Beamforming design and power control for spectrum sharing systems

Abstract

In order to provide wireless services for the current demand of high data rate mobile applications, more spectrally efficient systems are needed. As a matter of fact, the current wireless systems are limited by a frequency splitting spectrum management which in one hand minimizes the multiuser interference but; on the other hand, it precludes the use of wider bandwidth signals. As a more aggressive frequency reuse is targeted (ideally, all transmitters might eventually share the same frequency band), the use of multiple antennas for interference reliving, jointly with a smart power allocation is compulsory. In addition, novel spectrum management regulatory policies are required for ensuring a peaceful coexistence between adjacent spectrum sharing networks and for promoting their development. The aim of this dissertation is provide a beamforming and power allocation design for these novel spectrum sharing systems which are meant to exponentially increase the spectral efficiency of the systems. A mathematical framework based on multicriteria optimization for analyzing the beamforming design is provided which serves as a fundamental tool for describing the state-of-the-art studies in multiantenna interference networks. Indeed, the achievable rates are described and several ways of computing the Pareto rate region of MISO interference channel (i.e. the communication model that represents the spectrum sharing network when the transmitters use multiple antennas) are studied. Nevertheless, as the system designer aims to work in a single efficient rate point, the sum-rate optimal beamforming design is studied. Curiously, it results that under some realistic assumptions on both the desired and interference power levels, the obtained beamformer is the reciprocal version of a known receiving one and it optimizes a notion of antenna directivity for multiuser communications. Neverthelss, it is important to remark that the higher transmit power is used, the more interference dominated is the medium, not only within the wireless network, but also to eventually adjacent networks that might suffer from inter-network interference. In order to cope with this problem, a spectrum licensing system is revisited, namely time-area-spectrum license. Under this spectrum management mechanism, a license holder is able to radiate signals under a certain portion of time, within a concrete area and in a given band. Moreover, the amount of signal strength within the area is constraint by a certain value. Since controlling the signal power levels in a given area is cumbersome, we propose to restrict the receive power as an estimation of the overall accumulated signal strength. Therefore, the optimal transmit beamformers and power allocations are studied. Concretely, the achievable rates are derived and an operational working point is envisaged. In addition, a suboptimal yet low computationally complex and decentralized beamforming design is presented and it shows a good performance in front of other decentralized designs.