A selective cooperative spectrum sensing scheme based on improved evidence theory in cognitive radio networks

Due to providing ideal detection performance, cooperative spectrum sensing has been widely employed in cognitive radio networks. However, sensing results of the distributed second users have great uncertainty arising from channel conditions. In order to settle the problem, this paper proposes a selective cooperative spectrum sensing scheme based on improved Dempster-Shafer evidence theory. The scheme first selects second users with enough information to take part in combining at fusion center, which can save the bandwidth for reporting second users´ local observations. Afterwards we apply a method of basic probability assignment and three improved combination rules of evidence theory so as to combine the local observations of the selective second users. In this way, we efficaciously solve uncertainty problem and lower computational complexity when evidence increases. Simulation results show the improved combination rules supply preferable detection performance comparing with traditional Dempster-Shafer combination rule.