The agile mobility of Unmanned Aerial Vehicles (UAVs) makes them ideal for low-altitude edge computing. This paper proposes a novel multi-tier UAV edge computing system where lightweight Low-Tier UAVs (L-UAVs) function as edge servers for vehicle users, supported by a powerful High-Tier UAV (H-UAV) acting as a backup server. The objective is to minimize task execution delays while ensuring the long-term energy stability of the L-UAVs, despite unknown future system states. To this end, the problem is decoupled using Lyapunov optimization, which adaptively balances the priorities of task delays and L-UAV energy cost based on their real-time energy states. An efficient vehicle to L-UAV matching scheme is designed, and the joint optimization problem for task assignment, computing resource allocation, and trajectory control of L-UAVs and H-UAV is then solved via a Block Coordinate Descent (BCD) algorithm. Simulation results demonstrate a reduction in L-UAV transmission energy of over 26% and superior L-UAV energy stability compared to existing benchmarks.

翻译：无人机（UAV）的敏捷机动性使其成为低空边缘计算的理想平台。本文提出了一种新颖的多层级无人机边缘计算系统，其中轻量级的低层级无人机（L-UAV）作为车辆用户的边缘服务器，并由一个功能强大的高层级无人机（H-UAV）作为备份服务器提供支持。目标是在未来系统状态未知的情况下，最小化任务执行延迟，同时确保L-UAV的长期能量稳定性。为此，利用李雅普诺夫优化对该问题进行解耦，该优化方法根据L-UAV的实时能量状态，自适应地平衡任务延迟与L-UAV能量消耗的优先级。本文设计了一种高效的车辆到L-UAV匹配方案，并通过块坐标下降（BCD）算法，解决了针对L-UAV和H-UAV的任务分配、计算资源分配和轨迹控制的联合优化问题。仿真结果表明，与现有基准方法相比，所提方案能降低L-UAV超过26%的传输能量，并具有更优的L-UAV能量稳定性。