A blind spectrum-sensing method based on Bartlett decomposition

Spectrum sensing is a key element of Cognitive Radio communications. In this paper, a novel method is proposed based on the Bartlett decomposition of statistical covariance matrix. The proposed method exhibits good performance and robustness against noise uncertainty and without having to know any information of the primary signal and the transmission channel. Compared to other methods based on the statistical covariance matrix, the threshold of the proposed method is accurately calculated without assuming that the number of the signal samples is large enough. As a result, the performance of our method is superior to the others. Simulations based on wireless microphone signals are presented to verify our method.