Energy-Efficient Trajectory Design and Unsupervised Clustering for AAV-Aided Fair Data Collections With Dense Ground Users

In remote or high-demand wireless cellular networks, efficient data collection from ground users (GUs) with fixed infrastructure poses a significant challenge. Autonomous aerial vehicles (AAVs) have emerged as a promising solution due to their flexible deployment and cost-effectiveness. This article focuses on a AAV-aided wireless cellular communication system comprising a AAV and multiple adjacent GUs, where the mission of the AAV is to collect data from these GUs. The objective is to minimize AAV propulsion energy consumption while ensuring fair data uploading among all GUs. Due to the nonconvex and intractable nature of the above problem, we propose a novel real-time waypoint localization method based on the parallel projection method from a geometric perspective. By enhancing the projection process, this approach achieves energy-efficient and fair data collection, along with an efficient trajectory design algorithm. Further, considering the scenario of densely distributed GUs in large-scale areas, a GU-clustering algorithm is proposed based on Mean Shift. Additionally, this article categorizes GUs into homogeneous and heterogeneous scenarios and designs distinct trajectory designing algorithms to accommodate diverse real-world situations. Simulations and comparisons validate the effectiveness and efficiency of the proposed algorithms in tackling the AAV trajectory design challenges.