Near-Field Measurement System for the Upper Mid-Band

Ali Rasteh; Hao Guo; Marco Mezzavilla; Raghavendra Palayam Hari; Sundeep Rangan

REVIEW 3 major objections 4 minor 17 references

Near-Field Measurement System for the Upper Mid-Band

T0 review · 3 major / 4 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read This paper claims that near-field multipath parameters in the upper mid-band can be measured with a moving 2x2 antenna array by locating virtual image points behind reflectors via triangulation, avoiding large, costly arrays.

desk verdict A novel but unvalidated near-field measurement concept with a load-bearing delay-model inconsistency that undercuts the stated claims. read the letter →

arxiv 2412.02815 v1 pith:5Q2A7V7B submitted 2024-12-03 eess.SP

Ali Rasteh , Raghavendra Palayam Hari , Hao Guo , Marco Mezzavilla , Sundeep Rangan This is my paper · ORCID

classification eess.SP

keywords uppermid-bandFR3near-fieldchannelmeasurementreflectionmodelsyntheticapertureparameterextractiontriangulationnon-coherentmeasurements

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 argues that measuring near-field channels in the upper mid-band (FR3, 6-24 GHz) does not require a large, expensive antenna array. Instead, a 2x2 array moved along a linear track forms a synthetic aperture; each measurement is processed with a plane-wave model to extract angles and relative delays, and the reflection model's image points—the virtual transmitter positions behind reflecting surfaces—are located by triangulation. If correct, this gives a low-cost way to obtain the spherical-wave parameters needed for near-field MIMO models, which ordinary far-field angle-based parameters cannot provide.

What carries the argument

The load-bearing object is the reflection model's image-point characterization: the distance for path $\ell$ between a transmit point $x^t$ and receive point $x^r$ is $d_\ell(x^r,x^t)=\|x^r-z^t_\ell\|$, where $z^t_\ell$ is the reflected image of the transmit point under that path. The image point is an affine function of the transmit position, $z^t_\ell = z^t_{\ell 0} + s_\ell R(\alpha_\ell)(x^t-x^t_0)$, with a rotation $R(\alpha_\ell)$ and a parity $s_\ell=\pm 1$ for even or odd numbers of reflections. This representation turns near-field parameter extraction into image-point localization: plane-wave extraction gives the angles and relative delays, triangulation gives the distance to the image point, and the parity completes the reflection-model parameter vector $\theta^\mathrm{RM}_\ell=(g_\ell,\tau_\ell,\phi^r_\ell,\phi^t_\ell,s_\ell)$. The synthetic aperture is provided by a 2x2 array moved along linear tracks; the measurements need not be phase-coherent with each other.

What would settle it

In a controlled room with a single flat reflector at a known position, move the receiver between two positions separated by 0.6 m and compare the triangulated image point with the reflector's true image; the central claim fails if the estimated image point shifts by more than the error predicted from angular resolution, or if the relative delay between the direct and reflected paths changes by more than the delay resolution (about 2 ns for a 500 MHz bandwidth) across the two positions.

Watch

Extended reading notes

Core claim

On its own terms, the paper's central claim is that a multi-path near-field channel can be parameterized by the reflection model, in which every specular NLOS path is equivalent to a line-of-sight path to a virtual image point of the transmitter, and that these image points can be accurately estimated by triangulation from several wideband measurements made with only two transmit and two receive antennas. The estimation pipeline is: (1) move the small array to K positions to form a synthetic aperture; (2) at each position estimate the plane-wave parameters—angle of arrival, angle of departure, relative delay, and gain—via orthogonal matching pursuit; (3) triangulate the highest-SNR path's image point from two widely separated positions to recover the absolute time of flight, then recover the other paths' absolute delays from their relative delays. The reflection parity, a binary variable describing whether the path has an odd or even number of reflections, is read off from the model.

Load-bearing premise

The load-bearing premise is that the relative delays between paths stay constant as the receiver moves to widely separated positions, yet wide separation is exactly what triangulation needs to be accurate.

Editorial extensions

If this is right

Near-field channel sounding can be performed with a 2x2 array and mechanical motion instead of a large fixed array, lowering cost and complexity.
Phase coherence across synthetic-aperture measurements is not required, because each measurement's complex gain is treated as an arbitrary free parameter in the cost function.
Only the strongest path needs to be triangulated; the absolute delays of the remaining paths follow from the relative delays estimated in the plane-wave step.
Once the reflection-model parameters are known, the distance function predicts the full MIMO response over any aperture, so the channel can be extrapolated to other array geometries.
The proposed hardware prototype at 10 GHz with 500 MHz bandwidth is a working demonstration of this procedure, and controlled validation with known true parameters is identified as the next step.

Reading between the lines

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

The constant-relative-delay assumption in the plane-wave extraction step and the wide-baseline requirement of triangulation pull in opposite directions; a practical system must either keep the baseline small enough that delay drift stays below the delay resolution or estimate delays and positions jointly.
The same image-point triangulation idea could be inverted for indoor mapping and localization: a moving two-antenna receiver could estimate the virtual source positions of reflections, yielding a low-cost map of reflective surfaces.
Because the reflection model is exact for specular planar reflectors, its accuracy on curved or diffuse objects is untested; a controlled experiment with a curved metal surface would reveal the limits of the image-point representation.
The accuracy of triangulation depends on signal-to-noise ratio and angular resolution; computing a lower bound on image-point location error as a function of baseline and bandwidth would show how long the synthetic-aperture track needs to be.

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

Dear Colleague,

The short version: this paper has a genuinely new measurement concept, but it is not yet a valid result. The proposed combination of a 2x2 non-coherent array, a moving synthetic aperture, and reflection-model image-point triangulation is worth discussing. The hardware setup for FR3 is thoughtfully laid out, and the authors are honest that this is preliminary.

That said, the central claim—that image points can be accurately estimated via triangulation from many small-array measurements—is not supported by the evidence. The simulation in Section IV is a single qualitative example with no error metrics, and it generates the data from the same reflection model that the estimator assumes, so it can only test internal consistency, not model validity. Section V describes the apparatus but reports no measured results. The conclusion states the method is 'validated in a simulation and using a preliminary experimental procedure,' but the experiment is only a setup description.

More seriously, the stress-test note is right: there is an internal inconsistency in the delay model. In Step 2, Eq. (23) models every measurement in K0 as sharing a single relative-delay vector δℓ, and the cost function (24) sums over all k. That is only valid if the receiver positions are close enough that the distances from each image point change negligibly. Step 3 then requires two widely separated receiver positions to triangulate the image point. With the 0.6 m offsets in Section V and 500 MHz bandwidth, the relative delays between paths change by more than a cycle across the band. So the same measurements cannot satisfy both the constant-delay model and the large-baseline triangulation. If the authors intend to estimate PWA parameters per cluster, they need to explain how the per-cluster δℓ estimates are reconciled across different reference points. The paper doesn't do that. This is a load-bearing flaw, not a cosmetic one.

On the positive side, the reflection-model parametrization is clean, and the idea of non-coherent synthetic aperture for near-field is sensible and could be useful to the FR3 channel measurement community. The paper would benefit from a careful revision that either fixes the delay model or clearly separates the local and wide-aperture processing steps, and from real measured results against known ground truth.

For now, I would not cite this as a validated method, but I would not mind seeing it in a workshop or as a position paper. It deserves a serious referee because the concept is novel and the flaw is instructive, even if the current form should be rejected.

Referee Report

3 major / 4 minor

Summary. The manuscript proposes a near-field channel parameter extraction method for the upper mid-band (6–24 GHz), building on a previously developed reflection model. The key idea is to estimate the parameters of each multipath component in the reflection model by first estimating plane-wave-approximation (PWA) parameters from a small number of antennas (2×2) moved to form a synthetic aperture, and then recovering the absolute time of flight (hence the reflection image point) via triangulation from two widely separated receive positions. The paper claims accurate estimation of image points with this procedure, and it describes a preliminary hardware setup based on a Pi-Radio FR3 transceiver, an RFSoC, and linear tracks. A simulation in Section IV and a hardware description in Section V are presented in support of the claims.

Significance. The problem addressed is relevant: near-field multipath parameter extraction in the upper mid-band is important for 6G channel modeling, and a low-cost method using non-coherent measurements with a small antenna count would be practically valuable. The paper also brings a useful geometric interpretation via reflection image points and correctly identifies the limitations of standard plane-wave models in the near field. However, the central claim of accurate estimation is not supported by the evidence in the manuscript. The simulation is a qualitative, self-consistency test with no error metrics, and the experimental section contains no measured results. Equally important, the estimation algorithm as described contains an internal inconsistency between the joint PWA estimation step and the triangulation step. Given that the load-bearing claims are neither validated nor internally consistent, the paper in its current form cannot be accepted.

major comments (3)

[Section III, Steps 2 and 3 (Eqs. (23)–(25))] The joint PWA estimation in Eq. (24) uses a single relative delay δ_l for each path across all measurements k in the cost function. This model is valid only if the receive position changes negligibly relative to the path delay, because the relative delay between paths changes with the receiver position. Yet Step 3 explicitly instructs the use of two widely separated points for triangulation. For two receiver positions separated by Δ, the relative delay between paths l and 0 changes by roughly (u_0 − u_l)^T Δ / c, which is a frequency-dependent phase that cannot be absorbed by the per-measurement, frequency-independent gains g_{lk}. With the experimental parameters in Section V (0.6 m separation, 500 MHz bandwidth), this phase change can exceed 2π across the band. Thus the minimization (24) is based on an invalid model for the same large-baseline measurements that the triangulation requires. If the authors intend δ_l to be constant only within a local cluster, then Eq. (24) should sum over that cluster only, and the paper does not explain how the per-cluster δ_l estimates are combined in Eq. (25).
[Section IV (Simulations)] The simulation is qualitative only. Figure 1 shows heatmaps and estimated positions, but no quantitative metrics such as localization error, bias, variance, or success rate are reported. Moreover, the simulation generates the received data using the same reflection model that the estimator assumes, so it is a closed-loop self-consistency test rather than an independent validation of the model. The paper therefore does not provide evidence that the method works on data that are not already generated from the assumed model.
[Section V (Experiments)] Section V describes the hardware and the procedure but reports no measured results, no channel responses, no estimated image points, and no comparison with known ground-truth positions or delays. The conclusion explicitly states that future work will validate the experimental measurements in a controlled environment, which confirms that the experimental validation is absent. A system description without any measurement results cannot support the abstract's claim of accurate estimation via triangulation.

minor comments (4)

[Section II.C, Eq. (17)] The notation is inconsistent: Eq. (17) uses cτ0 on the right-hand side but τ0 is not explicitly defined as the absolute delay of a reference path; it should be τ_0 or cτ_0 consistently with Eqs. (9) and (18).
[Section III, Eq. (24)] The term 'a_{kn}' should be 'a^t_{kn}' for consistency with Eq. (23); this typo makes the formula harder to follow.
[Section IV, Figure 1] The caption states that 'red and green dots represent the actual and detected paths', but the associated plots are not self-explanatory; a legend and a quantitative evaluation of detection accuracy would be helpful.
[General] There are minor spacing and typographical errors, e.g., 'p = 2case' and 'PW A' with a stray space, which should be corrected.

Circularity Check

2 steps flagged · score 6.0 of 10

Closed-loop simulation and self-cited reflection model make the validation circular, though the triangulation algorithm itself has independent content.

self citation load bearing [Section I (Introduction) and Section II.A (Near-Field Channel Model), Eqs. (5)-(6)]
"we extend our recently-developed reflection model (RM) [8] ... To capture near-field propagation, we use the reflection model in [8]. ... Following a derivation similar to [8], it can be shown the reflection image of any point xt is given by zt = ztℓ0 + sℓR(αℓ)(xt − xt0)."

The target parameters of the entire extraction method — the image points ztℓ, absolute delay τℓ, angles (ϕr,ϕt), and parity sℓ — are defined by the reflection model taken from the authors' own prior paper [8] (Hu, Yin, Rangan, Mezzavilla). The present paper does not re-derive or independently validate that model; it only cites [8]. This is load-bearing because every later equation (13), (17), (20), and the triangulation step assume the RM parameterization. If [8]'s model were not a correct description of physical near-field paths, the extracted 'near-field parameters' would not describe the channel. The only validation offered (Section IV) assumes the same model, so the citation chain is not broken by independent evidence.
other [Section IV (Simulations), Fig. 1]
"The transmitter is presumed to be isotropic, resulting in three primary reflections from the respective walls, thereby creating three red virtual transmitters outside of the room. Each virtual transmitter constitutes a reflection of the original transmitter relative to the corresponding wall."

The simulation creates the ground-truth image points by applying the same specular-reflection/image-point model that the estimator in Section III assumes. The channel responses are recorded from these virtual transmitters, and then the algorithm estimates the locations of 'the primary transmitter and its corresponding reflection images.' Success in this setup only demonstrates that the estimator can invert its own assumed generative model; it is a closed-loop self-consistency test. It cannot validate the claim that real upper-mid-band multipath channels are described by the RM, nor that the 2x2 synthetic-aperture procedure extracts physical near-field parameters.

full rationale

The derivation of the triangulation procedure itself is not circular: given the RM parameterization, Eq. (13) expresses path distance in terms of image-point geometry, and the AoA triangulation in Step 3 is a standard geometric calculation. However, the paper's central validation is circular in two connected ways. First, the reflection model that defines the quantities to be estimated is imported from the authors' own prior work [8] and is not independently tested; it is the load-bearing premise of the entire method. Second, the only quantitative evidence that the method works is a simulation in which the ground-truth image points are generated by that same reflection model and then recovered by an estimator that assumes the same model. This is a closed-loop consistency test, not an external validation. The experimental section (V) describes hardware and procedure but presents no measured results, so it does not break the loop. There is also a non-circular but serious internal inconsistency: Step 2 (Eq. 23) assumes a single relative-delay vector δℓ across all measurements, while Step 3 requires widely separated receiver positions over which the relative delays necessarily change (per Eq. 17); this is a correctness risk, not a circularity, and does not by itself raise the circularity score. Overall, the algorithmic core has independent content, but the paper's advertised support for its strongest claim reduces to self-citation plus a self-consistent simulation, warranting a partial-circularity score of 6.

Assumptions & free parameters 0 free parameters · 4 assumptions · 1 invented entities

The method rests on the reflection model from [8], the local plane-wave approximation, the constancy of relative delays over a subset of synthetic apertures, and a 2D geometry. No numeric constants are fitted to data. The only introduced abstraction is the reflection image point, which is a reparameterization of the path geometry; no independent evidence for it is provided in this paper.

assumptions (4)

domain assumption Reflection model from [8]: every NLOS path can be represented by an image point with a rotation angle and parity, so the path distance equals the Euclidean distance from the receiver to the reflected image of the transmitter.
Inherited from prior work [8] by the same group; this paper does not re-derive or independently verify the model. Invoked in Section II.A, Eq. (5)-(6).
domain assumption The plane wave approximation is valid over the small (2x2) array within each measurement, so PWA parameters (angle and delay) are meaningful local estimates.
Assumed in Section III Step 2; the small array size makes the local PWA plausible, but no error analysis is provided.
ad hoc to paper The relative delays δℓ between paths are approximately constant across the subset K0 of measurements in the joint PWA estimation.
Eq. (23)-(24) use a single δℓ per path for all k; this contradicts a wide synthetic aperture and is unjustified; see red flag.
domain assumption All propagation paths are specular planar reflections in a 2D plane (p=2).
The paper restricts to 2D for presentation and measurement geometry; diffuse or out-of-plane scattering is not considered.

invented entities (1)

Reflection image point zt_ℓ0 and virtual transmitters
purpose: To reparameterize near-field NLOS paths as point sources behind reflecting surfaces, reducing parameter estimation to localization.
The image point is a mathematical construct from [8]. The paper attempts to locate it via triangulation, but provides no measured ground-truth validation, so no independent falsifiable handle is demonstrated in this paper.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Near-Field Measurement System for the Upper Mid-Band." pith.science (2026). https://pith.science/paper/5Q2A7V7B

@misc{pith2026241202815,
  author       = {Pith},
  title        = {Pith review of: Near-Field Measurement System for the Upper Mid-Band},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5Q2A7V7B}},
  note         = {Machine review of arXiv:2412.02815}
}

read the original abstract

The upper mid-band (or FR3, spanning 6-24 GHz) is a crucial frequency range for next-generation mobile networks, offering a favorable balance between coverage and spectrum efficiency. From another perspective, the systems operating in the near-field in both indoor environment and outdoor environments can support line-of-sight multiple input multiple output (MIMO) communications and be beneficial from the FR3 bands. In this paper, a novel method is proposed to measure the near-field parameters leveraging a recently developed reflection model where the near-field paths can be described by their image points. We show that these image points can be accurately estimated via triangulation from multiple measurements with a small number of antennas in each measurement, thus affording a low-cost procedure for near-field multi-path parameter extraction. A preliminary experimental apparatus is presented comprising 2 transmit and 2 receive antennas mounted on a linear track to measure the 2x2 MIMO channel at various displacements. The system uses a recently-developed wideband radio frequency (RF) transceiver board with fast frequency switching, an FPGA for fast baseband processing, and a new parameter extraction method to recover paths and spherical characteristics from the multiple 2x2 measurements.

Figures

Figures reproduced from arXiv: 2412.02815 by the authors.

**Figure 2.** Our designed FR3 near-field measurement system with Xilinx RFSoC 4x2, Pi-Radio FR3 transceiver, Vivaldi antennas, [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

17 extracted references · 11 canonical work pages

[1]

Cellular wireless networks in the upper mid-band,

S. Kang, M. Mezzavilla, S. Rangan, A. Madanayake, S. B. Venkatakrish- nan, G. Hellbourg, M. Ghosh, H. Rahmani, and A. Dhananjay, “Cellular wireless networks in the upper mid-band,” IEEE Open J. Commun. Soc., Mar. 2024

work page 2024
[2]

Angular Spread Statistics for 6.75 GHz FR1 (C) and 16.95 GHz FR3 Mid-Band Frequencies in an Indoor Hotspot Environment,

D. Shakya, M. Ying, and T. S. Rappaport, “Angular Spread Statistics for 6.75 GHz FR1 (C) and 16.95 GHz FR3 Mid-Band Frequencies in an Indoor Hotspot Environment,” arXiv preprint arXiv:2409.03013 , 2024

arXiv 2024
[3]

Propagation measurements and channel models in Indoor Environment at 6.75 GHz FR1 (C) and 16.95 GHz FR3 Upper-mid band Spectrum for 5G and 6G,

D. Shakya, M. Ying, T. S. Rappaport, H. Poddar, P. Ma, Y . Wang, and I. Al-Wazani, “Propagation measurements and channel models in Indoor Environment at 6.75 GHz FR1 (C) and 16.95 GHz FR3 Upper-mid band Spectrum for 5G and 6G,” arXiv preprint arXiv:2405.01358 , 2024

arXiv 2024
[4]

End-to-End Deep Learning for TDD MIMO Systems in the 6G Upper Midbands

J. Park, F. Sohrabi, A. Ghosh, and J. G. Andrews, “End-to-End Deep Learning for TDD MIMO Systems in the 6G Upper Midbands,” arXiv preprint arXiv:2402.01033, 2024

work page Pith review arXiv 2024
[5]

Near-field MIMO communications for 6G: Fundamentals, challenges, potentials, and future directions,

M. Cui, Z. Wu, Y . Lu, X. Wei, and L. Dai, “Near-field MIMO communications for 6G: Fundamentals, challenges, potentials, and future directions,” IEEE Commun. Mag. , vol. 61, no. 1, pp. 40–46, Jan. 2023

2023
[6]

Beam focusing for near-field multiuser MIMO communications,

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

work page 2022
[7]

Study on channel model for frequencies from 0.5 to 100 GHz (Release 16),

3GPP Technical Report 38.901, “Study on channel model for frequencies from 0.5 to 100 GHz (Release 16),” Dec. 2019

2019
[8]

Parametrization and estimation of high-rank line-of-sight MIMO channels with reflected paths,

Y . Hu, M. Yin, S. Rangan, and M. Mezzavilla, “Parametrization and estimation of high-rank line-of-sight MIMO channels with reflected paths,” IEEE Trans. Wireless Commun. , Apr. 2024

work page 2024

Show all 17 references

[9]

Efficient ray-tracing simulation for near-field spatial non-stationary mmWave massive MIMO channel and its experimental validation,

Z. Yuan, J. Zhang, V . Degli-Esposti, Y . Zhang, and W. Fan, “Efficient ray-tracing simulation for near-field spatial non-stationary mmWave massive MIMO channel and its experimental validation,” IEEE Trans. Wireless Commun., pp. 1–1, 2024

2024
[10]

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

Z. Yuan, J. Zhang, Y . Ji, G. F. Pedersen, and W. Fan, “Spatial non- stationary near-field channel modeling and validation for massive MIMO systems,” IEEE Trans. Antennas Propag. , vol. 71, no. 1, pp. 921–933, Jan. 2023

2023
[11]

A Frequency Hopping Software-Defined Radio Platform for Communications and Sensing in the Upper Mid-Band,

M. Mezzavilla, A. Dhananjay, M. Zappe, and S. Rangan, “A Frequency Hopping Software-Defined Radio Platform for Communications and Sensing in the Upper Mid-Band,” in Proc. IEEE International Workshop on Signal Processing Advances in Wireless Communications (SPAWC) , 2024, pp. 611–615

2024
[12]

Sub-terahertz near field channel measurements and analysis with beamforming and Bessel beams,

D. Bodet, V . Petrov, S. Petrushkevich, and J. M. Jornet, “Sub-terahertz near field channel measurements and analysis with beamforming and Bessel beams,” Scientific Reports, vol. 14, no. 1, p. 19675, 2024

2024
[13]

Near- field channel estimation for extremely large-scale Terahertz communi- cations,

S. Yang, Y . Peng, W. Lyu, Y . Li, H. He, Z. Zhang, and C. Yuen, “Near- field channel estimation for extremely large-scale Terahertz communi- cations,” Science China Information Sciences , vol. 67, no. 9, p. 192302, 2024

2024
[14]

An experimental multi-band channel characterization in the upper mid-band,

R. Bomfin, A. Bazzi, H. Guo, H. Lee, M. Mezzavilla, S. Rangan, J. Choi, and M. Chafii, “An experimental multi-band channel characterization in the upper mid-band,” arXiv preprint arXiv:2411.12888 , 2024

2024 arXiv
[15]

R. W. Heath Jr. and A. Lozano, Foundations of MIMO Communication. Cambridge University Press, 2018

2018
[16]

Signal recovery from random mea- surements via orthogonal matching pursuit,

J. A. Tropp and A. C. Gilbert, “Signal recovery from random mea- surements via orthogonal matching pursuit,” IEEE Transactions on information theory, vol. 53, no. 12, pp. 4655–4666, 2007

2007
[17]

Theoretical results on sparse representations of multiple-measurement vectors,

J. Chen and X. Huo, “Theoretical results on sparse representations of multiple-measurement vectors,” IEEE Transactions on Signal process- ing, vol. 54, no. 12, pp. 4634–4643, 2006

2006

Pith tools

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

[1] [1]

Cellular wireless networks in the upper mid-band,

S. Kang, M. Mezzavilla, S. Rangan, A. Madanayake, S. B. Venkatakrish- nan, G. Hellbourg, M. Ghosh, H. Rahmani, and A. Dhananjay, “Cellular wireless networks in the upper mid-band,” IEEE Open J. Commun. Soc., Mar. 2024

work page 2024

[2] [2]

Angular Spread Statistics for 6.75 GHz FR1 (C) and 16.95 GHz FR3 Mid-Band Frequencies in an Indoor Hotspot Environment,

D. Shakya, M. Ying, and T. S. Rappaport, “Angular Spread Statistics for 6.75 GHz FR1 (C) and 16.95 GHz FR3 Mid-Band Frequencies in an Indoor Hotspot Environment,” arXiv preprint arXiv:2409.03013 , 2024

arXiv 2024

[3] [3]

Propagation measurements and channel models in Indoor Environment at 6.75 GHz FR1 (C) and 16.95 GHz FR3 Upper-mid band Spectrum for 5G and 6G,

D. Shakya, M. Ying, T. S. Rappaport, H. Poddar, P. Ma, Y . Wang, and I. Al-Wazani, “Propagation measurements and channel models in Indoor Environment at 6.75 GHz FR1 (C) and 16.95 GHz FR3 Upper-mid band Spectrum for 5G and 6G,” arXiv preprint arXiv:2405.01358 , 2024

arXiv 2024

[4] [4]

End-to-End Deep Learning for TDD MIMO Systems in the 6G Upper Midbands

J. Park, F. Sohrabi, A. Ghosh, and J. G. Andrews, “End-to-End Deep Learning for TDD MIMO Systems in the 6G Upper Midbands,” arXiv preprint arXiv:2402.01033, 2024

work page Pith review arXiv 2024

[5] [5]

Near-field MIMO communications for 6G: Fundamentals, challenges, potentials, and future directions,

M. Cui, Z. Wu, Y . Lu, X. Wei, and L. Dai, “Near-field MIMO communications for 6G: Fundamentals, challenges, potentials, and future directions,” IEEE Commun. Mag. , vol. 61, no. 1, pp. 40–46, Jan. 2023

2023

[6] [6]

Beam focusing for near-field multiuser MIMO communications,

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

work page 2022

[7] [7]

Study on channel model for frequencies from 0.5 to 100 GHz (Release 16),

3GPP Technical Report 38.901, “Study on channel model for frequencies from 0.5 to 100 GHz (Release 16),” Dec. 2019

2019

[8] [8]

Parametrization and estimation of high-rank line-of-sight MIMO channels with reflected paths,

Y . Hu, M. Yin, S. Rangan, and M. Mezzavilla, “Parametrization and estimation of high-rank line-of-sight MIMO channels with reflected paths,” IEEE Trans. Wireless Commun. , Apr. 2024

work page 2024