Year: 2025
Type: article
Abstract: Unmanned Aerial Vehicles (UAVs) are increasingly utilized in wireless communication, yet accurate channel loss prediction remains a significant challenge, limiting resource optimization performance. This paper proposes a novel AIGC (Artificial Intelligence Generated Content)-driven framework that revolutionizes UAV communication through three key innovations: data augmentation, channel prediction, and trajectory optimization. First, aWasserstein Generative Adversarial Network (WGAN) is employed to generate high-quality synthetic channel data, addressing the time-consuming nature of data collection. Second, this augmented dataset trains a knowledge-driven Channel Knowledge Map (CKM) that achieves superior prediction accuracy by incorporating domain expertise. Finally, the enhanced CKM guides a reinforcement learning algorithm to generate optimal UAV trajectories. Experimental results demonstrate that our AIGC-powered approach significantly decreases train loss by 30.23% and reduces flight time from 81s to 37s compared to traditional methods, marking a substantial advancement in UAV-enabled wireless communications.
Authors: Chiya Zhang, Ting Wang, Rubing Han, Yuanxiang Gong
Institution: Harbin Institute of Technology
Language: English
Subfield: Aerospace Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W4408609488