Adaptive fusion based clustering approach in cognitive radio networks

Abstract

Recently, the cognitive radio (CR), which is a technology to effectively use the radio band, has been introduced to allow unlicensed users to use the spectrum efficiently. The key problem of the CR system is the overhead issue in control channel as the number of secondary users (SUs) is very high in the cooperative spectrum sensing (CSS) scenario. To alleviate this problem, the Cluster-based CSS solution is useful to prevent crowding on the control channel. Therefore, this paper uses a neural network (NN) approach based on a supervised learning algorithm at the fusion center (FC) for the suggested clustering technique. Scheme suggests a combination of hard decisions, using NN approach to increase the efficiency of CRNs with multiple clusters and to decrease the overhead issue between the FC to cluster head (CH). Also, under the Rayleigh fading channel, the proposed scheme is examined and simulated to test their detection efficiency and compare their performance with conventional schemes. The simulation results prove that proposed clustering approach enhances the accuracy of the system. Also, it is more robust than other conventional schemes.