A new method of wideband blind spectrum sensing for cognitive radio

Wideband spectrum sensing methods based on information theory criterion (ITC) solve the problem of noise uncertainty, so it has been extensively studied. But deducing the sensing decision rule using the information theory needs to assume that the received data vectors are statistical independent, and that the noise is Gaussian white noise. Meanwhile, this kind of methods possesses high implementation complexity. To overcome these limitations, this paper presented a new wideband blind spectrum sensing method based on the rank criterion. This proposed method firstly divides the sample covariance matrix of received signal into the “ideal” matrix having rank of q and the “disturbance” matrix, the rank criterion function is then used to search the optimal q value, which is used to determine the numbers and the locations of the occupied channels. The new method requires no a priori knowledge about the noise power, the statistical characteristics of the wireless channel and the primary user signal. Compared with the existing ITC based wideband sensing methods, the proposed method has a more concise decision expression and then has lower computational complexity. In addition, the new method shows excellent performance in colored noise sensing scenarios. Simulation results verified the effectiveness of the proposed method.