Energy efficiency of DMAs vs. conventional MIMO: a sensitivity analysis

Beatriz Soret; David Morales-Jim\'enez; Pablo Ram\'irez-Espinosa

REVIEW 4 major objections 5 minor 1 cited by

Energy efficiency of DMAs vs. conventional MIMO: a sensitivity analysis

T0 review · 4 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read This paper argues that dynamic metasurface antennas (DMAs) achieve higher energy efficiency than fully-digital and hybrid MIMO arrays in low-power, strongly coupled, and densely packed scenarios, and traces the advantage to the absence of…

desk verdict Useful framework and sensitivity analysis, but the hybrid-array power constraint is normalized wrong by a factor K, which inflates the DMA advantage and weakens the central scalability claim. read the letter →

arxiv 2506.09181 v1 pith:TBCEMZQ4 submitted 2025-06-10 eess.SP

Pablo Ramírez-Espinosa , David Morales-Jiménez , Beatriz Soret This is my paper

classification eess.SP

keywords energyefficiencydynamicmetasurfaceantennashybridMIMOfully-digitalarraysmutualcouplinginsertionlossestransmitWienerfilterpowerconsumptionmodel

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 that dynamic metasurface antennas (DMAs) can be more energy-efficient than fully-digital and hybrid arrays in several realistic conditions, and to show where that advantage comes from. Using a circuit-theoretic model that includes mutual coupling and insertion losses, it derives transmit Wiener filters for all three architectures. It then compares energy efficiency under linear and nonlinear amplifier models and varying supplied power. The central finding is that DMAs win when supplied power is low, when mutual coupling is strong, and when many radiating elements must fit in a small aperture, while classical arrays improve more when power is plentiful. The paper also warns that absolute efficiency numbers are sensitive to component consumption assumptions, so trends matter more than point values.

What carries the argument

The central object is the circuital (admittance-matrix) model inherited from [6,8]: every antenna is a z-oriented magnetic dipole on a conducting plane, with the array described by coupling matrix $\mathbf{Y}_{aa}$ and the DMA by waveguide coupling $\mathbf{Y}_{ss}$, a tunable load $\mathbf{Y}_s = R_s\mathbf{I}_N + i\mathbf{Y}^{\mathrm{im}}_s$, and mutual admittances $\mathbf{Y}_{st}$ and $\mathbf{Y}_{tt}$. These matrices yield closed-form equivalent channels $\mathbf{H}_a$ and $\mathbf{H}_d$, and the transmit Wiener filter is the linear precoder minimizing MSE under the constraint $\frac{1}{2}Y_g\mathrm{Tr}\{\mathbf{B}^H\mathbf{B}\} \le P^{\max}_g$. The gradient expressions for the DMA tuning parameters and for the hybrid phase-shifter matrix are what allow joint optimization of the analog and digital parts, and the energy-efficiency comparison rests on the total power model $P_{\mathrm{total}} = P_{bb} + N_t(2P_{dac}+P_{rf}) + P_a + N_{ps}P_{ps} + N_{var}P_{var}$.

What would settle it

A measurement campaign comparing a fabricated DMA prototype with a conventional phased array at the same frequency, aperture, and supplied power, measuring wall-plug energy per correctly received bit, would settle the ordering if the DMA's measured efficiency falls below the array once realistic waveguide attenuation and varactor losses are included; a simpler calculation is to re-run the paper's model with a lossy waveguide and check whether the DMA curves cross the hybrid curves at typical loss values.

Watch

Extended reading notes

Core claim

By modelling each antenna as a magnetic dipole over a conducting plane and writing the whole transceiver as admittance matrices, the authors obtain equivalent channels for fully digital, hybrid, and DMA topologies. Feeding these channels into a transmit Wiener filter under a supplied-power constraint, they find that DMA structures achieve higher bits-per-Joule than hybrid arrays in essentially all tested settings, and beat fully-digital arrays once the number of radiating elements per transmitter is large. The DMA advantage is traced to two mechanisms: it needs no phase-shifter hardware and its tunable reactive loads can be adjusted to reduce reflection caused by mutual coupling, so each added element contributes useful radiated power even at sub-wavelength spacing. The advantage is largest at low maximum supplied power; at high power, array solutions catch up because the DMA's Lorentzian amplitude-phase response limits how much it can exploit extra power.

Load-bearing premise

The circuit model treats every antenna as an ideal magnetic dipole on a perfectly conducting plane and assumes the feeding network, phase shifters, and DMA waveguides are lossless; if real hardware adds losses or different radiation patterns, the efficiency ordering between DMAs and arrays could change.

Editorial extensions

If this is right

With many antennas per transmitter in a constrained aperture, DMAs yield higher energy efficiency than hybrid arrays and, beyond a moderate array size, higher than fully-digital arrays, as shown in Figs. 2 and 5.
At low maximum supplied power the DMA is the most efficient architecture, but at high supplied power array-based solutions benefit more from the extra power, as shown in Fig. 4.
The choice of amplifier model (linear versus nonlinear) substantially shifts the efficiency ordering, so conclusions should be drawn from trends rather than from absolute efficiency values, as shown in Fig. 3.
Reducing inter-antenna spacing below about 0.4 wavelengths degrades conventional arrays, whereas DMAs keep improving because their tunable loads compensate the reflection caused by stronger mutual coupling, as shown in Fig. 5.
The DMA advantage over hybrid arrays persists when RF-chain consumption is raised from 40 mW to 400 mW, but the fully-digital architecture becomes comparatively less attractive at the higher RF-chain consumption.

Reading between the lines

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

If waveguide attenuation and feeding-network losses were included in the DMA model, the DMA efficiency advantage would likely shrink; a direct extension would be to re-run the analysis with a complex propagation constant in $\mathbf{Y}_{tt}$.
The Lorentzian amplitude-phase coupling in DMA elements explains why the DMA cannot fully exploit high supplied power, suggesting that hybrid arrays with active phase shifters may be the better choice in high-power regimes.
The same circuital Wiener-filter framework could be applied, with minimal changes, to other reconfigurable-surface architectures such as holographic MIMO or RIS-aided transmitters, giving a fairer comparison than codebook-based studies.
Because the paper shows that small changes in component consumption values shift efficiency rankings, a standardized power-consumption model would be needed before energy-efficiency numbers from different papers can be meaningfully compared.

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.

Desk Editor's Note

read the letter

The genuinely new piece here is the multi-user transmit Wiener filtering for DMA, hybrid, and fully-digital arrays under a circuit-aware model, plus the sensitivity analysis across supply power, spacing, amplifier models, and RF-chain consumption. The gradient expressions for hybrid and DMA tuning look new, and the paper is clearly written about its own caveats. That said, the stress-test note lands. For the hybrid, B = Q B_h with Q unit-modulus and K = N/N_t antennas per transmitter. Then Tr{B^H B} = K Tr{B_h^H B_h}, so the power constraint in (16b) is K times the actual transmitter-output power. A passive lossless phase-shifting network cannot provide this power gain; equal splitting requires entries of magnitude K^{-1/2}. The paper even uses the K factor in footnote 4 when computing amplifier output power, which confirms the error. The effect is that the hybrid is artificially handicapped by a factor K in the optimization, and its MSE and energy efficiency suffer exactly in the regime where the DMA advantage is most visible (Figs. 2 and 5). This is not a minor modeling detail; it directly biases the main scalability conclusion. The reader's concern about lossless feed networks and magnetic-dipole antennas is secondary — real, but the normalization problem should be settled first. Other soft spots are the omitted gradient proofs (acceptable in principle, harder to check without them) and the absence of error bars or code, which makes the numerical rankings hard to verify. The paper is a serious contribution in terms of scope and honesty, but the quantitative comparisons should not be taken at face value until the hybrid power constraint is corrected. I would still send it to peer review: the flaw is fixable and the framework is useful. But it needs heavy revision before the energy-efficiency ordering can be trusted. For me, it's not citable in its current form.

Referee Report

4 major / 5 minor

Summary. The paper develops a circuit-theoretic communication model for three MIMO architectures — fully digital arrays, hybrid arrays, and dynamic metasurface antennas — and derives transmit Wiener-filter precoders under a supplied-power constraint. It then evaluates energy efficiency as a function of the number of antennas per transmitter, maximum supplied power, amplifier model, RF-chain consumption, and inter-element spacing, using both linear and nonlinear PA models with a fixed baseline set of hardware parameters. The main conclusion is that DMAs become more energy-efficient than FD and hybrid arrays at low supply power, under strong mutual coupling, and when many radiating elements are packed into a fixed aperture, while the paper also cautions that absolute efficiency numbers are sensitive to the chosen hardware and consumption models.

Significance. If the results are correct, the paper provides a useful unified comparison of three MIMO front-ends that includes mutual coupling and insertion losses, and the closed-form Wiener-filter expressions for FD and DMA are useful building blocks. The paper is honest about the sensitivity of absolute EE values to hardware parameters, and it identifies low-power, strongly coupled, and densely packed regimes as those favoring DMAs. The use of the authors' own prior circuit model [6,8] is acceptable: that model is a published physical derivation and is not tailored to the energy-efficiency objective, so the circularity concern does not land. However, the quantitative support for the DMA-versus-hybrid ranking is compromised by an error in the hybrid PA power-consumption formula used in the simulations; because that error grows with the number of antennas per transmitter, it directly inflates the DMA advantage in the scalability plots. The gradient-based solutions are also stated without derivations, and the figures do not report Monte Carlo variability. These issues are fixable, but the numerical conclusions should be re-established after correction.

major comments (4)

[V, footnote 4, Eqs. (9), (16b)] The hybrid-array PA power computation in the non-linear model overestimates the PA consumption by a factor K=N/N_t. With B=Q B_h and K unit-modulus entries per column of Q, the true output power of the transmitter serving subarray S_t is P_out,t=(Y_g/2) \sum_{n\in S_t}|(j_g)_n|^2=(Y_g/2)K|(B_h)_t|^2, so the non-linear PA term should be \sum_{t=1}^{N_t} sqrt(P_out,t P_sat/\eta). The paper instead evaluates Y_g(N/N_t)|(j_g)_n|^2/2 for every one of the N antenna ports and sums over n, which gives K times the correct value. The inflation grows with K and therefore biases the DMA-versus-hybrid comparison in Figs. 2, 3, and 5 in the DMA's favor exactly in the scalability regime emphasized in the abstract. Please correct this and rerun, or explain why summing over antenna ports is physically intended.
[IV-B, Eq. (20), and IV-C, Eq. (25)] The gradients for the hybrid phase-shift matrix and for the DMA tunable admittance are stated without proof ('due to space constraints'), and Eq. (25) is the only optimization step for the DMA structure that determines the reported MSE/EE results. A journal paper should include these derivations, or at least a sketch in an appendix, especially because (25) contains an undefined dimension N_rf and the notation Yim_s is not formally introduced; as printed, N_rf should presumably be N_t, but this needs to be stated and the gradient expression verified.
[IV-A, Eqs. (14)-(15), (17)-(18), and IV-C, Eqs. (22)-(23)] The definition of A (and \tilde A) is inconsistent with the formulas that use it. If A is defined as (H_a^H H_a + \lambda I_N)^{-1} as printed, then B=\beta A^{-1} H_a^H equals \beta(H_a^H H_a + \lambda I_N)H_a^H, which is not the Wiener filter; the expressions only become correct if A denotes H_a^H H_a + \lambda I_N, or if the exponent -1 is moved elsewhere. Please rewrite these equations consistently, including the traces in (15), (18), and (23), so that the closed-form solutions can be checked.
[V, Figs. 2-5] The curves are presented without error bars or any statement of the number of channel realizations, despite the use of random Rayleigh-fading channels. Given that several comparisons rest on small efficiency differences (e.g., hybrid versus DMA in Fig. 3 at moderate N/N_t), the authors should report the number of Monte Carlo runs and add confidence intervals, or at least quantify the run-to-run variability, so that the statistical significance of the ordering is clear.

minor comments (5)

[II-C1] The sentence 'jg is the output of the transmitters' is inaccurate for hybrid arrays, where j_g=B x has length N while there are only N_t transmitters; please clarify that B_h is the digital signal and Q B_h is the antenna-port excitation.
[Eq. (25)] Use a consistent notation for the tunable admittance, e.g., Y_s^im or Y_s^{im}; the subscript/superscript placement varies between Section II-B and Eq. (25).
[V, Fig. 5 caption] State explicitly what is kept fixed (aperture, z-spacing) and the range of N/N_t, since the plot has multiple curves and the caption only says 'different antenna spacings.'
[Abstract] The phrase 'diverse tested scenarios' overstates the number of configurations; the paper tests a few parameter sets around one baseline at 10 GHz. Consider softening to 'several.'
[VI] There is a typo in 'divi ng further into the impact of power consumption models'; it should be 'diving' or 'delving.'

Circularity Check

1 steps flagged · score 6.0 of 10

Hybrid-array supplied-power normalization embeds an unphysical K-fold power gain, biasing the DMA-vs-hybrid efficiency comparison by construction.

other [Section II-C (Eq. 9), Section IV-B (Eq. 16b), and footnote 4 in Section V]
"Considering a lossless and perfectly matched phase shifting network, (Q)n,n′ = eiθn,n′ ... Using the relation jg = Bx, the supplied power is simply (4) Pg = 1/2 YgTr{BH B}. ... (16b) s.t. 1/2 YgTr{BH h QH QBh} ≤ Pmax g, |(Q)n,n′| = 1 ∀ (Q)n,n′ ≠ 0. ... The output power of n-th amplifier is calculated as Yg|(jg)n|2/2 for FD and DMA and as Yg N/Nt |(jg)n|2/2 for hybrid arrays."

For the hybrid architecture, B = QBh with Q having unit-modulus entries for every connected pair. If each of the Nt transmitters feeds K = N/Nt antennas and each antenna is connected to exactly one transmitter, then QH Q = K INt. Consequently, the 'supplied power' in Eq. (16b), equivalently Eq. (9), is K Tr{BhH Bh}, i.e., K times the actual transmitter-output power represented by Bh. Footnote 4 confirms this by multiplying the hybrid amplifier output power by N/Nt. A passive, lossless phase-shifting network cannot provide this K-fold power gain: for a 1-to-K power split, conservation requires per-path magnitudes of K^{-1/2}, not 1.

full rationale

The paper's electromagnetic modeling is largely self-contained in the sense that the circuit model is inherited from the authors' prior publications [6,8], which are published physical derivations not constructed to prove the energy-efficiency ordering; the Wiener-filter precoder solutions are obtained from standard results [14] and involve no fitted parameters renamed as predictions. No uniqueness theorem or load-bearing self-citation is used to force the conclusions, and the paper itself cautions that energy-efficiency conclusions are sensitive to modeling assumptions. However, the hybrid-array power constraint contains an internal normalization inconsistency: with the stated unit-modulus phase-shifting matrix Q and K antennas per transmitter, Eq. (16b) constrains K times the transmitter-output power, and footnote 4 explicitly multiplies the hybrid amplifier power by K. A passive lossless phase-shifting network cannot produce such a power gain, so the hybrid is penalized by construction relative to the DMA as K grows. Because the paper's strongest quantitative claim—DMA superiority under scalability—is the regime most affected by this factor, the main comparative result is partially forced by the model definition rather than by the underlying physics. Thus the circularity/construction score is 6, not 0 or 2.

Assumptions & free parameters 3 free parameters · 6 assumptions · 0 invented entities

The paper adds no new physical entities. Its results depend on previously published circuit models and standard hardware power parameters; the main chosen values are waveguide dimensions, supply admittance matching, and PA saturation margins.

free parameters (3)

Power amplifier saturation margin = Psat = 5Pmax_g/N (FD), 3Pmax_g/Nt (hybrid/DMA)
Chosen by simulation so each amplifier's max output is at least 3 dB below saturation; this hand-picked margin affects the nonlinear PA model and the energy efficiency comparison.
DMA generator admittance Yg = 35.33
Chosen to match the waveguide self-admittance in the DMA model; it directly influences the supplied power and thus the energy efficiency.
Waveguide dimensions (a, b, Lw) = 0.73λ, 0.17λ, Lw such that kx Lw = 3π/4
Adopted from [6] so that only the fundamental TE10 mode propagates and Ytt is normalized; these choices affect the channel matrix and power consumption.

assumptions (6)

domain assumption Antennas are modeled as magnetic dipoles on a conducting plane.
Inherited from [6] and used in Section II; if real antenna elements deviate, the coupling and radiation models change.
domain assumption The feeding network and DMA waveguides are lossless.
Stated in Section III; insertion losses in these structures would reduce DMA and hybrid efficiency, altering the comparison.
domain assumption Users' loads are conjugate-matched (Yr is a real diagonal matrix).
Assumed in Section II-A; mismatched loads would change the received power and the equivalent channel expressions.
domain assumption Perfect channel state information is available at the transmitter.
The Wiener filter derivation in Section IV relies on perfect CSI; training and feedback overhead are ignored.
domain assumption The fading covariance is Σ = ᾱ Re{Yaa}, scaled for unit average power.
Introduced in Section V; this specific scaling is a modeling choice that affects signal-to-noise ratio levels and the energy efficiency numbers.
standard math The transmit Wiener filter solution for FD arrays is taken from [14].
The KKT solution in (14)-(15) is a known result; the paper builds on it for the other topologies.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Energy efficiency of DMAs vs. conventional MIMO: a sensitivity analysis." pith.science (2026). https://pith.science/paper/TBCEMZQ4

@misc{pith2026250609181,
  author       = {Pith},
  title        = {Pith review of: Energy efficiency of DMAs vs. conventional MIMO: a sensitivity analysis},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TBCEMZQ4}},
  note         = {Machine review of arXiv:2506.09181}
}

read the original abstract

Motivated by the stringent and challenging need for `greener communications' in increasingly power-hungry 5G networks, this paper presents a detailed energy efficiency analysis for three different multi-antenna architectures, namely fully-digital arrays, hybrid arrays, and dynamic metasurface antennas (DMAs). By leveraging a circuital model, which captures mutual coupling, insertion losses, propagation through the waveguides in DMAs and other electromagnetic phenomena, we design a transmit Wiener filter solution for the three systems. We then use these results to analyze the energy efficiency, considering different consumption models and supplied power, and with particular focus on the impact of the physical phenomena. DMAs emerge as an efficient alternative to classical arrays across diverse tested scenarios, most notably under low transmission power, strong coupling, and scalability requirements.

Figures

Figures reproduced from arXiv: 2506.09181 by the authors.

**Figure 2.** Energy efficiency and average MSE achieved by the thre [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

**Figure 4.** Energy efficiency vs P max g for Nt = 10, N = 120, M = 8, spacing of λ/2 along the x axis and λ along the z axis, non-linear amplifier model, and E[|(Yra)m,n| 2 ]/σ2 n = 1. Solid/dashed lines correspond to Prf = 40 mW and Prf = 400 mW [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗

Figures from the paper (1 more)

**Figure 5.** Figure 5: Energy efficiency and MSE achieved by the three topolo [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]

Discussion (0). Sign in 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

Electromagnetics-Compliant Optimization of Dynamic Metasurface Antennas for Bistatic Sensing
eess.SP 2025-09 unverdicted novelty 6.0 of 10

Optimizes DMA-based beamforming for bistatic sensing with a physically consistent model including mutual coupling, yielding robust performance comparable to digital arrays via codebook search.

Reference graph

Works this paper leans on

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

[1]

GSMA, 5G energy efﬁciencies: green is the new black , Report, London, U.K., Nov. 2022. [Online]. Available: https://www .gsma.com/ betterfuture/climate-action/energy-efﬁciency

work page 2022
[2]

Energy efﬁciency c hallenges of 5G small cell networks,

X. Ge, J. Y ang, H. Gharavi, and Y . Sun, “Energy efﬁciency c hallenges of 5G small cell networks,” IEEE Commun. Mag. , vol. 55, no. 5, pp. 184–191, 2017

work page 2017
[3]

[Online]

GSMA, Going green: benchmarking the energy efﬁciency of mobile , Report, London, U.K., June 2021. [Online]. Available: https://data.gsmaintelligence.com/research/research/research- 2021/going-green-benchmarking-the-energy-efﬁciency- of-mobile

work page 2021
[4]

Energy efﬁciency maximization of massive MIMO communicat ions with dynamic metasurface antennas,

L. Y ou, J. Xu, G. C. Alexandropoulos, J. Wang, W. Wang, and X. Gao, “Energy efﬁciency maximization of massive MIMO communicat ions with dynamic metasurface antennas,” IEEE Trans. Wireless Commun. , vol. 22, no. 1, pp. 393–407, 2023

work page 2023
[5]

Hierarch ical codebook design with dynamic metasurface antennas for energy-efﬁci ent arrays,

J. Carlson, M. R. Castellanos, and R. W. Heath, “Hierarch ical codebook design with dynamic metasurface antennas for energy-efﬁci ent arrays,” IEEE Trans. Wireless Commun. , vol. 23, no. 10, pp. 14 790–14 804, 2024

work page 2024
[6]

Electromagnetic based communication model for dynamic me tasurface antennas,

R. J. Williams, P . Ram´ ırez-Espinosa, J. Y uan, and E. de C arvalho, “Electromagnetic based communication model for dynamic me tasurface antennas,” IEEE Trans. Wireless Commun. , vol. 21, no. 10, pp. 8616– 8630, 2022

work page 2022
[7]

Beam focusing for near-ﬁeld multiusercommunicati ons,

H. Zhang, N. Shlezinger, F. Guidi, D. Dardari, M. F. Imani , and Y . C. Eldar, “Beam focusing for near-ﬁeld multiusercommunicati ons,” IEEE Trans. Wireless Commun. , vol. 21, no. 9, pp. 7476–7490, 2022

work page 2022
[8]

Performance evaluation of dynamic metasurface antennas: Impact of insertion losses and coupling,

P . Ram´ ırez-Espinosa, R. J. Williams, J. Y uan, and E. De C arvalho, “Performance evaluation of dynamic metasurface antennas: Impact of insertion losses and coupling,” in IEEE Global Commun. Conf. (GLOBE- COM), 2022, pp. 2493–2498

work page 2022

Show all 15 references

[9]

Analysis of a waveguide-fed metasurface antenna,

D. R. Smith, O. Y urduseven, L. P . Mancera, P . Bowen, and N. B. Kundtz, “Analysis of a waveguide-fed metasurface antenna,” Phys. Rev. Applied, vol. 8, p. 054048, Nov 2017

2017
[10]

Cellular energy efﬁciency evaluation framework,

G. Auer et al. , “Cellular energy efﬁciency evaluation framework,” in IEEE V eh. Technol. Conf. (VTC Spring) , 2011, pp. 1–6

2011
[11]

Energy efﬁciency of mmWave massive MIMO precoding with low-resolu tion DACs,

L. N. Ribeiro, S. Schwarz, M. Rupp, and A. L. F. de Almeida , “Energy efﬁciency of mmWave massive MIMO precoding with low-resolu tion DACs,” IEEE J. Sel. Topics Signal Process. , vol. 12, no. 2, pp. 298– 312, 2018

2018
[12]

Hybrid MIMO architectures for millimeter wave comm unica- tions: Phase shifters or switches?

R. M´ endez-Rial, C. Rusu, N. Gonz´ alez-Prelcic, A. Alkhateeb, and R. W. Heath, “Hybrid MIMO architectures for millimeter wave comm unica- tions: Phase shifters or switches?” IEEE Access , vol. 4, pp. 247–267, 2016

2016
[13]

Ampliﬁer-a ware multiple- input multiple-output power allocation,

D. Persson, T. Eriksson, and E. G. Larsson, “Ampliﬁer-a ware multiple- input multiple-output power allocation,” IEEE Commun. Lett. , vol. 17, no. 6, pp. 1112–1115, 2013

2013
[14]

Linear transmit p rocessing in MIMO communications systems,

M. Joham, W. Utschick, and J. Nossek, “Linear transmit p rocessing in MIMO communications systems,” IEEE Trans. Signal Process. , vol. 53, no. 8, pp. 2700–2712, 2005

2005
[15]

Hjrungnes, Complex-V alued Matrix Derivatives: With Applications in Signal Processing and Communications , 1st ed

A. Hjrungnes, Complex-V alued Matrix Derivatives: With Applications in Signal Processing and Communications , 1st ed. USA: Cambridge University Press, 2011

2011

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.

[1] [1]

GSMA, 5G energy efﬁciencies: green is the new black , Report, London, U.K., Nov. 2022. [Online]. Available: https://www .gsma.com/ betterfuture/climate-action/energy-efﬁciency

work page 2022

[2] [2]

Energy efﬁciency c hallenges of 5G small cell networks,

X. Ge, J. Y ang, H. Gharavi, and Y . Sun, “Energy efﬁciency c hallenges of 5G small cell networks,” IEEE Commun. Mag. , vol. 55, no. 5, pp. 184–191, 2017

work page 2017

[3] [3]

[Online]

GSMA, Going green: benchmarking the energy efﬁciency of mobile , Report, London, U.K., June 2021. [Online]. Available: https://data.gsmaintelligence.com/research/research/research- 2021/going-green-benchmarking-the-energy-efﬁciency- of-mobile

work page 2021

[4] [4]

Energy efﬁciency maximization of massive MIMO communicat ions with dynamic metasurface antennas,

L. Y ou, J. Xu, G. C. Alexandropoulos, J. Wang, W. Wang, and X. Gao, “Energy efﬁciency maximization of massive MIMO communicat ions with dynamic metasurface antennas,” IEEE Trans. Wireless Commun. , vol. 22, no. 1, pp. 393–407, 2023

work page 2023

[5] [5]

Hierarch ical codebook design with dynamic metasurface antennas for energy-efﬁci ent arrays,

J. Carlson, M. R. Castellanos, and R. W. Heath, “Hierarch ical codebook design with dynamic metasurface antennas for energy-efﬁci ent arrays,” IEEE Trans. Wireless Commun. , vol. 23, no. 10, pp. 14 790–14 804, 2024

work page 2024

[6] [6]

Electromagnetic based communication model for dynamic me tasurface antennas,

R. J. Williams, P . Ram´ ırez-Espinosa, J. Y uan, and E. de C arvalho, “Electromagnetic based communication model for dynamic me tasurface antennas,” IEEE Trans. Wireless Commun. , vol. 21, no. 10, pp. 8616– 8630, 2022

work page 2022

[7] [7]

Beam focusing for near-ﬁeld multiusercommunicati ons,

H. Zhang, N. Shlezinger, F. Guidi, D. Dardari, M. F. Imani , and Y . C. Eldar, “Beam focusing for near-ﬁeld multiusercommunicati ons,” IEEE Trans. Wireless Commun. , vol. 21, no. 9, pp. 7476–7490, 2022

work page 2022

[8] [8]

Performance evaluation of dynamic metasurface antennas: Impact of insertion losses and coupling,

P . Ram´ ırez-Espinosa, R. J. Williams, J. Y uan, and E. De C arvalho, “Performance evaluation of dynamic metasurface antennas: Impact of insertion losses and coupling,” in IEEE Global Commun. Conf. (GLOBE- COM), 2022, pp. 2493–2498

work page 2022