Privacy-Preserving Federated Learning for UAV-Enabled Networks: Learning-Based Joint Scheduling and Resource Management

Unmanned aerial vehicles (UAVs) are capable of serving as flying base stations (BSs) for supporting data collection, machine learning (ML) model training, and wireless communications. However, due to the privacy concerns devices limited computation or communication resource UAVs, it is impractical send raw UAV servers training. Moreover, dynamic channel condition heterogeneous computing capacity in UAV-enabled networks, reliability efficiency sharing require be further improved. In this paper, we develop an asynchronous federated (AFL) framework multi-UAV-enabled which can provide distributed by enabling training locally without transmitting sensitive servers. The device selection strategy also introduced into AFL keep low-quality from affecting accuracy. propose advantage actor-critic (A3C) based joint selection, UAVs placement, management algorithm enhance convergence speed Simulation results demonstrate that our proposed achieve higher accuracy faster execution time compared other existing solutions.