Validating Properties of RIS Channel Models with Prototypical Measurements

TL;DR

The paper addresses the gap between theoretical RIS channel models and real-world measurements under varying incidence angles. It employs a 256-element RIS prototype on a turntable and a VNA-based setup at $f^{\mathrm{RIS}}=5.53$ GHz to quantify angle-dependent magnitude and phase changes in reflections. By introducing angle-dependent coefficients $\alpha_m(\nu)$ and solving a linear system $\hat{\mathbf{H}}\boldsymbol{\alpha}=\hat{\mathbf{y}}$ via least squares, the authors refine the channel model to closely match measurements, revealing attenuation up to $-14.5$ dB and per-element phase deviations. The results provide actionable calibration guidance for RIS-enabled 6G systems, highlighting the need to incorporate angle-dependent reflection properties into design and optimization workflows.

Abstract

The integration of Reconfigurable Intelligent Surfaces (RIS) holds substantial promise for revolutionizing 6G wireless networks, offering unprecedented capabilities for real-time control over communication environments. However, determining optimal RIS configurations remains a pivotal challenge, necessitating the development of accurate analytical models. While theoretically derived models provide valuable insights, their potentially idealistic assumptions do not always translate well to practical measurements. This becomes especially problematic in mobile environments, where signals arrive from various directions. This study deploys an RIS prototype on a turntable, capturing the RIS channels' dependency on the angle of incoming signals. The difference between theory and practice is bridged by refining a model with angle-dependent reflection coefficients. The improved model exhibits a significantly closer alignment with real-world measurements. Analysis of the reflect coefficients reveals that non-perpendicular receiver angles can induce an additional attenuation of up to -14.5dB. Additionally, we note significant phase shift deviations, varying for each reflect element.

Figures (9)

• Figure 1: Top view of the anechoic chamber, with the antenna positions at the respective distances and angles marked with crosses. The setting shown is for an RIS angle $\nu=45^\circ$.
• Figure 2: Photo of the RIS and antennas in the anechoic chamber.
• Figure 3: The logarithmic magnitude of the channel values $h^\mathsf{eff}$ measured, simulated (using $\theta_m$) and adapted (using $\tilde{\theta}_m$) for a determined RIS configuration reflecting towards $0^\circ$.
• Figure 4: The logarithmic magnitude of the channel values $h^\mathsf{eff}$ measured, simulated (using $\theta_m$) and adapted (using $\tilde{\theta}_m$) for a determined RIS configuration reflecting towards $30^\circ$.
• Figure 5: The logarithmic magnitude of the channel values $h^\mathsf{eff}$ measured, simulated (using $\theta_m$) and adapted (using $\tilde{\theta}_m$) for a determined RIS configuration reflecting towards $60^\circ$.