LAUSR

Correlation based detectors are very popular for spectrum sensing. In these detectors, temporal correlation is often employed to discriminate the signals of primary users from noises. However, large channel delay spread can significantly decrease the correlation of neighbouring samples of received signals, inducing severe performance degradation of temporal correlation based detectors. In this work, we show that the power spectral densities of different delayed versions of a primary signal have high correlation, despite unknown delays. Then, by exploiting the correlation of power spectral densities of received signals, we propose a spectral-correlation detector. To facilitate the practical application of the proposed detector, we also derive a theoretical expression for its decision threshold to achieve a given false alarm probability. Furthermore, in order to handle the case of channels with unknown delay profile, we employ the OR rule to combine the temporal- and spectral-correlation detectors, leading to a temporal-spectral-correlation detector. An expression for its decision threshold is also derived. Extensive numerical simulations are provided to verify the theoretical analysis and demonstrate the superior performance of the two proposed detectors compared to state-of-the-art temporal correlation based detectors.