RIS-Assisted Federated Learning Algorithm Based on Device Selection and Weighted Averaging

To protect user privacy and improve the transmitting environment of wireless communication, federated learning (FL) and reconfigurable intelligent surface (RIS) are proposed as promising technologies for future communication. Meanwhile, studies have proved that the combination of FL and RIS guarantees better performance for system models. However, the combined model still has problems such as high communication overhead and slow convergence speed. Therefore, in this paper, we proposed a channel quality based device selection and weighted averaging algorithm in a RIS-assisted federated learning model. Simulation results proved that the proposed algorithm outperforms the classic federated averaging (FedAvg) algorithm in convergence speed, test accuracy, and training loss.