Physical Layer Techniques for Massive MIMO Sub-6 GHz LoS and Millimetre-Wave Transmission

The explosive growth in data demand requires solutions with higher system throughput, lower energy consumption, and simultaneous support for many users. Massive multiple-input multiple-output (MIMO) and millimetre-wave (mmWave) techniques are promising candidates for next-generation wireless systems.
This thesis focuses on sub-6 GHz line-of-sight (LoS) transmissions in massive MIMO systems, which not only fulfil a variety of applications, such as small-cell back-haul but also provide a longer coherent time as the LoS channel varies more slowly and can be readily estimated compared with fading channels. This thesis also focuses on mmWave transmissions in massive MIMO systems since a large-scale antenna array can compensate for the strong pathloss of mmWave transmissions whilst the mmWave carrier frequencies enable compact BS configurations.
In this thesis, the fundamentals of the massive MIMO technique are studied comprehensively through theoretical analysis and simulations. The representative sub-6 GHz channel models of LoS and fading channels are considered. The characteristics of the LoS channel and the system performance of LoS transmissions are investigated and compared with fading channels along with the key factors that impact performance. The effective SINR expressions of the linear precoding schemes for LoS transmissions are presented. It is illustrated that the system performance of massive MIMO LoS transmissions is robust when the angles of departure are distributed within a wide range and the power of the LoS channel component is high.
The mmWave channel model and technical challenges are studied. The mmWave massive MIMO precoding problem is transformed into a beam-selecting problem. A novel channel deconstruction algorithm is proposed that enables the estimation of each received paths’ parameters from the perfect or Gaussian-perturbed channel state information. Utilising the estimated path parameters, new analogue and hybrid beam-selecting (ABS and HBS) linear precoding schemes are proposed that contribute substantially to system performance. The corresponding hardware architectures for the proposed schemes are demonstrated, which exploit low-complexity and low-cost signal processing with high energy efficiency. An enhanced hybrid beam-selecting precoding (E-HBS) scheme and hardware configuration are further proposed to achieve the optimal and near-optimal performance of digital baseband signal processing with low cost and high energy efficiency in massive MIMO systems. With E-HBS, the number of RF chains and the dimension of the baseband digital control is independent of the number of base station antennas, which is vital for massive MIMO systems.
Novel spatial user scheduling (SUS) schemes for sub-6 GHz LoS massive MIMO transmissions are proposed along with a capacity-enhancement check (CEC) scheme to further improve the system performance by mitigating the LoS channel cross-correlation.