XL-IRS Aided Near-Field Communications: A Two-Stage Hierarchical Approach

In the forthcoming sixth generation (6G) communication, it is more likely that intelligent reflecting surface (IRS) will evolve into extremely large-scale IRS (XL-IRS) to enhance the quality of wireless communications. However, the far-field methods of IRS cannot be directly extended to XL-IRS in near-field communications. In this paper, we analyze the XL-IRS aided communication scenario, and propose a two-stage hierarchical beam training scheme for XL-IRS based on the quartering method and non-uniform distance sampling. Specifically, we design a two-dimensional hierarchical training scheme to determine the user’s direction in the direction domain. Then, we utilize the non-uniform distance sampling method to determine the user’s distance in the distance domain. Simulation results demonstrate that our proposed two-stage beam training scheme can achieve the rate performance comparable to near-field 3D exhaustive search in the XL-IRS system, with lower training overhead.