Beamforming and Power Allocation in Overlay Cognitive Radio Network With Imperfect CSI

Abstract

In this paper, we investigate an overlay cognitive radio network comprising two primary users (PUs), K secondary users (SUs) and one secondary base station (SBS) equipped with a massive antenna array. The SBS helps the PUs to exchange information and at the same time serves the SUs by employing the beamforming technique. Assuming that the SBS has imperfect channel state information (CSI) of the primary links, we jointly optimize the beamforming vectors for the PUs and the power allocation to SUs by maximizing the minimum of the rates of the SUs under the constraint that a target rate is satisfied for the PUs. We utilize the semi-definite relaxation (SDR) to transform the problem to a solvable semi-definite program (SDP). Using the bisection search method and the randomization technique, we present an algorithm to achieve the solution of the optimization problem. From the numerical results, we conclude that as the number of antennas at the SBS increases (i.e. the massive scenario), the harmful effect of CSI error and the interfering effect of PUs and SUs on each other decrease.