Experimental Study of Channel Estimation Algorithms for General IRS Applications

Abstract

An intelligent reflecting surface (IRS) is a promising antenna array concept capable of realizing flexible electronically steerable beamforming. Depending on the location of the feeding antenna, IRS applications can be classified as either nearfield or farfield types. Using the IRS operating at 28 GHz, both application types are analyzed and compared in this paper. On the IRS testbed, channel estimation algorithms, including beam training and IRS unit channel state information (CSI) estimation, are implemented and analyzed for both application types. The results of the measurements indicate that the aforementioned algorithms are adaptable to both IRS application types. In addition, the IRS coding patterns designed using the estimated CSI at each IRS unit outperform the pre-designed patterns for beam training when constructing the wireless communication link. This paper provides evidence that IRS-based theoretical algorithms can be implemented in practice.