Joint beamforming and power control for throughput maximization in cognitive radio

Abstract

Joint beamforming and power control has been introduced into the cognitve radio network (CN) as a quite promising interference suppression technique in [1]–[4]. In this paper, subject to interference constraints of primary users (PU) as well as peak transmission power constraints of cognitive users (CU), the total throughput maximization of CN based on joint beamforming and power control is studied. In [4], a joint beamforming and power control algorithm for total throughput maximization in CN has been proposed, we note it as ZF-CML. ZF-CML is asymptotically optimal especially in both low and high SNR regimes. However, in ZF-CML, the number of active CU is required not to exceed the number of antenna at the cognitive base station (CBS), besides, the interference from primary network (PN) to CN has not been considered. The two limitations are ascribed to the ZF-DFE beamforming and may significantly deteriorate the throughput and fairness of CN. Therefore, in this paper, we replace ZF-DFE by MMSE-DFE beamforming, which is a SINR maximization algorithm and does not limit to the number of active users. However, with MMSE-DFE beamforming, there may be remnant inter-interference among CUs, which calls for new power control algorithm. When the inter-channel interference exists, iterative water-filling (IWF) <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">[5]</sup> is the representative power control algorithm, but IWF is usually not optimal for the total throughput of CN. In [6] and [7], an improved IWF power control algorithm has been proposed with the object of total throughput maximization, but it is proposed in non-CN scenario and can not be implemented directly in CN. Inspired by these considerations, in this paper, a best response (BR) based iterative power control algorithm is proposed, which could maximize the total throughput of CN. Our proposed joint beamforming and power control algorithm is implemented by the alternate iteration of MMSE-DFE beamforming and BR based power control, we note it as MMSE-BR. Compared with ZF-CML, MMSE-BR could achieve the optimal throughput performance under any interference situation, and does not limit to the number of CU, so is much more robust and practical.