Optimal Cooperator Set Selection in Social Cognitive Radio Networks
Salim Eryigit, Suzan Bayhan, Jussi Kangasharju, Tuna Tuğcu · IEEE Transactions on Vehicular Technology · 2015
While there is a huge body of research on cooperative spectrum sensing in cognitive radio networks (CRNs), incentives for cooperative behavior or the conditions under which cooperation is more likely have not been explored. We model cooperation among cognitive radios (CRs) as a function of social ties among CRs. In this social CRN , where CRs do not necessarily fulfill every cooperative sensing request, we focus on the cooperator set selection problem, i.e., in which CRs ask for cooperation so that the resulting throughput and sensing accuracy are maximized, subject to detection and false alarm probability constraints. For a single-channel scenario, we devise a multiobjective optimization model and obtain the solution using an evolutionary multiobjective algorithm. Our evaluations show that our solution is near optimal, in terms of throughput under legitimate operation, i.e., no malicious users, whereas it outperforms the expected-throughput-optimal scheme in the case of attackers. A numerical analysis demonstrates the robustness of our proposal with a slight loss in performance when the network is subject to common sensing attacks. In addition, our analysis underlines one significant drawback of existing works: assuming all CRs to be cooperative leads to a substantial overestimation of throughput capacity. Finally, to tackle the increasing complexity under a multichannel setting, we propose a heuristic for multichannel cooperative sensing for the considered social CRN.