Spectrum Sensing Technique for Cognitive Radio of Multiple OFDM Signals based on Singular Value Decomposition

This work addresses the Spectrum Sensing (SS) problem in Cognitive Radio (CR) context, where the CR allows SUs (Secondary Users, i.e. unlicensed users) to access P U s (Primary Users, i.e. licensed frequency bands) through dynamic spectrum access in order to reduce spectrum scarcity problem in wireless communication. We propose in this communication a new multi-antenna spectrum sensing technique of multiple OFDM (Orthogonal Frequency Division Multiplexing) PU transmitters Signals using only singular value decomposition of second-order statistics. Furthermore, we derived closed-form of the false alarm, the detection probability and the threshold. Computer simulations show that the proposed SS technique can obtain good performances in low SNR values, which is a challenging issue in spectrum sensing. This technique is also compared with energy detector and Maximum-Minimum Eigenvalue detector (MME).