Cooperative compressive spectrum sensing by sub-Nyquist sampling

Abstract

Compressive Sensing (CS) is a novel framework shows that a Qb-point discrete time signal that is k-sparse, can be exactly recovered by using small amounts of linear projections. In this paper, we propose an aliasing-based distributed compressive spectrum sensing technique for Cognitive Radio (CR) networks. We firstly model the spectrum aliasing phenomenon as a linear projection from the ideal sampled spectrum to the sub-sampled spectrum. Then the necessary conditions for jointly reconstructing the spectrum without aliasing are provided. Rather than using separate compression device, the Analog-to-Digital Converters (ADCs) in our proposed method perform data compression as well as sampling. More important, with multiple receivers operating at sub-Nyquist sampling rates, the fusion centre can effectively recover the spectrum without aliasing.