Pith. sign in

REVIEW 3 major objections 8 minor 1 cited by

Directivity-Aware Degrees of Freedom Analysis for Extremely Large-Scale MIMO

T0 review · 3 major / 8 minor · reviewed 2026-08-11 · deepseek-v4-flash

Pith's one-line read Antenna directivity remaps the spatial modes an XL-MIMO array can use.

desk verdict Useful EMCC pipeline for directivity-aware wavenumber-domain coupling, but Theorem 1's integration cell is shifted half a cell from the sampling grid and the EDoF threshold is unspecified, so the validation needs a closer look. read the letter →

arxiv 2412.14657 v2 pith:VBI5PLLC submitted 2024-12-19 cs.IT eess.SPmath.IT

classification cs.ITeess.SPmath.IT MSC 94A0578A50
keywords XL-MIMOeffectivedegreesoffreedomwavenumber-domainchannelantennadirectivitymutualcouplingergodiccapacityreconfigurableintelligentsurfaces
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper asks whether the directional radiation pattern of real antenna elements changes how many independent spatial channels an extremely large-scale MIMO (XL-MIMO) surface can support, and answers yes. Working in an isotropic Rayleigh-fading channel and using the wavenumber-domain channel representation, it shows that directivity reshapes the coupling coefficients that weight each plane-wave direction, and that the effective degrees of freedom (EDoF) and ergodic capacity are governed by the distribution of these coefficients rather than by aperture alone. The paper supplies a closed-form coupling coefficient for a general $\cos^m(\theta)$ radiation pattern, a numerical simulation pipeline for arbitrary antenna structures, and simulations showing that different antenna types and element spacings produce different EDoF and capacity. This matters because EDoF is the standard measure of how many data streams a large antenna surface can carry, so the result turns antenna-element design and spacing into first-order determinants of XL-MIMO performance.

What carries the argument

The wavenumber-domain channel representation: the spatial channel is expanded as a finite sum of plane waves whose wavevectors are sampled on a grid set by the array aperture, so channel gain toward each direction becomes a scalar coupling coefficient. The paper's load-bearing identity is Theorem 1's integral, which converts the angular radiation pattern and the isotropic angle distribution into a per-cell weight $(1-k_x^2-k_y^2)^{(m-1)/2}/(2\pi)$. Threshold-based EDoF in Eqs. (19)--(20) then counts how many of these weights are dominant, and the EMCC procedure obtains the weights numerically from full-wave element patterns by least-squares projection and variance fitting.

What would settle it

Measure the spatial channel of a fabricated $10\lambda \times 10\lambda$ patch or dipole array in an isotropic scattering chamber, compute the coupling coefficients from separately measured element patterns, and count dominant singular values; if the measured EDoF disagrees with the prediction of Eqs. (19)--(20) for any antenna type or spacing, the central claim fails.

Watch

Extended reading notes

Core claim

The central claim is Theorem 1: for an element with radiation pattern $G(\theta_T,\phi_T)=\cos^m(\theta_T)$ in an isotropic scattering Rayleigh-fading channel, the wavenumber-domain coupling coefficient equals $\sigma_T^2(m_x,m_y) = \frac{1}{2\pi}\iint_{K_T(m_x,m_y)} (1-k_x^2-k_y^2)^{(m-1)/2}\,dk_x\,dk_y$, with the integral over the wavenumber cell $K_T$ defined by the array aperture. This makes the coefficient distribution pattern-dependent: $m=1$ gives uniform coefficients, $m>1$ concentrates energy at broadside, and $m=0$ gives a bowl-shaped distribution. The paper then defines EDoF as the smallest number of largest coupling coefficients at each side that together contain a fraction $\gamma$ of the total coupling power, and verifies that this statistical definition matches deterministic singular-value counting in simulation. It also demonstrates that the EMCC numerical method recovers the same coefficients from full-wave simulated patterns, and that EDoF and capacity vary with both element type and spacing, with the considered RIS array reaching higher EDoF than patch or dipole arrays.

Load-bearing premise

The load-bearing assumption is that a unit-cell full-wave simulation with periodic boundaries gives the radiation pattern of every element in the finite XL-MIMO array, so that mutual coupling and edge effects enter only through one average pattern.

Editorial extensions

If this is right

  • For ideal elements with $\cos^m$ patterns, only $m=1$ yields a uniform coupling distribution; larger $m$ concentrates gain at broadside and reduces the number of EDoF for a fixed aperture.
  • Replacing the upper-bound degree of freedom $\eta_u$ with the threshold-based $\eta_e$ in the capacity expression reproduces the simulated capacity, so weak coupling coefficients can safely be dropped from the capacity sum.
  • Element spacing changes the realized pattern through mutual coupling, so EDoF and capacity are not monotonic in spacing; the best spacing depends on the element type, with the considered RIS optimum at about $0.4375\lambda$.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Beyond the paper: because the coupling-coefficient formula is written for a generic angle distribution $p(\theta,\phi)$, the same EDoF criterion should apply to measured or ray-traced non-isotropic scattering, though the paper only demonstrates the isotropic case.
  • Beyond the paper: the unit-cell assumption can be stress-tested by comparing arrays of very different physical size, since edge elements in a finite panel are expected to deviate from the periodic-boundary pattern; measuring per-element patterns would show how much EDoF shifts.
  • Beyond the paper: the evenness of the wavenumber coupling distribution, not realized gain alone, is the quantity that tracks EDoF, suggesting a design rule for element spacing that the paper does not state explicitly.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 8 minor

Summary. The paper analyzes the effect of antenna element directivity on the effective number of degrees of freedom (EDoF) and ergodic capacity of extremely large-scale MIMO (XL-MIMO) in isotropic Rayleigh fading. It introduces a wavenumber-domain coupling coefficient that incorporates the antenna radiation pattern, derives a closed-form expression (Theorem 1) for a cos^m(theta) pattern, defines EDoF from the cumulative distribution of sorted coupling coefficients, and proposes a full-wave-simulation-based numerical method (EMCC) to extract these coefficients. Numerical results are given for patch, dipole, RIS, and hypothetical antennas, showing how coupling coefficients, EDoF, and capacity vary with element spacing.

Significance. If the results are correct, the paper extends the wavenumber-domain framework for spatially stationary XL-MIMO channels to practical directional antenna elements and provides a numerical pipeline based on full-wave simulation. The derivation of Theorem 1 is self-contained and transparent, and the idea of using coupling-coefficient distributions to compute EDoF is natural. However, the paper currently contains a load-bearing definitional issue in the wavenumber-cell assignment and an unspecified threshold in the EDoF definition, both of which affect the numerical claims. The contribution is potentially useful, but the manuscript needs substantive correction and re-validation before it can be accepted.

major comments (3)
  1. [Section III-A, Eq. (13)] The cell K_T(m_x,m_y) is defined as [m_x lambda/L_x, (m_x+1)lambda/L_x] x [m_y lambda/L_y, (m_y+1)lambda/L_y], so its lower-left corner is the sampled wavenumber from Eq. (3). In the wavenumber-domain representation of a spatially stationary field on a finite aperture, the variance associated with sample (m_x,m_y) is the integral of the angular power spectrum over the cell centered at that sample, namely [(m_x-1/2)lambda/L_x, (m_x+1/2)lambda/L_x] x [(m_y-1/2)lambda/L_y, (m_y+1/2)lambda/L_y]. The definition in (13) is therefore shifted by half a cell in each dimension. For the even spectrum in Theorem 1 this is not a harmless relabeling: for m>1, the spectral peak at k_x=k_y=0 is split among four adjacent cells, so sigma_T^2(0,0) is undercounted and the neighboring coefficients are inflated. Since Eqs. (19)-(20) sort these coefficients and Eq. (10) uses their distribution, the EDoF and capacity results change. In addition, the LS estimator in Eq. (23) projects onto Fourier basis functions whose main lobes are centered on the sampled wavenumbers, so the EMCC method estimates centered-cell variances; the agreement between Eq. (15) and EMCC shown in Fig. 3 cannot be expected to hold for non-constant spectra. Please correct the cell definition (or provide a reference that justifies the lower-left cell) and recompute the affected figures.
  2. [Section III-B, Eqs. (19)-(20), and Fig. 5] The EDoF is defined as the number of largest coupling coefficients whose cumulative power reaches a fraction gamma, but no value of gamma is reported in Section IV. The SVD-based dotted curves in Fig. 5 also count 'dominant singular values' without a stated threshold criterion. Without these thresholds, the close agreement between the solid and dotted curves in Fig. 5 is not a falsifiable validation, because gamma can be chosen to match the SVD count. Please specify gamma, specify the SVD dominance criterion, and report sensitivity of EDoF and capacity to gamma.
  3. [Section III-C, Eq. (21), and Section IV] The EMCC method obtains a single realized gain pattern from a unit-cell periodic full-wave simulation and then simulates the spatial channel by weighting each multipath by sqrt(G(theta,phi)) independently for every element. For a finite XL-MIMO array with subwavelength spacing, edge effects and element-dependent mutual coupling can make the realized pattern position-dependent; the paper provides no comparison against a finite-array full-wave simulation or a measured pattern. The coupling coefficients in Fig. 4 and the EDoF/capacity curves in Figs. 5-6 therefore rely on an unverified approximation. Please either provide such a validation or explicitly state this approximation as a limitation and discuss its expected impact.
minor comments (8)
  1. [Fig. 5] The y-axis label 'EffecitveDoF' should be 'Effective DoF'.
  2. [Fig. 3 and Section III-C] Please specify the number of multipaths S, the number of LS runs I, and clarify the y-axis formatting ('2 10-3' should be 2 x 10^-3).
  3. [Section III-B and Fig. 6] The value of gamma in (19) and the SNR or noise power mu^2 in (10) are not given; both are needed to reproduce Fig. 6.
  4. [Eq. (22) and Section III-C] Eq. (22) states that ha is CN(0, 2 sigma_T^2), while Eq. (7) writes Ha ~ CN(0, sigma^2); please reconcile the factor of 2 and state whether the plotted coupling coefficients are sigma_T^2 or 2 sigma_T^2.
  5. [Eq. (10)] The notation diag(sigma_T ⊙ sigma_T) should be defined explicitly, since sigma_T ⊙ sigma_T is a vector of variances and the diagonal operator is overloaded.
  6. [Fig. 6] The caption does not explain which curves are solid and which are dotted; the text should clarify that solid curves use eta_e and dotted curves use eta_u.
  7. [Eq. (21)] Please state that G is a power pattern and that large-scale path loss is omitted; otherwise the normalization by 1/sqrt(S) and the units of h(x,y) are ambiguous.
  8. [Title/Abstract] There are several typographical artifacts ('Extreme ly', 'influenced', and similar spacing issues) throughout the manuscript; please proofread carefully.

Circularity Check

1 steps flagged · score 4.0 of 10

Theorem 1 is self-contained; partial circularity enters only through the threshold-based EDoF definition and its threshold-matched SVD validation.

  1. other [Section III-B Eq. (19)-(20) and Section IV Fig. 5]
    "where γ < 1 is a threshold to ensure that all dominant coupling coefficients are included in the calculation of ηe. ... The dotted curves are deterministic results via performing SVD to the MIMO channel and counting the number of dominant singular values [7]."

    ηe is defined by taking the smallest prefix of the sorted coupling-coefficient list whose cumulative power fraction reaches an unstated threshold γ, while the SVD benchmark counts 'dominant' singular values with the cutoff also unspecified. Both curves in Fig. 5 are therefore counts above an arbitrary cutoff of a sorted spectrum, so a match can be manufactured by choosing the two cutoffs. The agreement is a comparison of two threshold-based definitions rather than an external falsifiable prediction of EDoF. This partial circularity affects the validation claim, not Theorem 1, which is a self-contained change-of-variables evaluation of (12).

full rationale

The central analytical result, Theorem 1, is self-contained: it substitutes p(θ,φ)=sinθ/(2π), G(θ,φ)=cos^mθ, and the Jacobian (17) into definition (12) to obtain (15), with no fitted parameter. The wavenumber-domain representation is taken from an external source ([8]), not from the authors' own prior work, and the EMCC method is validated against the same definition (12), which is a legitimate implementation check rather than a circular prediction. The capacity results in Fig. 6 are internal consistency checks computed from the same coupling coefficients. The only concrete circularity-like step is the EDoF definition in (19)-(20) and its Fig. 5 validation: both the proposed EDoF and the SVD-based benchmark require arbitrary cutoffs (γ for one, 'dominant' singular values for the other), so their agreement is partly a matter of choosing thresholds rather than an independent confirmation. The half-cell convention in K_T of (13) is a possible correctness issue, but it is an independent modeling choice and not a circularity. Overall, the derivation chain is mostly self-contained, with a partial circularity in the EDoF validation.

Assumptions & free parameters 2 free parameters · 6 assumptions · 0 invented entities

The paper rests on the wavenumber-domain representation and separability assumptions of Pizzo et al. [8], an idealized isotropic scattering model, and a unit-cell full-wave approximation for mutual coupling. The main new free parameter is the EDoF threshold gamma, which is not specified; the directivity coefficient m in Theorem 1 is a hand-chosen model parameter for the analytical example.

free parameters (2)
  • gamma (EDoF threshold) = not stated in the paper
    In Eq. (19), EDoF is defined as the number of largest coupling coefficients whose cumulative power reaches gamma of the total. The value of gamma is not reported, and the EDoF results in Fig. 5 depend on it, so the comparison with SVD is not fully external.
  • m (antenna directivity coefficient) = 0, 1, 2 in Fig. 2
    Used in Theorem 1 to model element patterns as G(theta,phi)=cos^m(theta) for the analytical example; it is a hand-chosen model parameter, not fitted to the physical antennas used in the simulations.
assumptions (6)
  • domain assumption Isotropic scattering Rayleigh channel with angular distribution p(theta,phi)=sin(theta)/(2 pi) over the half-space
    Used in the proof of Theorem 1 (Eq. 16) and in the EMCC channel simulation (Eq. 21).
  • standard math Spatially stationary complex Gaussian wavenumber-domain representation with independent coefficients, from Pizzo et al. [8]
    Invoked in Section II-B to express H = sqrt(N_T N_R) Phi_R H_a Phi_T^H and to model H_a entries as zero-mean Gaussians.
  • domain assumption Separability of coupling coefficients as sigma^2(m_x,m_y;l_x,l_y)=sigma_T^2(m_x,m_y) sigma_R^2(l_x,l_y) in isotropic scattering
    Invoked after Eq. (7) in Section II-B and used to write the capacity expression (10).
  • domain assumption Antenna directivity enters the channel only as a per-multipath multiplicative gain sqrt(G(theta,phi))
    This is the channel model in Eq. (21) used by the EMCC method; it assumes no pattern distortion beyond the element radiation pattern.
  • domain assumption Unit-cell periodic-boundary full-wave simulation captures the mutual coupling of the finite array
    Section III-C step i relies on CST unit-cell simulation to obtain realized gain; edge effects are not validated.
  • ad hoc to paper EDoF can be defined as the number of coupling coefficients carrying a chosen fraction gamma of total coupling power (Eq. 19)
    This is a proposed definition rather than a derived theorem; gamma is left unspecified, so all EDoF numbers depend on a free threshold.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Directivity-Aware Degrees of Freedom Analysis for Extremely Large-Scale MIMO." pith.science (2026). https://pith.science/paper/VBI5PLLC

@misc{pith2026241214657,
  author       = {Pith},
  title        = {Pith review of: Directivity-Aware Degrees of Freedom Analysis for Extremely Large-Scale MIMO},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VBI5PLLC}},
  note         = {Machine review of arXiv:2412.14657}
}
read the original abstract

Extremely large-scale multiple-input multiple-output (XL-MIMO) communications, enabled by numerous antenna elements integrated into large antenna surfaces, can provide increased effective degree of freedom (EDoF) to achieve high diversity gain. However, it remains an open problem that how the EDoF is influenced by the directional radiation pattern of antenna elements. In this work, empowered by the wavenumber-domain channel representation, we analyze the EDoF in a general case where the directivity of antennas, determined by the antenna structure and element spacing, is considered. Specifically, we first reveal the uneven distribution of directivity-aware wavenumber-domain coupling coefficients, i.e., channel gain towards different directions, in the isotropic Rayleigh fading channel. EDoF is then calculated based on such distribution of coupling coefficients. A numerical method is also provided to obtain coupling coefficients via electromagnetic full-wave simulations. Due to the influence of antenna directivity, how EDoF and ergodic channel capacity vary with the element spacing are explored via simulations for different antenna types.

Figures

Figures reproduced from arXiv: 2412.14657 by the authors.

Figure 1
Figure 1. The sizes of the transmit and receive antenna arrays [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 1
Figure 1. An XL-MIMO communication system in isotropic scatte [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Distribution of coupling coefficients for different [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figures from the paper (4 more)
Figure 3
Figure 3. Figure 3: Coupling Coefficients of different sizes of antenna ar [PITH_FULL_IMAGE:figures/full_fig_p004_3.png]
Figure 4
Figure 4. Figure 4: The distributions of coupling coefficients of differ [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: EDoF vs. element spacing of different antenna types. [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]
Figure 6
Figure 6. Figure 6: Ergodic channel capacity vs. element spacing of diffe [PITH_FULL_IMAGE:figures/full_fig_p005_6.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Effective DoF-Oriented Optimal Antenna Spacing in Near-Field XL-MIMO Systems

    eess.SP 2025-01 conditional novelty 6.0 of 10

    Near-field XL-MIMO effective degrees of freedom peak at antenna spacing d = sqrt(lambda*L/sqrt(N)), the same spacing where the array gain to the nearest antenna first becomes zero.

Reference graph

Works this paper leans on

14 extracted references · 14 canonical work pages · cited by 1 Pith paper

  1. [8]

    Spatially -stationary model for holographic MIMO small-scale fading,

    A. Pizzo, T. L. Marzetta, and L. Sanguinetti, “Spatially -stationary model for holographic MIMO small-scale fading,” IEEE J. Sel. Areas Commun. , vol. 38, no. 9, pp. 1964-1979, Sept. 2020

  2. [1]

    Non-stationarities in extra-large-scale massive MIMO,

    E. D. Carvalho, A. Ali, A. Amiri, M. Angjelichinoski and R . W. Heath, “Non-stationarities in extra-large-scale massive MIMO,” IEEE Wireless Commun., vol. 27, no. 4, pp. 74-80, Aug. 2020

  3. [2]

    Deep learning based be am training for extremely large-scale massive MIMO in near-field domain ,

    W. Liu, H. Ren, C. Pan, and J. Wang, “Deep learning based be am training for extremely large-scale massive MIMO in near-field domain ,” IEEE Commun. Lett. , Jan. 2023, vol. 27, no. 1, pp. 170-174

  4. [3]

    Reconfigurable holographic surfaces for ultra-massive MIMO in 6G: Practical design, optimization and implementat ion,

    R. Deng et al. , “Reconfigurable holographic surfaces for ultra-massive MIMO in 6G: Practical design, optimization and implementat ion,” IEEE J. Sel. Areas Commun. , vol. 41, no. 8, pp. 2367-2379, Aug. 2023

  5. [4]

    RIS-based IMT-2030 testbed for MmWave multi-stream ultra-massive MIMO communications,

    S. Zeng et al. , “RIS-based IMT-2030 testbed for MmWave multi-stream ultra-massive MIMO communications,” IEEE Wireless Commun , vol. 31, no. 3, pp. 375-382, June 2024

  6. [5]

    On Landau’s eigenvalue theorem f or line-of- sight MIMO channels,

    A. Pizzo and A. Lozano, “On Landau’s eigenvalue theorem f or line-of- sight MIMO channels,” IEEE Wireless Commun. Lett. , vol. 11, no. 12, pp. 2565-2569, Dec. 2022

  7. [6]

    Elect romagnetic effective-degree-of-freedom prediction with parabolic e quation method,

    Y . Shen, Z. He, W. E. I. Sha, S. S. A. Y uan and X. Chen, “Elect romagnetic effective-degree-of-freedom prediction with parabolic e quation method,” IEEE Trans. Antennas Propagat. , vol. 71, no. 4, pp. 3752-3757

  8. [7]

    Communicating with large intelligent surf aces: Fundamental limits and models,

    D. Dardari, “Communicating with large intelligent surf aces: Fundamental limits and models,” IEEE J. Sel. Areas Commun. , vol. 38, no. 11, pp. 2526–2537, Nov. 2020

Show all 14 references
  1. [9]

    Orbital-angul ar-momentum versus MIMO: Orthogonality, degree of freedom, and capacit y,

    H. Jing, W. Cheng, X. -G. Xia and H. Zhang, “Orbital-angul ar-momentum versus MIMO: Orthogonality, degree of freedom, and capacit y,” 2018 IEEE 29th Annu. Int. Symp. Personal, Indoor Mobile Radio Com mun. (PIMRC), Bologna, Italy, Sept., 2018, pp. 1-7

  2. [10]

    W. L. Stutzman and G. A. Thiele, Antenna Theory and Design , 3rd ed., New Y ork, USA: Wiley, 2012

  3. [11]

    Performance analysis and low-complexity design for XL-MIMO with near-field spatial non-stationarities,

    K. Zhi et al. , “Performance analysis and low-complexity design for XL-MIMO with near-field spatial non-stationarities,” IEEE J. Sel. Areas Commun., vol. 42, no. 6, pp. 1656-1672, June 2024

  4. [12]

    Small-scale spatial-temporal corre lation and de- grees of freedom for reconfigurable intelligent surfaces,

    S. Sun and H. Y an, “Small-scale spatial-temporal corre lation and de- grees of freedom for reconfigurable intelligent surfaces,” IEEE Wireless Commun. Lett. , vol. 10, no. 12, pp. 2698-2702, Dec. 2021

  5. [13]

    An overview of eq uivalent circuit modeling techniques of frequency selective surfac es and metasur- faces,

    F. Costa, A. Monorchio, and G. Manara, “An overview of eq uivalent circuit modeling techniques of frequency selective surfac es and metasur- faces,” Appl. Comput. Electromagn. Soc. J. , vol. 29, no. 12, pp. 1-15, Dec. 2014

  6. [14]

    Active RIS versus passive RIS: Which is superior with the same power budget?,

    K. Zhi et al., “Active RIS versus passive RIS: Which is superior with the same power budget?,” IEEE Commun. Lett., vol. 26, no. 5, pp. 1150-1154, May 2022

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.