Embedded primary users identification and channel estimation for underlay cognitive radio network based on compressive sensing

Because of its spectrum-sharing nature, a CR network inevitably operates in interference-intensive environments. One challenge is how to maintain the interferences generated by the cognitive transmissions to the primary network below an acceptable threshold level. In this paper, we firstly propose an efficient primary users identification, using compressive sensing (CS). Our focus is on the angular sparsity of the received signal given an unknown number of primary user source signals impinging upon the antenna array from different directions of arrival (DOA). Given multiple snapshots, multiple measurement vectors (MMV) are available at the secondary base station and considered for primary channel detection over the angular domain using the regularized M-FOCUSS algorithm. Then, we develop novel methods for paths separation and primary channels estimation based on their autocorrelation matrix properties. Through simulations, we show that the recovery performance of the proposed approach in terms of false alarm and average minimum square error (MES) between the true and the estimated primary channel, is better than the conventional maximum to minimum eigenvalue (MME) detector.