Low-Complexity Near-Field Channel Estimation for Hybrid RIS Assisted Systems

TL;DR

This paper proposes two low-complexity channel estimation algorithms for near-field hybrid RIS systems, exploiting block-sparsity, with one outperforming baselines in accuracy and the other in computational efficiency.

Contribution

It introduces two novel CE algorithms tailored for near-field RIS systems, leveraging block-sparsity and regularization techniques to improve efficiency and accuracy.

Findings

TVR algorithm outperforms baselines in high SNR regimes

BESVR algorithm achieves similar accuracy with minimal CPU time

Proposed methods effectively exploit near-field channel properties

Abstract

We investigate the channel estimation (CE) problem for hybrid RIS assisted systems and focus on the near-field (NF) regime. Different from their far-field counterparts, NF channels possess a block-sparsity property, which is leveraged in the two developed CE algorithms: (i) boundary estimation and sub-vector recovery (BESVR) and (ii) linear total variation regularization (TVR). In addition, we adopt the alternating direction method of multipliers to reduce their computational complexity. Numerical results show that the linear TVR algorithm outperforms the chosen baseline schemes in terms of normalized mean square error in the high signal-to-noise ratio regime while the BESVR algorithm achieves comparable performance to the baseline schemes but with the added advantage of minimal CPU time.