DMA Reception for Simultaneous Area-Wide Sensing and Multi-User Uplink Communications

George C. Alexandropoulos; Ioannis Gavras

REVIEW 4 major objections 6 minor 24 references

DMA Reception for Simultaneous Area-Wide Sensing and Multi-User Uplink Communications

T0 review · 4 major / 6 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 A single DMA receiver can be tuned to sense a whole region while keeping uplink users connected, and it beats localization-first designs in simulation.

desk verdict Solid area-wide PEB extension of the authors' DMA work; the SDR rank-one/projection step needs verification before the QoS guarantee is credible. read the letter →

arxiv 2504.18843 v1 pith:AOM4PYST submitted 2025-04-26 eess.SP

Ioannis Gavras , George C. Alexandropoulos This is my paper · ORCID

classification eess.SP

keywords dynamicmetasurfaceantennasintegratedsensingandcommunicationCramér-Raoboundpositionerrornear-fieldlocalizationanalogbeamformingsemidefiniterelaxationmulti-useruplink

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 show that a Dynamic Metasurface Antenna used as a receiver can be set to do two jobs at once: estimate the positions of radar targets anywhere inside a defined Area of Interest, and keep a set of uplink users connected. Its route is to derive the Cramér–Rao Bound and the Position Error Bound for DMA reception, then minimize the area-wide PEB over a grid of points subject to per-user signal-to-noise-ratio constraints. Because the DMA's tunable elements are arranged in microstrips (a partially-connected architecture), the otherwise nonconvex weight-design problem is recast as convex semidefinite programs. The simulations claim all three proposed designs—direct PEB minimization, a lower-bound approximation, and a closed-form solution—outperform a DMA-based CRB-minimizing receiver that is even given the exact target positions.

What carries the argument

The load-bearing object is the partially-connected DMA reception model: $N_{\mathrm{RF}}$ microstrips, each with $N_E$ Lorentzian-constrained elements whose analog weights belong to $\mathcal{W}=\{0.5(j+e^{j\varphi}):\varphi\in[-\pi/2,\pi/2]\}$, plus the near-field channel and steering-vector model that turns element responses into a Fisher Information Matrix. The PEB, defined as the trace of the inverse FIM, is the objective; semidefinite relaxation with rank-one matrices $\mathbf{Q}_i$ and the projection $\mathbf{q}_i^{\mathrm{opt}}=\exp(j\angle[\mathbf{u}_i]_{1:N_E})$ converts the nonconvex codebook design into convex problems P1 and P2. The lower-bound trick $\operatorname{Tr}\{\mathbf{I}^{-1}\}\ge(\operatorname{Tr}\{\mathbf{I}\})^{-1}$, drawn from the harmonic-geometric mean inequality, is what produces the cheaper P2 and the closed-form solution.

What would settle it

Run the paper's setup ($N_{\mathrm{RF}}=8$ microstrips, $N_E=64$ elements, $\gamma_u=30$ dB, $|\mathcal{A}|=8$ points) and, after solving P1 or P2, compute the actual per-user SNR $\Gamma_u$ using the projected weights $\mathbf{q}_i^{\mathrm{opt}}=\exp(j\angle[\mathbf{u}_i]_{1:N_E})$ rather than the relaxed matrices. If any $\Gamma_u$ falls below $\gamma_u$ or the simulated RMSE departs from the SDR's predicted PEB, the advertised sensing/communication trade-off is not delivered.

Watch

Extended reading notes

Core claim

In the paper's own terms, the central claim is that minimizing the Position Error Bound across a discretized Area of Interest, rather than minimizing the Cramér–Rao Bound for known target coordinates, is the right objective for a dual-function DMA receiver. The paper derives the Fisher Information Matrix for the DMA-received signal, expresses the PEB as $\mathrm{PEB}(\mathbf{W}_{\mathrm{RX}};\xi)=\sqrt{\operatorname{Tr}\{\mathbf{I}_\xi^{-1}\}}$ for both active users and passive targets in the AoI, and optimizes the Lorentzian phase-constrained analog weights under per-user SNR constraints $\Gamma_u\ge\gamma_u$. The reported simulations show the localization RMSE converging to the corresponding PEB and all proposed solutions outperforming the benchmark DMA-based CRB-minimizing design, even when that benchmark knows all target positions.

Load-bearing premise

The design's promised guarantee depends on the assumption that the relaxed optimization, after projecting its solution back onto the allowed antenna phases, still gives every user the required signal strength.

Editorial extensions

If this is right

If the area-wide PEB result holds, a DMA receiver can monitor a region without any prior target locations, whereas localization-first designs appear to need that prior knowledge to stay competitive.
The convex formulations P1 and P2 mean the beamforming weights can be computed by standard convex programming solvers, making the approach a practical alternative to existing DMA designs.
The lower-bound solution P2 and the closed-form solution offer sensing accuracy comparable to direct PEB minimization at lower computational cost, so a system designer can trade complexity against accuracy.
As the required communication SNR grows, the optimized sensing accuracy degrades at a similar rate for P1 and P2, while the closed-form solution holds its PEB roughly constant, showing that the QoS constraint is what consumes sensing performance.

Reading between the lines

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

An ablation that runs the same area-wide objective without the SNR constraints would separate the gain due to area-wide sensing from the gain due to the QoS-aware formulation, since the benchmark comparison is the paper's strongest test.
Because the rank-one tightness of the semidefinite relaxation is imported from a general theory result rather than verified for this specific multi-constraint problem, a direct numerical check of each user's SNR after the Lorentzian projection is a natural prerequisite for deployment.
The same PEB-over-AoI formulation could be extended to moving targets by replacing the static grid $\mathcal{A}$ with a trajectory-weighted set, or to wideband operation by treating each subcarrier's FIM independently.

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

Short version: this is a legitimate extension of the authors' own DMA localization work. The area-wide PEB objective over an AoI grid, with per-user SNR constraints and a low-complexity closed-form alternative, is the new piece. The model and CRB derivation are standard but carefully laid out, and the simulations are plausible: P1 beats the benchmark [12] even when the benchmark knows all target positions, which makes the case that area-wide sensing is not just a gimmick.

What I trust: the partially-connected DMA structure is exploited to turn the Lorentzian phase constraint into a rank-one PSD form; the trace expressions are consistent; and the CFS from the principal singular vectors of the B_i matrices gives a sensible fallback with much lower complexity. The benchmark choice is fair—it gives the reference method an advantage and the proposed method still wins. The RMSE-to-PEB convergence in Fig. 1a is the kind of sanity check that makes the simulation credible.

The weak spot is exactly the one flagged in the stress test. Section III-B recovers q_i^opt = exp(j∠[u_i]_{1:NE}) from the SDR solution and claims via [22] that the relaxation is rank-one tight. No argument maps the general tightness conditions in [22] to this problem, which couples NRF matrices through U SNR constraints and Schur-complement LMIs. The projection is also a heuristic: even if each Q_i is rank-one, phase-only extraction need not stay in the Lorentzian codebook (1), and it can reduce the SNR term. The paper never checks, after projection, that Γ_u ≥ γ_u holds over any Monte Carlo trial. Since the abstract and conclusion promise 'guaranteed' communication performance, that promise currently rests on an unexamined assumption. The FIM is a second, softer issue: ξ contains only positions, not the passive-target reflection coefficients β_k, so the CRB is optimistic if those are truly unknown.

Bottom line: conditional. The central idea is sound and the simulation trend is believable; the missing feasibility check for the projection is removable by adding a constraint-recovery step plus a Monte Carlo histogram of post-projection SNR. I'd send it to peer review, and ask the authors to verify or repair the rank-one/feasibility claim or soften the guarantee language. A reader working on metasurface ISAC will get value; I'd likely cite the area-wide PEB formulation.

Referee Report

4 major / 6 minor

Summary. This paper studies receive beamforming design for a dynamic metasurface antenna (DMA) that simultaneously supports multi-user uplink communications and wide-area radar sensing. The authors derive a Cramér-Rao Bound (CRB) and the corresponding Position Error Bound (PEB) for estimating the positions of active users and passive targets in the near field, and then use the PEB as the design objective for the DMA's Lorentzian-constrained analog weights under per-user SNR constraints. The nonconvex design problem is relaxed via semidefinite relaxation (SDR) into problems P1 and P2, and a closed-form solution is also proposed. Numerical results over an area of interest indicate that the proposed designs outperform the CRB-minimizing DMA benchmark of [12]. The paper also includes a complexity analysis of the proposed algorithms.

Significance. The paper addresses a timely problem—ISAC with extremely large DMA apertures—and contains several useful ingredients: a closed-form CRB tailored to the partially-connected DMA structure, a convex SDR formulation that exploits the block structure of the codebook, and a low-complexity alternative. If the technical issues below are fixed, the area-wide PEB design can be a meaningful contribution to DMA-based ISAC. The stated guarantees, however, currently depend on assumptions that are not proven or tested: the tightness of the SDR and the treatment of unknown data symbols and reflection coefficients in the CRB. The simulation study is honest in comparing against a benchmark with full target knowledge, but it does not currently validate the QoS claims after the projection step.

major comments (4)

[Section III-B, P1 and P2] The SDR is not closed under the actual variable Q_i=[q_i;1][q_i;1]^H. As written, the relaxed constraints only impose Q_i⪰0 and do not include the unit-modulus constraints on q_i (equivalently the diagonal constraints on Q_i) that follow from q_{i,n}=e^{jφ_{i,n}} in the Lorentzian codebook (1). With only PSD constraints, the problems are unbounded below: scaling each Q_i by a common factor t scales the FIM I_η by t and the SNR left-hand sides by t, so PEB tends to 0 while the SNR constraints remain feasible. A normalization constraint (e.g., diag(Q_i)=1) must be included and the optimization problem restated.
[Section III-B, recovery of q_i^opt] The recovery q_i^opt = exp(j∠[u_i]_{1:N_E}) from the principal singular vector u_i of the SDR solution is asserted to be exact by citing [22], but the general rank-one guarantee of [22] is not verified for this problem, which couples N_RF blocks through U SNR constraints and 3|A| Schur-complement LMIs. The phase-only projection can leave the Lorentzian codebook W in (1) and can reduce the per-user SNR below the threshold. Section IV reports PEB and RMSE but not the post-projection SNR or codebook membership; thus the 'guaranteed' multi-user QoS conclusion is currently unsupported. Please provide either a correctness argument, numerical verification of the constraints after projection, or a qualification of the claim.
[Section III-A, Eqs. (4)–(7)] The FIM is computed from the conditional mean µ = vec{W^H P^H ∑_u h_u s_u}, which treats the uplink data symbols s_u and the reflection coefficients β_k (via h_R,u) as known. In the actual uplink ISAC scenario these are unknown nuisance parameters; ignoring them yields a PEB that is a lower bound for an estimator with additional side information. The manuscript neither includes these nuisance parameters in the FIM nor states this caveat. The sensing performance claims in Section IV should be re-evaluated or explicitly restricted to the pilot-assisted/data-aided case.
[Section III-B, P1] The objective PEB(WRX;η) is not well defined as a straightforward substitution of the 3|A| grid coordinates into the 3K-dimensional FIM of Section III-A. The received signal in (4) depends only on the K actual targets/UEs; for a grid point at which no target is present, the derivatives ∂h_u/∂[η]_i vanish, making the corresponding rows and columns of I_η zero and the 'PEB' singular. If a virtual target is postulated at every grid point, the signal model in (4) and h_R,u must be modified accordingly, and the evaluation of the PEB in the simulations must be aligned with that model. Please clarify the definition of the area-wide PEB and how it is used in the simulations.

minor comments (6)

[Introduction] There is a typo in 'Quality-of-Sevice' in the third paragraph; it should be 'Quality-of-Service'.
[Section III-B] The word 'Relexation' in the first paragraph of Section III-B should be 'Relaxation'.
[Section III-C] The sentence 'where Bi ∈ C^{(N_E+1)×(N_E+1)} has a similar structure to Hu,i, but is defined with respect to the the matrix B' contains the duplicated word 'the the'.
[Section III-B, P1] The notation 'Iη ea' in the LMI constraint is hard to read; it should be written as [I_η e_a; e_a^T b_a] ⪰ 0 with clear subscripts.
[References] Reference [23] is a paper on the physics of the near-field and does not appear to be the source of the harmonic-geometric mean inequality used in Section III-C; please replace it with a standard matrix inequality reference.
[Section IV] It would help the reader if the figures reported the feasibility rate of the projected solutions (e.g., the fraction of Monte Carlo trials in which the projected q_i^opt satisfies both the codebook and the SNR constraints), since this directly bears on the QoS claim.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity; the CRB/PEB derivation and area-wide optimization are self-contained, with self-citations limited to model components and a comparison baseline.

full rationale

The paper's central derivation constructs the Fisher Information Matrix and Position Error Bound from the stated near-field channel and received-signal models in Eqs. (2)-(8), then optimizes the DMA weights against that PEB subject to SNR constraints; no parameter is fitted to the simulation data and no prediction is equivalent to an input by construction. The claimed comparison in Section IV is against the authors' earlier CRB-minimizing design [12], but [12] is used as an external baseline rather than to justify the proposed formulation; the AoI-wide PEB objective and the P1/P2 relaxations stand on the equations given in the paper. The SDR rank-one recovery cites the external general result [22] and is a correctness assumption rather than a self-consistent circular step: even if the projection fails to preserve the Lorentzian codebook or SNR constraints, that would be a validity gap, not a reduction of the output to the input. Self-citations to [10] for channel coefficients and the MUSIC estimator supply standard components and do not carry the central claim. Overall, the derivation chain is self-contained, so the circularity score is minimal.

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

The paper's central optimization rests on standard CRB and SDR machinery plus several unproven or inherited assumptions: rank-one recovery of SDR, known data symbols and reflection coefficients in the FIM, and the near-field DMA channel model from the authors' prior work. No parameters are fitted to data; all simulation settings are stated.

assumptions (6)

standard math The FIM/CRB formalism for complex Gaussian observations with known mean applies to the DMA received signal.
Used in Section III-A to compute PEB; standard estimation theory from [18].
ad hoc to paper SDR of the nonconvex QCQP yields rank-one optimal solutions satisfying the QoS constraints.
Asserted in Section III-B via [22]; not proven for this multi-constraint problem.
ad hoc to paper The uplink data symbols s_u and reflection coefficients β_k are known or conditioned when computing the FIM.
Enter Eqs. (6)-(7); the paper does not model them as nuisance parameters or random.
domain assumption The near-field channel model and steering vectors from [10] accurately describe DMA propagation and target reflections.
Used in Section II-A; inherited from the authors' prior work.
domain assumption The Lorentzian-constrained phase response codebook W with φ in [-π/2,π/2] is achievable for each metamaterial.
Used in Eq. (1) and enforced in P1/P2.
domain assumption Accurate channel estimates for all UEs are available per coherent block via SRS.
Stated in Section II-B to enable SNR computation and LoS channel composition.

how reviews work

0 comments

Cite this review

Pith. "Pith review of DMA Reception for Simultaneous Area-Wide Sensing and Multi-User Uplink Communications." pith.science (2026). https://pith.science/paper/AOM4PYST

@misc{pith2026250418843,
  author       = {Pith},
  title        = {Pith review of: DMA Reception for Simultaneous Area-Wide Sensing and Multi-User Uplink Communications},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/AOM4PYST}},
  note         = {Machine review of arXiv:2504.18843}
}

read the original abstract

The recent surge in deploying extremely large antenna arrays is expected to play a vital role in future sixth generation wireless networks, enabling advanced radar target localization with enhanced angular and range resolution. This paper focuses on the promising technology of Dynamic Metasurface Antennas (DMAs), integrating numerous sub-wavelength-spaced metamaterials within a single aperture, and presents a novel framework for designing its analog reception beamforming weights with the goal to optimize sensing performance within a spatial Area of Interest (AoI), while simultaneously guaranteeing desired multi-user uplink communication performance. We derive the Cramer-Rao Bound (CRB) with DMA-based reception for both passive and active radar targets lying inside the AoI, which is then used as the optimization objective for configuring the discrete tunable phases of the metamaterials. Capitalizing on the DMA partially-connected architecture, we formulate the design problem as convex optimization and present both direct CRB minimization approaches and low complexity alternatives using a lower-bound approximation. Simulation results across various scenarios validate the effectiveness of the proposed framework, showing it consistently outperforms existing state-of-the-art methods.

Figures

Figures reproduced from arXiv: 2504.18843 by the authors.

**Figure 1.** Sensing and communications performances with the proposed dual-functional RX DMA framework, considering a panel with NRF = 8 microstrips, each hosting NE = 64 phase-tunable metamaterials. we can formulate the following optimization problem for the intended dual-functional RX DMA design as follows: P2 : max {Qi} NRF i=1 N XRF i=1 Tr{BiQi} s.t. N XRF i=1 Tr{Hu,iQi} ≥ γu ∀u, Qi ⪰ 0 ∀i, where Bi ∈ C (NE+1)×(NE+1) has a … view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

24 extracted references · 12 canonical work pages

[12]

Near-field localization with dynamic metasurface antennas at THz: A CRB minimizing approach,

——, “Near-field localization with dynamic metasurface antennas at THz: A CRB minimizing approach,”IEEE Wireless Commun. Lett., early access, 2025

work page 2025
[22]

Semidefinite relaxation of quadratic optimization problems,

Z.-Q. Luo et al. , “Semidefinite relaxation of quadratic optimization problems,” IEEE Signal Process. Mag. , vol. 27, no. 3, pp. 20–34, 2010

work page 2010
[1]

A tutorial on extremely large-scale MIMO for 6G: Fundamentals, signal processing, and applications,

Z. Wang et al. , “A tutorial on extremely large-scale MIMO for 6G: Fundamentals, signal processing, and applications,” IEEE Commun. Surveys & Tuts. , early access, 2024

2024
[2]

Holographic MIMO communications: Theoretical foundations, enabling technologies, and future directions,

T. Gong et al. , “Holographic MIMO communications: Theoretical foundations, enabling technologies, and future directions,” IEEE Commun. Surveys & Tuts. , vol. 26, no. 1, pp. 196–257, 2024

work page 2024
[3]

Near-field integrated sensing and communication with extremely large-scale antenna array,

H. Hua et al. , “Near-field integrated sensing and communication with extremely large-scale antenna array,” arXiv preprint arXiv:2407.17237 , 2024

arXiv 2024
[4]

Towards distributed and intelligent integrated sensing and communications for 6G networks,

E. Calvanese Strinati et al. , “Towards distributed and intelligent integrated sensing and communications for 6G networks,” IEEE Wireless Commun., vol. 32, no. 1, pp. 60–67, 2025

2025
[5]

Dynamic metasurface antennas for 6G extreme massive MIMO communications,

N. Shlezinger et al. , “Dynamic metasurface antennas for 6G extreme massive MIMO communications,” IEEE Wireless Commun. , vol. 28, no. 2, pp. 106–113, 2021

2021
[6]

Near-field wideband extremely large-scale MIMO transmissions with holographic metasurface-based antenna arrays,

J. Xu et al. , “Near-field wideband extremely large-scale MIMO transmissions with holographic metasurface-based antenna arrays,” IEEE Transactions on Wireless Communications , vol. 23, no. 9, pp. 12 054– 12 067, 2024

work page 2024

Show all 24 references

[7]

Beam focusing for near-field multiuser MIMO communications,

H. Zhang et al. , “Beam focusing for near-field multiuser MIMO communications,” IEEE Trans. Wireless Commun. , vol. 21, no. 9, pp. 7476–7490, 2022

2022
[8]

Autoregressive attention neural networks for non-line-of-sight user tracking with dynamic metasurface antennas,

K. Stylianopoulos et al. , “Autoregressive attention neural networks for non-line-of-sight user tracking with dynamic metasurface antennas,” in Proc. IEEE CAMSAP , Los Sueños, Costa Rica, 2023

2023
[9]

Near-field beam tracking with extremely massive dynamic metasurface antennas,

P. Gavriilidis and G. C. Alexandropoulos, “Near-field beam tracking with extremely massive dynamic metasurface antennas,” IEEE Trans. Wireless Commun., early access, 2025

2025
[10]

Full duplex holographic MIMO for near-field integrated sensing and communications,

I. Gavras et al., “Full duplex holographic MIMO for near-field integrated sensing and communications,” in Proc. EUSIPCO , Helsinki, Finland, 2023

2023
[11]

Simultaneous near-field THz communications and sensing with full duplex metasurface transceivers,

I. Gavras and G. C. Alexandropoulos, “Simultaneous near-field THz communications and sensing with full duplex metasurface transceivers,” in Proc. IEEE SPA WC, Lucca, Italy, 2024

2024
[13]

Extremely large full duplex MIMO for simultaneous downlink communications and monostatic sensing at sub-THz frequencies,

G. C. Alexandropoulos and I. Gavras, “Extremely large full duplex MIMO for simultaneous downlink communications and monostatic sensing at sub-THz frequencies,” arXiv preprint arXiv:2502.10693 , 2025

2025 arXiv
[14]

How do microstrip losses impact near-field beam depth in dynamic metasurface antennas?

P. Gavriilidis and G. C. Alexandropoulos, “How do microstrip losses impact near-field beam depth in dynamic metasurface antennas?” arXiv preprint arXiv:2503.12280, 2025

2025 arXiv
[15]

A survey of new mid-band/FR3 for 6G: Channel measurement, characterization and modeling in outdoor environment,

H. Miao et al. , “A survey of new mid-band/FR3 for 6G: Channel measurement, characterization and modeling in outdoor environment,” arXiv preprint arXiv:2504.06727 , 2025

2025 arXiv
[16]

Near-field communications: A tutorial review,

Y . Liu et al., “Near-field communications: A tutorial review,”IEEE Open J. Commun. Soc. , vol. 4, pp. 1999–2049, 2023

1999
[17]

A tutorial on 5G positioning,

L. Italiano et al., “A tutorial on 5G positioning,” IEEE Commun. Surveys & Tuts., early access, 2024

2024
[18]

S. M. Kay, Fundamentals of Statistical Signal Processing: Estimation Theory. Prentice-Hall, Inc., 1993

1993
[19]

Optimal spatial signal design for mmWave positioning under imperfect synchronization,

M. F. Keskin et al. , “Optimal spatial signal design for mmWave positioning under imperfect synchronization,” IEEE Trans. V eh. Technol., vol. 71, no. 5, pp. 5558–5563, 2022

2022
[20]

Metasurface-based receivers with 1-bit ADCs for multi-user uplink communications,

P. Gavriilidis et al. , “Metasurface-based receivers with 1-bit ADCs for multi-user uplink communications,” in Proc. IEEE ICASSP, Seoul, South Korea, 2024

2024
[21]

CVX: Matlab software for disciplined convex programming, version 2.0,

I. CVX Research, “CVX: Matlab software for disciplined convex programming, version 2.0,” http://cvxr.com/cvx, 2012

2012
[23]

The physics of the near-field,

C. Girard, C. Joachim, and S. Gauthier, “The physics of the near-field,” Reports on Progress in Physics , vol. 63, no. 6, p. 893, 2000

2000
[24]

Interior point polynomial time methods in convex programming,

A. Nemirovski, “Interior point polynomial time methods in convex programming,” Lecture notes, vol. 42, no. 16, pp. 3215–3224, 2004

2004

Pith tools

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

[1] [12]

Near-field localization with dynamic metasurface antennas at THz: A CRB minimizing approach,

——, “Near-field localization with dynamic metasurface antennas at THz: A CRB minimizing approach,”IEEE Wireless Commun. Lett., early access, 2025

work page 2025

[2] [22]

Semidefinite relaxation of quadratic optimization problems,

Z.-Q. Luo et al. , “Semidefinite relaxation of quadratic optimization problems,” IEEE Signal Process. Mag. , vol. 27, no. 3, pp. 20–34, 2010

work page 2010

[3] [1]

A tutorial on extremely large-scale MIMO for 6G: Fundamentals, signal processing, and applications,

Z. Wang et al. , “A tutorial on extremely large-scale MIMO for 6G: Fundamentals, signal processing, and applications,” IEEE Commun. Surveys & Tuts. , early access, 2024

2024

[4] [2]

Holographic MIMO communications: Theoretical foundations, enabling technologies, and future directions,

T. Gong et al. , “Holographic MIMO communications: Theoretical foundations, enabling technologies, and future directions,” IEEE Commun. Surveys & Tuts. , vol. 26, no. 1, pp. 196–257, 2024

work page 2024

[5] [3]

Near-field integrated sensing and communication with extremely large-scale antenna array,

H. Hua et al. , “Near-field integrated sensing and communication with extremely large-scale antenna array,” arXiv preprint arXiv:2407.17237 , 2024

arXiv 2024

[6] [4]

Towards distributed and intelligent integrated sensing and communications for 6G networks,

E. Calvanese Strinati et al. , “Towards distributed and intelligent integrated sensing and communications for 6G networks,” IEEE Wireless Commun., vol. 32, no. 1, pp. 60–67, 2025

2025

[7] [5]

Dynamic metasurface antennas for 6G extreme massive MIMO communications,

N. Shlezinger et al. , “Dynamic metasurface antennas for 6G extreme massive MIMO communications,” IEEE Wireless Commun. , vol. 28, no. 2, pp. 106–113, 2021

2021

[8] [6]

Near-field wideband extremely large-scale MIMO transmissions with holographic metasurface-based antenna arrays,

J. Xu et al. , “Near-field wideband extremely large-scale MIMO transmissions with holographic metasurface-based antenna arrays,” IEEE Transactions on Wireless Communications , vol. 23, no. 9, pp. 12 054– 12 067, 2024

work page 2024