Deadline-Aware and Energy-Efficient Distributed Scheduling Algorithm in Ground-Air Cooperative Networks

The scarcity of energy resources and computational capabilities poses significant challenges for Unmanned Aerial Vehicles (UAVs) when tasked with executing time-sensitive and complex demands, such as those necessary for Artificial Intelligence (AI)-enabled applications. Integrating Unmanned Ground Vehicles (UGVs) in Ground-Air Cooperative Systems offers a promising solution by providing additional computational and storage capacities near the point of need. However, these systems' distributed, dynamic, and resource-constrained nature presents a critical challenge for efficient task scheduling. This study introduces a novel two-tiered distributed online task scheduling approach aimed at optimizing the resource utilization of UGVs in edge computing environments. Our methodology encompasses two algorithms: a task dispatching algorithm designed to identify a UGV server that meets the deadline requirements while minimizing energy consumption and a computing scheduling algorithm that employs the Earliest Deadline First (EDF) strategy to refine the sequence in which tasks are processed, thereby reducing the average task completion time. We thoroughly assess the performance of the suggested algorithm with state-of-the-art methods using the OMNeT++ network simulator. The outcomes indicate that the proposed method significantly reduces the average task latency and energy consumption while meeting task deadline requirements.