LAUSR

Rooted in the compressed sensing theory, sub-Nyquist spectrum sensing (SNSS) has been considered as a promising approach to dealing with difficulties and limitations of conventional wideband spectrum sensing in cognitive radio (CR) networks. Most existing SNSS methods require some prior knowledge of the monitored frequency bands, such as the spectrum occupancy/sparsity level and/or the noise power, to determine a termination condition used by an underlying iterative signal recovery process. However, such prior knowledge may be difficult to acquire in practical CR scenarios. To address this problem, we propose a blind SNSS algorithm, referred to as the residual energy ratio based detector (RERD), which bypasses the need for the above-mentioned prior knowledge and performs spectrum sensing in a more autonomous way. The RERD algorithm, which is based on the modulated wideband converter (MWC) sub-Nyquist sampling framework, employs energy ratios of adjacent channels of the MWC as test statistics. We derive closed-form expressions of the decision threshold and the false alarm probability following the Neyman–Pearson criterion. Simulation results show that, without requiring the aforementioned prior knowledge, the RERD algorithm can accurately determine the support of a multiband signal contaminated by background noise in a wide range of signal-to-noise ratio. Moreover, the RERD algorithm is shown to be robust to a range of sparsity orders and different number of sampling channels.