Cooperative Spectrum Sensing under Correlated Shadowing based on Fuzzy Logic in Multi-Hop Cognitive Radio Networks

Abstract

Cognitive Radio has been invented to provide wireless communications with efficient radio spectrum utilization. The secondary users (SUs) can therefore, access opportunistically licensed band by sensing spectrum holes without interfering with primary users (PU) or keeping the interference, if it happens below a tolerable threshold. Main functions of cognitive radio are: Spectrum sensing, spectrum management, spectrum mobility and spectrum sharing. This research focuses on sensing and spectrum access optimization in cooperative multi-hop cognitive radio networks. In a cooperative spectrum sensing, nodes located in their vicinities can experience spatially correlated fading and it leads to a degraded detection performance. To combat that effect, it is has been demonstrated in several works that by selecting only spatially independent nodes, good results can be obtained. A fuzzy-based user selection is investigated to cope with the aforementioned issues in a multi-hop clustered cooperative spectrum sensing architecture. Moreover, many researchers have based their work on single-hop cognitive radio network architecture. However, this architecture does not portray practical environments whereby nodes can be located far from each other and will hence need their communication to be forwarded through relays. By optimizing fuzzy inputs, we are able to achieve a high detection performance in a multi-hop architecture by selecting only less correlated users to cooperate. By considering uncorrelated users individually, the developed fuzzy based detection system outperforms the distance based one by providing probability of detection 40% more than the distance-based one when the decorrelation distance is 30 meters. On the other hand, when the decorrelation distance takes respectively values of 65m and 100m, the system does not show any gain compared to the distance-based one. Finally, when using only uncorrelated users, the detection performance is approximately the double of the one when using correlated users.