Transmitting Data Through Reconfigurable Intelligent Surface: A Spatial Sigma-Delta Modulation Approach

Wai-Yiu Keung, Hei Victor Cheng, Wing-Kin Ma

This paper, recently published by researchers from The Chinese University of Hong Kong and Aarhus University, presents a clever way to turn a passive Reconfigurable Intelligent Surface (RIS) into an active data transmitter without the typical computational headaches.

TL;DR

The authors propose using spatial Sigma-Delta ($\Sigma\Delta$) modulation to design the reflection phases of an RIS, effectively turning a single-antenna transmitter into a massive MIMO system. This approach replaces complex, high-latency optimization algorithms with a low-complexity, "noise-shaping" feedback loop that pushes quantization errors away from users.

What Problem Does This Paper Address?

Reconfigurable Intelligent Surfaces (RIS) are arrays of passive elements that can shift the phase of incoming electromagnetic waves. While usually used as "smart mirrors" to improve coverage, they can also be used to transmit data by modulating their phases.

The challenge? To perform multi-user precoding (sending different data to different people simultaneously), you need to find the exact phase for hundreds or thousands of RIS elements. This usually requires solving "constant modulus" optimization problems—mathematically "hard" problems that are computationally expensive to solve in real-time. Furthermore, hardware often limits these elements to a few discrete phase levels (e.g., 2-bit or 3-bit quantization), which adds another layer of complexity and potential signal distortion.

Methods

The researchers move away from optimization and toward signal processing heuristics. Their core idea is to adapt Sigma-Delta ($\Sigma\Delta$) modulation—a technique traditionally used in hardware like ADCs to shape temporal noise—and apply it spatially across the physical elements of the RIS.

Key components of their approach include:

  • Spatial Noise Shaping: Instead of treating quantization error as random noise that hits every user, the $\Sigma\Delta$ loop "pushes" the quantization noise into specific spatial frequencies (angles) where no users are located.
  • Angle-Steering: They introduce a phase-shifting term in the feedback loop. This allows them to "steer" the noise-free "notch" of the modulator toward the specific angular sector where the users are located.
  • $\Sigma\Delta$-ZF Precoding: They take a standard Zero-Forcing (ZF) precoding vector (which assumes perfect, continuous amplitudes and phases) and pass it through the spatial $\Sigma\Delta$ modulator to generate the actual discrete phase shifts for the RIS.
  • Dithering and Overloading: To handle very low-resolution hardware (like 2-bit phases), they use "subtractive dithering" (adding and then removing controlled noise) and "overloading" (purposely exceeding the modulator's stable range) to prevent the error floor from flattening out.

Datasets

The paper does not use external real-world datasets. Instead, the authors evaluate their method using Monte-Carlo simulations based on standard wireless channel models:

  • Configuration: A Uniform Linear Array (ULA) with $N=512$ RIS elements serving $K=8$ users.
  • Modulation: 16-QAM symbols over block-fading channels.
  • Geometry: Users were randomly placed within a specific angular sector ($20^\circ$ to $60^\circ$).

Key Insights and Ideas

  • The Space-Time Duality: The most elegant insight is treating the physical sequence of antennas like a time sequence. Just as a temporal $\Sigma\Delta$ filter shapes noise in the frequency domain, a spatial $\Sigma\Delta$ filter shapes noise in the angular domain.
  • Optimization-Free Design: By using a recursive feedback loop (Equation 5 in the paper), the phase for each element is calculated based only on the previous element's error. This is significantly faster than iterative optimization.
  • The "No-Overload" Condition: The authors derive a mathematical bound (Fact 1) for the input signal amplitude. If the input stays within this bound, the quantization noise remains stable and predictable.
  • Clustered Users are a Benefit: The method works best when users are located in a relatively narrow angular sector, as it allows the "noise notch" to cover all users effectively.

Key Results

  • Near-Optimal Performance: With 16 discrete phase levels ($L=16$), the proposed $\Sigma\Delta$ approach achieves a Bit Error Rate (BER) almost identical to an "ideal" unquantized system.
  • 1-Bit/2-Bit Success: Even with very coarse quantization (e.g., $L=4$, which is 2-bit), the combination of overloading and dithering allowed the system to avoid the "error floor" that usually plagues simple quantization methods.
  • Complexity: The authors demonstrate that their approach provides a "negligible complexity" solution compared to the high computational cost of symbol-level optimization.

Limitations and Future Work

  • 1D vs. 2D: The current study assumes a Uniform Linear Array (1D). Real-world RIS units are usually rectangular planar arrays (2D). Extending the $\Sigma\Delta$ logic to 2D spatial structures is a necessary next step.
  • Angular Constraints: The noise-shaping works best for users in a specific sector. If users are scattered $360^\circ$ around the RIS, the "notch" might not be wide enough to protect everyone from quantization noise.
  • Direct Link Assumption: The paper assumes the direct path between the Base Station and the user is blocked. Future work will need to account for the interference or constructive combination of the direct signal and the RIS-reflected signal.

Summary updated

Abstract

Transmitting data using the phases on reconfigurable intelligent surfaces (RIS) is a promising solution for future energy-efficient communication systems. Recent work showed that a virtual phased massive multiuser multiple-input-multiple-out (MIMO) transmitter can be formed using only one active antenna and a large passive RIS. In this paper, we are interested in using such a system to perform MIMO downlink precoding. In this context, we may not be able to apply conventional MIMO precoding schemes, such as the simple zero-forcing (ZF) scheme, and we typically need to design the phase signals by solving optimization problems with constant modulus constraints or with discrete phase constraints, which pose challenges with high computational complexities. In this work, we propose an alternative approach based on Sigma-Delta ($ΣΔ$) modulation, which is classically famous for its noise-shaping ability. Specifically, first-order $ΣΔ$ modulation is applied in the spatial domain to handle phase quantization in generating constant envelope signals. Under some mild assumptions, the proposed phased $ΣΔ$ modulator allows us to use the ZF scheme to synthesize the RIS reflection phases with negligible complexity. The proposed approach is empirically shown to achieve comparable bit error rate performance to the unquantized ZF scheme.
BibTeX
@inproceedings{keung2024transmitting,
  author = {Wai-Yiu Keung and Hei Victor Cheng and Wing-Kin Ma},
  title = {{Transmitting Data Through Reconfigurable Intelligent Surface: A Spatial Sigma-Delta Modulation Approach}},
  booktitle = {IEEE International Conference on Acoustics, Speech and Signal Processing},
  year = {2024},
  doi = {10.1109/icassp48485.2024.10448246},
}

Loading related content...

Figure