Abstract

This paper addresses the challenge of optimizing communication coverage in multi-unmanned aerial vehicle (UAV) networks. Our primary focus is maximizing the secrecy rate by jointly optimizing UAV trajectories and transmission power within a defined timeframe. We employ a dual-method approach to tackle this complex, non-convex problem. Firstly, we apply advanced iterative methods, including the Gradient Descent technique, to identify the most effective strategy. Secondly, we introduce a novel machine learning-based model as an alternative approach. This model aims to enhance optimization, offering faster and more efficient solutions than traditional methods. We assess both strategies regarding computational efficiency, memory usage, and effectiveness in solving the optimization challenge. The performance of both methods is critically evaluated in terms of computational speed, memory demands, and optimization effectiveness. This approach demonstrates a potential for improved solutions in UAV network operations compared to traditional optimization techniques.

Keywords: Unmanned Aerial Vehicles (UAVs), Trajectory Optimization, Communication Coverage, Gradient Descent (GD), Machine Learning, Non-Convex Optimization, Power Transmission Optimization, Computational Efficiency.