PhD defence of Asil Koc - Hybrid Beamforming Techniques in Full-Duplex/Half-Duplex Massive MIMO Wireless Communications
The massive multiple-input multiple-output (mMIMO) technology brings great potential for supporting various emerging applications by means of large antenna arrays, high beamforming gain, enhanced spectral efficiency, and coverage. Two-stage hybrid beamforming (HBF) has been proposed as a promising architecture for a reduced hardware cost/complexity in mMIMO systems. HBF employs a low number of radio-frequency (RF) chains to interconnect analog RF beamformer and digital baseband (BB) precoder/combiner. In this thesis, we propose novel HBF techniques for various mMIMO systems for half-duplex or full-duplex communications. We aim to address a set of versatile objectives, such as user grouping, beamforming optimization, interference mitigation, sum-rate maximization, fairness, energy efficiency, self-interference cancellation (SIC), and low channel estimation overhead.
First, we investigate the half-duplex mMIMO wireless communications. For the single-cell downlink transmission, we initially develop a user-grouping algorithm for efficient clustering in the multi-user mMIMO (MU-mMIMO) systems. Afterwards, the proposed RF beamformer design aims to jointly reduce hardware cost/complexity, maximize the beamforming gain in the intended direction, mitigate the inter-group interference, and lower the channel estimation overhead size. For this purpose, we build the RF beamformer via slow time-varying channel state information (CSI), i.e., angular parameters. Then, the BB-stage is constructed via the reduced-size effective instantaneous CSI. During the BB-stage design, in addition to equal power allocation, we propose to apply artificial intelligence (AI)-based solutions for sum-rate maximization and fairness objectives. According to the cloud radio access network (C-RAN), we also investigate cooperation strategies among the base stations in the multi-cell downlink transmission, where our primary objectives include mitigating the effect of inter-cell interference and enhancing the sum-rate capacity, especially for cell-edge users. Later on, we analyze the effect of low-resolution hardware components in HBF for the point-to-point mMIMO (P2P-mMIMO) systems.
Second, we explore the full-duplex mMIMO wireless communications. Full-duplex technology further extends the impacts of mMIMO systems by enabling simultaneous transmission/reception over the same frequency to theoretically double the spectral efficiency of the conventional half-duplex operation. However, strong self-interference is potentially the major limiting factor in the full-duplex operation. For this, we introduce HBF techniques for full-duplex P2P-mMIMO and MU-mMIMO systems. Along with the above-mentioned objectives in the RF beamformer design, we also aim to successfully cancel the strong self-interference by proposing and developing both orthogonal and non-orthogonal beamforming schemes. Illustrative results indicate that the proposed full-duplex HBF techniques achieve large SIC sufficient to double the capacity of their half-duplex counterparts.