SPECTRUM SENSING TECHNIQUES IN WIRELESS COMMUNICATION NETWORKS

Abstract

Recent studies have confirmed that the currently employed spectrum management schemes have brought major inefficiencies in the usage of the frequency spectrum. As a result, dynamic spectrum access (DSA) technologies have been proposed to enable intelligent reuse of the frequency spectrum by secondary users (SUs). However, the SUs should be capable of sensing the surrounding wireless scene, in order to avoid harmful interference to the primary users (PUs) of the network. Several challenges that are generally encountered when conducting spectrum sensing have been identified. Thus, the main focus of this thesis is to address the highlighted coexistence and spectrum sensing challenges, while trying to overcome the shortcomings inherent within the selected sensing technique.The contributions of this study are explained as follows. The energy detector (ED) was examined under co-channel interference (CCI). The obtained mathematical models are derived using a moment generating function (MGF) approach. This approach provides several advantages over direct approaches, one of which is that it greatly simplifies averaging-out the random variables involved. Furthermore, the problem of detecting spread spectrum users was tackled. This issue was investigated by examining the efficiency of EDs to exploit opportunities resulting from the use of slow frequency hopping by the PUs. In addition, in order to curb the hidden node problem and due to the fact that coexistence etiquettes are not perfect against interference, the efficiency of EDs was further investigated under two simple multiple access protocols that do not require any centralised control. Furthermore, three other detector types were investigated. Firstly, the power law detector, which is a generalisation of the ED, for which no known exact PDFs exist for the statistics of its decision variable. Hence, approximations are generally relied upon. Here, novel precise approximations for the PDF and CDF of the PLD's decision variable were presented. Secondly, the performance of Kay's detector for unknown deterministic signals was analysed under AWGN, Rayleigh and Nakagami fading channels by deriving the relevant closed-form analytical expressions. Finally, the robustness of a matched filter detector in the presence of CCI was investigated. For all the detectors, accurate mathematical transformations were derived and their performance compared against the ED.

Date of Award	31 Dec 2013
Original language	English
Awarding Institution	The University of Manchester
Supervisor	Khairi Hamdi (Supervisor)