Abstract

Unmanned aerial vehicles (UAVs) are becoming an integral part of numerous commercial and military applications. In many of these applications, the UAV is required to self-navigate in highly dynamic urban environments. Existing localization and trajectory planning techniques, which rely mainly on the Global Positioning System (GPS), do not provide an effective real-time solution for self-localization and path planning, particularly in dense urban environments. The purpose of this thesis is to study the localization and trajectory planning of UAVs independent of GPS systems or other detectable mobile signals. We propose to utilize the broadcast signals from existing cellular networks to localize and navigate the UAV from a given source to a destination. This simply implies that the drone needs to rely on the signals of the surrounding cellular base stations without having to interact with these base stations. The applications of this include, but are not limited to, mission-critical applications in which the unreliable GPS signal detection may compromise the mission. The use of AI-based techniques will provide near-optimal location and path determination, while providing a practical real-time calculation that is needed in such dynamic applications. For this purpose, we first address the cellular network-based autonomous 3-D UAV localization problem. Our objective is to propose an effective alternative solution to enable the UAV to autonomously determine its location independent of the GPS and without message exchanges. We formulate the UAV localization problem to minimize the error of the RSSI measurements from the surrounding cellular base stations. While exact optimization techniques can be applied to accurately solve such a problem, they cannot provide the real-time calculation that is needed in such dynamic applications. Machine-learning based techniques are strong candidates as an attractive alternative to provide a near-optimal localization solution with the needed practical real-time calculation. Accordingly, we propose two machine learning-based approaches, namely, deep neural network and reinforcement learning based approaches, to solve the formulated UAV localization problem in real time. We then provide a detailed comparative analysis for each of the proposed localization techniques along with a comparison with the optimization-based techniques as well as other techniques from the literature. Next, we address the autonomous UAV trajectory planning problem. For this purpose, we formulate the UAV trajectory planning problem as a joint objective optimization problem to minimize a composite cost metric that we also introduce. The computational complexity involved in exact optimization techniques hinders obtaining the real-time calculation requirement that is needed due to the dynamic nature of the UAV operation and the environment. To overcome this complexity, we utilize machine-learning based techniques to solve the formulated trajectory planning problem. Specifically, we propose two machine learning-based techniques, namely, reinforcement learning and the deep supervised learning-based approaches. We then analyze the performance of each of the proposed techniques as compared to the optimization-based approaches and other solutions from the literature.

Moreover, we propose a reliable cellular network supported trajectory planning solution independent of transmissible detectable signals and the GPS system under realistic channel propagation conditions assumptions. The reliability of the UAV trajectory planning solutions is crucial for mission critical applications. As such, we derive a UAV trajectory reliability model as pertains to UAV motion uncertainty along the trajectory. We then formulate the reliable UAV navigation problem as an optimization problem to maximize the probability of minimum error along the path of the UAV. We utilize conventional optimization methods to determine the optimal bound of the solution of the reliable UAV navigation problem. To navigate the UAV reliably and autonomously in real time, we propose a machine learning based technique. Specifically, we propose double deep Q-Learning (DDQN) as a framework to solve the formulated reliable UAV navigation problem. We provide an in-depth evaluation to assess the performance of our proposed DDQN-based solution as compared to the presented optimization methods and other representative techniques from the literature. Finally, we address the smoothness of our proposed trajectory planning approaches to realize a practical algorithm implementation. We provide a comparative evaluation of the set of algorithms that we proposed and demonstrate the recommended use case scenario for each of these trajectory planning solutions. Our simulation results show that our proposed machine learning-based approaches provide near optimal solutions to the formulated UAV localization and trajectory planning problems, with comparable accuracy to the optimal bound while meeting the real-time calculation requirement.

School

School of Sciences and Engineering

Department

Electronics & Communications Engineering Department

Degree Name

PhD in Engineering

Graduation Date

Winter 1-31-2024

Submission Date

7-11-2023

First Advisor

Yasser Gadallah

Committee Member 1

Karim Seddik

Committee Member 2

Tamer El Batt

Extent

188 p.

Document Type

Doctoral Dissertation

Institutional Review Board (IRB) Approval

Not necessary for this item

Download

Available for download on Friday, July 11, 2025

Share

COinS