Robust sensor selection for collaborative spectrum sensing with attacks under correlated shadowing

Collaborative spectrum sensing is an effective technique to improve the performance of spectrum sensing in cognitive radio. However, collaborative gains are limited by attacks and correlated shadowing. In this paper we propose a robust sensor selection algorithm to select reliable and less correlated sensors under both independent and collaborative attacks in correlated environment. Firstly, we iteratively exclude sensors whose reports have abnormal high similarities since the reports of collaborative attackers have the highest similarities. Secondly, we select the desired number of sensors which are reliable and less correlated from the remained sensors based on a well designed cost matrix which measures sensors' reliabilities and correlations. The sensor selection is modeled as a cost sum optimization problem and a suboptimal solution is found by applying greedy algorithm. It is shown by simulations that the proposed robust sensor selection algorithm has a good performance.