Joint Channel Estimation Algorithm for IRS-Assisted Multi-User MIMO Systems

Abstract

Intelligent reflective surface (IRS) is considered as a promising technology for massive MIMO systems. To harvest the benefit of IRS, accurate channel state information (CSI) is essentially required. However, existing methods simplified the multi-user channel estimation problem into independent subproblems, resulting in significant performance loss. To overcome this limitation, we propose a novel framework for multi-user joint estimation that explicitly exploits the common partial channel property. Incorporating the framework, we present a two-level expectation propagation approximation (EPA) method to reduce cumulative error by compressing the decoupling process. Simulation results validate the superiority of the proposed method.