Modeling and Analysis of Hybrid Aerial-Terrestrial Networks: A Stochastic Geometry Approach

Abstract

The ever-increasing demand for better mobile experiences is propelling the research communities to look ahead at how future networks can be geared up to meet such demands. It is likely that the next-generation of wireless communications will be revolutionary, outpacing the current systems capabilities in terms connectivity, reliability and intelligence. These trends and predictions will cause a revolutionary change in the wireless communications. In this context, the concept of Ultra-Dense Network (UDN) is poised to be the cornerstone of the development of fifth generation(5G) systems, whereby a massive number of base stations (BSs) are deployed for enhancing the network performance metrics. Though such densification might be economically viable in urban areas, it is mostly unfavorable in rural ones due to the sheer complexity and the various factors involved the planning and installation processes; all of which trigger the need for cost-effective, flexible and easily-implementable solutions. As a result, unmanned aerial vehicles (UAVs) emerge as a promising alternative solution for enhancing wireless coverage. Due to their mobility capabilities, UAVs are of particular importance in events of (i) terrestrial-based cellular systems dilapidation, (ii) infrastructure absence in remote and suburban areas, or (iii) limited-duration events or activities wherein there is a short-term need for supplementary network resources to handle the overload. While a growing body literature works towards characterizing and providing insights into the performance of UAVs-only networks (serving the first two purposes), understanding the performance of such networks when coupled with existing terrestrial BSs remains a challenging, yet interesting, open research venue. Towards this direction, this thesis provides a rigorous analysis of the downlink coverage probability of hybrid aerial-terrestrial networks using tools from Stochastic Geometry. The thesis presents a mathematical model that characterizes the coverage probability metric under different network environments. The proposed model is validated against intensive simulations so as to substantiate the analytical results. The developed work is essential to understanding the premises of one possible solution to the UDNs of tomorrow, capture its key performance metrics and, most importantly, to uncover key design insights and reveal new directions for the wireless communication industry.