Analysis of Covariance Ratio Based Detection Techniques for Cognitive Radio Networks

Abstract

The conventional method of spectrum sensing such as Energy detection, matched filter and cyclostationary methods require prior knowledge of primary user. Also, these methods are highly vulnerable to noise uncertainty. In this paper statistical covariance based matrix methods such as CVD, CFN, MME, EME and MSEE are analysed. These methods employ the sample covariance matrix for obtaining the test statistic and threshold. Moreover these methods are robust to noise uncertainty and also perform well at low SNR.

Keywords: Spectrum sensing, Covariance based matrix, SNR, noise uncertainty