Wireless Communication with Extremely Large-Scale Intelligent Reflecting Surface

Chao Feng, Haiquan Lu, Yong Zeng, Shi Jin, Rui Zhang

Introduction

Intelligent Reflecting Surfaces (IRS) have emerged as a transformative technology for 6G wireless networks, offering the ability to program the propagation environment. By adjusting the phase shifts of numerous passive reflecting elements, an IRS can constructively combine signals at a receiver, thereby enhancing coverage and spectral efficiency. However, the inherent "double path loss" associated with the transmitter-IRS-receiver link necessitates the use of a very large number of elements to achieve significant gains. This has led to the concept of Extremely Large-Scale IRS (XL-IRS), where surfaces may cover entire building facades.

Conventional performance analysis of IRS-aided systems typically relies on the Uniform Plane Wave (UPW) model. This model assumes that the distances between the IRS and the transceivers are sufficiently large such that the signal amplitude and angle of arrival are uniform across the entire surface. As IRS dimensions grow to the "extremely large" scale, these assumptions often fail, as users frequently operate within the near-field region of the array.

System Model and Methodology

To accurately characterize XL-IRS performance, this work moves beyond the simplified UPW model to a rigorous element-based approach that accounts for the physical realities of large apertures.

System model of wireless communication aided by a large-scale intelligent reflecting surface (XL-IRS).

The proposed model incorporates several critical factors:

  1. Spherical Wavefronts: Instead of flat waves, the model accounts for the exact distance from the transmitter and receiver to each individual element on the IRS.
  2. Amplitude and Aperture Variation: The model integrates the inverse-square law for signal decay at each element. Crucially, it also accounts for the "projected aperture" effect, where the effective area of an element decreases when viewed at an oblique angle.
  3. Mathematical Framework: For a general Uniform Planar Array (UPA), the authors derive tight upper and lower bounds for the received signal-to-noise ratio (SNR) using integral approximations. For the Uniform Linear Array (ULA) case, a closed-form solution is derived using incomplete Elliptic Integrals.

Key Insights: The Myth of Quadratic Scaling

A cornerstone of early IRS research was the "square power scaling law," which suggested that the SNR increases quadratically ($M^2$) with the number of reflecting elements $M$. This work demonstrates that this law is an artifact of the UPW assumption and does not hold for XL-IRS.

As the physical size of the IRS increases, the contribution of peripheral elements diminishes. This is due to two factors: the increased distance from the transceivers (leading to higher path loss) and the sharper angles of incidence (reducing the effective aperture). Consequently, the SNR does not grow indefinitely but instead reaches a physical saturation point.

Performance Analysis

Numerical simulations validate the derived analytical bounds and highlight the discrepancies of the conventional model.

SNR performance versus IRS size, demonstrating the saturation effect and the inaccuracy of the conventional UPW model at large scales.

As shown in the results, the UPW model significantly overestimates performance as the IRS size $L$ increases. While the UPW model predicts unbounded SNR growth, the actual SNR (and the derived bounds) asymptotically approaches a finite limit. In scenarios where the transmitter is close to a large IRS, the UPW model can overestimate the SNR by more than 25 dB, leading to highly unrealistic system designs.

Conclusion and Future Directions

This research provides a fundamental shift in how we model and understand large-scale intelligent surfaces. By proving the existence of an SNR saturation limit, it sets realistic expectations for the deployment of XL-IRS in future 6G networks.

Future work should extend these findings to more complex environments, including non-Line-of-Sight (NLoS) channels, hardware constraints like discrete phase shifts, and multi-user interference scenarios. By grounding IRS theory in physical reality, this work ensures that the next generation of wireless infrastructure is built on a solid mathematical foundation.

Summary updated

BibTeX
@inproceedings{feng2021wireless,
  author = {Chao Feng and Haiquan Lu and Yong Zeng and Shi Jin and Rui Zhang},
  title = {{Wireless Communication with Extremely Large-Scale Intelligent Reflecting Surface}},
  booktitle = {ICCC Workshops},
  year = {2021},
  doi = {10.1109/icccworkshops52231.2021.9538864},
}
Figure