Low-Latency Resource Allocation for Store-and-Forward Transmission in UAV-Aided LEO Communication

In this paper, we address a radio resource allocation challenge for multiple unmanned aerial vehicles (UAVs) connected to low Earth orbit (LEO) satellites, supporting ground users through store-and-forward transmission. Unlike existing integrated systems assuming seamless fusion of LEOs and UAVs, our practical approach treats UAVs and LEOs as nodes of distinct systems. UAVs act as relay nodes, receiving and forwarding data packets to LEO with possibly decapsulation and encapsulation. Our proposed resource allocation minimizes maximum transmission delays or buffer lengths of UAVs, formulated as a convex optimization problem. Given unknown arrival rates and channel conditions, we explore reinforcement learning for learning these critical parameters in resource allocation.