| dc.description.abstract | The wide spectrum available in the millimetre-wave band is key to provide enhanced capacity for the demands of the fifth-generation (5G) of cellular networks.
However, the usage of millimetre-wave frequencies introduces a series of challenges to the system design and performance.
The millimetre-wave signal does not propagate well from outdoor to indoor environments. Thus independent deployments are required to provide coverage in indoor venues, making indoor deployments an important use case for 5G networks.
Moreover, the millimetre-wave propagation is easily obstructed by human bodies, which are ubiquitous in the mobile broadband environment and therefore impose a significant problem for 5G networks.
Therefore, the goal of this thesis is to study the impact of human blockage on the design and performance of indoor millimetre-wave networks.
We test system parameters such as deployment density, antenna beamwidth and transmission frame, aiming to find design solutions to effectively cope with adverse blockage effects on the millimetre-wave communication channel.
Firstly, we address the problem of human blockage in provisioning coverage in indoor networks. Our work focuses on solutions in the network deployment design to mitigate the blockage effects, using highly-dense ceiling-mounted access points with fixed directional antennas facing downward.
Highly-dense network deployments may lead to excessive interference from neighbouring cells, causing deterioration of the system performance.
However, the usage of this ceiling-mounted setup may reduce the interference by illuminating selected spots on the ground with the use of fixed directional antennas, confining the signal energy.
The effectiveness of this approach on millimetre-wave networks is not clear in the literature as there is a lack of investigation on this type of setup when considering the blockage effects in the signal propagation.
Thus, in our first research question, we study the impact of blockage on the performance and design of indoor millimetre-wave networks with access points installed on the ceiling.
We develop a body blockage model and derive an analytical expression for the blockage probability, which allows us to evaluate the system for a set of body blockage scenarios.
We also develop a simulation framework to test design parameters, such as the deployment density and the antenna beamwidth, to find the best configuration to cope with the human blockage problem.
We find that there exists a set of optimal density-beamwidth configurations for each blockage scenario, and there exists a trade-off in this configuration, in which a certain configuration optimises either coverage or spectral efficiency, but not both. The optimal configuration takes different values depending on the severity of the blockage scenario.
Furthermore, we develop an analytical framework to evaluate the expected data rate and test the design of transmission frame parameters given by recent 5G standardisation such as the transmission time interval.
We show that there is a trade-off between the probability of blockage and the transmission efficiency when choosing the best transmission time interval for different blockage scenarios.
Therefore, the system design configuration should carefully consider the intended scenario since each scenario needs a different configuration in terms of deployment density, antenna beamwidth and transmission interval to achieve the best system performance.
Secondly, we address the problem of human blockage in the efficiency of transmission resource allocation. In this second research question, we investigate how resource allocation can help mitigate the blockage effects on transmission efficiency.
Most of the existing efforts on resource allocation techniques aiming to mitigate blockage effects consist in handing-off the communication to another spatial resource (other cell or antenna beam direction).
However, the hand-off procedure has an inherent re-connection latency and overhead that may degrade the efficiency of the blockage mitigation.
Moreover, another cell or beam might not always be available, and thus other resource allocation techniques that do not depend on other spatial resources should be investigated.
We focus our study on the resource allocation provided by scheduling algorithms.
First, we analyse how conventional algorithms, such as the Proportional Fair scheduler, behave under blockage events.
We show that such algorithms disfavour the allocation for users suffering from blockage, leading to poor data rate performance.
Then, we show that to improve the performance of such users, the system needs a proactive scheduling approach aided by blockage prediction mechanisms.
Hence, the scheduler can preemptively allocate the transmission resources before a blockage event and release them after that.
We conclude that a resource allocation approach that allows for adaptive scheduling intervals can help mitigate blockage, and has to consider the specific blockage scenario and the predicted blockage event to be effective. | en |
