CSI Estimation for IRS-Assisted Massive MIMO Wireless Communication Systems: A Low Complexity Greedy Pursuit Approach

The future wireless networks are capable of responding to the huge demand for connectivity if they employ the new technologies in physical and network layers. The intelligent reflecting surface (IRS) is an emerging technology introduced for the physical layer of such networks. It fairly controls the multipath environment by reflecting or transmitting the signal, thereby alleviating the problem of signal blockage, specifically in millimeter wave (mm-wave) systems. However, channel state information (CSI) estimation is one of the fundamental challenges of IRS technology when it operates in passive mode. In this paper, we propose a new greedy algorithm to estimate the uplink cascade channel between the base station (BS), IRS, and users. This algorithm utilizes the inherent spare property of the channel and divides it into two general forms, that is, row and column sparsity, searches for the location of the whole non-zero coefficients in the rows, and then refines the locations by identifying the non-zero ones in the columns. Finally, the coefficient values are calculated by solving a simple least square problem. The simulation results demonstrate that the proposed approach substantially outperforms state-of-the-art algorithms regarding computational complexity, pilot compression rate, and mean square error.