The Doctoral School in Science and Engineering is happy to invite you to Asad MAHMOOD’s defence entitled
Communication Technologies for UAV-Assisted 5G and Beyond Wireless Networks
Supervisor: Prof. Dr Björn OTTERSTEN
Recent advances in sixth-generation (6G) communication systems mark a critical evolution towards widespread connectivity and ultra-reliable, low-latency communication for numerous devices engaged in real-time data collection. Unmanned aerial vehicles (UAVs) have emerged as a practical solution in this context, offering on-demand wireless services and flexible deployment for critical applications such as enhancing coverage and capacity and supporting the internet of things (IoT). Crucially, UAVs are invaluable in emergency scenarios, disaster relief, and regions lacking infrastructure by enabling rapid, efficient, cost-effective deployment for search, rescue, and information dissemination. Despite their benefits, UAV-assisted networks face significant challenges including optimal 3D placement for static users, accurate channel modeling, and the limitations imposed by finite battery life affecting flight duration. For mobile users, these networks must carefully plan UAV trajectories and manage interference to ensure continuous, robust service. This thesis presents innovative design strategies that refine UAV trajectory, resource allocation, and user-to-UAV association while integrating groundbreaking technologies like intelligent reconfigurable surfaces (IRS) within mobile edge computing (MEC) networks to boost connectivity and service quality across diverse environments.
Firstly, we investigate an orthogonal frequency division multiple access (OFDMA)-enabled multi-UAV system to optimize 3D UAV placement, user association, and radio resource allocation. By decoupling the problem into $3$D placement/user association and sum-rate maximization subproblems, the proposed iterative algorithm achieves efficient resource utilization, reduces inter-cell interference, and enhances sum-rate performance.
Secondly, to improve connectivity in urban and remote areas, we introduce Beyond Diagonal IRS (BD-IRS) integrated with UAVs in MEC networks, referred to as BD-IRS-UAV. This system enables remote users to offload tasks to MEC servers, addressing both battery and resource constraints. Joint optimization of BD-IRS-UAV deployment, computational resource allocation, and communication resources yields substantial performance gains, with a $17.75%$ rate increase over traditional IRS designs and up to 25.43% improvement over fixed IRS placements. Moreover, the proposed approach achieves an $8.09\%$ reduction in worst-case latency over binary offloading and $16.92\%$ over fixed resource allocations, enhancing system responsiveness.
Lastly, we develop a predictive dynamic 3D trajectory design for UAVs, utilizing a Markov chain-based model to estimate user locations with minimal computational overhead accurately. Clustering users by location and demand enables real-time trajectory adjustments, maximizing minimum data rates through optimized resource allocation and adaptive user association. Simulation results reveal notable improvements in path loss reduction, fairness, and system robustness, significantly outperforming static and non-predictive models. The proposed approach ensures zero outages up to 30% faster than conventional methods and converges within 10 iterations, offering a scalable, resilient framework for UAV-assisted networks in complex and dynamic environments.