Doctor of Philosophy
School of Electrical, Computer & Telecommunications Engineering
Large-scale multiple-input multiple-output (MIMO) is a promising technology for nextgeneration wireless communication as it can offer a high data rate, high link reliability, and better quality-of-service (QoS). In the meantime, due to the heavy traffic in the sub-6 GHz band which has been used for almost all the wireless communications nowadays, the millimeter wave (mmWave) band ranging from 30-300 GHz is explored to support the vast amount of connectivity required by next-generation wireless communications. Therefore, large-scaleMIMOandmmWave communications are two of the most important techniques for future wireless communications. To make the best advantage of large-scale MIMO systems, the transmitter or receiver should obtain good enough channel state information (CSI) to perform precoding and other operations. The transmitter and receiver in largescale MIMO systems are equipped with antenna arrays with tens or hundreds of antennas, causing the size of the channel larger than that in conventional MIMO systems. Traditional training pilots, which are mutually orthogonal between the antennas, for the conventional MIMO systems may be no longer feasible for large-scale MIMO systems as this training scheme requires more training resources for a larger channel. Meanwhile, among the many types of large-scale MIMO systems, the hybrid systems with a reduced number of radio frequency (RF) chains are suggested for mmWave communications as they are more energy-efficient. In hybrid systems, obtaining a large number of samples is even more time consuming because of the reduced number of RF chains. Therefore, for large-scale MIMO systems, CSI acquisition techniques with low training overhead are required. Moreover, large-scale MIMO systems can be more vulnerable to hardware imperfections as they are more complicated than conventional MIMO systems. For this reason, possible hardware imperfections should be considered when developing CSI acquisition techniques.
In this thesis, we consider CSI acquisition for large-scale MIMO systems with reduced training overhead. The CSI considered in this thesis includes the channel matrix and channel covariance matrix. The work reported in this thesis contains the design of covariance matrix estimation methods that can be applied to channel estimation for fully digital largescale MIMO systems, the design of channel estimator for switch-based hybrid large-scale MIMO systems and the design of channel estimator as well as channel covariance matrix estimator for phase shifter-based fully connected hybrid large-scale MIMO systems.
Hu, Rui, Channel State Information Acquisition for Large-Scale MIMO Systems with Low Training Overhead, Doctor of Philosophy thesis, School of Electrical, Computer & Telecommunications Engineering, University of Wollongong, 2019. https://ro.uow.edu.au/theses1/726
Unless otherwise indicated, the views expressed in this thesis are those of the author and do not necessarily represent the views of the University of Wollongong.