A Reliable Optimal Hybrid Spectrum Sensing Algorithm with Hardware Impairments for Cognitive Radio Network

This is the authors' abstract. We don't add key points for this paper.

Spectrum sensing algorithms exploit partial knowledge about the signal structure. A typical strategy for doing this is feature matching, i.e. having prior knowledge about some features of the signal; the detector makes a decision based on whether the feature is present in the input. Maximizing the detection probability for a provided false alarm rate is a hectic challenge for most of the spectral sensing methods. This paper presents a reliable, optimal hybrid spectrum sensing scheme (ROHSS) with hardware impairments for cognitive radio networks. The proposed two-stage ROHSS algorithm utilizes two detectors for low and high Signal Noise Ratio (SNR) bands. In the first stage, a double-threshold improved energy detector is used for the high SNR band and an anti eigen value-based detector is used for the low SNR band based on their merits and complexities. In the second stage, the Student-Teacher Neural Network (STNN) based detector utilizes the estimated energy and eigenvalue of the signal and gives a decision. The main objective of the proposed ROHSS algorithm is to sense the vacant frequency slots and allocate them to the Primary User's (PUs) quickly in order to reduce the delay caused by the efficient operation of the fusion center. The proposed ROHSS algorithm is implemented in both MATLAB and Xilinx simulation tools and the performance is compared with the existing state-of-art algorithms.