Eigenvector Based Wideband Spectrum Sensing with Sub-Nyquist Sampling for Cognitive Radio

In the Cognitive Radio (CR) technology, fast and precise spectrum sensing is essential, so that the Secondary Users (SUs) can quickly adapt their parameters by dynamic monitoring of the spectrum, enabling them to utilize the available spectrum, and more importantly to prevent interference with Primary Users (PUs). To this effect the implementation of the classical spectrum sensing methods in a wideband scenario is a great challenge. This is because the classical methods need sampling rates greater than or equal to the Nyquist rate. Modern Compressive Sensing (CS) techniques exploit the sparseness of a typical wideband spectrum. In this paper a subNyquist sampling based sensing technique is studied. The correlation matrix of a limited number of samples containing noise is constructed and the Eigenvector (EV) estimator is used to discern the functional channels of the spectrum. The performance of this technique is assessed by calculating the probability of detection of the occupied signal as a function of the number of samples and the SNR parameters of random input. Simulation results show that a robust detection is possible, even with less number of samples and at low SNR.