Joint Motion, Angle, and Range Estimation in Near-Field under Array Calibration Imperfections

Abdulkadir Celik; Ahmed Hussain; Ahmed M. Eltawil; Asmaa Abdallah

REVIEW 4 major objections 7 minor 11 references

Joint Motion, Angle, and Range Estimation in Near-Field under Array Calibration Imperfections

T0 review · 4 major / 7 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read The paper shows that in the near field of an ultra-massive antenna array, a target's angle, range, radial velocity, and transverse velocity can be read jointly from the width and center of its 2D-DFT angle-Doppler response, then refined…

desk verdict The paper's headline Doppler-spread claim is unsupported by its own Eq. (13), which is a spatial projection, not a Doppler spectrum; the idea is plausible and simulations are clean, but the core derivation must be redone. read the letter →

arxiv 2507.13463 v1 pith:WEIMOOHB submitted 2025-07-17 eess.SP

Ahmed Hussain , Asmaa Abdallah , Abdulkadir Celik , Ahmed M. Eltawil This is my paper · ORCID

classification eess.SP

keywords near-fieldsensingUM-MIMO2D-DFTangularspreadDopplertransversevelocityMUSICrefinementarraycalibrationimperfections

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 claims that in the radiative near field of an ultra-massive multiple-input multiple-output (UM-MIMO) array, a single moving target's angle, range, radial velocity, and transverse (sideways) velocity can be estimated jointly by examining how its response spreads in the two-dimensional discrete Fourier transform (2D-DFT) angle-Doppler domain. The angular spread is centered at the true angle and its width is set by range, while the Doppler spread is centered at the true radial velocity and its width grows with transverse velocity. Exploiting these relations, the paper develops a two-stage estimator: coarse 2D-DFT measurements of the two spreads, followed by one-dimensional multiple signal classification (MUSIC) refinement of each parameter in sequence. This replaces a four-dimensional search by one-dimensional scans. Simulations report a normalized mean squared error of about -40 dB for all four parameters at high signal-to-noise ratio, with lower complexity than gradient-based maximum likelihood, and the method is valid only inside the effective beamfocusing Rayleigh distance where the spreads exist.

What carries the argument

The load-bearing object is the 2D-DFT angle-Doppler response of the space-time steering matrix $\mathbf{V}(\theta,r,v_r,v_\theta)=\sqrt{\xi_t}(\mathbf{A}\odot\mathbf{B})$, where $\mathbf{A}$ carries the near-field spherical-wave phases and $\mathbf{B}$ carries element-wise Doppler shifts. The paper's central identity is the Fresnel-integral form of the angular gain, $G_{\mathrm{ang}}\approx\left|\left(C(\gamma_1,\gamma_2)+jS(\gamma_1,\gamma_2)\right)/(2\gamma_2)\right|^2$, where $C$ and $S$ are Fresnel integrals, with $\gamma_1=\sqrt{r_F/(d\cos^2\theta_u)}(\sin\theta_n-\sin\theta_u)$ and $\gamma_2=(N_{\mathrm{BS}}/2)\sqrt{d\cos^2\theta_u/r_F}$; it shows the response is centered on the true angle and its width shrinks with range. The analogous Doppler-domain expression shows a linear phase slope whose width is set by $v_\theta\cos\theta_u/(r_F f_r)$ and whose center is set by $v_r$. Offline lookup tables map measured 3 dB spread widths to range and transverse velocity, and 1D MUSIC on the sample covariance matrix then refines each parameter separately.

What would settle it

Simulate one target whose transverse velocity is large enough that its Doppler spread covers several Fourier bins, run the proposed estimator, and compare the 1D-MUSIC-refined estimates against a full four-dimensional maximum-likelihood search; if the refined estimates become biased as the spread widens, the single-shape refinement assumption is wrong.

Watch

Extended reading notes

Core claim

The central discovery is that the two-dimensional DFT of the near-field space-time snapshot decouples location from motion: the angular spread width maps to range, the Doppler spread width maps to transverse velocity, and the centers of the two spreads give angle and radial velocity. Because the spherical wavefront makes the per-antenna Doppler shift vary linearly across the array, transverse motion appears not as a single Doppler tone but as a spread whose phase slope is proportional to $v_\theta\cos\theta_u / r_F$. The paper derives this from a Fresnel approximation of the array response and shows that a coarse 2D-DFT stage plus per-parameter 1D MUSIC refinement reaches an NMSE of $-40$ dB under calibration imperfections while avoiding the exponential four-dimensional search.

Load-bearing premise

The refinement stage assumes that after the coarse Fourier step, each parameter can be recovered by matching a single idealized response shape to the measured signal, even though the moving target's Doppler spread makes the measured signature a mix of many shapes; the paper does not say how large the error from that mix can be.

Editorial extensions

If this is right

A monostatic UM-MIMO base station can estimate transverse velocity from a single coherent processing interval, something far-field planar-wavefront radars cannot do without multistatic geometry.
The two-stage estimator replaces a four-dimensional grid search by one-dimensional scans, with complexity about $O(N_{\mathrm{BS}}M(\log N_{\mathrm{BS}}+\log M)) + O((N_{\mathrm{BS}}M)^3 + G N_{\mathrm{BS}}M)$, avoiding the $O(TG^4N_{\mathrm{BS}}M)$ cost of the gradient-based search.
The angular and Doppler spreads shrink as range grows, so the method's validity is bounded by the effective beamfocusing Rayleigh distance; beyond it the near-field spreads disappear and the approach becomes inapplicable.
The coarse DFT stage already gives usable angle, radial-velocity, range, and transverse-velocity estimates, so the 1D MUSIC refinement only needs to search a small neighborhood around them.

Reading between the lines

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

Beyond the paper, the same width-to-range and width-to-transverse-velocity mappings suggest that a target crossing boresight produces the widest Doppler support, so sensitivity to transverse velocity is highest exactly where conventional radial-Doppler sensing is blind; the paper does not analyze this angular dependence of the resolution limit.
A natural untested extension is multi-target operation: when two targets' spreads overlap in the angle-Doppler plane, the lookup-table correlation step would need to be replaced by a sparse-recovery or iterative cancellation routine, since the paper's single-target model would otherwise merge the two spreads.
The offline lookup tables couple spread width to range and transverse velocity on a discrete grid; in practice the achievable resolution is set by how finely the 3 dB spread changes with each parameter, so a Cramér-Rao-style bound on spread-width estimation would predict where the coarse stage forces the MUSIC stage to work hardest.

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

Read this one before you trust the abstract. The central analytical claim — that a near-field target's transverse velocity creates a Doppler spread whose center gives radial velocity and width gives transverse velocity — is not derived correctly in the paper. Equation (13), which is supposed to be the Doppler-domain gain, is actually a spatial inner product between the Doppler steering vector b and a DFT codebook f, summed over the antenna index n. There's no temporal DFT over the symbol index m, so it cannot produce a Doppler spectrum. Worse, the radial velocity vr appears inside the phase as a constant that factors out of the squared magnitude, so the pattern is not shifted by vr at all. The text immediately after Eq. (13) claims the opposite.

That's the load-bearing flaw. If the Doppler-spread-to-vθ relationship is real, it is not established by this derivation. The authors might be seeing an effect in simulation that they haven't correctly characterized analytically.

Credit where due: the idea itself is new relative to the cited near-field estimation works [6]-[8], which don't exploit Doppler spread across the array for transverse velocity. The two-stage DFT-coarse/1D-MUSIC-fine workflow is sensible, and the simulation results are internally consistent, reaching -40 dB NMSE at high SNR. The calibration error model is reasonable.

Other soft spots, in decreasing order: the signal model in Eq. (7) has a dimension mismatch that makes it hard to verify the data generation; the 1D MUSIC refinement is applied without analyzing the multi-rank spatial signature induced by Doppler spread; the lookup tables are populated by simulating the same forward model used for the test data, so the coarse estimates are matched to a self-consistent dictionary; and the complexity expression includes an O((NBSM)^3) SVD term that overcounts the actual NBS×NBS covariance — conservative, but sloppy.

Who should read it: researchers in near-field ISAC or automotive radar who want a low-complexity estimator for transverse velocity. They'll get the idea and maybe the simulation approach, but they'll have to redo the derivation themselves.

Recommendation: worth sending to peer review because the observation is potentially valuable and the simulations could be reproduced. But I would not accept it without a corrected Doppler-spread analysis — either a proper temporal DFT derivation or an explicit statement that the relationship is empirical. For your own work, I wouldn't cite it until the math is fixed.

Referee Report

4 major / 7 minor

Summary. The paper proposes a two-stage estimator for the joint location and velocity of a moving target in the radiative near field of a UM-MIMO array. The first stage applies a 2D-DFT to the space-time snapshot and exploits the angular spread and the Doppler spread of the response to obtain coarse estimates of angle, range, radial velocity, and transverse velocity. The second stage refines each parameter with a separate 1D MUSIC scan. Simulation results with a 256-antenna array at 28 GHz report NMSEs approaching -40 dB at high SNR, comparable to polar-codebook benchmarks and better than gradient-based ML, at lower nominal complexity.

Significance. The problem is timely and important for near-field ISAC, and the idea of decoupling the four parameters through spectral spread characteristics is appealing. If the analytical characterization were correct, the proposed two-stage method would be a valuable low-complexity alternative to 4D ML search. The paper also explicitly models array calibration errors and shows graceful degradation. However, the central Doppler-spread analysis is currently mis-stated, and the range/transverse-velocity estimates rely on lookup tables generated from the same forward model used in the test simulations, so the evidence for the claimed spectral relationships is partly circular. The manuscript does not provide code or machine-checked proofs; the numerical claims are not independently reproducible from the text alone.

major comments (4)

[III-B] Equation (13) is not a Doppler spectrum. G_dopp(vr,vtheta; theta_n) is defined as an inner product over the antenna index n between a single-symbol Doppler steering vector b(vr,vtheta) and a spatial DFT codebook f(theta_n); it does not involve a temporal DFT over the symbol index m. Consequently, it cannot characterize the Doppler spread of the space-time snapshot. Moreover, the vr-dependent term inside the exponent in Eq. (13) is independent of n and therefore disappears from the squared magnitude, so the statement that the gain pattern is shifted by vr and that the midpoint of the Doppler spread gives vr does not follow from the displayed expression. Since Eqs. (14)-(15) and the lookup table K_v are built on this relationship, the analytical foundation of the velocity estimation is unsupported as written. Please provide a correct derivation of the 2D-DFT response of the space-time snapshot in Eq. (7), or present the approach as purely empirical and remove the claimed analytical relationships.
[II-B] The signal model in Eq. (7) is dimensionally inconsistent. Y is defined as [V^1 s(1), ..., V^M s(M)] with V in C^{NBS x M} and s(m) in C^{NBS x 1}; the product V^m s(m) is not defined (V^m is not specified, and the dimensions do not match a standard narrowband receive model). This is load-bearing because the entire derivation and the definition of the space-time snapshot depend on it. Please clarify whether s(m) is a scalar transmitted symbol, whether V^m denotes a column or a diagonal matrix, and how the transmit vector is combined with the steering matrix.
[IV-A] The lookup tables K_a and K_v are populated by simulating the same forward model that is later used to generate the test data. The coarse range and transverse-velocity estimates therefore measure agreement with a self-consistent dictionary, not with an independent closed-form mapping. The claim that angular-spread width is uniquely linked to range and Doppler-spread width to transverse velocity needs to be supported by an independent analytical derivation (e.g., a closed-form relation between the 3-dB spread and the parameters) or by validation on data generated from a different forward model. As it stands, the simulations demonstrate self-consistency rather than predictive accuracy of the proposed spectral relationships.
[IV-B] The 1D MUSIC refinement scans a single parameter-specific steering vector a(x) while keeping other parameters fixed. In the NF regime with Doppler spread, the spatial signature across the array can be multi-rank, and a single steering vector may not lie in the signal subspace. The paper does not bound or analyze the bias this introduces, and it is not obvious that sequential 1D scans over theta, r, vr, vtheta converge to the true parameters. Please provide a theoretical justification or a numerical subspace-distance analysis supporting the single-vector MUSIC scan in the multi-rank regime.

minor comments (7)

[III-A] The summation limits in Eq. (10) are written as sum_{n=-NBS/2}^{NBS/2}, which is off by one for even NBS; please correct to sum_{n=-(NBS-1)/2}^{(NBS-1)/2} or an equivalent integer range.
[III-B] The Doppler phase in Eq. (3) uses -j pi m omega^(n); a standard narrowband Doppler phase is -j 2 pi m omega^(n) (or +j 2 pi). Please resolve the factor-of-2 discrepancy or justify the definition of omega^(n).
[II-A2] The approximation v_r^(n) approximately v_r is described as neglecting O(1/r_F^2) terms, but it also discards the NF coupling of the radial velocity to the antenna index; please state explicitly that this is an additional modeling assumption beyond the Taylor expansion.
[V] The complexity expression O(NBS M (log NBS + log M)) + O((NBS M)^3 + G NBS M) is ambiguous: if the SVD is performed on the covariance R = Y Y^H in C^{NBS x NBS}, the cost should be O(NBS^3), not O((NBS M)^3). Please clarify the implementation and update the complexity comparison accordingly.
[IV-A] The lookup table description does not specify the grid sizes, the resolution of the angle/range/velocity grids, or the interpolation method used when matching measured spreads to table entries; please add these details for reproducibility.
[V] The NMSE definition averages over 1000 iterations, but reporting only NMSE can conceal bias; please also report per-parameter RMSE or bias to help interpret the -40 dB values.
[III-A] There are several typos, including 'the the system behavior' in Section III-A; please proofread the manuscript carefully.

Circularity Check

1 steps flagged · score 4.0 of 10

Range-estimation leg rests on a load-bearing self-citation to the authors' prior work [9]; the Doppler-spread and MUSIC stages are new/standard, so the circularity is partial.

self citation load bearing [Section III-A (Eqs. (10)-(12)); premise introduced in Section I; used in Section IV-A-2, Eq. (16)]
"As shown in [9], the true angle lies at the center of this spread, and its width is determined by the user's range. ... The width of Ω3dB is uniquely linked to the user's range rF — it is broader for users located closer to the array and narrows down as the range increases. In particular, noticeable angular spreading occurs only when the user is located within the effective beamfocusing Rayleigh distance (EBRD), characterized by rF < rRD cos2 θ/10 for a ULA [9]."

The coarse range/angle estimates in Eq. (14) and Eq. (16) invert exactly this cited spread-center/width relationship, but the relationship is not derived in this paper; it is imported from the authors' own prior work [9] (an arXiv preprint by the same group). Thus the 'analysis reveals' step for range reduces to a load-bearing self-citation rather than an independently established result. The Doppler-spread-to-velocity claim is new to this paper and the MUSIC refinement is standard, so the circularity is partial rather than total.

full rationale

The paper's central novelty is the Doppler-spread characterization and the two-stage DFT+MUSIC estimator; neither reduces to its inputs by construction. The lookup tables (Sec. IV-A-2) are forward-model inversions populated from simulations, not fits to the test targets, so I do not treat them as circular. The main circularity is the angular-spread-to-range relationship, which is load-bearing for the coarse range estimate but is justified only by citation to the authors' own prior work [9] rather than derived here. That is a self-citation chain for one leg of the algorithm, giving a score of 4. I note that the skeptical concern about Eq. (13) (spatial projection, vr phase cancelling in magnitude) is a correctness issue, not a circular-equivalence issue, and is not scored as circularity under the hard rules.

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

The method depends on the near-field spherical wavefront model, several approximations (second-order Taylor range profile, neglect of O(1/r^2) velocity terms), and a lookup-table dictionary generated from the same forward model. The 3-dB threshold, the lookup table grids, and the MUSIC grid resolution are free design choices that influence accuracy and complexity but are not specified. No new physical entities are introduced.

free parameters (4)

3-dB threshold factor = 0.5
Defines angular and Doppler spreads as the set of points above half the peak gain; chosen as a standard 3-dB width, not derived from an optimality criterion.
Lookup table grid resolution = not specified
The lookup tables Ka and Kv require a predefined grid over angles, ranges, and velocities; grid density is a design choice that trades accuracy against memory and is not given.
MUSIC fine grid resolution = not specified
The 1D MUSIC scans a fine grid for each parameter; the step size determines accuracy and complexity and is not specified.
EBRD boundary coefficient = rF < rRD cos^2(theta) / 10
Adopted from [9] to define the valid near-field region; the factor 10 is a heuristic boundary that is not independently justified.

assumptions (5)

domain assumption Spherical wavefront propagation model with a point target (Eq. 1)
The entire steering vector and spread analysis assumes exact spherical wavefronts from a point target at range r and angle theta.
domain assumption Second-order Taylor expansion of the range profile for analysis (Eq. 9)
Used to derive angular and Doppler spread expressions; higher-order terms are neglected and may matter for very close targets.
ad hoc to paper Neglect of O(1/r^2) terms in local velocity approximation (Section III-B)
Simplifies the Doppler spread derivation but may not hold for close targets or large array apertures.
domain assumption Single target with constant parameters during one CPI
Standard for radar sensing, but restricts applicability to a single moving target.
domain assumption Complex Gaussian noise and known transmitted symbols
Used in the likelihood formulation and MUSIC spectrum; if the transmitted symbols are unknown, the signal model changes.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Joint Motion, Angle, and Range Estimation in Near-Field under Array Calibration Imperfections." pith.science (2026). https://pith.science/paper/WEIMOOHB

@misc{pith2026250713463,
  author       = {Pith},
  title        = {Pith review of: Joint Motion, Angle, and Range Estimation in Near-Field under Array Calibration Imperfections},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WEIMOOHB}},
  note         = {Machine review of arXiv:2507.13463}
}

read the original abstract

Ultra-massive multiple-input multiple-output MIMO (UM-MIMO) leverages large antenna arrays at high frequencies, transitioning communication paradigm into the radiative near-field (NF), where spherical wavefronts enable full-vector estimation of both target location and velocity. However, location and motion parameters become inherently coupled in this regime, making their joint estimation computationally demanding. To overcome this, we propose a novel approach that projects the received two-dimensional space-time signal onto the angle-Doppler domain using a two-dimensional discrete Fourier transform (2D-DFT). Our analysis reveals that the resulting angular spread is centered at the target's true angle, with its width determined by the target's range. Similarly, transverse motion induces a Doppler spread centered at the true radial velocity, with the width of Doppler spread proportional to the transverse velocity. Exploiting these spectral characteristics, we develop a low-complexity algorithm that provides coarse estimates of angle, range, and velocity, which are subsequently refined using one-dimensional multiple signal classification (MUSIC) applied independently to each parameter. The proposed method enables accurate and efficient estimation of NF target motion parameters. Simulation results demonstrate a normalized mean squared error (NMSE) of -40 dB for location and velocity estimates compared to maximum likelihood estimation, while significantly reducing computational complexity.

Figures

Figures reproduced from arXiv: 2507.13463 by the authors.

**Figure 1.** NF system model. range. Similarly, transverse velocity induces a spread in the Doppler frequency across multiple bins, which we refer to as Doppler support. We demonstrate that the true radial velocity lies at the center of this Doppler spread, and that its width is directly related to the target’s transverse velocity. By leveraging these relationships, we develop a low-complexity algorithm that provides coarse esti… view at source ↗

**Figure 2.** Angle Doppler response for a NF target with parameters as [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

**Figure 5.** NMSE of estimated radial velocity vs. SNR [PITH_FULL_IMAGE:figures/full_fig_p005_5.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

11 extracted references · 11 canonical work pages

[6]

M. D. Sharp and J. A. Nanzer, ``Accuracy of angle rate measurements using a distributed radar with a correlation receiver,'' IEEE Antennas Wireless Propag. Lett., vol. 17, no. 2, pp. 209--212, 2018

work page 2018
[8]

Near-Field Sensing Enabled Predictive Beamforming: From Estimation to Tracking

H. Jiang, Z. Wang, and Y. Liu, ``Near-field sensing enabled predictive beamforming: From estimation to tracking,'' arXiv preprint arXiv:2408.02027, 2024

work page Pith review arXiv 2024
[1]

11em plus .33em minus .07em 4000 4000 100 4000 4000 500 `\.=1000 = #1 \@IEEEnotcompsoconly \@IEEEcompsoconly #1 * [1] 0pt [0pt][0pt] #1 * [1] 0pt [0pt][0pt] #1 * \| ** #1 \@IEEEauthorblockNstyle \@IEEEcompsocnotconfonly \@IEEEauthorblockAstyle \@IEEEcompsocnotconfonly \@IEEEcompsocconfonly \@IEEEauthordefaulttextstyle \@IEEEcompsocnotconfonly \@IEEEauthor...

work page
[2]

Jiang, B

W. Jiang, B. Han, M. A. Habibi, and H. D. Schotten, ``The road towards 6 G : A comprehensive survey,'' IEEE Open J. Commun. Soc., vol. 2, pp. 334--366, Feb. 2021

work page 2021
[3]

Near-Field ISAC: Synergy of Dual-Purpose Codebooks and Space-Time Adaptive Processing

A. Hussain, A. Abdallah, A. Celik, and A. M. Eltawil, ``Near-field ISAC : Synergy of dual-purpose codebooks and space-time adaptive processing,'' arXiv preprint arXiv:2501.08776, 2025

work page Pith review arXiv 2025
[4]

Abdallah, A

A. Abdallah, A. Hussain, A. Celik, and A. M. Eltawil, ``Exploring frontiers of polar-domain codebooks for near-field channel estimation and beam training: A comprehensive analysis, case studies, and implications for 6 G ,'' IEEE Signal Process. Mag., vol. 42, no. 1, pp. 45--59, 2025

work page 2025
[5]

J. A. Nanzer and M. D. Sharp, ``On the estimation of angle rate in radar,'' IEEE Trans. Antennas Propag., vol. 65, no. 3, pp. 1339--1348, 2017

work page 2017
[7]

Z. Wang, X. Mu, and Y. Liu, ``Near-field velocity sensing and predictive beamforming,'' IEEE Trans. Veh. Technol., vol. 74, no. 1, pp. 1806--1810, 2025

work page 2025

Show all 11 references

[9]

C. Meng, Z. Wang, Z. Wei, Y. Liu, and Z. Feng, ``Near-field motion parameter estimation: A variational bayesian approach,'' arXiv preprint arXiv:2502.14193, 2025

2025 arXiv
[10]

Hussain, A

A. Hussain, A. Abdallah, A. Celik, and A. M. Eltawil, ``Near-field beam prediction using far-field codebooks in ultra-massive MIMO systems,'' arXiv preprint arXiv:2503.14317, 2025

2025 arXiv
[11]

Hussain, A

A. Hussain, A. Abdallah, and A. M. Eltawil, ``Redefining polar boundaries for near-field channel estimation for ultra-massive MIMO antenna array,'' Submitted to IEEE Trans. Wireless Commun., 2024. [Online]. Available: https://repository.kaust.edu.sa/handle/10754/702050

2024

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.

[1] [6]

M. D. Sharp and J. A. Nanzer, ``Accuracy of angle rate measurements using a distributed radar with a correlation receiver,'' IEEE Antennas Wireless Propag. Lett., vol. 17, no. 2, pp. 209--212, 2018

work page 2018

[2] [8]

Near-Field Sensing Enabled Predictive Beamforming: From Estimation to Tracking

H. Jiang, Z. Wang, and Y. Liu, ``Near-field sensing enabled predictive beamforming: From estimation to tracking,'' arXiv preprint arXiv:2408.02027, 2024

work page Pith review arXiv 2024

[3] [1]

11em plus .33em minus .07em 4000 4000 100 4000 4000 500 `\.=1000 = #1 \@IEEEnotcompsoconly \@IEEEcompsoconly #1 * [1] 0pt [0pt][0pt] #1 * [1] 0pt [0pt][0pt] #1 * \| ** #1 \@IEEEauthorblockNstyle \@IEEEcompsocnotconfonly \@IEEEauthorblockAstyle \@IEEEcompsocnotconfonly \@IEEEcompsocconfonly \@IEEEauthordefaulttextstyle \@IEEEcompsocnotconfonly \@IEEEauthor...

work page

[4] [2]

Jiang, B

W. Jiang, B. Han, M. A. Habibi, and H. D. Schotten, ``The road towards 6 G : A comprehensive survey,'' IEEE Open J. Commun. Soc., vol. 2, pp. 334--366, Feb. 2021

work page 2021

[5] [3]

Near-Field ISAC: Synergy of Dual-Purpose Codebooks and Space-Time Adaptive Processing

A. Hussain, A. Abdallah, A. Celik, and A. M. Eltawil, ``Near-field ISAC : Synergy of dual-purpose codebooks and space-time adaptive processing,'' arXiv preprint arXiv:2501.08776, 2025

work page Pith review arXiv 2025

[6] [4]

Abdallah, A

A. Abdallah, A. Hussain, A. Celik, and A. M. Eltawil, ``Exploring frontiers of polar-domain codebooks for near-field channel estimation and beam training: A comprehensive analysis, case studies, and implications for 6 G ,'' IEEE Signal Process. Mag., vol. 42, no. 1, pp. 45--59, 2025

work page 2025

[7] [5]

J. A. Nanzer and M. D. Sharp, ``On the estimation of angle rate in radar,'' IEEE Trans. Antennas Propag., vol. 65, no. 3, pp. 1339--1348, 2017

work page 2017

[8] [7]

Z. Wang, X. Mu, and Y. Liu, ``Near-field velocity sensing and predictive beamforming,'' IEEE Trans. Veh. Technol., vol. 74, no. 1, pp. 1806--1810, 2025

work page 2025