Dimension estimation based detector for multiple-antenna cognitive radio networks

Abstract

In this paper, a spectrum sensing approach for multiple-antenna cognitive radio networks will be proposed. Specifically, we will apply dimension estimation based spectrum sensing technique using minimum description length (MDL) criteria to sense the presence of primary users signals. Therefore, we consider a cognitive radio network with cognitive user equipped with M received antennas and K (M >; K) primary users and we try to estimate the number of primary users present in this network using MDL criteria. We detect then presence of primary signals for a fixed false alarm probability derived analytically using the distribution of Tracy-Widom of order two. Performance study of the proposed detector using different properties DVB-T systems will be presented. Simulation results show high performances detection even with a very small number of antennas.