Channel Prediction-based Robust Multibeam Precoding Design for LEO Satellite Internet of Things

Multibeam precoding is widely recognized as an enabling technique to support wide-area massive connectivity with limited spectrum in LEO satellite Internet of Things (IoT). However, timely channel state information (CSI) acquisition for multibeam precoding is a critical issue due to fast mobility of LEO satellites. To this end, we propose a novel channel prediction method to obtain current CSI by using past estimated CSI through an autoregressive model. Then, we design an effective robust multibeam precoding algorithm for LEO satellite IoT by maximizing the weighted sum rate of IoT devices based on predicted CSI. Both theoretical analysis and simulation results confirm the effectiveness of the proposed algorithm.