Optimal Beamforming Strategies At Mmwaves For 5G+ Networks

Beyond fifth generation (5G+) communication aims to provide a 3-dimensional ubiquitous network to support high data rates caused by device proliferation. To aid 5G+ communications, three spectrum bands are being investigated: sub-6 GHz, millimeterwave (mmWave), and terahertz frequencies. Higher frequency range has the advantage of increased bandwidth availability, but it also has its own set of challenges such as increased signal attenuation, reduced cell range, increased computational and hardware complexities, and increased hardware cost. However, substantial research has been conducted to demonstrate the feasibility of communication at higher frequencies. This dissertation focuses specifically on mWave-enabled communications for serving users under different 2-dimensional as well as 3-dimensional network. The number of radio frequency (RF) chains that can be deployed at a device is a bottleneck in the mmWave range due to high hardware power consumption and hardware complexity. This, in turn, presents challenges in designing precoders and combiners, parameter estimation, signal processing, and user scheduling. Therefore, the objective of this dissertation is to investigate and design energy and spectrally efficient reduced complexity terrestrial and aerial mmWave communications frameworks/architectures with a limited number of RF chains. The first topic investigates sectored-cell framework for a 2-dimensional multi-user terrestrial mmWave communications when the user population exceeds the number of RF chains available at the gNodeB. This framework is analyzed in the context of a single RF chain that serves all sectors of a cell in a round-robin manner with equal sector sojourn time. The idea is to partition the cell into identical sectors and serve multiple users simultaneously that fall within each sector at a given epoch using a single steerable beam generated from a uniform linear antenna array coupled to an RF chain at the gNodeB. Consequently, a large number of users are served with a limited number of RF chains. An optimization problem is formulated for combined resource allocation of orthogonal frequency division multiple (OFDM) symbols to users in a sector and sector beamwidth optimization in order to maximize average long-term user rate and system energy efficiency. Through simulation results, it is verified that serving multiple users over orthogonal frequency division multiple access with a single RF chain a sectored cell system model achieves better performance than serving single user per RF chain at a time. Additionally, the effect of localization error on the optimum beamwidth, which results from position estimation, is also studied. The second part of the dissertation extends the study of the sectored-cell framework with a single RF chain to the case of multiple RF chains at gNodeB. Because of its low complexity, the sub-array based or partially connected hybrid precoder is considered, in which each RF chain, connected to a separate antenna array, generates one steerable beam. Notably, the presence of sidelobes causes inter-beam interference when concurrent beams are generated from multiple RF chains. Therefore, the optimal beamwidth is estimated while accounting for inter-beam interference and beam squint. Furthermore, an optimal number of RF chains at the gNodeB is estimated by accounting for power waste in RF units. A variable time frame structure for the sectored-cell framework is also proposed for a standalone mmWave communications system with variable transmission time interval units as short as one OFDM symbol long. The frame structure for enhanced mobile broadband (eMBB) services is typically made up of slots with a fixed number of OFDM symbols, and the smallest transmission time interval unit is equivalent to one slot duration, as defined by the 3rd Generation Partnership Project (3GPP) New Radio (NR) in Release 15. Furthermore, 3GPP guidelines have suggested a separate beam management phase for the initial gNodeB-user beam pair to establish the best gNodeB-user beam pair to be used for subsequent data transfer in data transmission mode. During the beam management phase, narrow beams must scan the area multiple times before granting users channel access. As a result, using a fixed frame structure along with separate beam management and data transmission phase causes significant initial access delays. Furthermore, because beam training overhead is often low for beam search operations, the wideband mmWave channel is underutilized during the beam training phase. Therefore, a modified sectorwise initial access procedure is proposed, which offers decreased initial access delay and increased bandwidth utilization. The proposed variable time frame structure, along with the modified initial access procedure, offers an improved average and the geometric mean of long-run user rates, especially in the case of non-homogeneous user distribution in the area. The third part of the dissertation studies in detail the effect of beam squint in wideband mmWave communication. The beam squint effect is caused by the use of a large number of antenna elements connected per RF chain to generate a narrow beam in order to overcome high attenuation at mmWaves. The direction of maximum beam gain in beam squint varies with frequency. Based on this observation, a new reduced complexity joint OFDM resource allocation and beamforming strategy is proposed, which employs beam squint to maximize system performance using sub-array hybrid precoder at the transmitter to serve a clustered user population greater than the number of RF chains. Numerical results demonstrate that the proposed sequence for designing RF precoder, baseband precoder, and subcarrier allocation using beam squint provides a greater increase in spectral efficiency at mmWaves than at sub-6 GHz. The fourth part of the dissertation explores the feasibility of using an unmanned aerial vehicle (UAV) as a fronthaul unit at mmWaves and contrasts its performance with that at the sub-6 GHz range for ad hoc communication. The ideal user grouping method in a 3-dimensional environment with UAV-assisted mmWave communications is investigated for a system with a user population much higher than the number of RF chains available at the UAV. Similar to the analysis of multiple RF chains in terrestrial communication, multiple users per beam are served employing OFDMA. The optimal number of RF chains at the UAV is estimated to maximize the sum rate while satisfying minimum user rate specifications with a given UAV power constraint. Since UAV performance is constrained by battery size, so it is assumed that the UAV is equipped with a solar panel to harvest solar energy in order to increase operational hours. The effect of additional weight because of the solar panel on the performance of the UAV is also analyzed. Finally, the dissertation's last part studies the feasibility of using an existing backscatter device infrastructure in an indoor environment to provide data support to an obstructed user at mmWaves. When in idle mode, the backscatter device is used to reflect the incoming signal to desired direction without modulating the data stream. This analogous to a distributed reconfigurable intelligent surface (RIS) system in which the tags acts as distributed RIS in indoor environment. Both users and backscatter devices in the mmWave range will have multiple antenna elements, increasing the system’s complexity. The challenge, however, is to estimate the angle’s direction using a single RF chain at the user. Therefore, a link establishment procedure is proposed that includes estimation of angle of arrival using the retro-reflective property of the antenna array, beamforming design at the user, and beamforming at the backscatter devices.