Performance analysis of a new two-stage tunable detector

Abstract

This study focuses on the construction of a two-stage detector for homogeneous Gaussian noise with an unknown covariance matrix. A Durbin detector and an Enhanced RAO detector are cascaded to design the proposed detector. Analyses are used to establish the analytical expression for the probability of false alarm and detection. The new detector's performance is evaluated during the analysis stage. To compare performance in the situations of matched and mismatched signals, many classical detectors were employed. The results suggest that the proposed detector can offer improved rejection capabilities in the case of mismatched signals at the expense of a little detection loss for matching signals.