Diversity and Multiplexing for Continuous Aperture Array (CAPA)-Based Communications

Chongjun Ouyang, Zhaolin Wang, Xingqi Zhang, Yuanwei LiuIEEE Trans. Wireless Communications 2024

TL;DR

This paper provides a rigorous theoretical framework for Continuous-Aperture Arrays (CAPAs)—the ultimate evolution of MIMO where antennas are packed with infinitesimal spacing—operating in realistic multipath fading environments. By applying Landau’s eigenvalue theorem, the authors prove that CAPAs significantly outperform traditional discrete arrays in both reliability (diversity) and data rate (multiplexing), establishing the fundamental performance limits for next-generation holographic surfaces.

What Problem Does This Paper Address?

In the quest for higher wireless capacity, we have moved from single antennas to Massive MIMO. The logical conclusion of this trend is the Continuous-Aperture Array (CAPA), a spatially continuous electromagnetic surface. While previous research on CAPAs often assumed simple Line-of-Sight (LoS) conditions, real-world wireless channels are messy, filled with multipath scattering and fading.

The challenge is that traditional discrete-signal processing doesn't neatly apply to a continuous surface. We need to know:

  1. How do we model fading between two continuous surfaces that aren't perfectly parallel?
  2. What are the fundamental trade-offs between reliability (diversity gain) and speed (multiplexing gain) in these systems?
  3. Does a continuous surface actually offer a tangible benefit over a very dense "standard" antenna array?

Methods

The authors tackle these questions using a mix of electromagnetic theory and stochastic analysis:

  • Wavenumber-Domain Modeling: Instead of looking at individual antenna "ports," they model the channel as a continuous integral kernel. They use Weyl’s identity to decompose the field into an angular spectrum of plane waves.
  • Landau’s Eigenvalue Theorem: This is the mathematical "secret sauce." The authors use it to analyze the autocorrelation of the spatial response. It reveals that even though a CAPA has an "infinite" number of points, it possesses a finite number of significant Degrees of Freedom (DoFs) determined by the aperture size and wavelength ($\lambda$).
  • Transmission Framework: For MIMO scenarios, they propose a framework that modulates data into source currents using a linear combination of Fourier bases, effectively "sampling" the continuous space in the wavenumber domain.
  • Asymptotic Analysis: They derive closed-form expressions for Outage Probability (OP) and Ergodic Channel Capacity (ECC) in the high Signal-to-Noise Ratio (SNR) regime to define the Diversity-Multiplexing Trade-off (DMT).

Datasets

As a theoretical communication theory paper, no external real-world datasets were used. The findings are validated through extensive computer simulations (Monte Carlo methods) that model isotropic scattering environments and compare the analytical derivations against simulated CAPA and Spatially-Discrete Array (SPDA) performance.

Key Insights and Ideas

  • The "Step-Like" Behavior: The most critical insight is that the eigenvalues of the channel autocorrelation function follow a step-like pattern. There is a specific number of "strong" channels (DoFs); beyond that number, the channel strength drops off precipitously. This means we can treat a continuous surface as having a discrete set of "virtual" parallel pipes.
  • CAPA as the Performance Ceiling: A CAPA represents the absolute upper bound of what an antenna array can achieve in a given physical footprint.
  • The $\lambda/2$ Threshold: The paper confirms that a traditional discrete array (SPDA) can match the multiplexing capability of a CAPA only if its elements are spaced at half a wavelength ($\lambda/2$) or less. However, even then, the CAPA maintains a superior array gain because it captures the entire electromagnetic field without the "gaps" present in discrete arrays.
  • Non-Parallelism Matters: Unlike many models that assume TX and RX arrays are perfectly aligned, this model accounts for arbitrary orientations, making it much more applicable to mobile users.

Key Results

  • Higher Capacity: CAPAs consistently achieve a higher Ergodic Channel Capacity (ECC) than discrete arrays across all SNR ranges.
  • Lower Outage: CAPAs provide a lower Outage Probability, meaning more reliable connections in fading environments.
  • DMT Superiority:
    • If an SPDA has spacing $>\lambda/2$, it loses to CAPA in both diversity and multiplexing.
  • If an SPDA has spacing $\le \lambda/2$, it matches the CAPA's diversity/multiplexing slope but suffers from a lower array gain (a constant power offset).
  • Scaling Laws: The average received SNR scales linearly with the product of the transmit and receive aperture areas.

Limitations and Future Work

  • Mutual Coupling (MC): The authors acknowledge that this study excludes electromagnetic mutual coupling between infinitesimal elements, which is a significant physical challenge in real-world CAPA/holographic implementations. They point to preliminary work in this area but leave a full integration for the future.
  • Evanescent Waves: The model focuses on the "far-field" components of the angular spectrum. While they mention evanescent waves (which decay rapidly), these could play a larger role in very-near-field "extreme" holographic communications.
  • Hardware Complexity: The paper focuses on the information-theoretic limits. Building the feed networks and transceivers capable of driving a truly continuous current distribution remains a massive engineering hurdle.

Summary updated

Abstract

A general fading model for multipath channels between two non-parallel continuous-aperture arrays (CAPAs) is proposed. Building on this model, the performance of diversity and multiplexing achieved by CAPAs over fading channels is analyzed. i) For multiple-input single-output (MISO) and single-input multiple-output (SIMO) channels, Landau's eigenvalue theorem is applied to analyze the autocorrelation of the spatial response. Closed-form expressions are derived for the outage probability (OP) and ergodic channel capacity (ECC). Asymptotic analyses in the high signal-to-noise ratio (SNR) regime are conducted to reveal the maximal achievable diversity and multiplexing gains. The diversity-multiplexing trade-off (DMT) is characterized, along with the array gain within the DMT framework. ii) For multiple-input multiple-output (MIMO) channels, a wavenumber-domain-based transmission framework is proposed to leverage the spatial degrees of freedom offered by CAPAs. Asymptotic approximations for the OP and ECC are derived, and the DMT is explored. The performance of CAPAs is further compared with that of conventional spatially-discrete arrays (SPDAs). Analytical and numerical results demonstrate that: i) CAPAs achieve a lower OP and higher ECC than SPDAs; ii) CAPAs achieve the same DMT as SPDAs with antenna spacing no larger than half a wavelength while attaining a higher array gain; and iii) CAPAs outperform SPDAs with antenna spacing greater than half a wavelength in terms of DMT.
BibTeX
@article{ouyang2024diversity,
  author = {Chongjun Ouyang and Zhaolin Wang and Xingqi Zhang and Yuanwei Liu},
  title = {{Diversity and Multiplexing for Continuous Aperture Array (CAPA)-Based Communications}},
  journal = {IEEE Trans. Wireless Communications},
  year = {2024},
}
Figure