Channel Reconstruction in Intelligent Surface aided Communications

Abstract

Intelligent reflecting surface (IRS) has recently emerged as a promising candidate technology for enhancing capacity, coverage, and energy efficiency in future wireless communication systems. An IRS comprises of many low-cost antenna elements that can be programmed to impact and direct impinging electromagnetic waves in a beneficial manner. However, performance enhancements are subject to availability of accurate channel estimates, which are especially hard to obtain for an IRS with all passive elements. In this work, we propose novel channel reconstruction schemes which do not require any active elements on the IRS panel. Our novel formulation combines low-rank matrix completion with subspace side information and also exploits sparsity. It can account for non-ideal IRS elements and leads to an implementable algorithm. This algorithm in turn yields a reconstructed effective channel vector, and as a byproduct, an optimized IRS pattern that is well suited for facilitating data transmission. Furthermore, we propose an algorithm for subspace estimation and also provide a recipe for designing IRS pattern vectors that can be gainfully used during the training phase. Results generated using the opensource SimRIS tool demonstrate substantial spectral efficiency gains with a significantly reduced pilot overhead.