Wideband Spectrum Sensing in Cognitive Radio with Double-Level Binary Tree Bayesian Compressed Sensing

Based on the fact that the spectrum in cognitive radio system is typically sparse, a novel wideband spectrum sensing algorithm is proposed taking advantage of Bayesian compressed sensing. Under our proposed scheme, the signal of interest can be sampled at sub-Nyquist rate so relaxing the sampling tension of front-end hardware. Furthermore, the block structure of the spectrum molded by a set of double-level binary tree is also exploited in the spectrum sensing process so promoting the spectrum sensing accuracy with lower sampling rate. Taking the MCMC sampling method as efficient inference, experimental results show the validity of our proposed method.