Performance analysis of covariance based detection in cognitive radio

In cognitive radio networks, the first cognitive task preceding any form of dynamic spectrum management is the spectrum sensing and identification of spectrum holes in wireless environment. In cognitive radio, spectrum sensing is fundamental crucial task. Energy detection method is a basic method, which requires knowledge of noise power but suffers from noise uncertainty problem. Covariance based detection exploits space-time signal correlation that does not require the knowledge of noise and signal power. The covariances of signal and noise are generally different which can be used in detection of licensed user. However, there are not many studies that show the feasibility of the detectors and analyze their performance under fading channels. In this paper, we analyzed the detector performance exploiting TV White Space under Rayleigh and Rician fading channel by setting probabilities of false alarm and measuring probability of detection. We further analyze the effect of smoothing factor and overall correlation coefficient on the performance of covariance based detector. Covariance based detector outperformed the energy detector with noise uncertainty even under the time-varying fading channels.