Realistic Optimization Framework for UAV-BS Deployment in Integrated Aerial-Terrestrial Networks

In mobile network operations, managing Key Performance Indicators (KPIs), like user data rates, to ensure satisfactory Quality of Service ($\mathbf{Q o S}$) is crucial. Terrestrial cellular networks often struggle to deliver adequate data rates in remote regions with insufficient infrastructure or in densely populated urban areas with high user concentration. This presents a significant challenge in maintaining high-quality network service across diverse environments. Non-Terrestrial Networks (NTNs), utilizing Unmanned Aerial Vehicle Base Stations (UAV-BSs), with their ability to move and be deployed flexibly, present a viable solution for enhancing data rates and improving overall network performance. Despite their potential, the challenge lies in effectively positioning UAV-BSs within these NTNs to maximize data transmission rates while managing operational constraints such as battery life and interference caused by proximity between UAV-BSs. To effectively manage these challenges, this paper introduces a realistic optimization framework that combines Constrained K-means clustering with the Particle Swarm Optimization (PSO) algorithm for optimizing the positioning of UAVBSs in NTNs. This strategy considers operational constraints and the necessity for nearby recharging stations. Our simulation results show notable improvements in average data rates for mobile users compared to conventional approaches, such as PSO-based UAV-BS positions. Through comparative analysis, this study underscores significant improvements in network performance, demonstrating the effectiveness of our resource optimization strategy in modern NTNs.