Cost-efficient hierarchical federated edge learning for satellite-terrestrial Internet of Things

With the widespread deployment of dense Low Earth Orbit (LEO) constellations, satellites can serve as an alternative solution to the lack of proximal multi-access edge computing (MEC) servers for mobile Internet of Things (IoT) devices in remote areas. Simultaneously, the implementation of LEO on-board federated learning (FL), which can meet the data privacy requirements of user devices, makes it sensible to offloading data processing tasks to intelligent edge devices equipped with adaptive learning capabilities. However, traditional satellite on-board federated learning may encounter challenges due to the limited satellite resources. Hence, we propose a cost-efficient satellite-terrestrial assisted hierarchical federated edge learning (STA-HFEL) architecture and an innovative communication scheme between satellites based on Intra-plane ISLs in this paper. Accordingly, managing CPU resources for local training and allocating bandwidth for learning information upload among battery-limited mobile devices is crucial. With this in mind, we define a joint computation and communication resource optimization problem for device users to achieve global cost minimization. A distributed Jacobi-Proximal ADMM (JPADMM) algorithm is used to tackle the newly formulated problem iteratively. Extensive performance evaluations demonstrate that the potential of STA-HFEL as a cost-efficient and privacy-preserving approach for machine learning tasks across distributed remote environments.