Deep Multi-Stage CSI Acquisition for Reconfigurable Intelligent Surface Aided MIMO Systems

This article aims to reduce huge pilot overhead when estimating the reconfigurable intelligent surface (RIS) relayed wireless channel. Motivated by compelling grasp of deep learning in tackling nonlinear mapping problems, proposed approach only activates a part RIS elements and utilizes corresponding cascaded channel estimate predict another part. Through synthetic neural network (DNN), direct active are first estimated sequentially, followed prediction for inactive elements. A three-stage training strategy is developed this DNN. From simulation results, based effective reducing guaranteeing reliable estimation accuracy.