UAV-Mounted Intelligent Reflecting Surface-Assisted Edge Content Delivery Networks Maximized Sum Rate with Deep Learning

Unmanned aerial vehicles (UAV) and intelligent reflecting surfaces (IRS) have become the important parts besides the emerging technologies, architectures, and advanced applications and services (A&S) required to meet the groundbreaking key performance indicators and goals in 6G networks. This paper studies a system of UAV-mounted IRS (UI) assisted edge content delivery networks in which the contents of advanced A&Ss are originally transmitted from the base stations (BS) to mobile users (MU) via the UIs. The system is formulated to propose a phase-shift and altitude (PSA) optimization problem. The PSA problem is solved for the optimal values of phase-shift of reflecting elements and the optimal altitudes of the UIs to maximize the sum rate received at the MUs. A convolutional neural network based deep learning (CNN-based DL) model is designed to capture the rapid changes of the environment and derive the optimal working parameters immediately to serve the MUs the contents at the maximum sum rate. Simulation results demonstrate not only the convergence and accuracy of the CNN-based DL model but also the benefits of PSA solution compared to other benchmarks.