Dynamic clustering scheme for distributed spectrum sensing in cognitive radio networks. (1st January 2014)

• Record Type: Journal Article
• Authors: Joshi, Pratik P.; Karia, Deepak C.; Rawat, Pratibha V.; Thakur, Meghna G.
• Is Part Of: International journal of communication networks and distributed systems. Volume 12:Number 2(2014)
• Page Start: 169
• Page End: 182
• Publication Date: 2014-01-01
• Subjects: distributed spectrum sensing -- DSS -- dynamic clustering -- fading channels -- detection threshold -- fusion -- energy efficiency -- censoring -- primary user -- PU; Computer networks -- Periodicals; Telecommunication systems -- Periodicals; Electronic data processing -- Distributed processing -- Periodicals; 004.6
• Languages: English
• ISSNs: 1754-3916

Abstract

The detection performance of cooperative spectrum sensing in cognitive radio networks is compromised by shadowing, multi-path fading and receiver uncertainty problems when the sensing results are forwarded to the fusion centre. In this paper, we propose a dynamic clustering scheme for a two-level hierarchical cognitive radio network to improve the sensing efficiency. By dynamic optimal clustering of all the cognitive radio users and selection of the user with highest instantaneous channel gain and highest detection probability for making cluster decision, our proposed model exploits spatial diversity and improves local sensing accuracy. Further, our scheme also exploits user selection diversity by forwarding results to the fusion centre using channel with highest instantaneous SNR. Moreover, by dynamic switching in the fusion centre along with forwarding of the censored results, our scheme achieves increased energy efficiency and reduced control channel bandwidth. Furthermore, consideration of fading channels within clusters, makes our scheme more generic than existing works. Performance evaluations evaluate and validate the improved spectrum sensing accuracy and energy efficiency of our proposed scheme compared to the conventional spectrum sensing methods.