Intelligent 3-D Trajectory and Resource Allocation for UAV Communications Under a Blockage-Aware Channel Model

In this paper, we propose a novel deep learning (DL)-based framework for three-dimensional (3-D) trajectory design and resource allocation in unmanned aerial vehicle (UAV)-enabled wireless networks to maximize the minimum average spectral efficiency (SE) among moving ground nodes (GNs). The proposed framework incorporates an obstacle avoidance algorithm for 3-D trajectory planning and employs a practical blockage-aware channel model that accounts for the blockage effects on air-to-ground links caused by obstacles, while also accounting for the stochastic movement of GNs. To this end, we develop a new mathematical approximation based on a point cloud method to efficiently determine whether the channel between the UAV and the GN is blocked by obstacles and whether the trajectory of the UAV intersects with obstacles. Subsequently, we introduce a DL framework that integrates deep neural network (DNN) structures designed to solve the formulated problem with an unsupervised learning-based training methodology. This approach enables efficient modeling of 3-D trajectory and resource allocation while also facilitating the effective training of the DNN without the need for labeled data. Through performance evaluations, we demonstrate that the proposed scheme accurately accounts for the location-dependent channel blockage effects caused by obstacles and successfully avoids potential UAV collisions with obstacles. Furthermore, the proposed scheme outperforms baseline schemes in achieving the minimum average SE by jointly optimizing the 3-D trajectory and resource allocation while maintaining low computation times for real-time operation.