Trace Based Semi-Blind and Blind Spectrum Sensing Schemes for Cognitive Radio

To make full use of the correlations among the received signal samples, new spectrum sensing schemes based on trace of noise sample covariance matrix are introduced. The new schemes utilize the fact, the statistical covariance matrices of received signal samples and noise samples are different with high probability, to detect whether the primary signal exists or not. Both of them can execute spectrum sensing without the information about primary signal and communication channel. Furthermore, to eliminate the negative effect of noise uncertainty, we modify the proposed model to adapt the case when the noise variance is completely unavailable. In addition, an optimal trace based semi-blind sensing scheme by maximizing the received signal-to-noise ratio (SNR) is also demonstrated. Multivariate statistical theories are used to derive the theoretical decision thresholds. Simulation results verify the effectiveness of the proposed schemes.