AHSWN Home · Issue Contents · Forthcoming Papers

Aero-CSN: A Mobility-Aware Wireless Ad Hoc Network Architecture for Hierarchical Low-Altitude AI Services
Liang Guo, Jizhuang Zhao, Wei Quan, Shaopeng Wang, and Jie Li

The rapid expansion of the low-altitude economy relies on Unmanned Aerial Vehicles (UAVs) forming dynamic Wireless Ad Hoc Networks. However, supporting computationally intensive Hierarchical AI services over these resource-constrained topologies remains challenging due to coverage gaps and link instability in terrestrial Mobile Edge Computing (MEC). We propose A-CSN (Aero Computing Star Network), a mobility-aware wireless ad hoc architecture that extends the Computing Power Network (CPN) into the low-altitude domain. By utilizing terrestrial MEC nodes as wireless anchors to organize the volatile ad hoc topology, A-CSN mitigates link intermittency via an infrastructure-assisted hierarchy combining resilient control signaling with stable optical backhaul. Furthermore, we introduce the Mobility-Aware Deterministic (MAD) orchestration protocol—comprising Pessimistic Reservation (PR-Arch), the Dynamic-Stable Predictive Planner (D-SPP), and the Hybrid MEC-Regional Strategy (H-MRS)—to resolve spatiotemporal resource contention. Simulations demonstrate that A-CSN achieves a 95% service acceptance rate and a 5.3× throughput gain over state-of-the-art baselines.

Keywords: wireless ad hoc networks, unmanned aerial vehicles (UAVs), hierarchical AI, computing power network, computing star network, mobility management, Aero-CSN