Primary User Transmit Mode Classification Based Spectrum Sensing in Cognitive Radio Network

Machine learning (ML) aided spectrum sensing is viewed as an alternative promising solution for cognitive radio (CR) systems in recent years. In this paper, a novel spectrum sensing scheme is proposed on the basis of various machine learning algorithms. Driven by the large number of spectrum observations collected from secondary user equipments (SUEs) in the cellular cognitive radio networks (CCRN), the proposed scheme first identify, by means of various ML classification algorithms, the transmit mode (TM) of the primary user (PU) transmitters in the CCRN to gain a coarse knowledge of the PU behaviors over the licensed spectrum. With the TM of the primary user transmitters (PUTs) ascertained, geo-location oriented spectrum opportunities can be subsequently pinpointed for SUEs in the CCRN. Simulations verify that the proposed scheme increases the spectrum access opportunity under the premise of minimizing the interference to the PUs.