User scheduling and beamforming design for millimeter wave MIMO communications

Abstract

Millimeter wave (mmWave) beamforming (BF) technique, which uses a large bandwidth, is expected to play a key role in 5G and beyond-5G mobile communications. However, the use of a large number of antennas to compensate the high path-loss of high-frequency channels causes several problems in practice. First, a lot of radio frequency (RF) chains are needed as many as antennas, which requires considerable implementation cost and power consumption. To address this, digital-analog hybrid BF technique, which significantly reduces the number of RF chains by introducing analog BF, should be used. Second, although the mmWave BF systems are expected to increase the areal capacity in ultra-dense network (UDN), the feedback overhead required for channel information exchange may be problematic. In particular, the size of the channel matrix increases depending on the number of antennas, and the amount of resources required for feedback increases along with the number of users or base stations, resulting in a significant overhead which degrades the throughput. This dissertation deals with the BF systems with limited feedback information to reduce implementation complexity and feedback overhead. Also, since all the users in UDN cannot be simultaneously supported by a single transmission, we propose a joint design of user scheduling and hybrid BF, and address different network scenarios. In addition, the performances of the proposed schemes are analyzed for each scenario, and the design guidelines for choosing the system parameters are provided.