Throughput analysis of multi-antenna cognitive broadcast networks

In this study, the authors analyse the throughput of random beamforming (RBF) for a multi-antenna cognitive radio (CR) broadcast network. Using extreme value theory, the asymptotic average throughput of a single-beam RBF and a multiple-beam RBF are derived from the limiting distribution of the sample maximum of received signal-to-noise ratio (signal-to-interference-plus-noise ratio) with maximal ratio combiner at the secondary user. It is shown that the throughput of the single-beam RBF scales as logarithm of the number of users, which stands in contrast to the double logarithm of the number of users in non-CR networks. They also found that the single-beam RBF is always much preferable to the multi-beam RBF on the throughput in the CR networks, which is opposite to the previous results in the non-CR networks. Simulation results show that the author´s approximated expressions match with the simulation results very well.