Task-Driven Synchronization-Assisted TDMA for Multi-priority Long-Range UAV Swarms

This paper studies a task-driven adaptive networking protocol for long-range unmanned aerial vehicle (UAV) swarms operating over distances up to several thousand kilometres. We first construct a unified “task–topology–link” model that distinguishes on-board control, urgent alarms and bulk data, and captures the asymmetric star-shaped A–B cluster topology and large propagation delay of long-range air–air links. Building on GNSS-based coarse synchronization and RTT-based fine ranging, we design a synchronization-assisted TDMA superframe with beacon, random access, grant and data/command phases. Random access employs hash-thinned contention and frame-level adaptation of RA length to cope with varying access loads, while data/command transmission uses per-task frame quotas and a debt-based scheduler to balance delay and throughput across task classes. We then develop a queueing-theoretic model of the access and data subsystems, derive approximate stability conditions and delay expressions, and identify how key parameters shape the joint stability region. Finally, an NS-3 and MATLAB co-simulation platform is built to compare the proposed protocol with representative long-range MAC baselines. Results show that the proposed design provides higher access success probability, lower control and alarm delay, and a larger stable throughput region under high load and very long propagation delays.