Priority-Aware Task Offloading in UAV-Assisted Satellite MEC Networks

Satellite communication networks are crucial for enabling uninterrupted coverage and supporting collaborative computing, making them essential for remote applications such as transportation, environmental monitoring, and disaster management. However, traditional terrestrial networks are hindered by coverage gaps, high latency, and significant energy consumption. To address these limitations, this study integrates Mobile Edge Computing (MEC) with Unmanned Aerial Vehicles (UAVs) and Low Earth Orbit (LEO) satellites within a Space-Air-Ground Integrated Network (SAGIN). This approach leverages ad-hoc networks to establish self-organising, infrastructure-free connectivity, particularly suited for communication in remote and disaster-affected areas. We propose an Attention-Enhanced Proximal Policy Optimisation (A-PPO) algorithm to optimise task offloading by reducing latency and energy consumption while ensuring task-to-service alignment based on the processing capabilities of offloading locations. Simulation results demonstrate that the proposed algorithm achieves at least 28.3%, 17.2%, 22.9%, 15.8%, and 13.9% reductions in total offloading costs across various task batches compared to baseline methods.