Online Federated Learning for Air-Ground Edge Intelligence

Traditional federated learning (FL) usually relies on static ground base stations (BSs) for model aggregation. Unmanned aerial vehicles (UAVs), due to their flexible 3D deployment, can effectively complement ground BSs, enabling the establishment of air-ground integrated FL to boost ubiquitous edge intelligence. However, for online FL under such an air-ground setting, resource-limited users need to determine the resource allocation and sample selection to enhance training performance and achieve high energy efficiency. This paper studies how to minimize the tradeoff between energy consumption and training performance of air-gound online FL, and proposes an efficient algorithm to solve the aforementioned problem. Simulation results validate the effectiveness of our proposed algorithm in terms of both training effect and energy consumption.