Fuzzy C-means and spatial correlation based clustering for cooperative spectrum sensing

Abstract

A cluster formation algorithm is proposed to save the wastage of energy in cooperative spectrum sensing (CSS), in which small number of groups called clusters are made using fuzzy c-means (FCM). Based on spatial correlation, only limited number of SUs are selected from each cluster, whose sensing information is forwarded to their cluster head (CHs). The primary goal of cognitive radio network is spectrum sensing, that determines whether a particular spectrum hole is free or not, so that secondary users (SUs) can utilize that spectrum, if it is not occupied. Many factors such as shadowing and multipath fading problem may degrade the detection performance of spectrum sensing. To overcome these problems CSS schemes were proposed, in which different number of SUs collaborate with each other and share their sensing information with fusion center (FC) to make the final decision of PU's activity. This may increase the overhead for the FC when a huge number of SUs are involved in sensing. The proposed algorithm consumes small energy and achieves the highest efficiency as compare to non-clustering and conventional clustering techniques.