Nagwa Shaghluf

Topic

Adaptive and Efficient Resource Allocation for Cognitive Radio Networks

Department of Electrical and Computer Engineering

Date & location

• Monday, June 22, 2026

• 11:00 A.M.

• Virtual Defence

Reviewers

Supervisory Committee

• Dr. T. Aaron Gulliver, Department of Electrical and Computer Engineering, UVic (Supervisor)

• Dr. Xiaodai Dong, Department of Electrical and Computer Engineering, UVic (Member)

• Dr. Andrew Rowe, Department of Mechanical Engineering, UVic (Outside Member) 

External Examiner

• Dr. Brent Petersen, Department of Electrical and Computer Engineering, University of New Brunswick 

Chair of Oral Examination

• Dr. Raad Rashmi, Department of Biology, UVic

   

Abstract

The increasing demand for wireless connectivity and data-intensive applications has intensified the need for efficient spectrum utilization in modern communication systems. Although radio spectrum resources are limited, their underutilization due to static allocation policies highlights the importance of dynamic and intelligent spectrum access mechanisms. Cognitive Radio Networks (CRNs) provide a promising solution by enabling Secondary Users (SUs) to opportunistically access underutilized spectrum while protecting Primary Users (PUs) from harmful interference. However, achieving efficient and scalable resource allocation in CRNs remains a significant challenge due to dynamic environments, interference coupling, and limited channel state information. 

This dissertation develops a unified framework for intelligent and adaptive resource allocation in CRNs by integrating predictive spectrum sensing, Multi-Agent Deep Reinforcement Learning (MADRL), and Non-Orthogonal Multiple Access (NOMA). First, a Predictive Cooperative Spectrum Sensing (PCSS) approach is proposed to enhance spectrum awareness using statistical and Machine Learning (ML) models. By combining prediction with cooperative decision-making, the framework improves Spectrum Efficiency (SE) while reducing sensing overhead and energy consumption. 

Building on this foundation, a MADRL-based framework is developed for Cognitive Device-to-Device (C-D2D) communication, enabling distributed and adaptive resource allocation through joint channel selection and power control. The problem is formulated as a mixed-integer nonlinear optimization task and addressed using a Proximal Policy Optimization (PPO)-based approach, allowing agents to learn efficient policies under dynamic channel conditions and imperfect information. 

Finally, the framework is extended by integrating NOMA into C-D2D networks, resulting in a unified multi-agent learning approach for joint optimization of spectrum reuse, power allocation, interference management, and fairness. The proposed method effectively addresses the challenges of highly coupled and non-convex optimization problems while maintaining scalability under realistic constraints. 

Simulation results demonstrate that the proposed framework significantly improves spectrum efficiency, energy efficiency, system throughput, and fairness compared to conventional optimization and learning-based methods. Overall, this work provides a scalable and adaptive solution for next-generation wireless networks and contributes to the development of intelligent Cognitive Radio (CR) systems capable of autonomous decision-making in complex and dynamic environments.