Performance evaluation of cognitive radio spectrum sensing using multitaper-singular value decomposition

Cognitive radio (CR) is a promising technology for the efficient use of available spectrum. Reliable detection of primary user (PR) signals and spectrum sensing are the most important challenges which meet the implementation of the CR. In CR's communications, the PR user must be protected from the interference from the CR users. This can be achieved only by using an optimal spectrum sensing. In this paper we apply the multitaper method with singular value decomposition (MTMSVD) in the estimation of the power spectrum density and the decision about the presence or the absence of PR transmission. Our system model consists of a licensed PR user transmits QPSK-OFDM signal in a specific band, and a number of CR's sensors that detect the PR user's signal and this information is used by a CR base station to facilitate the CR user's communications through AWGN channel. We will compare the MTM-SVD technique with the conventional method of power spectrum density estimation (periodogram). Our results show that the MTM-SVD has a better performance in low SNR compared with the energy detection method. Furthermore, MTM-SVD is a powerful method that combines the mutual information using a number of sensors and different tapers. The decision process has a large threshold margin, which increases the probability of the correct decision. Increasing the sensors improves the performance of the spectral sensing in a low SNR environment.