Interference Detection in Centralized Cooperative Spectrum Sensing from Sub-Nyquist Samples

Wideband spectrum sensing plays a crucial role in a number of applications among which the cognitive radio (CR) is one of the most prominent. In this work we consider a scenario where the wide band of interest is comprised of multiple communication channels occupied by several independent transmissions. Due to the propagation conditions, some of the transmissions can cause potential interference by occupying the same channel at different locations. In order to alleviate the effect of such interference, we employ a network of distributed sensing nodes that sample the wideband signal at sub-Nyquist rate and share the acquired data with the fusion center. We show that using the structure of the correlations between the sub-Nyquist samples obtained at different sensors, we can detect the presence of potential in-channel interference. We present a concrete approach for estimation of the central frequencies of the channels in which it occurs and demonstrate its effectiveness in simulations.