Effect of Primary User Traffic on Largest Eigenvalue Based Spectrum Sensing Technique

In this paper, the effect of primary user (PU) traffic on the performance of largest eigenvalue based spectrum sensing technique (Roy’s Largest Root Test (RLRT)) is investigated. A simple and realistic discrete time modeling of PU traffic is considered which is only based on the discrete time distribution of PU free and busy periods. Furthermore, in order to analyze the effect of PU traffic on the detection performance, analytical expressions for the probability density functions of the decision statistic are derived and validated by Monte-Carlo simulations. Numerical results demonstrate that the sensing performance of RLRT is no more monotonically increasing with the length of the sensing duration and also with SNR which contrasts with the common property of the spectrum sensing techniques under known PU traffic scenario. Furthermore, it is shown that the performance gain due to the multiple antennas in the sensing unit is significantly suppressed by the effect of the PU traffic when the frequency of the PU traffic transitions is higher