Gelecek Nesil Haberleşme Sistemlerinde İnsansız Hava Araçları için Kaynak Yönetim Teknikleri

Abstract

The use of unmanned aerial vehicles as flying base stations in 5G and beyond networks has emerged as a promising solution to some of the problems faced by the terrestrial communication network. These problems include blind spots, low coverage, low service quality and sustainable service. It is anticipated that unmanned aerial vehicles can cope with these problems thanks to their high mobility, adjustable heights, and easy and quick deployment. In particular, due to the high possibility of establishing a line-of-sight link, these vehicles are expected to provide a sustainable coverage in regions with service demand such as high data rate, high reliability and low latency etc. Despite these advantages of UAV communications, there are still problems which require further research on, e.g. low flight times, optimal three-dimensional deployment, route planning and energy-efficiency. The main goal of this dissertation is to examine resource management techniques in UAV communications in order to find solutions to these problems. The topic of resource management deals with the optimization of the resources that the flying access point (UxNB) consumes to fly and communicate while performing its task. Thus, unlike the common literature, this study brings a new perspective to low flight time, route planning, optimal 3D deployment, backhaul problems and performance analysis.

First, the power dissipated by a rotary-wing UAV to fly is mathematically analyzed using the momentum and blade element theories. The results obtained show that the power spent to fly can be nearly hundreds of times higher than the power spent for communication. Therefore, flight dynamics should also be included in communication problems.

The problem of energy-efficient 3D deployment for single and multiple UAVs is discussed in different context in different chapters. First, the deployment of a single UxNB is examined in a communication scenario under flight dynamics and quality of service constraints of the ground users. Next, the deployment problem of the multiple UAV base stations that maximizes the network endurance and the coverage density in a natural disaster scenario is examined where circle packing theory is utilized. Following that, performance analysis of the airborne communication network created with different types of UAVs is presented. For the solution of the deployment problems specific to each UAV type, an algorithm that gives the suboptimal result using K-means clustering and DC programming is proposed and with the help of this algorithm, a detailed comparison framework is provided. To the best of our knowledge, this comprehensive study is the first for the UAV communications field.

Although the idea of using UAV base stations in intelligent transportation systems, which is one of the smart city applications, has been proposed before in the literature, it is the first time that in this study they are used as flying roadside units in a V2X communication scenario. In a communication scenario with the latency and the backhaul constraints, UxNB establishes the vehicle-to-infrastructure (V2I) connection as the flying roadside unit. If UAVs are used to increase the data rate in the network, the backhaul capacity of the network poses a significant limitation. In order to investigate this, the thesis discusses the energy-efficient deployment problem in an enhanced mobile broadband communication scenario. Upon analytically showing that the cost function of this problem is unimodular, a heuristic algorithm is provided to obtain the solution.

UAVs, which will operate as flying base stations, forms one of the innovative approaches in 5G technology. Hence, a UxNB network should be able to implement network slicing as well. In the last chapter of the dissertation, a communication scenario with users having different service demands is discussed. A stochastic time-averaged problem is proposed to minimize the total power consumed by the UxNB. In the proposed problem, there are two network slices where one of these slices belongs to users demanding ultra-reliable low-latency communication, and the other one represents users demanding high data rate. In this scenario, an iterative algorithm is proposed in which dynamic radio resource allocation and UxNB's route planning are jointly handled.

In summary, the analytical and numerical foundations obtained in this dissertation constitute a general framework about the design and application areas of non-terrestrial networks created by unmanned aerial vehicles, which will have an important place in wireless communication in the future.