Base Station Coordination in Multicell MIMO Networks

Sammanfattning: The use of multi-antenna technology, also referred to as multiple-input multipleoutput (MIMO), has been shown to improve both the achievable data rates and the link reliability in single-cell wireless systems without a need for extra power or bandwidth. The promised gains of MIMO techniques are, however, severely degraded in a multicell environment due to the presence of intercell interference, especially for users at the cell edge. One efficient technique to combat intercell interference is via exploiting coordination among multiple base stations, which is known as multicell processing or simply base station coordination. This thesis investigates the design and the performance of practically implementable base station coordination schemes. The main contribution of this thesis is to formally study different types of coordination, to develop analytical tools for their performance evaluation, and to propose simple algorithms for their implementation. First, we focus on the most complex form of coordination, namely the network MIMO. In this scheme all coordinating base stations share the data and the channel state information of all users, and act as a single distributed multi-antenna transmitter to serve them. We develop an analytical framework to facilitate the ergodic rate analysis of such a system under linear precoding. We also propose a simple scheduling algorithm, which only requires the knowledge of long-term channel statistics. In the next stage, we consider a simpler form of coordination in which the data of each user is served only by one base station. The scheduling and beamforming design, however, can be shared among the coordinating base stations. For this scheme, we propose a low-complexity joint user scheduling and beamforming strategy selection which requires a limited level of inter-base station information exchange, while providing significant performance improvement over non-coordinated systems. Finally, we investigate the effect of the antenna elevation tuning parameter, referred to as antenna tilt, on the performance of multicell multiple-input singleoutput (MISO) systems. We propose a framework in which multiple base stations jointly adjust their tilt angles based on the location of the scheduled users to maximize their sum throughput. We also provide an analytical expression for the sum throughput, enabling the decentralized implementation of the proposed scheme at each base station.