Subspace Fitting Approach for Wideband Near-Field Localization

Ruiyun Zhang; Yuanwei Liu; Zehui Xiong; Zhaolin Wang; Zhiqing Wei; Zhiyong Feng

REVIEW 3 major objections 3 minor 12 references

Subspace Fitting Approach for Wideband Near-Field Localization

T0 review · 3 major / 3 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read Two subspace-fitting MUSIC algorithms localize multiple wideband emitters in the near field, one estimating distance and angle jointly and a Fresnel variant decoupling them for lower complexity.

desk verdict Plausible wideband near-field localization paper, but the supplied full text is unreadable mojibake, so nothing can be verified. read the letter →

arxiv 2508.04169 v1 pith:2IFD55PQ submitted 2025-08-06 eess.SP

Ruiyun Zhang , Zhaolin Wang , Zhiqing Wei , Yuanwei Liu , Zehui Xiong , Zhiyong Feng This is my paper · ORCID

classification eess.SP

keywords near-fieldlocalizationwidebandarrayprocessingsubspacefittingMUSICFresnelapproximationjointrange-angleestimationfrequency-domainsignalmodelmulti-target

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 tackles a localization problem: when a wideband emitter sits in the near field of an antenna array, its distance and angle are entangled in the curvature of the wavefront, so the far-field habit of estimating angle first and range second fails. The authors derive a frequency-domain signal model for multiple wideband near-field targets and fit its signal subspace with a MUSIC-style search that scores candidate distance-angle pairs together. Because that joint two-dimensional search is expensive, they add a second estimator that replaces the exact spherical wavefront with a Fresnel quadratic approximation, which lets distance and angle be searched separately. Numerical results are reported as confirming that both algorithms localize the targets, with the Fresnel version trading a modest accuracy loss for lower complexity.

What carries the argument

The engine is the frequency-domain near-field array steering vector, whose phase contains both the angle and the radial distance of each source through the spherical wavefront, combined with the MUSIC subspace-fitting criterion that searches this two-parameter manifold for the targets. In the reduced-complexity variant, the Fresnel approximation expands the spherical phase to quadratic order, making the array response separable into an angle-only factor and a distance-only factor; that separability is what converts the joint two-dimensional search into two one-dimensional searches. The model also collects observations across multiple frequency bins, so the wideband signal contributes additio

What would settle it

Run both estimators with a single wideband source placed progressively closer than the usual Fresnel boundary, say $r < 2D^2/\lambda$ where $D$ is the array aperture and $\lambda$ the carrier wavelength, and compare range estimates with the true positions. If the Fresnel-variant bias grows sharply while the joint estimator stays accurate, the decoupling premise is the limit; if both fail equally, the frequency-domain subspace model itself is at fault.

Watch

Extended reading notes

Core claim

The central claim is that wideband near-field localization can be cast as a subspace-fitting problem in the frequency domain, and that the resulting MUSIC estimator jointly recovers the range and angle of several targets. The paper asserts that, unlike far-field propagation, near-field propagation makes the array response depend on both the angle and the radial distance of each source, because the phase across the array follows a sphere rather than a plane. It therefore builds a multi-target wideband model in the frequency domain, forms a covariance across antennas and frequency bins, and searches over candidate ranges and angles for the set whose modeled steering vectors best occupy the sig

Load-bearing premise

For the low-complexity variant, the load-bearing premise is that the Fresnel shortcut represents the true curved wavefront faithfully across the whole tested distance range; for both variants, the number of targets must be known in advance so the signal and noise subspaces can be separated.

Editorial extensions

If this is right

Distance and angle can be recovered together for several wideband targets in the near field, which far-field estimators cannot do directly.
Wideband signals become an asset: each frequency bin supplies extra data to the subspace fit rather than requiring narrowband preprocessing.
The Fresnel variant offers a concrete complexity-accuracy trade-off: two one-dimensional searches instead of one two-dimensional search, at some accuracy loss.
The joint estimation principle applies to any array geometry whose steering vectors can be parameterized by range and angle.
The numerical results reported by the paper support the conclusion that both algorithms localize the targets under the tested conditions.

Reading between the lines

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

The paper leaves implicit an adaptive strategy: use the cheap Fresnel search for distant targets and switch to the joint search only when the estimated range enters the very near field, where the quadratic phase model loses accuracy.
Because the covariance is assembled from all frequency bins, a model-order selection step on its eigenvalues could remove the need to know the target count in advance, which the paper assumes is given.
The decoupling trick is really a separability condition on the steering manifold, so it may transfer to other wave-based localization settings—acoustic, ultrasound, or millimeter-wave—where spherical wavefronts couple the same two parameters.

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

Punchline: the abstract outlines a coherent, plausible contribution—wideband multi-target near-field localization using subspace fitting, with a MUSIC-type joint range-angle estimator plus a Fresnel-approximated decoupled variant for lower complexity. That is a natural extension of an established program, and the decoupling idea is the one piece that could turn out genuinely useful.

The problem framing is right: spherical wave propagation couples distance and angle, so a joint approach makes sense. The complexity-motivated Fresnel variant is the most interesting claim; if the approximation holds across a useful near-field region, it gives a real trade-off. Credit where due: the abstract is clear, modest, and does not oversell.

The problem: the supplied PDF body is almost entirely replacement characters. I can inspect the abstract, but not the signal model derivation, the subspace fitting formulation, the Fresnel expansion, the complexity analysis, or any simulation figure. The text even carries a watermark from a different arXiv ID. This is mechanical corruption, not evidence of author misconduct, but it makes the paper unauditable. I cannot confirm the derivations, the baselines, or the numerical results. The verification is presumably in-loop synthetic data from the same model, which is common in this literature, but I cannot check whether external baselines are included.

If a clean copy shows up, the main things I'd probe: (1) the range of validity of the Fresnel approximation—if targets sit in the extreme near field, the decoupled range estimates will be biased and the complexity-accuracy trade-off fails; (2) how the algorithm deals with unknown target count, which MUSIC-type subspace fitting usually needs; (3) whether the experiments compare against a credible baseline or just the paper's own algorithms.

Bottom line: plausible incremental contribution for the array-processing and 6G positioning audience. It deserves a serious referee once a readable manuscript is available. I would not cite it or bring it to reading group yet.

Referee Report

3 major / 3 minor

Summary. The manuscript (arXiv:2508.04169, eess.SP) proposes two wideband near-field localization approaches: a subspace-fitting MUSIC method that jointly estimates distance and angle from a derived frequency-domain spherical-wave model, and a reduced-complexity Fresnel-approximation MUSIC variant that decouples distance and angle. The abstract claims numerical verification of both. However, the supplied full text is almost entirely corrupted: the body consists of replacement-character mojibake, so the signal model, derivations, complexity analysis, and numerical results cannot be inspected. Only the abstract is readable.

Significance. If the derivations and simulations were accessible and correct, the contribution would be plausibly useful: wideband near-field localization with joint range/angle estimation is an active area, and a Fresnel-based decoupling variant is a reasonable complexity-reduction idea. The manuscript does not, in the supplied form, provide auditable support for these claims. No machine-checked proofs, reproducible code, or readable numerical results are present. Consequently, the significance cannot be assessed from the current file; the claims are plausible but unverified.

major comments (3)

[Full text (all sections after abstract)] The entire body of the manuscript is unreadable mojibake. Equations, the derived frequency-domain model, the subspace-fitting formulation, the Fresnel decoupling, and the simulation figures are inaccessible. This is a load-bearing issue: the central claims cannot be checked. I cannot verify the derivation or the numerical results.
[Embedded watermark] The supplied text contains the line 'arXiv:2508.04170v1 [eess.SY] 6 Aug 2025', which is a different arXiv identifier and different subject class from the claimed paper 'arXiv:2508.04169 (eess.SP)'. This provenance inconsistency must be resolved before review can proceed.
[Abstract, last sentence] The abstract states that 'Numerical results verify the effectiveness of both proposed approaches,' but the readable portion contains no error metrics, no comparison baselines, no SNR or distance ranges, and no number-of-target specifications. Because the body is corrupted, there is no way to assess whether the numerical verification is statistically meaningful or whether it is in-loop on the same model used by the algorithms.

minor comments (3)

[Abstract] The abstract should state whether the number of targets is assumed known and should specify the validity region of the Fresnel approximation. MUSIC-type subspace methods require the target count to separate signal and noise subspaces, and the Fresnel variant's accuracy depends on the tested near-field range.
[Abstract] No equations or symbol definitions are given in the abstract. This makes it hard to judge what exactly is being estimated (e.g., range vs. distance, angular coordinate convention).
[File metadata] The arXiv ID and subject class embedded in the file should be corrected to match the claimed submission: 2508.04169 and eess.SP.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity can be established: the manuscript body supplied for review is unreadable mojibake, and the readable abstract contains no equations or self-citation chain that would reduce a prediction to its inputs.

full rationale

The only readable portion of the manuscript is the abstract. The full text is corrupted mojibake, with replacement characters throughout and a watermark indicating a different arXiv identifier ('arXiv:2508.04170v1 [eess.SY] 6 Aug 2025'). Because no equations, derivations, algorithm statements, or numerical settings are legible, it is impossible to exhibit any specific reduction of a claimed result to a fitted parameter, a self-definition, or a self-citation chain. The abstract's claim—that a subspace-fitting MUSIC method jointly estimates distance and angle, and that a Fresnel-approximation variant decouples them—is not accompanied by any discernible derivation that could be checked for circularity. Testing an algorithm on synthetic data generated from the same model is common practice and is not, by itself, a circularity argument under the stated rules. No load-bearing self-citation, imported uniqueness theorem, or ansatz-smuggling citation can be identified because the reference list and citations are unreadable. Accordingly, the appropriate finding is no significant circularity, with the caveat that the manuscript text provided was not auditable in its current corrupted form.

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

All entries are inferred from the abstract's statements. The spherical-wave and Fresnel assumptions are the modeling backbone; the frequency-domain decomposition and known-target-count assumptions are standard but unstated prerequisites for the MUSIC-type machinery. None of these can be verified against the derivation because the body text is corrupted.

assumptions (5)

domain assumption Spherical wavefront propagation model with range-dependent phase that couples distance and angle
The entire approach rests on the near-field spherical wave model stated in the abstract: 'spherical wave propagation in near-field systems couples these parameters'.
domain assumption Frequency-domain decomposition of wideband signals into narrowband components
The abstract says a 'frequency-domain near-field signal model for multi-target wideband systems' is derived; this assumes a valid frequency-domain representation, typically via DFT, which the abstract does not justify.
domain assumption Fresnel expansion is an accurate model of the spherical wavefront over the tested near-field localization range
The low-complexity algorithm 'decouples the distance and angle parameters' via a Fresnel approximation; the accuracy of the decoupling depends on the truncated expansion being valid over the tested range.
domain assumption Number of targets is known or correctly estimated so the noise subspace can be separated
MUSIC-based subspace fitting requires the number of sources to estimate the noise subspace; the abstract does not state how this is handled for multi-target systems.
standard math Signal and noise subspaces are orthogonal under the sample covariance eigendecomposition
Underlies any MUSIC-type subspace fitting method; assumed but not stated in the abstract.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Subspace Fitting Approach for Wideband Near-Field Localization." pith.science (2026). https://pith.science/paper/2IFD55PQ

@misc{pith2026250804169,
  author       = {Pith},
  title        = {Pith review of: Subspace Fitting Approach for Wideband Near-Field Localization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2IFD55PQ}},
  note         = {Machine review of arXiv:2508.04169}
}

read the original abstract

Two subspace fitting approaches are proposed for wideband near-field localization. Unlike in conventional far-field systems, where distance and angle can be estimated separately, spherical wave propagation in near-field systems couples these parameters. We therefore derive a frequency-domain near-field signal model for multi-target wideband systems and develop a subspace fitting-based MUSIC method that jointly estimates distance and angle. To reduce complexity, a Fresnel approximation MUSIC algorithm is further introduced to decouple the distance and angle parameters. Numerical results verify the effectiveness of both proposed approaches.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

12 extracted references · 11 canonical work pages

[1]

Bj \"o rnson, Y

E. Bj \"o rnson, Y. C. Eldar, E. G. Larsson, A. Lozano, and H. V. Poor, ``Twenty-five years of signal processing advances for multiantenna communications: From theory to mainstream technology,'' IEEE Signal Process. Mag. , vol. 40, no. 4, pp. 107--117, 2023

work page 2023
[2]

F. Liu, Y. Cui, C. Masouros, J. Xu, T. X. Han, Y. C. Eldar, and S. Buzzi, ``Integrated sensing and communications: Toward dual-functional wireless networks for 6 G and beyond,'' IEEE J. Sel. Areas Commun. , vol. 40, no. 6, pp. 1728--1767, 2022

work page 2022
[3]

Z. Wang, J. Zhang, H. Du, D. Niyato, S. Cui, B. Ai, M. Debbah, K. B. Letaief, and H. V. Poor, ``A tutorial on extremely large-scale MIMO for 6 G : Fundamentals, signal processing, and applications,'' IEEE Commun. Surv. Tutor., vol. 26, no. 3, pp. 1560--1605, 2024

work page 2024
[4]

Z. Wang, X. Mu, and Y. Liu, ``Performance analysis of near-field sensing in wideband MIMO systems,'' 2025. [Online]. Available: https://arxiv.org/abs/2404.05076

arXiv 2025
[5]

Wang and K

Y. Wang and K. Ho, ``Unified near-field and far-field localization for AOA and hybrid AOA-TDOA positionings,'' IEEE Trans. Wireless Commun. , vol. 17, no. 2, pp. 1242--1254, 2017

work page 2017
[6]

X. Chen, Z. Feng, Z. Wei, X. Yuan, P. Zhang, J. A. Zhang, and H. Yang, ``Multiple signal classification based joint communication and sensing system,'' IEEE Trans. Wireless Commun. , vol. 22, no. 10, pp. 6504--6517, 2023

work page 2023
[7]

Cui and L

M. Cui and L. Dai, ``Channel estimation for extremely large-scale MIMO : Far-field or near-field?'' IEEE Trans. Commun. , vol. 70, no. 4, pp. 2663--2677, 2022

work page 2022
[8]

Z. Wang, X. Mu, and Y. Liu, ``Rethinking integrated sensing and communication: When near field meets wideband,'' IEEE Communications Magazine, vol. 62, no. 9, pp. 44--50, 2024

work page 2024

Show all 12 references

[9]

H. Luo, F. Gao, W. Yuan, and S. Zhang, ``Beam squint assisted user localization in near-field integrated sensing and communications systems,'' IEEE Trans. Wireless Commun. , vol. 23, no. 5, pp. 4504--4517, 2024

2024
[10]

Z. Wang, P. Ramezani, Y. Liu, and E. Björnson, ``Near-field localization and sensing with large-aperture arrays: From signal modeling to processing,'' IEEE Signal Process. Mag. , vol. 42, no. 1, pp. 74--87, 2025

2025
[11]

Bohme, ``Estimation of source parameters by maximum likelihood and nonlinear regression,'' in ICASSP IEEE Int

J. Bohme, ``Estimation of source parameters by maximum likelihood and nonlinear regression,'' in ICASSP IEEE Int. Conf. Acoust. Speech Signal Process., vol. 9. 1em plus 0.5em minus 0.4em IEEE, Mar. 1984, pp. 271--274

1984
[12]

Pesavento, M

M. Pesavento, M. Trinh-Hoang, and M. Viberg, ``Three more decades in array signal processing research: An optimization and structure exploitation perspective,'' IEEE Signal Process. Mag. , vol. 40, no. 4, pp. 92--106, 2023

2023

Pith tools

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

[1] [1]

Bj \"o rnson, Y

E. Bj \"o rnson, Y. C. Eldar, E. G. Larsson, A. Lozano, and H. V. Poor, ``Twenty-five years of signal processing advances for multiantenna communications: From theory to mainstream technology,'' IEEE Signal Process. Mag. , vol. 40, no. 4, pp. 107--117, 2023

work page 2023

[2] [2]

F. Liu, Y. Cui, C. Masouros, J. Xu, T. X. Han, Y. C. Eldar, and S. Buzzi, ``Integrated sensing and communications: Toward dual-functional wireless networks for 6 G and beyond,'' IEEE J. Sel. Areas Commun. , vol. 40, no. 6, pp. 1728--1767, 2022

work page 2022

[3] [3]

Z. Wang, J. Zhang, H. Du, D. Niyato, S. Cui, B. Ai, M. Debbah, K. B. Letaief, and H. V. Poor, ``A tutorial on extremely large-scale MIMO for 6 G : Fundamentals, signal processing, and applications,'' IEEE Commun. Surv. Tutor., vol. 26, no. 3, pp. 1560--1605, 2024

work page 2024

[4] [4]

Z. Wang, X. Mu, and Y. Liu, ``Performance analysis of near-field sensing in wideband MIMO systems,'' 2025. [Online]. Available: https://arxiv.org/abs/2404.05076

arXiv 2025

[5] [5]

Wang and K

Y. Wang and K. Ho, ``Unified near-field and far-field localization for AOA and hybrid AOA-TDOA positionings,'' IEEE Trans. Wireless Commun. , vol. 17, no. 2, pp. 1242--1254, 2017

work page 2017

[6] [6]

X. Chen, Z. Feng, Z. Wei, X. Yuan, P. Zhang, J. A. Zhang, and H. Yang, ``Multiple signal classification based joint communication and sensing system,'' IEEE Trans. Wireless Commun. , vol. 22, no. 10, pp. 6504--6517, 2023

work page 2023

[7] [7]

Cui and L

M. Cui and L. Dai, ``Channel estimation for extremely large-scale MIMO : Far-field or near-field?'' IEEE Trans. Commun. , vol. 70, no. 4, pp. 2663--2677, 2022

work page 2022

[8] [8]

Z. Wang, X. Mu, and Y. Liu, ``Rethinking integrated sensing and communication: When near field meets wideband,'' IEEE Communications Magazine, vol. 62, no. 9, pp. 44--50, 2024

work page 2024