Spectrum Sensing With Sparsity Estimation For Cognitive Radio Systems

Cognitive Radio (CR) is an important wireless system since it opportunistically utilizes the spectrum via sensing the state of each channel in the environment. The key functional block of CR is the spectrum sensing as it senses the spectrum holes. Recently, Compressive Sensing (CS) is utilized in CR to determine the exact state of the channels of wideband spectrum. A real challenge is to sense the spectrum using CS without knowing the sparsity level of the active primary users. In this paper, an algorithm is proposed to blindly sense and accurately find out the exact busy/active primary channels with a small number of measurements below Nyquist rate. The proposed algorithm solves an optimization problem and utilizes the QR decomposition for the sensing matrix and applies a threshold to jointly get the sparsity ratio and the spectrum holes. The results present the strength of the proposed algorithm, since it provides good performance for both probabilities (detection and false alarm). Furthermore, it requires a sensing time below the acceptable range in CR systems.