Cooperative spectrum sensing algorithm based on cognitive users selection and Adaboost

Abstract

In cognitive radio networks, it is a challenge to increase the performance of cooperative spectrum sensing when the number of cognitive users become large. In this paper, a novel spectrum sensing algorithm based on the cognitive users selection strategy and Adaboost classification is proposed by adaptively selecting uncorrelated cognitive radio users with the difference of the maximum degree based on the K‐medoids clustering. Since the proposed users selection strategy is based on the correlation evaluation among the cognitive radio users, it is mandatory to have parameters able to measure the correlation among them and divide into distinct clusters. For this, the spatial correlation coefficient and the clustering quality coefficient are proposed to express the correlation characteristics of cognitive radio users in sensing environments. Performance evaluation is conducted through simulations, and the results are shown to the superior detection performance of the proposed CR user selection strategy based on the K‐medoids clustering for cooperative spectrum sensing.