Soft-Limited Polarity-Coincidence-Array Spectrum Sensing in the Presence of Non-Gaussian Noise

A new spectrum sensing scheme is proposed for detecting a primary signal corrupted by generalized Gaussian noise (GGN). It is assumed that the power of the primary signal is much smaller than noise power and that the secondary users (SUs) take advantage of multiple receiver antennas. The proposed scheme, which is referred to as soft-limited polarity coincidence array (SL-PCA), is characterized by a slope parameter that controls the operation of its limiter. It is shown that by optimizing the slope parameter, the proposed scheme outperforms its predecessor (i.e., PCA), as well as the well-known energy detection (ED) scheme in the presence of heavy-tailed GGN. It is also observed that the SL-PCA scheme can achieve superior performance with a single receiver antenna, whereas the PCA scheme cannot be used when the SUs are equipped with only one receiver antenna. Finally, the sensitivity of the proposed scheme to the slope parameter is investigated for several scenarios. It is observed that even when the slope parameter is not optimum, SL-PCA can still outperform the PCA and ED schemes, provided that the slope parameter is appropriately chosen.