Dual-Layer Dual-Timescale Collaborative Deployment for Dynamic UAV Ad Hoc Networks

Abstract

Dynamic deployment in unmanned aerial vehicle ad hoc networks (UANETs) is challenging because wireless coverage, aerial backhaul connectivity, and user mobility are tightly coupled over time. Centralized methods can exploit broader network information but become costly under continuously changing network states, whereas purely local adaptation is responsive yet often lacks sufficient global coordination. To address this issue, this paper proposes a dual-layer dual-timescale collaborative deployment framework for dynamic UANETs. The macro layer operates at the window level to aggregate system status and generate guidance parameters for the next window, while the local layer performs slot-level closed-loop adaptation over UAV position, transmit power, and user association using local observations and one-hop neighbor references. Based on this architecture, we further develop two macro-level implementations, namely one-shot optimization and iterative optimization. Simulation results in dynamic user-mobility scenarios with obstacle blocking and aerial backhaul constraints show that window-level macro guidance improves global coverage over the Local-Only baseline while maintaining high network connectivity.