Stacked Intelligent Metasurfaces-Based Electromagnetic Wave Domain Interference-Free Precoding

Hetong Wang, Yashuai Cao, Tiejun Lv, Jintao Wang, Ni Wei, Jiancheng An, Chau YuenIEEE Trans. Wireless Communications 2026 · DOI

TL;DR

This paper introduces a novel interference-free precoding technique for wireless communication systems using stacked intelligent metasurfaces (SIMs). By performing precoding directly in the electromagnetic wave domain, this approach compensates for hardware impairments like nonlinear distortion and significantly outperforms existing methods in terms of signal detection reliability and data rates.

What Problem Does This Paper Address?

Future wireless communication systems, like 6G, demand higher spectral efficiency, energy efficiency, and faster transmission speeds. While intelligent reflecting surfaces (IRS) and stacked intelligent metasurfaces (SIMs) offer promising solutions for manipulating electromagnetic waves, existing methods for interference management often rely on complex and power-hungry hardware. Furthermore, practical hardware limitations, such as nonlinear distortion (NLD) from power amplifiers, can degrade signal quality. This paper tackles the challenge of developing an interference-free precoding scheme that leverages the unique capabilities of SIMs to perform precoding in the analog wave domain, effectively compensating for NLD and enhancing system performance without the need for traditional digital pre-distortion.

Methods

The core of this work is a Stacked Intelligent Metasurface (SIM)-based Interference Exploitation Precoding (IEP) architecture. Unlike traditional methods that aim to suppress interference, this approach cleverly exploits interference to improve signal detection. Key methodological contributions include:

  • Wave-Domain Precoding: The precoding operations are performed directly within the analog wave domain using SIMs, bypassing the need for complex digital signal processing and expensive hardware components.
  • NLD Compensation: A novel approach is presented to compensate for nonlinear distortion (NLD) introduced by power amplifiers. This is achieved by optimizing SIM phase shifts in the wave domain, effectively "aware" of the NLD.
  • Frame-Level IEP: To reduce computational complexity, the precoding is performed on a frame-by-frame basis rather than symbol-by-symbol.
  • Recursive Oblique Manifold (ROM) Algorithm: A novel algorithm is developed to optimize the SIM phase shifts. This algorithm addresses the complex coupling between phase shifts across multiple SIM layers and the non-convex nature of the optimization problem by decoupling it into layer-wise subproblems and employing Riemannian manifold optimization.
  • Greedy Safety Margin-Based Antenna Selection (AS) and Power Allocation (PA): To further enhance performance and adapt to dynamic environments, the paper proposes a greedy AS algorithm and integrates it with PA optimization within a ROM-based Alternating Optimization (ROM-AO) framework. This jointly optimizes antenna selection, SIM phase shifts, and power allocation.

Datasets

Not discussed in the paper. The paper relies on simulated data generated based on the described system model and parameters.

Key Insights and Ideas

  • Interference as a Resource: The paper demonstrates that interference, often seen as a nuisance, can be constructively utilized to improve signal detection and enhance system performance.
  • Analog Wave-Domain Processing Power: SIMs offer a powerful platform for performing complex signal processing tasks, like precoding and NLD compensation, directly in the analog domain, leading to significant efficiency gains.
  • NLD Compensation in the Analog Domain: Compensating for NLD directly in the wave domain is a crucial advancement, eliminating the need for traditional, power-hungry digital pre-distortion techniques.
  • ROM Algorithm's Effectiveness: The proposed ROM algorithm effectively tackles the challenging optimization problem of SIM phase shifts, even with multiple layers and complex coupling, by leveraging Riemannian manifold optimization.
  • Synergy of AS, SIM, and PA: Jointly optimizing antenna selection, SIM phase shifts, and power allocation yields substantial performance improvements, highlighting the importance of a holistic system design.

Key Results

  • Significant Performance Gains: The proposed SIM-enabled frame-level IEP scheme significantly outperforms benchmarks.
  • NLD Compensation Effectiveness: The NLD-aware IEP strategy achieves a 20 dB performance gain compared to the case without any strategy (with AS and PA) at a 12 dB SNR.
  • Improved Safety Margin: The NLD-aware approach leads to a larger safety margin, indicating better signal detection reliability.
  • Enhanced SER and Sum Rate: Simulations show substantial improvements in Signal Error Rate (SER) and sum rate compared to single-layer RIS and traditional BLP schemes. For instance, at 10 dB SNR, the NLD-aware SIM achieves a 13.07 dB SER performance gain and approximately doubles the sum rate compared to NLD-aware RIS.
  • Robustness of ROM Algorithm: The ROM algorithm demonstrates monotonic and stable convergence, with fewer iterations required as the number of meta-atoms increases.

Limitations and Future Work

  • Hardware Impairments: While the paper addresses NLD, other hardware impairments are not explicitly considered.
  • Channel Estimation: The performance relies on accurate channel state information (CSI), and the paper assumes existing CSI estimation techniques are sufficient.
  • Complexity of SIM Design: Fabricating and controlling multi-layer SIMs can still be complex and costly.

Future work could explore the impact of other hardware impairments, investigate more advanced CSI estimation techniques for SIM-enabled systems, and explore the scalability and practical deployment challenges of multi-layer SIMs in real-world scenarios. The paper also suggests that further hardware resources could be more effectively allocated to other system parameters beyond a certain number of SIM layers.

Summary updated

BibTeX
@article{wang2026stacked,
  author = {Hetong Wang and Yashuai Cao and Tiejun Lv and Jintao Wang and Ni Wei and Jiancheng An and Chau Yuen},
  title = {{Stacked Intelligent Metasurfaces-Based Electromagnetic Wave Domain Interference-Free Precoding}},
  journal = {IEEE Trans. Wireless Communications},
  year = {2026},
  doi = {10.1109/twc.2026.3656012},
}
Figure