Multi-Antenna Techniques for Next Generation Cellular Communications

Future cellular communications are expected to offer substantial improvements for the pre- existing mobile services with higher data rates and lower latency as well as pioneer new types of applications that must comply with strict demands from a wider range of user types. All of these tasks require utmost efficiency in the use of spectral resources. Deploying multiple antennas introduces an additional signal dimension to wireless data transmissions, which provides a significant alternative solution against the plateauing capacity issue of the limited available spectrum. Multi-antenna techniques and the associated key enabling technologies possess unquestionable potential to play a key role in the evolution of next generation cellular systems. Spectral efficiency can be improved on downlink by concurrently serving multiple users with high-rate data connections on shared resources. In this thesis optimized multi-user multi-input multi-output (MIMO) transmissions are investigated on downlink from both filter design and resource allocation/assignment points of view. Regarding filter design, a joint baseband processing method is proposed specifically for high signal-to-noise ratio (SNR) conditions, where the necessary signaling overhead can be compensated for. Regarding resource scheduling, greedy- and genetic-based algorithms are proposed that demand lower complexity with large number of resource blocks relative to prior implementations. Channel estimation techniques are investigated for massive MIMO technology. In case of channel reciprocity, this thesis proposes an overhead reduction scheme for the signaling of user channel state information (CSI) feedback during a relative antenna calibration. In addition, a multi-cell coordination method is proposed for subspace-based blind estimators on uplink, which can be implicitly translated to downlink CSI in the presence of ideal reciprocity. Regarding non-reciprocal channels, a novel estimation technique is proposed based on reconstructing full downlink CSI from a select number of dominant propagation paths. The proposed method offers drastic compressions in user feedback reports and requires much simpler downlink training processes. Full-duplex technology can provide up to twice the spectral efficiency of conventional resource divisions. This thesis considers a full-duplex two-hop link with a MIMO relay and investigates mitigation techniques against the inherent loop-interference. Spatial-domain suppression schemes are developed for the optimization of full-duplex MIMO relaying in a coverage extension scenario on downlink. The proposed methods are demonstrated to generate data rates that closely approximate their global bounds