Wireless communication networks are emerging fast with a lot of challenges and ambitions. Requirements that are expected to be delivered by modern wireless networks are complex, multi-dimensional, and sometimes contradicting. In this thesis, we investigate several types of emerging wireless networks and tackle some challenges of these various networks. We focus on three main challenges. Those are Resource Optimization, Network Management, and Cyber Security. We present multiple views of these three aspects and propose solutions to probable scenarios. The first challenge (Resource Optimization) is studied in Wireless Powered Communication Networks (WPCNs). WPCNs are considered a very promising approach towards sustainable, self-sufficient wireless sensor networks. We consider a WPCN with Non-Orthogonal Multiple Access (NOMA) and study two decoding schemes aiming for optimizing the performance with and without interference cancellation. This leads to solving convex and non-convex optimization problems. The second challenge (Network Management) is studied for cellular networks and handled using Machine Learning (ML). Two scenarios are considered. First, we target energy conservation. We propose an ML-based approach to turn Multiple Input Multiple Output (MIMO) technology on/off depending on certain criteria. Turning off MIMO can save considerable energy of the total site consumption. To control enabling and disabling MIMO, a Neural Network (NN) based approach is used. It learns some network features and decides whether the site can achieve satisfactory performance with MIMO off or not. In the second scenario, we take a deeper look into the cellular network aiming for more control over the network features. We propose a Reinforcement Learning-based approach to control three features of the network (relative CIOs, transmission power, and MIMO feature). The proposed approach delivers a stable state of the cellular network and enables the network to self-heal after any change or disturbance in the surroundings. In the third challenge (Cyber Security), we propose an NN-based approach with the target of detecting False Data Injection (FDI) in industrial data. FDI attacks corrupt sensor measurements to deceive the industrial platform. The proposed approach uses an Autoencoder (AE) for FDI detection. In addition, a Denoising AE (DAE) is used to clean the corrupted data for further processing.


School of Sciences and Engineering


Electronics & Communications Engineering Department

Degree Name

PhD in Engineering

Graduation Date

Fall 12-2021

Submission Date


First Advisor

Karim Seddik

Committee Member 1

Yasser Gadallah

Committee Member 2

Tamer Elbatt


141 p.

Document Type

Doctoral Dissertation

Institutional Review Board (IRB) Approval

Not necessary for this item