Tri-Hybrid Multi-User Precoding Based on Electromagnetically Reconfigurable Antennas

Anas Chaaban; Md. Jahangir Hossain; Pinjun Zheng; Tareq Y. Al-Naffouri; Yuchen Zhang

REVIEW 3 major objections 4 minor 1 cited by

Tri-Hybrid Multi-User Precoding Based on Electromagnetically Reconfigurable Antennas

T0 review · 3 major / 4 minor · reviewed 2026-08-16 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Optimizing radiation patterns in a spherical-harmonics basis boosts tri-hybrid precoding sum rate, but projecting onto realizable ERA patterns erases most of the gain.

desk verdict Useful cautionary study of tri-hybrid ERA precoding, but the central negative result likely conflates physical realizability with a mismatched evaluation protocol. read the letter →

arxiv 2505.02254 v1 pith:FJBLGMHZ submitted 2025-05-04 eess.SP

Pinjun Zheng , Yuchen Zhang , Tareq Y. Al-Naffouri , Md. Jahangir Hossain , Anas Chaaban This is my paper · ORCID

classification eess.SP

keywords tri-hybridprecodingelectromagneticallyreconfigurableantennasradiationpatternoptimizationsphericalharmonicsweightedMMSEMISOdownlinksumrate

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 tries to establish how much of the tri-hybrid precoding gain promised by electromagnetically reconfigurable antennas survives contact with real hardware. It models each antenna's radiation pattern with a truncated real spherical-harmonic expansion, jointly optimizes the EM-domain, analog, and digital precoders through an alternating WMMSE solver with closed-form updates, and then projects the optimized patterns onto a set of 64 realizable patterns from an ERA prototype. Unconstrained spherical-harmonic optimization produces large sum-rate gains over conventional hybrid precoding, but after projection the gain shrinks markedly and at the highest transmit power tested the projected scheme falls below the conventional baseline. This matters because ERA-based tri-hybrid MIMO is being proposed as a way to increase data rates without extra spectrum, and the result suggests that realized gains depend on hardware expressiveness and projection quality, not just on the optimization algorithm.

What carries the argument

The load-bearing object is the truncated real spherical-harmonic representation of each antenna's radiation pattern, $G^{(n)}(\theta,\phi)\approx b(\theta,\phi)^{\mathsf T}c^{(n)}$, collected into the block-diagonal EM precoder $F_{\mathrm{EM}}$. That representation turns each antenna's pattern into a vector of coefficients, so the channel factorizes as $h_k=F_{\mathrm{EM}}^{\mathsf T}h_k^{\mathrm{EM}}$ and the radiation pattern can be optimized jointly with the analog and digital precoders. The solution alternates WMMSE updates for auxiliary variables, a fictitious fully digital precoder $F_D$, and each coefficient vector $c^{(n)}$, with Theorem 1 guaranteeing two locally optimal Lagrange multipliers that can be found by bisection. A final projection step replaces each optimized pattern by the nearest physically realizable pattern from a 64-pattern hardware library, which is what exposes the gap between relaxed and realizable performance.

What would settle it

Directly comparing measured or full-wave-simulated sum rates for the 3×3 ERA array under the unconstrained spherical-harmonic patterns and under the projected discrete patterns would settle the claim. If the unconstrained patterns also fail to deliver the modeled gain, the spherical-harmonic channel model is the source of the overestimate; if they match the model but the projected patterns lose the gain, the projection library is the bottleneck.

Watch

Extended reading notes

Core claim

The paper's central discovery is that the apparent benefit of reconfiguring each antenna's radiation pattern in tri-hybrid precoding depends almost entirely on how much freedom the pattern model has. When the radiation pattern is optimized over a truncated spherical-harmonic basis, the sum-rate gain over conventional hybrid precoding is large. But the paper then projects each optimized pattern onto the closest of 64 radiation patterns measured from a real ERA prototype, and most of that gain disappears; at the highest power simulated ($P_{\max}=20$ dBm), the projected tri-hybrid scheme falls below the conventional hybrid baseline. The authors read this as evidence that relaxed pattern optimization overstates ERA benefits, and that both better projection methods and richer reconfigurable-antenna hardware are needed before the tri-hybrid architecture delivers its theoretical promise.

Load-bearing premise

The whole comparison rests on the channel model in (3) and the truncated spherical-harmonic representation (12), which assume each antenna's far-field gain pattern stays valid when mounted in the 3×3 array and that mutual coupling and polarization effects can be ignored.

Editorial extensions

If this is right

With the 64-pattern realizable library used in the paper, the unconstrained spherical-harmonic gains are mostly out of reach, so ERA hardware that offers richer or more application-tuned pattern sets would be needed to realize them.
At high transmit power (such as $P_{\max}=20$ dBm), the projected tri-hybrid scheme can fall below the conventional hybrid baseline, so realizability should enter the optimization as a constraint rather than as a post-hoc repair.
Because the channel model covers both far-field and near-field paths, the same alternating solver applies to near-field XL-MIMO scenarios without needing a separate formulation.
The monotonicity result for the two Lagrange-multiplier intervals makes each coefficient-vector update computationally cheap via bisection, so the method scales to arrays larger than the 3×3 test case.

Reading between the lines

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

A testable extension would be to optimize directly over the convex hull of the 64 realizable patterns instead of projecting after the fact; the negative result suggests this would recover much of the lost gain if the bottleneck is the projection, and would not if the bottleneck is the limited pattern library.
If the same relaxed-then-projected pattern holds for other reconfigurable-antenna architectures, the lesson generalizes: theoretical pattern degrees of freedom are not a reliable proxy for achievable rate until hardware expressiveness is priced in.
The 64-pattern library comes from one antenna prototype; interpolating or learning a continuous manifold between those patterns would separate hardware limits from projection losses and could guide hardware designers on which patterns to add.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Desk Editor's Note

read the letter

Quick take: the genuinely new thing is taking the spherical-harmonics EM-domain channel model, solving the tri-hybrid WMMSE in closed form, then projecting onto the 64 measured patterns from the authors' own ERA hardware and showing the gain shrinks. That last comparison is the paper's punchline, and it is where the trouble is.

What is good: the spherical-harmonics truncation is sensible; the alternating WMMSE derivation is standard but clean, and Theorem 1 gives a practical bisection for the two local optima of each EM-precoder update. Using a real measured pattern set as the projection target is a real step beyond the discretized angular-domain models in [8]. The paper says plainly that relaxed pattern optimization overstates gains, and that honest statement deserves credit.

The load-bearing flaw: after projecting the optimized patterns onto the realizable set, the paper does not re-solve for F_RF and F_BB. Algorithm 1 optimizes all three jointly. If you swap in the nearest realizable patterns and keep the old analog/digital precoders, sum rate can only drop, even when the realizable patterns are perfectly good for the problem. So Fig. 2(b)'s collapse at high power, which the abstract and conclusion lean on, measures the cost of not re-optimizing rather than the intrinsic cost of realizability. This is not a minor detail; it is the central empirical claim. The authors call the projection "suboptimal" in the conclusion, but they present the result as evidence that real ERA hardware erases the gain. That inference is not established by this comparison.

Other soft spots are secondary: no Monte Carlo runs or error bars; channel generation details are thin; the positivity guarantee via a fixed eta is asserted without quantification. Both local optima are evaluated, which is fine.

Bottom line: the optimization machinery is competent and the projection idea is worth publishing, but the negative result needs to be re-tested by re-running Algorithm 1 with projected patterns, or at least by matching benchmarks to the same realizable set. A serious referee should ask for that. I'd send it to review, conditional on fixing the evaluation protocol.

Referee Report

3 major / 4 minor

Summary. The paper considers a tri-hybrid multi-user precoding architecture in which a base station with electromagnetically reconfigurable antennas (ERAs) jointly optimizes baseband, RF, and EM-domain precoders. Each antenna's radiation pattern is represented by a truncated spherical-harmonic expansion, and the weighted sum-rate problem is solved via an alternating WMMSE-type algorithm with closed-form updates. After optimization, the coefficient vectors are projected onto a set of 64 experimentally validated radiation patterns taken from the authors' earlier hardware work. Simulations show large sum-rate gains for the unconstrained spherical-harmonic optimization, but after projection the gains shrink and can become negative at high transmit power, leading the authors to conclude that relaxed pattern optimization overstates the benefit of ERAs.

Significance. If the conclusion were established, the paper would be a useful cautionary result for the tri-hybrid MIMO literature, which often relies on relaxed or discrete radiation-pattern models. The algorithmic contribution has merit: the WMMSE reformulation is standard, the per-antenna EM update in Theorem 1 is a plausible closed-form solution with a practical bisection procedure, and the use of a real hardware pattern set is a strength. However, the load-bearing comparison in Fig. 2(b) is compromised by the evaluation protocol, so the main message is not yet convincing. With a fair comparison and full simulation details, the paper could become a solid contribution.

major comments (3)

[Section IV-C, Eqs. (28)-(29); Section V, Fig. 2(b)] After replacing each optimized coefficient vector c_opt^(n) with a realizable pattern via (28)-(29), the manuscript does not state that F_RF and F_BB are re-optimized. Since Algorithm 1 solves F_RF, F_BB, and F_EM jointly, reusing the F_RF and F_BB obtained for the unconstrained F_EM with a different, projected F_EM yields a suboptimal combination; the resulting sum-rate drop measures the cost of not re-solving the tri-hybrid problem for the realizable patterns, not the intrinsic cost of physical realizability. The abstract and conclusion rely on exactly this comparison to claim the gain is "notably reduced or even negligible." Please either re-optimize F_RF and F_BB (for example, by running Algorithm 1 with F_EM fixed to the selected realizable patterns) or explicitly present the curves as the performance of the relaxed solution after a one-shot pattern replacement.
[Section IV-B, Eqs. (18)-(19); Section V] The claim that fixing c_DC^(n) = eta in (0, sqrt(4*pi)) with a "sufficiently large" eta ensures strictly positive radiation patterns is not quantified. For the truncated expansion G^(n)(theta,phi) = Y_0^0(theta,phi)*eta + b_AC(theta,phi)^T c_AC^(n) with ||c_AC^(n)|| = sqrt(4*pi - eta^2), the minimal value over the sphere depends on the sup-norm of the truncated harmonics. The chosen eta = sqrt(2*pi) in Section V is not justified by any analytic bound or by a numerical positivity check. Without this, the unconstrained upper bound may be physically meaningless even before projection. Please provide a lower bound on eta for the given U or verify positivity of the optimized patterns on a dense angular grid.
[Section V] The simulation setup is insufficiently specified for the quantitative claims. The paper does not report the number of Monte Carlo realizations, the distribution of the complex gains alpha_{k,ell}, the path angles and their distributions, the array geometry (element spacing, near-field versus far-field assumption), or the radiation pattern used for the conventional hybrid baseline. Consequently, the magnitudes in Fig. 2(b), including the reported crossover at P_max = 20 dBm, cannot be reproduced or independently assessed. Please provide complete parameters, run counts, and error bars.

minor comments (4)

[Throughout] There are several typographical errors, including "preconding" in Section II, "recorders" instead of "precoders" in Section I, and the heading "CONSLUSION" in Section VI.
[Figure 2(a)] The axis label reads "Decomposition (Step 5 in Algo. 1)", but the text describes the plot as sum rate versus iteration number of Algorithm 1; the label should be corrected.
[Section III-B.1] The truncation order U is not stated explicitly; please state that T = 25 corresponds to U = 4, and clarify the choice of truncation length.
[Section IV-C] The projection criterion in (29) is based on Euclidean distance at the sampled path angles only; a pattern close at those angles may differ substantially elsewhere. A brief justification of this metric, or a discussion of its limitations, would strengthen the paper.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the derivation is self-contained and the post-projection performance drop is an acknowledged evaluation limitation, not a fitted prediction.

full rationale

The paper's derivation is self-contained and does not reduce any central result to its own inputs by construction. The channel factorization h_k = F_EM^T h_EM^k in Section III-B is a representation identity built from the spherical-harmonics expansion (12), not a fitted relation; the EM precoder F_EM is an optimization variable whose coefficients are updated from the KKT conditions in (24)-(26). The WMMSE equivalence of (2) and (14) and the closed-form updates (15)-(17) are standard results from [17], and the comparison baselines [10] and [17] are external benchmarks. The only self-citations are [7], which supplies the spherical-harmonics background and the 4π normalization, and [3], which supplies the 64 realizable radiation patterns. Both are inputs or standard mathematics rather than conclusions of this paper: the pattern set is experimentally characterized hardware data, and the spherical-harmonics basis is a standard orthonormal basis also cited to [9]. The projection step (28)-(29) is explicitly treated as a suboptimal heuristic; the paper states that 'the adopted projection steps (28) and (29) are suboptimal,' so the post-projection performance drop in Fig. 2(b) is an acknowledged evaluation limitation rather than a fitted input renamed as a prediction. No equation in the paper is equivalent to its own inputs, and no load-bearing argument depends on an unverified self-citation.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

No new physical entities are introduced. The EM precoder and spherical-coefficient vectors are mathematical constructs. The central claim rests on the channel model, the SH truncation, the hand-set DC coefficient, and the hardware pattern set, all listed above.

free parameters (2)

T (spherical harmonics truncation length) = 25
Hand-picked in Section V. It controls the expressiveness of the radiation pattern model and the dimension of the EM precoder.
eta (DC spherical harmonic coefficient) = sqrt(2*pi)
Fixed in Section IV-B1 and in simulation to help ensure strictly positive radiation patterns. No quantitative bound is given to show this value is sufficient for all channels.

assumptions (5)

standard math Real spherical harmonics form a complete orthonormal basis on the sphere, and truncation at T=25 captures the needed radiation patterns.
Used in (9)-(12) to represent any radiation pattern and to derive the EM-domain channel representation.
domain assumption The channel is the Hadamard product of complex gain, antenna gain, and array response as in (3), with antenna gain evaluated at each path's AoD and no mutual coupling or polarization effects.
This is load-bearing for the entire EM-domain channel model in Section III.
standard math The WMMSE reformulation (14) is equivalent to the weighted sum-rate problem (2), and the alternating updates converge to a useful stationary point.
This is the basis of Algorithm 1 and all closed-form update rules.
ad hoc to paper Fixing a sufficiently large DC coefficient eta ensures strictly positive radiation patterns throughout the optimization.
Stated in Section IV-B1 without a quantitative bound; eta=sqrt(2*pi) is used in simulation.
domain assumption The 64 measured patterns from [3] are representative of real ERA hardware in a 3x3 array configuration.
Used as the realizable set in Section V. If these patterns are not representative or change when mounted in the array, the projection result changes.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Tri-Hybrid Multi-User Precoding Based on Electromagnetically Reconfigurable Antennas." pith.science (2026). https://pith.science/paper/FJBLGMHZ

@misc{pith2026250502254,
  author       = {Pith},
  title        = {Pith review of: Tri-Hybrid Multi-User Precoding Based on Electromagnetically Reconfigurable Antennas},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FJBLGMHZ}},
  note         = {Machine review of arXiv:2505.02254}
}

read the original abstract

The tri-hybrid precoding architecture based on electromagnetically reconfigurable antennas (ERAs) is a promising solution for overcoming key limitations in multiple-input multiple-output communication systems. Aiming to further understand its potential, this paper investigates the tri-hybrid multi-user precoding problem using pattern reconfigurable ERAs. To reduce model complexity and improve practicality, we characterize each antenna's radiation pattern using a spherical harmonics decomposition. While mathematically tractable, this approach may lead to over-optimized patterns that are physically unrealizable. To address this, we introduce a projection step that maps the optimized patterns onto a realizable set. Simulation results demonstrate that spherical harmonics-based radiation pattern optimization significantly enhances sum rate performance. However, after projection onto a realizable set obtained from real ERA hardware, the performance gain is notably reduced or even negligible, underscoring the need for more effective projection techniques and improved reconfigurable antenna hardware.

Figures

Figures reproduced from arXiv: 2505.02254 by the authors.

**Figure 2.** Evaluation of (a) sum-rate versus iteration in Algorithm 1 and [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗

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

Tri-Hybrid Multi-User Precoding Using Pattern-Reconfigurable Antennas: Fundamental Models and Practical Algorithms
eess.SP 2025-05 conditional novelty 5.0 of 10

A tri-hybrid MIMO design with pattern-reconfigurable antennas and spherical-harmonics based optimization lifts weighted sum-rate by up to 5.8 bps/Hz and can cut RF chains or antennas, while current 64-state hardware d...

Reference graph

Works this paper leans on

18 extracted references · 13 canonical work pages · cited by 1 Pith paper

[8]

Tri-timescale beamforming design for tri- hybrid architectures with reconfigurable antennas,

M. Liu, M. Liet al., “Tri-timescale beamforming design for tri- hybrid architectures with reconfigurable antennas,”arXiv preprint arXiv:2503.03620, 2025

arXiv 2025
[1]

A survey on beyond 5G network with the advent of 6G: Architecture and emerging technologies,

A. Dogra, R. K. Jhaet al., “A survey on beyond 5G network with the advent of 6G: Architecture and emerging technologies,”IEEE Access, vol. 9, pp. 67 512–67 547, 2021

work page 2021
[2]

Massive MIMO for next generation wireless systems,

E. G. Larsson, O. Edforset al., “Massive MIMO for next generation wireless systems,”IEEE Communications Magazine, vol. 52, no. 2, pp. 186–195, 2014

work page 2014
[3]

Electromagnetically reconfigurable fluid antenna system for wireless communications: Design, modeling, algorithm, fabrication, and experiment,

R. Wang, P. Zhenget al., “Electromagnetically reconfigurable fluid antenna system for wireless communications: Design, modeling, algorithm, fabrication, and experiment,” 2025. [Online]. Available: https://arxiv.org/abs/2502.19643

arXiv 2025
[4]

Embracing reconfigurable antennas in the tri-hybrid MIMO architecture for 6G,

M. R. Castellanos, S. Yanget al., “Embracing reconfigurable antennas in the tri-hybrid MIMO architecture for 6G,”preprint arXiv:2501.16610, 2025

arXiv 2025
[5]

Linear polarization optimization for wideband MIMO systems with reconfigurable arrays,

M. R. Castellanos and R. W. Heath, “Linear polarization optimization for wideband MIMO systems with reconfigurable arrays,”IEEE Transactions on Wireless Communications, vol. 23, no. 3, pp. 2282–2295, 2024

work page 2024
[6]

Energy-efficient tri-hybrid precoding with dynamic metasurface antennas,

M. R. Castellanos, J. Carlsonet al., “Energy-efficient tri-hybrid precoding with dynamic metasurface antennas,” in57th Asilomar Conference on Signals, Systems, and Computers, 2023, pp. 1625–1630

work page 2023
[7]

Enhanced beampattern synthesis using electromagnetically reconfigurable antennas,

P. Zheng, M. J. Hossainet al., “Enhanced beampattern synthesis using electromagnetically reconfigurable antennas,” 2025. [Online]. Available: https://arxiv.org/abs/2503.03856

arXiv 2025

Show all 18 references

[9]

Reconfigurable massive MIMO: Precoding design and channel estimation in the electromagnetic domain,

K. Ying, Z. Gaoet al., “Reconfigurable massive MIMO: Precoding design and channel estimation in the electromagnetic domain,”IEEE Transactions on Communications, pp. 1–1, 2024, early access

2024
[10]

Hybrid digital and analog beamforming design for large-scale antenna arrays,

F. Sohrabi and W. Yu, “Hybrid digital and analog beamforming design for large-scale antenna arrays,”IEEE Journal of Selected Topics in Signal Processing, vol. 10, no. 3, pp. 501–513, 2016

2016
[11]

Channel estimation for hybrid architecture-based wideband millimeter wave systems,

K. Venugopal, A. Alkhateebet al., “Channel estimation for hybrid architecture-based wideband millimeter wave systems,”IEEE Journal on Selected Areas in Communications, vol. 35, no. 9, pp. 1996–2009, 2017

1996
[12]

Spatially sparse precoding in millimeter wave MIMO systems,

O. E. Ayach, S. Rajagopalet al., “Spatially sparse precoding in millimeter wave MIMO systems,”IEEE Transactions on Wireless Communications, vol. 13, no. 3, pp. 1499–1513, 2014

2014
[13]

TeraMIMO: A channel simulator for wideband ultra-massive MIMO terahertz communications,

S. Tarboush, H. Sarieddeenet al., “TeraMIMO: A channel simulator for wideband ultra-massive MIMO terahertz communications,”IEEE Transactions on V ehicular Technology, vol. 70, no. 12, pp. 12 325–12 341, 2021

2021
[14]

Coverage analysis of joint localization and communication in THz systems with 3D arrays,

P. Zheng, T. Ballalet al., “Coverage analysis of joint localization and communication in THz systems with 3D arrays,”IEEE Transactions on Wireless Communications, vol. 23, no. 5, pp. 5232–5247, 2024

2024
[15]

Spatial non-stationary near-field channel modeling and validation for massive MIMO systems,

Z. Yuan, J. Zhanget al., “Spatial non-stationary near-field channel modeling and validation for massive MIMO systems,”IEEE Transactions on Antennas and Propagation, vol. 71, no. 1, pp. 921–933, 2023

2023
[16]

A tutorial on near-field XL-MIMO communications toward 6G,

H. Lu, Y . Zenget al., “A tutorial on near-field XL-MIMO communications toward 6G,”IEEE Communications Surveys & Tutorials, vol. 26, no. 4, pp. 2213–2257, 2024

2024
[17]

An iteratively weighted MMSE approach to distributed sum-utility maximization for a MIMO interfering broadcast channel,

Q. Shi, M. Razaviyaynet al., “An iteratively weighted MMSE approach to distributed sum-utility maximization for a MIMO interfering broadcast channel,”IEEE Transactions on Signal Processing, vol. 59, no. 9, pp. 4331–4340, 2011

2011
[18]

Hybrid precoder design for angle- of-departure estimation with limited-resolution phase shifters,

H. Huang, M. F. Keskinet al., “Hybrid precoder design for angle- of-departure estimation with limited-resolution phase shifters,”IEEE Transactions on Communications, 2024, early access

2024

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.

[1] [8]

Tri-timescale beamforming design for tri- hybrid architectures with reconfigurable antennas,

M. Liu, M. Liet al., “Tri-timescale beamforming design for tri- hybrid architectures with reconfigurable antennas,”arXiv preprint arXiv:2503.03620, 2025

arXiv 2025

[2] [1]

A survey on beyond 5G network with the advent of 6G: Architecture and emerging technologies,

A. Dogra, R. K. Jhaet al., “A survey on beyond 5G network with the advent of 6G: Architecture and emerging technologies,”IEEE Access, vol. 9, pp. 67 512–67 547, 2021

work page 2021

[3] [2]

Massive MIMO for next generation wireless systems,

E. G. Larsson, O. Edforset al., “Massive MIMO for next generation wireless systems,”IEEE Communications Magazine, vol. 52, no. 2, pp. 186–195, 2014

work page 2014

[4] [3]

Electromagnetically reconfigurable fluid antenna system for wireless communications: Design, modeling, algorithm, fabrication, and experiment,

R. Wang, P. Zhenget al., “Electromagnetically reconfigurable fluid antenna system for wireless communications: Design, modeling, algorithm, fabrication, and experiment,” 2025. [Online]. Available: https://arxiv.org/abs/2502.19643

arXiv 2025

[5] [4]

Embracing reconfigurable antennas in the tri-hybrid MIMO architecture for 6G,

M. R. Castellanos, S. Yanget al., “Embracing reconfigurable antennas in the tri-hybrid MIMO architecture for 6G,”preprint arXiv:2501.16610, 2025

arXiv 2025

[6] [5]

Linear polarization optimization for wideband MIMO systems with reconfigurable arrays,

M. R. Castellanos and R. W. Heath, “Linear polarization optimization for wideband MIMO systems with reconfigurable arrays,”IEEE Transactions on Wireless Communications, vol. 23, no. 3, pp. 2282–2295, 2024

work page 2024

[7] [6]

Energy-efficient tri-hybrid precoding with dynamic metasurface antennas,

M. R. Castellanos, J. Carlsonet al., “Energy-efficient tri-hybrid precoding with dynamic metasurface antennas,” in57th Asilomar Conference on Signals, Systems, and Computers, 2023, pp. 1625–1630

work page 2023

[8] [7]

Enhanced beampattern synthesis using electromagnetically reconfigurable antennas,

P. Zheng, M. J. Hossainet al., “Enhanced beampattern synthesis using electromagnetically reconfigurable antennas,” 2025. [Online]. Available: https://arxiv.org/abs/2503.03856

arXiv 2025