interpretation of the classical matched filter using a phase of estimation. Moreover some adaptive detectors were suggested without effective calculation or simulations concerning their performances. In (3) some adaptive algorithms were presented in order to introduce the concept of recursivity.

Laboratoire des Signaux et Systèmes du Centre National de la Recherche Scientifique et de l’Ecole Superieure d’Electricité, associé a l’Universite de Paris-Sud.