Spectrum Sensing and Power Classification in Spatially Correlated Noise Scenarios

In this paper, we propose a spectrum sensing and power classification scheme in hybrid interweave-underlay cognitive radio networks, considering that the primary system is with multiple transmission powers and the noise at the secondary user (SU) is spatially correlated. The primary target is to detect the presence of the primary user (PU), while the secondary target is to classify the transmission power of the PU, such that the SU can switch to underlay model with a flexible transmission power for fully exploring the spectrum access opportunities. The proposed scheme is a non-coherent detection scheme, where the weighted energy of the received signals serves as the decision metric. We derive the optimal sensing threshold for detecting the "on/off" state of the PU as well as closed-form decision thresholds for classifying the PU's transmission power. The proposed scheme can efficiently identify the PU's transmission power by leveraging the correlation information of the noise observations. Simulation results are provided to evaluate the proposed scheme.