Spectrum sensing method without the impact of noise uncertainty

In cognitive radio, the spectrum sensing plays a key role in determining the performance of both the primary and the secondary networks. The eigenvalue based detection (EbD) algorithm has received broad attentions, since it shows significant robustness to the noise uncertainty problem. However, in EbD algorithm, it is quite difficult to obtain the distribution for the eigenvalues of the statistic covariance matrix, causing the false alarm or detection probability fail to be controlled efficiently. To this end, in this paper, we reinvestigate the eigenvalues from the perspective of the circulant matrix and propose a power spectrum density (PSD) based detection (PbD) algorithm. The proposed algorithm not only shares the same principle as the EbD algorithm, but also its error probability can be calculated efficiently. The closed-form expressions for the sensing threshold and the detection probability of our PdD algorithm are also provided. Simulation results validate our theoretical analysis well.