Document Type


Date of Award


Degree Name

Doctor of Philosophy in Electrical Engineering - (Ph.D.)


Electrical and Computer Engineering

First Advisor

Nirwan Ansari

Second Advisor

Roberto Rojas-Cessa

Third Advisor

Abdallah Khreishah

Fourth Advisor

Qing Gary Liu

Fifth Advisor

Jing Li


In the past few years, unmanned aerial vehicle (UAV)-assisted heterogeneous network has attracted significant attention due to its wide range of applications, such as disaster rescue and recovery, ground macro base station (MBS) traffic offloading, communications for temporary events, and data collection for further processing in Internet of Things (IoT). A UAV can act as a flying base station (BS) to quickly recover the communication coverage in the disaster area when the regular terrestrial infrastructure is malfunctioned. The UAV-assisted heterogeneous network can effectively provision line of sight (LoS) communication links and therefore can mitigate potential signal shadowing and blockage. The regulation relaxation and cost reduction of UAVs as well as communication equipment miniaturization make the practical deployment of highly mobile wireless relays more feasible than before. In fact, the 3GPP Rel-16 has included UAV-enabled wireless communications in the new radio standard, aiming to boost capacity and coverage of fifth generation (5G) wireless networks. However, the performance of UAV-assisted communications is greatly affected by the resource allocation scheme, user association policy and the UAV placement strategy. Also, the limited on-board energy and flight time of the UAV poses a great challenge on designing a robust and reliable UAV-enabled IoT network.

To maximize the throughput in the UAV-assisted mobile access network, an optimization problem which determines the 3D UAV deployment and resource allocation in a given hotspot area under the constraints of user Quality of Service (QoS) requirements and total available resources is formulated. First, the primal problem is decomposed into two subproblems, i.e., the 3D UAV placement problem and the resource allocation problem. Second, a cyclic iterative algorithm which solves the two sub-problems separately and uses the output of one as the input of the other is proposed.

An optimization problem that aims to minimize the average latency ratio of all users is formulated by determining the 3D location of the UAV, the user association and the bandwidth allocation policy between the MBS and the drone base station (DBS) with the constraint of each user’s QoS requirement and total available bandwidth. The formulated problem is a mixed integer non-convex optimization problem, a very challenging and difficult problem. To make formulated problem tractable, it is decomposed into two subproblems, i.e., the user association and bandwidth allocation problem and the 3D DBS placement problem. These two subproblems are alternatively optimized until no performance improvement can be further achieved.

To address the challenge of limited on-board battery capacity and flight time, a tethered UAV (TUAV)-assisted heterogeneous network where the aerial UAV is connected with a ground charging station (GCS) through a tether is proposed. The objective of the formulated problem is to maximize the sum rate of all users by jointly optimizing the user association, resource allocation and placement of the GCSs and the aerial UAVs, constrained by each user’s QoS requirement and the total available resource. Since the primal problem is highly non-convex and non-linear and thus challenging to solve, it is decomposed into three subproblems, i.e., the TUAV placement problem, the resource allocation problem and the user association problem. Then, the three sub-problems are alternately and iteratively optimized by using the outputs of the first two as the input for the third.

The future work comprises two parts. First, IoT devices usually are generally deployed at remote areas with limited battery capacities and computing power. Therefore, the generated data needs to be offloaded to a more powerful computing server for further processing. Unfortunately, the trajectory design in UAV data collection is generally NP-hard and difficult to obtain the optimal solution. Advances of machine learning (ML) provide a promising alternative approach to solve such problems that cannot be solved by traditional optimization methods. Hence, deep reinforcement learning (DRL) is proposed to be explored to obtain a near optimal solution. Second, the low earth orbit (LEO) satellite networks will revolutionize traditional communication networks with their promising benefits of service continuity, wide-area coverage, and availability for critical communications and emerging applications. However, the integration of LEO satellite networks and terrestrial networks will be another future research endeavor.



To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.