Cooperative composite sequential detection and its application in spectrum sensing

Abstract

In this study, the authors present a method to derive sequential detector (SD) for a general class of composite hypothesis problems. The authors first explain how the SD is constructed by employing a `weight function'. Then, the authors employ this method in the cooperative spectrum sensing (SS) in cognitive radio networks where the primary user transmits a phase shift keying (PSK) signal with unknown complex amplitude in additive white Gaussian noise. The noise power is assumed known in the first scenario and unknown in the second one. To evaluate the performance of the resulting SDs, the authors obtain the required average sample number (ASN) function to meet the bounds of false alarm and missed detection probabilities through some numerical evaluations. The results illustrate that the average sensing delay of the proposed SDs are less than the required number of observations in the traditional fixed sample size detectors. In the proposed SDs, the authors also demonstrate that the increase of signal-to-noise ratio leads to decrease of ASN.