Pith. sign in

REVIEW 3 major objections 6 minor 2 cited by

A Novel Cost-Effective MIMO Architecture with Ray Antenna Array for Enhanced Wireless Communication Performance

T0 review · 3 major / 6 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 A ray antenna array made of directly connected uniform linear subarrays can steer beams with switches instead of phase shifters, and the paper argues this gives finer uniform angular resolution, higher gain, and lower cost than a…

desk verdict Promising phase-shifter-free array architecture with sound math, but the headline gains rest on an unequal antenna budget and hand-picked element patterns. read the letter →

arxiv 2505.23394 v1 pith:WH72FAOA submitted 2025-05-29 cs.AR

Zhenjun Dong , Zhiwen Zhou , Yong Zeng This is my paper · ORCID
classification cs.AR
keywords rayantennaarrayphase-shifter-freebeamforminghybridmmWaveMIMOangularresolutionselectionnetworkcost-effectiveterahertzcommunications
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

This paper proposes a multi-antenna architecture called the ray antenna array (RAA) that does beamforming without any phase shifters. The array is built from many small uniform linear arrays (sULAs), each with its antennas wired directly together so that it naturally forms a beam along its physical orientation; a switch network then connects the best rays to the radio-frequency chains. The paper's central claim is that, compared with a conventional uniform linear array using hybrid analog/digital beamforming, the RAA has three advantages at the same array gain factor: uniform and finer angular resolution across the coverage range, higher beamforming gain because each ray can use more directional antenna elements, and drastically lower hardware cost. The argument matters because phase shifters are expensive and hard to build at millimeter-wave and terahertz frequencies, so a phase-shifter-free architecture could make large antenna arrays practical in 6G base stations. Simulations with a 47.2 GHz channel model show the RAA beating the ULA baseline in single-user and multi-user uplink and downlink scenarios.

What carries the argument

The load-bearing object is the sULA: M antennas spaced half a wavelength apart and wired directly to one port, so their signals add coherently only for a wave arriving along the line of the array. Its array response is $f(\phi,\eta_n)=M H_M(\sin(\phi-\eta_n))\,b(\phi-\eta_n)$, with $H_M$ the Dirichlet kernel; this identity carries the argument because it shows the beam direction is set by geometry, not by tunable phases. The orientation rule $\eta_n=n\arcsin(2/M)$ and the count $N=2\lfloor \eta_{\max}/\arcsin(2/M)\rfloor+1$ tile the coverage range with equal beamwidths, and the radiation-pattern condition in Theorem 1, $\phi_{3\mathrm{dB}}\ge\arcsin(2/M)$ with $G(0.5\phi_{3\mathrm{dB}})\ge(\varepsilon/|H_M(\sin(0.5\arcsin(2/M)))|)^2$, is what lets the array use directional elements. The ray selection network, a binary matrix $S\in\{0,1\}^{N_{\mathrm{RF}}\times N}$ that chooses which sULAs reach the RF chains, converts the geometric beams into flexible MIMO transmission.

What would settle it

A concrete test: simulate or build the 38 GHz example with a ULA using the same total element budget of N×M=25,728 antennas (or the same physical aperture) under the same channel, and compare achievable rates and hardware cost; if that equally financed ULA matches or beats the RAA, the claimed advantage over fairly compared alternatives fails. Alternatively, measure the realized antenna element pattern of a fabricated RAA ray at $\phi_{3\mathrm{dB}}=0.3\pi$ and check whether the assumed directivity, and hence the 5 dB SNR gain, is physically obtained.

Watch

Extended reading notes

Core claim

The paper's central claim is that N simple uniform linear arrays, each with M directly connected antennas and oriented to cover a different slice of the angular range, can replace an M-element ULA with hybrid beamforming while improving performance. For a ray with orientation $\eta_n$, the sULA output for a path at angle $\phi$ is $f(\phi,\eta_n)=M H_M(\sin(\phi-\eta_n))\,b(\phi-\eta_n)$, where $H_M$ is the Dirichlet kernel; the beam peaks exactly when $\phi=\eta_n$, with no phase shifters. The design sets adjacent ray orientations $|\eta_n-\eta_{n-1}|=\arcsin(2/M)$ so the null of one ray aligns with the peak of its neighbor, giving every ray the same null-to-null beamwidth $\phi_{\mathrm{BW}}=2\arcsin(2/M)$. Because each ray covers only a fraction of the total angle range, the antenna elements can have 3dB beamwidth $\phi_{3\mathrm{dB}} \ge \arcsin(2/M)$ instead of the ULA's $\phi^{\mathrm{ULA}}_{3\mathrm{dB}}\ge 2\phi_{\max}$, and this narrower element pattern raises the peak gain through the conservation relation $G(0)\approx G_{\mathrm{sum}}/(1.066\,\phi_{3\mathrm{dB}})$. The paper proves the resulting beamwidth is never wider than the ULA's DFT-codebook beamwidth and is strictly narrower away from boresight, and it gives a cost comparison showing RAA at roughly 17% of the ULA-HBF hardware cost in its example.

Load-bearing premise

The comparison gives the RAA N×M antenna elements while the ULA gets only M, and it assumes the extra antennas cost about a cent each and take negligible space, so the cost and resolution advantages rest on antennas being far cheaper and physically smaller than phase shifters.

Editorial extensions

If this is right

  • A 128-element-class base station with $N_{\mathrm{RF}}=16$ chains can drop from 2048 phase shifters to a switch network, cutting the quoted hardware cost from about 268,700 USD to about 46,300 USD in the paper's 38 GHz example.
  • Angular resolution no longer degrades for users at the edge of the cell: every ray has the same beamwidth, whereas a ULA's DFT beam widens away from boresight.
  • Because each sULA covers only a small angular sector, antenna elements can be made more directional, which the simulations turn into roughly 5 dB higher SNR for single-user uplink.
  • The greedy ray selection for uplink sum rate and the alternating optimization for downlink max-min SINR reach near-optimal performance in the paper's tests, so the architecture's gains do not require exhaustive search.
  • Isotropic-element RAA still matches or edges out ULA at high SNR in downlink, showing the resolution advantage alone carries value even without directional elements.

Reading between the lines

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

  • Editorial inference: the comparison is normalized to the same array gain factor M, not the same total element count; if the ULA were allowed N×M elements or the same physical aperture, the resolution and gain advantages would shrink, so the practical merit depends on the cost and size of the extra antennas being as negligible as assumed.
  • Editorial inference: the architecture's beam directions are discrete, set by the ray orientations, so angular alignment error relative to a user's true direction is bounded by half the ray spacing; a testable extension is to quantify how this quantization loss behaves when N is small or when M is modest.
  • Editorial inference: the same phase-shifter-free idea could apply to sensing and localization at mmWave and terahertz bands, since uniform angular resolution is valuable for angle-of-arrival estimation; the paper only mentions sensing as future 3D-RAA work.
  • Editorial inference: the directivity gain at $\phi_{3\mathrm{dB}}=0.3\pi$ is assumed achievable at no extra fabrication cost; a hardware prototype measuring the realized element pattern would settle whether the SNR gains survive in practice.
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 / 6 minor

Summary. The paper proposes a new multi-antenna architecture, the ray antenna array (RAA), in which N directly-connected uniform linear subarrays (sULAs) are arranged with carefully chosen physical orientations. Steering is performed by selecting sULAs through a switch network rather than by phase shifters, which the authors argue is much cheaper at mmWave/THz frequencies. The paper derives the sULA beam pattern, the orientation spacing and array-count formulas, and sufficient conditions on the element pattern (Theorems 1–3). It then formulates uplink and downlink joint beamforming and ray-selection problems, proposes a greedy uplink algorithm and an alternating downlink algorithm, and reports simulations showing a roughly 5 dB single-user SNR gain over a conventional ULA with hybrid beamforming, as well as large hardware-cost reductions.

Significance. The architectural idea is genuinely interesting: replacing phase shifters with physically oriented subarrays and switches could be a useful low-cost option for high-frequency arrays, and the core sULA beam-pattern mathematics (Dirichlet kernel analysis, beamwidth, orientation spacing) is correct and clearly presented. The greedy uplink algorithm is shown to match exhaustive search in a small example, which is a solid algorithmic contribution. However, the central comparative claims are currently weakened by a resource-budget asymmetry (RAA uses N×M elements versus M for the ULA), by an element-pattern assumption whose physical cost and realizability are not analyzed, and by a downlink algorithm that relies on exhaustive search. If these issues are rigorously addressed, the paper could be a valuable contribution to the mmWave/THz architecture literature.

major comments (3)
  1. [Section III-C, Eqs. (12), (28), and Section VI] The performance comparison normalizes by the array gain factor M rather than by total element count or aperture. In the large example (M=128, N=201), the RAA deploys 25,728 antenna elements in a 2D fan, while the ULA baseline has 128 elements on a line. If the ULA were allowed the same total element budget (or the same aperture), its beamwidth would be far narrower than 2 arcsin(2/M) and its array gain correspondingly larger, so the claimed advantages in angular resolution and beamforming gain would not hold in an equal-resource comparison. The paper should add such a comparison, or explicitly justify why equal-array-gain-factor normalization is the appropriate fairness criterion for the claims in the abstract.
  2. [Section III-C, Eqs. (19), (21), (31), and Section VI, Fig. 8] The roughly 5 dB advantage in Fig. 8 is largely an input choice rather than a derived architectural property: the simulations set the RAA element 3 dB beamwidth to phi_3dB=0.3pi (peak gain 5.13 dB) and the ULA element beamwidth to pi (0 dB), and Eq. (21) then converts that beamwidth difference into the gain difference under equal G_sum. The paper does not establish that high-directivity RAA elements with phi_3dB around 0.3pi can be realized at the assumed $0.01 unit cost and without increasing element size, mutual coupling, or footprint, nor does it quantify the larger size admitted in Section VII. A feasibility analysis or a sensitivity study over phi_3dB and element cost should be included.
  3. [Section V, Algorithm 2, Problem (47)] The downlink alternating optimization algorithm uses exhaustive search over all N choose N_RF sULA subsets in every iteration of Problem (47). Its complexity is combinatorial in N_RF and N, so it is not efficient for the large configuration (N=201, N_RF=8) used elsewhere in Section VI. The downlink simulations consequently use only M=6 and N_RF=3, where the search is feasible. The abstract's claim of efficient algorithms for multi-user RAA communications is therefore not supported for the downlink case; a polynomial-complexity ray-selection method (e.g., a greedy scheme analogous to Algorithm 1) is needed, or the complexity and scalability limitations should be clearly stated.
minor comments (6)
  1. [Appendix A] The statement that the function sqrt(G(phi-eta_n*)) |H_M(sin(phi-eta_n*))| is 'a periodic function' is inaccurate; the subsequent argument works because every angular segment has the same functional form up to a shift. Please rephrase.
  2. [Section III-A, Fig. 4] The caption states that G(zeta)=1 for the plotted beam patterns, but this is an idealized element pattern; the later RAA design uses the 3GPP pattern in Eq. (20). Clarify in the caption that Fig. 4 isolates the array-factor behavior.
  3. [Section III-B, Eqs. (13), (14)] The floor operation in Eq. (13) gives a slightly conservative number of sULAs; stating the resulting coverage radius explicitly (for example, as (floor(eta_max/s)+1)*s) would make the coverage argument easier to follow.
  4. [Section VI] The simulations use only 50 channel realizations; adding error bars or confidence intervals, especially in the multi-user sum-rate and max-min SINR plots, would help assess the statistical reliability of the observed gains.
  5. [Section III-C, cost comparison] The cost model counts N_RF*N switches for the RSN and M*N_RF phase shifters for the ULA, but the physical topology of the RSN (e.g., whether each RF chain needs a full N-way selector or a smaller network) is not described in detail; a sentence describing the switching topology and its control overhead would clarify the cost comparison.
  6. [Section IV, first paragraph] The phrase 'Noted that' should be 'Note that', and the statement about 'prohibitive complexity' of exhaustive search in Section V could be quantified by giving the exact order of the search space.

Circularity Check

1 steps flagged · score 4.0 of 10

The RAA's angular-resolution and element-directivity arguments derive independently from array geometry (Corollaries 1-2, Theorems 1-3), but the ~5 dB SNR advantage in Fig. 8 reduces by construction through Eq. (21) to the beamwidth inputs chosen in Section VI, so the numerical 'superior performance' validation is partly input-as-output.

  1. fitted input called prediction [Section VI (Fig. 8), with Section IV (single-path SNR relation) and Section III-C Theorem 2 / Eq. (21)]
    "For the antenna element pattern, we set ϕ3dB = 0.3π and G(0) = 5.1335dB for the proposed RAA, and ϕULA 3dB = π and GULA = 0dB for the conventional ULA, ensuring that Gsum = GULA sum. … It can be observed from Fig. 8 that, when the directional antenna elements are employed, RAA achieves an SNR approximately 5 dB higher than that of the ULA across different transmit SNR levels. … In contrast, when isotropic antennas are used, ULA slightly outperforms RAA, as RAA’s finer and uniform angular resolution makes it more challenging to align with the signal directions."

    By construction, the gain reduces to the inputs: the single-path maximum SNR derived in Section IV equals M·P̄t·|α|²·G(0), i.e., it scales linearly with the chosen element peak gain, while Eq. (21) fixes G(0) ≈ Gsum/(1.066·ϕ3dB). With Gsum set equal to GULA sum, the SNR ratio is forced to ϕULA3dB/ϕ3dB = π/(0.3π) = 3.33, i.e., 5.2 dB. The 'approximately 5 dB' advantage in Fig. 8 is therefore the chosen beamwidth ratio propagated through a linear formula and the equal-Gsum normalization, not a quantity produced by the channel, ray-selection, or baseband simulation. The isotropic-element control ('ULA slightly outperforms RAA') confirms the entire simulated advantage is the input element-pattern difference.

full rationale

Core derivation is non-circular. The sULA output model (8), the identical mainlobe beamwidth 2 arcsin(2/M) (Corollary 1), the orientation design and sULA count (13)-(14), the element-pattern sufficiency conditions for RAA (Theorem 1, ϕ3dB ≥ arcsin(2/M)) and for ULA (Theorem 3, ϕULA3dB ≥ 2ϕmax), and the peak-gain identity G(0) ≈ Gsum/(1.066ϕ3dB) (Eq. 21) form a self-contained mathematical chain from stated geometric and physical assumptions. Nothing is fitted to data, and no load-bearing result is imported from the authors' prior work; citations [1], [20], [26], and [27] are contextual and do not carry the derivation. The resolution comparison (Corollaries 1 and 2) is a sound derived inequality under the paper's disclosed 'same array gain factor M' normalization, so the finer-and-uniform-resolution claim has independent content. One partial input-as-output step is flagged: in Section VI the element patterns are chosen with ϕ3dB = 0.3π (G(0) = 5.13 dB) versus π (0 dB) at equal Gsum, and Fig. 8 then reports the ~5 dB SNR advantage; because the single-path SNR is proportional to G(0) and G(0) is fixed by Eq. (21), the gap equals 10 log10(π/0.3π) ≈ 5.2 dB. The isotropic-element control, where ULA slightly outperforms RAA, confirms that the entire simulated advantage is the assumed element-directivity difference; only the entitlement to use the narrower element is independently derived (Theorems 1 and 3). The hardware-cost 'advantage' is likewise an accounting exercise resting on the stated unit prices (pps ≈ 131.2, psw ≈ 14.31, pant ≈ 0.01) and on the unquantified assumption, conceded only as 'requires a larger size' in Section VII, that the extra N×M = 25,728 elements and 2D footprint are nearly free; this is a resource-budget fairness concern rather than definitional circularity. Overall score 4: partial self-confirmation in the gain validation, but the central architectural derivation is self-contained and non-circular.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The RAA, sULA, and RSN are architectural arrangements of standard antennas, switches, and RF chains; the paper introduces no new physical particles, forces, or conserved quantities. The free parameters and axioms above capture the hand-chosen design values and physical idealizations on which the central comparison rests.

free parameters (4)
  • RAA element 3 dB beamwidth phi_3dB = 0.3 pi (Section VI)
    Hand-chosen in simulations. The sufficient condition (19) only requires phi_3dB >= arcsin(2/M), about 0.005 pi for M=128, so the simulated gain advantage over ULA is directly controlled by this choice.
  • ULA element 3 dB beamwidth phi_3dB^ULA = pi (Section VI)
    Set to a wide-coverage element, giving ULA peak gain 0 dB. Together with equal total gain, this defines the baseline for the performance comparison in Figs. 8-11.
  • Total antenna power gain G_sum = unspecified; set equal for RAA and ULA
    The comparison G(0)/G_ULA(0) = phi_3dB^ULA / phi_3dB follows from this equality, so the peak-gain ratio is fixed by the normalization assumption.
  • Coverage threshold epsilon_0 = not specified
    The design criterion (18) depends on an epsilon_0 threshold, but the simulations never state it or verify that the chosen element patterns satisfy condition (19).
assumptions (5)
  • domain assumption Far-field, narrowband plane-wave propagation with half-wavelength spacing; each sULA output is the coherent sum of its M element responses.
    Used throughout the channel and array model in Section III-A, equations (5)-(8).
  • domain assumption Antenna element patterns are independent, have no mutual coupling, and rotate rigidly with the sULA orientation; direct connection is ideal.
    The pattern G(phi-eta_n) enters the model in (7), and the directivity design assumes no coupling, matching loss, or power-divider loss.
  • ad hoc to paper High-directivity RAA antenna elements can be realized at the same negligible per-element cost (about $0.01) without increasing size or power consumption.
    The enhanced-beamforming-gain and cost advantages both depend on this. Section III-C uses p_ant about $0.01, and Section VI chooses phi_3dB=0.3 pi without a realizability argument.
  • domain assumption Commercial prices for phase shifters, switches, and antennas quoted from Qorvo in 2025 are representative.
    The hardware cost comparison in Section III-C relies entirely on these prices from reference [30].
  • domain assumption The 3GPP TR 38.901 UMa NLoS channel model is an adequate testbed for the claims.
    Section VI uses the channel parameters in Table I from [28] for all simulations.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Novel Cost-Effective MIMO Architecture with Ray Antenna Array for Enhanced Wireless Communication Performance." pith.science (2026). https://pith.science/paper/WH72FAOA

@misc{pith2026250523394,
  author       = {Pith},
  title        = {Pith review of: A Novel Cost-Effective MIMO Architecture with Ray Antenna Array for Enhanced Wireless Communication Performance},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WH72FAOA}},
  note         = {Machine review of arXiv:2505.23394}
}
read the original abstract

This paper proposes a novel multi-antenna architecture, termed ray antenna array (RAA), which practically enables flexible beamforming and also enhances wireless communication performance for high frequency systems in a cost-effective manner. RAA consists of a large number of inexpensive antenna elements and a few radio frequency (RF) chains. These antenna elements are arranged in a novel ray like structure, where each ray corresponds to one simple uniform linear array (sULA) with a carefully designed orientation. The antenna elements within each sULA are directly connected, so that each sULA is able to form a beam towards a direction matching the ray orientation without relying on any analog or digital beamforming. By further designing a ray selection network (RSN), appropriate sULAs are selected to connect to the RF chains for subsequent baseband processing. Compared to conventional multi-antenna architectures such as the uniform linear array (ULA) with hybrid analog/digital beamforming (HBF), the proposed RAA enjoys three appealing advantages: (i) finer and uniform angular resolution for all signal directions; (ii) enhanced beamforming gain by using antenna elements with higher directivity, as each sULA is only responsible for a small portion of the total angle coverage range; and (iii) dramatically reduced hardware cost since no phase shifters are required, which are expensive and difficult to design in high-frequency systems such as mmWave and THz systems. To validate such advantages, we first present the input-output mathematical model for RAA-based wireless communications. Efficient algorithms for joint RAA beamforming and ray selection are then proposed for single-user and multi-user RAA-based wireless communications. Simulation results demonstrate that RAA achieves superior performance compared to the conventional ULA with HBF, while significantly reducing hardware cost.

Figures

Figures reproduced from arXiv: 2505.23394 by the authors.

Figure 1
Figure 1. MIMO architectures with NRF chains. (a) The conventional fully-connected HBF architecture based on phase shifters. (b) The proposed RAA architecture without any phase shifters. Specifically, simulation results verify that the proposed RAA achieves finer and uniform angular resolution com￾pared to the conventional ULA with the same array gain factor. Moreover, the RAA system yields notable improvements in communicati… view at source ↗
Figure 2
Figure 2. Uplink multi-user communications based on the pro [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. The proposed RAA architecture is composed of [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (9 more)
Figure 4
Figure 4. Figure 4: Beam patterns [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Orientation design principle for adjacent sULAs in [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]
Figure 6
Figure 6. Figure 6: The overall procedure of RAA parameter design. [PITH_FULL_IMAGE:figures/full_fig_p006_6.png]
Figure 7
Figure 7. Figure 7: Beam patterns |f(ϕ, ηn)| in (8) and |fULA(ϕ, ϕn)| in (25), where M = 8 and ϕmax = 0.499π. The beam patterns of the proposed RAA |f(ϕ, ηn)| in (8) and the conventional ULA |fULA(ϕ, ϕn)| in (25) are illustrated in [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]
Figure 8
Figure 8. Figure 8: Single-user uplink maximum SNR (dB) in (34) versus [PITH_FULL_IMAGE:figures/full_fig_p010_8.png]
Figure 10
Figure 10. Figure 10: Illustration of the convergence behavior of the pro [PITH_FULL_IMAGE:figures/full_fig_p011_10.png]
Figure 9
Figure 9. Figure 9: Multi-user uplink maximum sum rate Rsum (bps/Hz) in (35) versus P¯ t for the RAA and ULA, where (a) M = 6, NRF = 3 and K = 3, (b) M = 128, NRF = 8 and K = 8. and ULA architectures. These results validate the convergence and effectiveness of the proposed alternating opt…
Figure 12
Figure 12. Figure 12: The angle ηn∗ = n ∗ × arcsin(2/M) with n ∗ being the optimal solution to (17) vs any given signal direction ϕ. Under the two assumptions (i) and (ii) stated in Theorem 1, substituting (7) and (8) into (17) leads to a reformulation that holds under the condition ϕ3dB ≥…
Figure 13
Figure 13. Figure 13: DFT codeword sin ϕn∗ with n ∗ being the optimal solution to (29) vs the signal direction ϕ. Thus, (59) can be equivalently rewritten as N′ segmented functions, as follows min ∀ϕ∈Ωn∗ ,∀n∗∈N′ p GULA(ϕ)|HM(sin ϕ − sin ϕn∗ )| ≥ ε. (60) Under the given assumption (iii) in …

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

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

  1. Complementary RL: Towards Efficient Experience-Driven Agent Learning

    cs.LG 2026-03 unverdicted novelty 5.0 of 10

    Complementary RL co-trains a policy actor on sparse outcome rewards and an experience extractor on whether its distilled experiences help the actor succeed, yielding about 10% gains over experience-free agentic RL.

  2. Flexible XL-MIMO via Array Configuration Codebook: Codebook Design and Array Configuration Training

    cs.IT 2025-08 conditional novelty 5.0 of 10

    A codebook of pre-designed array configurations (compact, sparse, modular, nested, co-prime) enables low-overhead antenna activation training for XL-MIMO communication and localization, with two-stage scanning approac...

Reference graph

Works this paper leans on

28 extracted references · 23 canonical work pages · cited by 2 Pith papers

  1. [1]

    Ray antenna array: A novel cost- effective multi-antenna architecture for enhanced wireless communica- tion,

    Z. Dong, Z. Zhou, and Y . Zeng, “Ray antenna array: A novel cost- effective multi-antenna architecture for enhanced wireless communica- tion,” in2025 IEEE 101st V ehicular Technology Conference-VTC2025- Spring, accepted

    work page

  2. [2]

    R. W. Heath Jr and A. Lozano,F oundations of MIMO communication. 1st ed. Cambridge, U.K., Cambridge University Press, Dec. 2018

    work page 2018

  3. [3]

    An overview on integrated localization and communication towards 6G,

    Z. Xiao and Y . Zeng, “An overview on integrated localization and communication towards 6G,”Sci. China Inf. Sci., vol. 65, no. 3, pp. 1–46, Dec. 2021

    work page 2021

  4. [4]

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

    H. Lu, Y . Zeng, C. Youet al., “A tutorial on near-field XL-MIMO communications towards 6G,”IEEE Commun. Surveys Tuts., vol. 26, no. 4, pp. 2213–2257, Apr. 2024

    work page 2024

  5. [5]

    Cell-free massive MIMO versus small cells,

    H. Q. Ngo, A. Ashikhmin, H. Yanget al., “Cell-free massive MIMO versus small cells,”IEEE Trans. Wireless Commun., vol. 16, no. 3, pp. 1834–1850, Jan. 2017

    work page 2017

  6. [6]

    Sparse MIMO for ISAC: New opportu- nities and challenges,

    X. Li, H. Min, Y . Zenget al., “Sparse MIMO for ISAC: New opportu- nities and challenges,”IEEE Wireless Commun. Mag., 2024, accepted

    work page 2024

  7. [7]

    A tutorial on environment-aware communications via channel knowledge map for 6G,

    Y . Zeng, J. Chen, J. Xuet al., “A tutorial on environment-aware communications via channel knowledge map for 6G,”IEEE Commun. Surveys Tuts, no. 3, pp. 1478–1519, Feb. 2024

    work page 2024

  8. [8]

    Beam codebook based beamforming protocol for multi-Gbps millimeter-wave WPAN systems,

    J. Wang, Z. Lan, C.-w. Pyoet al., “Beam codebook based beamforming protocol for multi-Gbps millimeter-wave WPAN systems,”IEEE J. Sel. Areas Commun., vol. 27, no. 8, pp. 1390–1399, 2009

    work page 2009

Show all 28 references
  1. [9]

    Millimeter wave beamforming for wireless backhaul and access in small cell networks,

    S. Hur, T. Kim, D. J. Loveet al., “Millimeter wave beamforming for wireless backhaul and access in small cell networks,”IEEE Trans. Commun., vol. 61, no. 10, pp. 4391–4403, 2013

    2013

  2. [10]

    Spatially sparse precoding in millimeter wave MIMO systems,

    O. El Ayach, S. Rajagopal, S. Abu-Surraet al., “Spatially sparse precoding in millimeter wave MIMO systems,”IEEE Trans. Wireless Commun., vol. 13, no. 3, pp. 1499–1513, 2014

    2014

  3. [11]

    An overview of signal processing techniques for millimeter wave MIMO systems,

    R. W. H. Jr., N. Gonz ´alez-Prelcic, S. Ranganet al., “An overview of signal processing techniques for millimeter wave MIMO systems,”IEEE J. Sel. Top. Signal Process., vol. 10, no. 3, pp. 436–453, 2016

    2016

  4. [13]

    Energy-efficient hybrid analog and digital precoding for mmWave MIMO systems with large antenna arrays,

    X. Gao, L. Dai, S. Hanet al., “Energy-efficient hybrid analog and digital precoding for mmWave MIMO systems with large antenna arrays,”IEEE J. Sel. Areas Commun., vol. 34, no. 4, pp. 998–1009, 2016

    2016

  5. [14]

    A survey on hybrid beam- forming techniques in 5G: Architecture and system model perspectives,

    I. Ahmed, H. Khammari, A. Shahidet al., “A survey on hybrid beam- forming techniques in 5G: Architecture and system model perspectives,” IEEE Commun. Surv. Tutorials, vol. 20, no. 4, pp. 3060–3097, Jun. 2018

    2018

  6. [15]

    Hybrid beamforming for terahertz wireless communications: Challenges, architectures, and open problems,

    C. Han, L. Yan, and J. Yuan, “Hybrid beamforming for terahertz wireless communications: Challenges, architectures, and open problems,”IEEE Wireless Commun., vol. 28, no. 4, pp. 198–204, Aug. 2021

    2021

  7. [16]

    A 28-GHz CMOS phased-array transceiver based on LO phase-shifting architecture with gain invariant phase tuning for 5G new radio,

    J. Pang, R. Wu, Y . Wanget al., “A 28-GHz CMOS phased-array transceiver based on LO phase-shifting architecture with gain invariant phase tuning for 5G new radio,”IEEE J. Solid-State Circuits, vol. 54, no. 5, pp. 1228–1242, Jun. 2019

    2019

  8. [17]

    Hybrid MIMO architectures for millimeter wave communications: Phase shifters or switches?

    R. M ´endez-Rial, C. Rusu, N. Gonz ´alez-Prelcicet al., “Hybrid MIMO architectures for millimeter wave communications: Phase shifters or switches?”IEEE Access, vol. 4, pp. 247–267, 2016

    2016

  9. [18]

    Phase shifters versus switches: An energy efficiency perspective on hybrid beamform- ing,

    S. Payami, N. M. Balasubramanya, C. Masouroset al., “Phase shifters versus switches: An energy efficiency perspective on hybrid beamform- ing,”IEEE Wirel. Commun. Lett., vol. 8, no. 1, pp. 13–16, Jun. 2018

    2018

  10. [19]

    Hybrid precoding for mixture use of phase shifters and switches in mmWave massive MIMO,

    C. Qi, Q. Liu, X. Yuet al., “Hybrid precoding for mixture use of phase shifters and switches in mmWave massive MIMO,”IEEE Trans. Commun., vol. 70, no. 6, pp. 4121–4133, May 2022

    2022

  11. [20]

    Millimeter wave MIMO with lens antenna array: A new path division multiplexing paradigm,

    Y . Zeng and R. Zhang, “Millimeter wave MIMO with lens antenna array: A new path division multiplexing paradigm,”IEEE Trans. Commun., vol. 64, no. 4, pp. 1557–1571, Apr. 2016

    2016

  12. [21]

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

    M. R. Castellanos, S. Yang, C.-B. Chaeet al., “Embracing reconfigurable antennas in the tri-hybrid MIMO architecture for 6G,”IEEE Trans. Commun., 2025, invited paper

    2025

  13. [22]

    Fluid antenna systems,

    K.-K. Wong, A. Shojaeifard, K.-F. Tonget al., “Fluid antenna systems,” IEEE Trans. Wireless Commun., vol. 20, no. 3, pp. 1950–1962, Nov. 2020

    1950

  14. [23]

    A tutorial on fluid antenna system for 6G networks: Encompassing communication theory, optimization methods and hardware designs,

    W. K. New, K.-K. Wong, H. Xuet al., “A tutorial on fluid antenna system for 6G networks: Encompassing communication theory, optimization methods and hardware designs,”IEEE Commun. Surv. Tutorials, 2024, doi: 10.1109/COMST.2024.3498855

    2024

  15. [24]

    Movable antennas for wireless commu- nication: Opportunities and challenges,

    L. Zhu, W. Ma, and R. Zhang, “Movable antennas for wireless commu- nication: Opportunities and challenges,”IEEE Commun. Mag., vol. 62, no. 6, pp. 114–120, Jun. 2024

    2024

  16. [25]

    A tutorial on movable antennas for wireless networks,

    L. Zhu, W. Ma, W. Meiet al., “A tutorial on movable antennas for wireless networks,”IEEE Commun. Surv. Tutorials, 2025, doi: 10.1109/COMST.2025.3546373

    2025

  17. [26]

    Movable antenna for wireless com- munications: Prototyping and experimental results,

    Z. Dong, Z. Zhou, Z. Xiaoet al., “Movable antenna for wireless com- munications: Prototyping and experimental results,” submitted to IEEE Trans. Wireless Commun., 2024. Avaliable online:arXiv:2408.08588

    2024 arXiv

  18. [27]

    Group movable antenna with flexible sparsity: Joint array position and sparsity optimization,

    H. Lu, Y . Zeng, S. Jinet al., “Group movable antenna with flexible sparsity: Joint array position and sparsity optimization,”IEEE Wireless Commun. Lett., vol. 13, no. 12, pp. 3573–3577, Dec. 2024. [28]5G; Study on Channel Model for Frequencies from 0.5 to 100 GHz, document 38...

    2024

  19. [29]

    S. U. Pillai,Array signal processing. Springer Science & Business Media, 2012

    2012

  20. [30]

    Qorvo official website,

    Qorvo, “Qorvo official website,” https://www.qorvo.com/, 2025, ac- cessed: 2025-05-17

    2025

Pith tools

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