On the Performance of Cooperative Spectrum Sensing in Random Cognitive Radio Networks

This paper investigates the performance of cooperativespectrum sensing in cognitive radio networks using thestochastic geometry tools. In order to cope with the diversityof received signal-to-noise ratios (SNRs) at secondary users, apractical and efficient cooperative spectrum sensing model isproposed and investigated based on the generalized likelihoodratio test (GLRT) detector. In order to investigate the cooperativespectrum sensing system, the theoretical expressions of theprobabilities of false alarm and detection of the local decision arederived. The optimal number of cooperating secondary users isthen investigated to achieve the minimum total error rate of thefinal decision by assuming that the secondary users follow a homogeneousPoisson point process (PPP).Moreover, the theoreticalexpressions for the achievable ergodic capacity and throughput ofthe secondary network are derived. Furthermore, the techniqueof determining an appropriate number of cooperating secondaryusers is proposed in order to maximize the achievable ergodiccapacity and throughput of the secondary network based on atarget total error rate requirement. The analytical and simulationresults validate the chosen optimal number of collaboratingsecondary users in terms of spectrum sensing, achievable ergodiccapacity and throughput of the secondary network.