Multi-antenna energy detector under unknown primary user traffic

In cognitive radio (CR) networks, the knowledge of primary user (PU) traffic plays a crucial role in designing the sensing slot duration and synchronization with PU traffic. However, the secondary user (SU) sensing unit usually does not have the knowledge of the exact time slot structure in the primary network. Moreover, it is also possible that the communication among PUs are not based on synchronous schemes at all. In this paper, the effect of unknown primary user (PU) traffic on the performance of multi-antenna spectrum sensing is evaluated under a flat fading channel. In contrast to the commonly used continuous time Markov model of the existing literature, a realistic and simple PU traffic model is proposed which is based only on the discrete time distribution of PU free and busy periods. Furthermore, in order to assess the effect of PU traffic on the detection performance, analytical expressions for the probability density functions of the decision statistic are derived considering Energy Detection (ED) test as spectrum sensing method. It is shown that the time varying PU traffic severely affects the spectrum sensing performance. Most importantly, our results show that the performance gain due to multiple antennas in the sensing unit is significantly reduced by the effect of PU traffic when the mean lengths of free and busy periods are of the same order of magnitude of the sensing slot.