OPTIMIZED RESOURCE ALLOCATION AND SECURITY ENHANCEMENT IN DRONE-ASSISTED NETWORKS

Abstract

The rapid evolution of Open Radio Access Networks (O-RAN) has introduced a transformative architecture that promotes openness, flexibility, and cost efficiency in modern telecommunications. However, the open and multi-vendor nature of O-RAN presents significant challenges in resource allocation and security, particularly in dynamic and heterogeneous environments. This thesis addresses these challenges by developing an intelligent framework for secure and efficient resource allocation in O-RAN, leveraging Unmanned Aerial Vehicle (UAV) as assisting relays, and Reinforcement Learning (RL) as a technique for efficient mobile nodes management in such highly dynamic systems. We first tackle the problem of resource allocation in O-RAN by formulating a multi-objective optimization framework that balances efficiency, latency, and security. This framework dynamically adapts to real-time network conditions, ensuring optimal resource utilization while mitigating security risks such as eavesdropping and data breaches. By incorporating advanced encryption techniques, we demonstrate how resource allocation decisions can be optimized to enhance network performance and resilience in open and multi-vendor environments. Next, we extend this framework by integrating UAV-assisted relays, which act as mobile nodes to enhance network coverage and flexibility. UAVs are particularly valuable in scenarios where traditional infrastructure is unavailable or compromised, such as disaster-stricken areas or high-density events. We apply RL-based techniques to optimize UAV positioning and resource allocation, ensuring efficient and secure communication in dynamic environments. Additionally, we address the energy efficiency of UAVs by minimizing their mechanical energy consumption, enabling sustained operation without compromising network performance. We evaluate the effectiveness of our approach by conducting extensive experiments in simulated O-RAN environments. The results demonstrate that the proposed RL-based solution significantly improves resource allocation efficiency, security, and adaptability compared to traditional methods. Furthermore, the integration of UAV relays provides a scalable and flexible solution for extending network coverage and addressing connectivity challenges in complex environments.