A blind source separation technique for spectrum sensing in cognitive radio networks based on Kurtosis metric

Cognitive radio is proposed as a key solution to overcome the inefficient spectrum usage and spectrum sensing is an important part of cognitive radio networks. In classical methods of spectrum sensing, when the secondary user analyzes the channel to find out if primary users are in operation or not, secondary users are not allowed to transmit. In this paper, we propose a spectrum sensing method based on blind source separation technique by using Kurtosis metrics that can sense the spectrum while secondary transmitters are in operation. We compare our proposed channel sensing method with classical ones. Simulation results provided in terms of receiver operating characteristic (ROC) curves indicate that the proposed method outperforms the conventional spectrum sensing techniques, considerably.