Exploiting STAR-RISs in Near-Field Communications

Jiaqi Xu, Xidong Mu, Yuanwei Liu

TL;DR

This paper moves beyond traditional "patch-array" models for Reconfigurable Intelligent Surfaces (RIS) by treating them as continuous metasurfaces. By applying a Green’s function-based channel model, the authors reveal that near-field communications offer significantly higher degrees of freedom and a unique linear power scaling law, providing a roadmap for ultra-high-capacity 6G networks.

What Problem Does This Paper Address?

As wireless communications push into millimeter-wave (mmWave) and Terahertz (THz) bands, the "near-field" region—where the wavefront is curved rather than flat—expands significantly. Traditional models treat RIS as an array of discrete patches, each with a single phase shift. However, modern metasurfaces have elements so small (millimeter or molecular scale) that they behave more like a continuous electromagnetic curtain.

Existing research often ignores two things:

  1. The Continuous Nature of Metasurfaces: Discrete models fail to capture the microscopic electric current distributions that drive these surfaces.
  2. The Receiver's Role in the Near-Field: Most models define the near-field boundary based only on the size of the transmitter, ignoring the physical dimensions of the receiver.

The authors aim to bridge this gap by providing a mathematically rigorous framework for both standard RIS and Simultaneously Transmitting and Reflecting (STAR)-RIS in the near-field.

Methods

The researchers employ the Green’s function method to model the channel. Instead of using a simple phase-shift matrix, they characterize the RIS elements using a continuous electric current distribution.

Key methodological steps include:

  • Kernel Function Derivation: They developed a kernel function to measure the coupling strength between source points on the RIS and field points at the receiver.
  • Paraxial Approximation: To make the complex integrals tractable, they used approximations that hold true when the RIS and receiver are far apart relative to their individual sizes, but still within the radiating near-field.
  • STAR-RIS Strategies: For multi-user scenarios, they proposed and analyzed three configuration strategies:
  • Power Splitting (PS): Every element serves all users simultaneously.
  • Selective Element Grouping (SEG): Elements are assigned to the user they are closest to.
  • Random Element Grouping (REG): Elements are assigned to users randomly.

Datasets

Not discussed in the paper. This is a theoretical and analytical work; the results are validated through numerical simulations based on electromagnetic theory rather than empirical datasets.

Key Insights and Ideas

  • The "Aha" Boundary: The near-field/far-field boundary ($r_b$) isn't just about the RIS size ($L$); it depends on the aperture of both the RIS and the receiver. In high-frequency scenarios like Li-Fi, the near-field can extend for kilometers.
  • Degrees of Freedom (DoF) Boost: In the far-field, the channel is usually "rank-one" (like a single pipe). In the near-field, because path lengths vary so much across the surface, the channel acts like multiple parallel pipes, offering much higher DoF for multiplexing.
  • The Power Scaling Shift:
    • In the far-field, received power scales quadratically ($M^2$) with the number of elements.
  • In the near-field, it scales linearly ($M$). This is a crucial distinction for engineers designing large-scale surfaces.
  • STAR-RIS Superiority: STAR-RIS can eliminate "blind zones" (e.g., behind a wall) by simultaneously reflecting signals to outdoor users and transmitting them to indoor users, outperforming traditional reflecting-only surfaces.

Key Results

  • Scaling Laws: Numerical results confirmed that as the number of elements ($M$) increases, the quadratic growth of the far-field eventually hits a "ceiling" and transitions into linear growth as the user enters the near-field.
  • Strategy Comparison: For STAR-RIS, SEG provides the highest channel gain because it optimizes for proximity, but PS yields the highest DoF, making it better for serving many users with independent data streams.
  • Coverage Improvement: In a hybrid indoor/outdoor simulation (30 GHz), the STAR-RIS improved signal strength in indoor "blind zones" by approximately 5 dB compared to a simple open window and significantly more compared to a solid wall.

Limitations and Future Work

  • Hardware Complexity: While the paper proves that near-field channels have high DoF, exploiting them requires sophisticated RF chains or "holographic" beamforming hardware that is currently expensive and difficult to implement.
  • Channel Estimation: The SEG strategy requires precise positional information for all users. The authors note that the overhead for this channel estimation could be significant in highly mobile environments.
  • Scalar Simplification: The current model focuses on scalar currents. Future work could extend this to vector fields to account for complex polarization effects.

Summary updated

Abstract

The reconfigurable intelligent surface (RIS) is a promising technology to provide smart radio environment. In contrast to the well-studied patch-array-based RISs, this work focuses on the metasurface-based RISs and simultaneously transmitting and reflecting (STAR)-RISs where the elements have millimeter or even molecular sizes. For these meticulous metasurface structures, near-field effects are dominant and a continuous electric current distribution should be adopted for capturing their electromagnetic response instead of discrete phase-shift matrices. Exploiting the electric current distribution, a Green's function method based channel model is proposed. Based on the proposed model, performance analysis is carried out for RISs and STAR-RISs. 1) For the transmitting/reflecting-only RIS-aided single-user scenario, closed-formed expressions for the near-field/far-field boundary and the end-to-end channel gain are derived. Then, degrees-of-freedom (DoFs) and the power scaling laws are obtained. It is proved that the near-field channel exhibits higher DoFs than the far-field channel. It is also confirmed that when communication distance increases beyond the field boundary, the near-field power scaling law degrades to the well-known far-field result. 2) For the STAR-RIS-aided multi-user scenario, three practical STAR-RIS configuration strategies are proposed, namely power splitting (PS), selective element grouping (SEG), and random element grouping (REG) strategies. The channel gains for users are derived within both the pure near-field regime and the hybrid near-field and far-field regime. Finally, numerical results confirm that: 1) for metasurface-based RISs, the field boundary depends on the sizes of both the RIS and the receiver, 2) the received power scales quadratically with the number of elements within the far-field regime and scales linearly within the near-field regime.
BibTeX
@article{xu2024exploiting,
  author = {Jiaqi Xu and Xidong Mu and Yuanwei Liu},
  title = {{Exploiting STAR-RISs in Near-Field Communications}},
  journal = {IEEE Trans. Wireless Communications},
  year = {2024},
  doi = {10.1109/twc.2023.3296191},
}
Figure