Pith. sign in

REVIEW 2 major objections 2 minor 36 references

G-iMUSIC: Greedy Iterative MUSIC Algorithms for Multi-Target DoA Estimation

T0 review · 2 major / 2 minor · reviewed 2026-06-29 · grok-4.3

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

Pith's one-line read G-iMUSIC algorithms combine greedy selection with subspace methods using a single initial eigenvalue decomposition for multi-target DoA estimation.

desk verdict G-iMUSIC links greedy and subspace methods to cut iMUSIC down to one EVD with FFT options for ULAs, and the simulations claim gains, but the correlation handling is the part to verify first. read the letter →

arxiv 2605.26875 v2 pith:5K52BALG submitted 2026-05-26 eess.SP

Martin Willame , Gilles Monnoyer , François Horlin , Jérôme Louveaux This is my paper · ORCID
classification eess.SP
keywords DoAestimationMUSICgreedyalgorithmsOMPOLSarraysignalprocessingeigenvaluedecompositionmulti-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 develops OMP-iMUSIC and OLS-iMUSIC to estimate directions of arrival for several targets at once. It creates a unified approach that joins the iterative selection of greedy algorithms with the angular resolution of MUSIC while needing only one eigenvalue decomposition at the start. This removes the need to repeat decompositions in each step and supports fast FFT versions on uniform linear arrays. Monte Carlo tests show higher detection rates and better location accuracy than plain OMP, OLS, or MUSIC, particularly when targets are close in angle or their signals are correlated.

What carries the argument

The unified framework linking subspace and greedy estimations, which permits a single initial eigenvalue decomposition to support the entire iterative selection process without recomputation.

What would settle it

Monte Carlo trials on an OFDM radar array in which OMP-iMUSIC or OLS-iMUSIC show no gain in detection probability or root-mean-square error over standard MUSIC when two targets lie within one beamwidth and their signals are fully correlated.

Watch

Extended reading notes

Core claim

The proposed G-iMUSIC algorithms, namely OMP-iMUSIC and OLS-iMUSIC, are derived from a unified framework that links subspace and greedy estimations; unlike prior iterative MUSIC methods they require only one initial EVD, avoid eigendecomposition at each iteration, admit FFT-accelerated implementations for ULAs, and yield improved detection and precision over conventional OMP, OLS, and MUSIC in Monte Carlo simulations across signal correlation and angular proximity regimes.

Load-bearing premise

A single unified framework can connect subspace and greedy methods so that one initial eigenvalue decomposition remains sufficient and accurate even when targets are angularly close or their signals are strongly correlated.

Editorial extensions

If this is right

  • Only one eigenvalue decomposition is performed instead of one per iteration.
  • FFT acceleration becomes available for uniform linear arrays, lowering computational cost.
  • Detection and localization improve relative to standalone OMP, OLS, or MUSIC for correlated or proximate targets.
  • Diagnostic metrics separate performance across correlation and angular-separation regimes.

Reading between the lines

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

  • The single-EVD property may extend the approach to arrays other than uniform linear ones if an equivalent fast transform exists.
  • The diagnostic metrics could serve as a general tool to predict when greedy-subspace hybrids outperform pure methods in other estimation tasks.
  • Reduced per-iteration cost could enable real-time operation in dynamic scenarios where targets move between snapshots.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

2 major / 2 minor

Summary. The paper proposes two G-iMUSIC algorithms (OMP-iMUSIC and OLS-iMUSIC) for multi-target DoA estimation. These are derived from a unified framework linking greedy selection (OMP/OLS) with subspace methods (MUSIC), requiring only a single initial EVD rather than per-iteration decompositions, admitting FFT acceleration for ULAs, and claiming superior detection/precision over standard OMP, OLS, and MUSIC in Monte Carlo trials, plus diagnostic metrics for correlation and angular-proximity regimes.

Significance. If the central claims hold, the work provides a computationally attractive bridge between greedy and subspace estimators that avoids repeated EVDs while retaining super-resolution capability. The FFT implementation path and diagnostic metrics for correlation/proximity regimes are concrete strengths that could aid practical deployment in OFDM radar and similar array-processing settings.

major comments (2)
  1. [§3] §3 (unified framework derivation): the assertion that a single initial noise-subspace projector suffices for all greedy iterations without recomputation must be supported by an explicit error bound or invariance condition when the signal covariance is rank-deficient (high correlation) or when sources are proximate; the skeptic concern about error propagation in the selection step is load-bearing for the complexity-reduction claim.
  2. [§4] §4 (Monte Carlo results): the reported gains in detection and precision over OMP/OLS/MUSIC are central, yet the text must specify the number of trials, exact SNR and snapshot counts, angular-separation grid, and correlation-coefficient sweep so that the single-EVD claim can be verified against the regimes where subspace leakage is expected.
minor comments (2)
  1. Notation for the initial EVD and the reused projector should be introduced once with a clear equation reference rather than repeated descriptive phrases.
  2. The diagnostic metrics are introduced in the abstract but their definitions and interpretation should appear in a dedicated subsection with explicit formulas.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments, which help clarify the presentation of the unified framework and simulation details. We respond to each major comment below.

read point-by-point responses

  1. Referee: [§3] §3 (unified framework derivation): the assertion that a single initial noise-subspace projector suffices for all greedy iterations without recomputation must be supported by an explicit error bound or invariance condition when the signal covariance is rank-deficient (high correlation) or when sources are proximate; the skeptic concern about error propagation in the selection step is load-bearing for the complexity-reduction claim.

    Authors: Section 3 derives the fixed noise-subspace projector from the initial EVD under the standard array model, with greedy steps updating only the signal component. This yields the claimed single-EVD property. We agree an explicit invariance condition or error bound for rank-deficient covariance and proximate sources is not derived in the current text. We will add a brief discussion of the approximation's validity range and error-propagation considerations in the revised §3. revision: partial

  2. Referee: [§4] §4 (Monte Carlo results): the reported gains in detection and precision over OMP/OLS/MUSIC are central, yet the text must specify the number of trials, exact SNR and snapshot counts, angular-separation grid, and correlation-coefficient sweep so that the single-EVD claim can be verified against the regimes where subspace leakage is expected.

    Authors: We will revise the Monte Carlo section to list the exact number of trials, SNR range, snapshot count, angular-separation values, and correlation-coefficient sweep used in the experiments. revision: yes

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: derivation links existing methods without self-referential reduction

full rationale

The paper presents G-iMUSIC (OMP-iMUSIC and OLS-iMUSIC) as derived from a unified framework that links subspace-based (MUSIC) and greedy (OMP/OLS) estimations. The abstract explicitly states the algorithms 'require only one initial EVD' as a consequence of this linking, not as a definitional input. No equations or steps in the provided text reduce a claimed result to a fitted parameter or self-citation by construction. Prior iMUSIC variants are referenced only for contrast, not as load-bearing justification. Monte Carlo simulations serve as external empirical check. The derivation chain remains self-contained against external benchmarks with no quoted reduction of outputs to inputs.

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

Abstract-only review; no explicit free parameters, axioms, or invented entities are identifiable from the provided text.

how reviews work

0 comments
Cite this review

Pith. "Pith review of G-iMUSIC: Greedy Iterative MUSIC Algorithms for Multi-Target DoA Estimation." pith.science (2026). https://pith.science/paper/5K52BALG

@misc{pith2026260526875,
  author       = {Pith},
  title        = {Pith review of: G-iMUSIC: Greedy Iterative MUSIC Algorithms for Multi-Target DoA Estimation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5K52BALG}},
  note         = {Machine review of arXiv:2605.26875}
}
read the original abstract

This paper presents novel algorithms for multi-target direction-of-arrival (DoA) estimation in array signal processing. Although the maximum likelihood estimator (MLE) asymptotically attains the Cram\'er-Rao bound, its exponential complexity motivates practical alternatives, such as greedy or subspace-based methods. In this context, greedy methods such as orthogonal matching pursuit (OMP) and orthogonal least squares (OLS) are sensitive to early selection errors, especially for angularly proximate targets, whereas subspace-based methods such as multiple signal classification (MUSIC) present angular super-resolution capabilities but degrade under strong inter-target signal correlation. To overcome these limitations, we propose two greedy iterative MUSIC (G-iMUSIC) algorithms, namely OMP-iMUSIC and OLS-iMUSIC, derived from a unified framework that links subspace and greedy estimations. Unlike prior iMUSIC approaches, the proposed methods require only one initial eigen value decomposition (EVD) and avoid computing eigendecomposition at each iteration. They also admit Fast Fourier Transform (FFT)-accelerated implementations for uniform linear arrays (ULAs), enabling low-complexity operation. Monte Carlo simulations demonstrate improved detection and precision over conventional OMP, OLS, and MUSIC, as well as reduced processing time compared to greedy baselines. Finally, we introduce diagnostic metrics that interpret performance across signal correlation and angular proximity regimes, supporting generalization beyond the specific orthogonal frequency-division multiplexing (OFDM) radar scenario considered.

Figures

Figures reproduced from arXiv: 2605.26875 by the authors.

Figure 1
Figure 1. Youden J statistic versus (a) SNR, (b) number of targets [PITH_FULL_IMAGE:figures/full_fig_p010_1.png] view at source ↗
Figure 2
Figure 2. RMSE versus (a) SNR, (b) number of targets [PITH_FULL_IMAGE:figures/full_fig_p010_2.png] view at source ↗
Figure 3
Figure 3. Mean processing time versus (a) number of targets [PITH_FULL_IMAGE:figures/full_fig_p011_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

36 extracted references · 4 canonical work pages

  1. [1]

    Three More Decades in Array Signal Processing Research: An optimization and structure exploitation perspective,

    M. Pesavento, M. Trinh-Hoang, and M. Viberg, “Three More Decades in Array Signal Processing Research: An optimization and structure exploitation perspective,”IEEE Signal Processing Magazine, vol. 40, no. 4, pp. 92–106, Jun. 2023

    2023

  2. [2]

    Direction of arrival estimation: A tutorial survey of classical and modern methods,

    A. A. Salama, “Direction of arrival estimation: A tutorial survey of classical and modern methods,” 2025. [Online]. Available: https://arxiv.org/abs/2508.11675

    work page arXiv 2025

  3. [3]

    Maximum-likelihood direction-of- arrival estimation in the presence of unknown nonuniform noise,

    M. Pesavento and A. Gershman, “Maximum-likelihood direction-of- arrival estimation in the presence of unknown nonuniform noise,”IEEE Transactions on Signal Processing, vol. 49, no. 7, pp. 1310–1324, 2001

    2001

  4. [4]

    M. K. B. Jeffrey Foutz, Andreas Spanias,Narrowband Direction of Arrival Estimation for Antenna Arrays. Springer Cham, 2008

    2008

  5. [5]

    Orthogonal matching pursuit: recursive function approximation with applications to wavelet decomposition,

    Y . Pati, R. Rezaiifar, and P. Krishnaprasad, “Orthogonal matching pursuit: recursive function approximation with applications to wavelet decomposition,” inProceedings of 27th Asilomar Conference on Signals, Systems and Computers, Nov. 1993, pp. 40–44 vol.1, iSSN: 1058-6393

    1993

  6. [6]

    Greed is good: algorithmic results for sparse approximation,

    J. Tropp, “Greed is good: algorithmic results for sparse approximation,” IEEE Transactions on Information Theory, vol. 50, no. 10, pp. 2231– 2242, Oct. 2004

    2004

  7. [7]

    Orthogonal least squares methods and their application to non-linear system identification,

    S. Chen, S. A. Billings, and W. Luo, “Orthogonal least squares methods and their application to non-linear system identification,”International Journal of Control, vol. 50, no. 5, pp. 1873–1896, Nov. 1989, publisher: Taylor & Francis eprint: https://doi.org/10.1080/00207178908953472

    work page doi:10.1080/00207178908953472 1989

  8. [8]

    On the Difference Between Orthogonal Matching Pursuit and Orthogonal Least Squares,

    T. Blumensath and M. Davies, “On the Difference Between Orthogonal Matching Pursuit and Orthogonal Least Squares,” Mar. 2007

    2007

Show all 36 references
  1. [9]

    Sparse de- convolution: Comparison of statistical and deterministic approaches,

    S. Bourguignon, C. Soussen, H. Carfantan, and J. Idier, “Sparse de- convolution: Comparison of statistical and deterministic approaches,” in 2011 IEEE Statistical Signal Processing Workshop (SSP), Jun. 2011, pp. 317–320, iSSN: 2373-0803

    2011

  2. [10]

    Orthogonal Least Squares Detector for Generalized Spatial Modulation,

    J. Wen, J. Li, H. Ge, Z. Zhou, and W. Luo, “Orthogonal Least Squares Detector for Generalized Spatial Modulation,”IEEE Transactions on Wireless Communications, vol. 20, no. 8, pp. 5071–5082, Aug. 2021

    2021

  3. [11]

    Signal Recovery in Uncorrelated and Correlated Dictionaries Using Orthogonal Least Squares,

    S. Mukhopadhyay, P. V . and, and M. Chakraborty, “Signal Recovery in Uncorrelated and Correlated Dictionaries Using Orthogonal Least Squares,” Jul. 2016, arXiv:1607.08712 [cs]

    2016 arXiv

  4. [12]

    Joint K-Step Analysis of Orthogonal Matching Pursuit and Orthogonal Least Squares,

    C. Soussen, R. Gribonval, J. Idier, and C. Herzet, “Joint K-Step Analysis of Orthogonal Matching Pursuit and Orthogonal Least Squares,”IEEE Transactions on Information Theory, vol. 59, no. 5, pp. 3158–3174, May 2013

    2013

  5. [13]

    Multi target localization with block orthogonal least squares for multi- static mimo radars,

    M. Willame, G. Monnoyer, H. C. Yildirim, F. Horlin, and J. Louveaux, “Multi target localization with block orthogonal least squares for multi- static mimo radars,”IEEE Signal Processing Letters, vol. 32, pp. 1990– 1994, 2025

    1990

  6. [14]

    How to beat the Rayleigh resolution limit: A lecture demonstration,

    W. Rueckner and P. Costas, “How to beat the Rayleigh resolution limit: A lecture demonstration,”American Journal of Physics, vol. 70, no. 6, pp. 587–594, 2002

    2002

  7. [15]

    Multiple emitter location and signal parameter estimation,

    R. Schmidt, “Multiple emitter location and signal parameter estimation,” IEEE Transactions on Antennas and Propagation, vol. 34, no. 3, pp. 276–280, 1986

    1986

  8. [16]

    MUSIC, maximum likelihood, and cramer- rao bound: further results and comparisons,

    P. Stoica and A. Nehorai, “MUSIC, maximum likelihood, and cramer- rao bound: further results and comparisons,”IEEE Transactions on Acoustics, Speech, and Signal Processing, vol. 38, no. 12, pp. 2140– 2150, 1990

    1990

  9. [17]

    MUSIC and improved MUSIC algorithm to estimate direction of arrival,

    P. Gupta and S. Kar, “MUSIC and improved MUSIC algorithm to estimate direction of arrival,” in2015 International Conference on Communications and Signal Processing (ICCSP), 2015, pp. 0757–0761

    2015

  10. [18]

    Multistatic OFDM radar fusion of MUSIC-based angle estimation,

    M. Willame, H. C. Yildirim, L. Storrer, F. Horlin, and J. Louveaux, “Multistatic OFDM radar fusion of MUSIC-based angle estimation,” in2024 18th European Conference on Antennas and Propagation (EuCAP), 2024, pp. 1–5

    2024

  11. [19]

    Iterative MUSIC: Coherent signal estimation, performance analysis,

    Z. Tian, “Iterative MUSIC: Coherent signal estimation, performance analysis,” in2002 IEEE International Conference on Acoustics, Speech, and Signal Processing, vol. 3, 2002, pp. III–3041–III–3044

    2002

  12. [20]

    Iterative MUSIC for single-look SAR tomography,

    N. Haddad, L. Harkati, K. Hadj-Rabah, A. Bouaraba, and A. Mesloub, “Iterative MUSIC for single-look SAR tomography,” in2024 IEEE Mediterranean and Middle-East Geoscience and Remote Sensing Sym- posium (M2GARSS), 2024, pp. 201–204

    2024

  13. [21]

    Iterative root-MUSIC algorithm for DoA estimation,

    M. Shaghaghi and S. A. V orobyov, “Iterative root-MUSIC algorithm for DoA estimation,” in2013 5th IEEE International Workshop on Com- putational Advances in Multi-Sensor Adaptive Processing (CAMSAP), 2013, pp. 53–56

    2013

  14. [22]

    DoA estimation with a combination of iterative MUSIC and SBL in low snr conditions,

    M. Zhou, H. Tian, and S. Fan, “DoA estimation with a combination of iterative MUSIC and SBL in low snr conditions,” in2021 IEEE Wireless Communications and Networking Conference Workshops (WCNCW), 2021, pp. 1–6

    2021

  15. [23]

    An efficient MUSIC algorithm enhanced by iteratively estimating signal subspace and its applications in spatial colored noise,

    X. Zhang and D. Feng, “An efficient MUSIC algorithm enhanced by iteratively estimating signal subspace and its applications in spatial colored noise,”Remote Sensing, vol. 14, no. 17, 2022. [Online]. Available: https://www.mdpi.com/2072-4292/14/17/4260

    2022

  16. [24]

    Indoor tracking of multiple individuals with an 802.11ax Wi-Fi-based multi-antenna passive radar,

    L. Storrer, H. C. Yildirim, M. Crauwels, E. I. P. Copa, S. Pollin, J. Louveaux, P. De Doncker, and F. Horlin, “Indoor tracking of multiple individuals with an 802.11ax Wi-Fi-based multi-antenna passive radar,” IEEE Sensors Journal, vol. 21, no. 18, pp. 20 462–20 474, 2021

    2021

  17. [25]

    Maximum likelihood localization of multiple sources by alternating projection,

    I. Ziskind and M. Wax, “Maximum likelihood localization of multiple sources by alternating projection,”IEEE Transactions on Acoustics, Speech, and Signal Processing, vol. 36, no. 10, pp. 1553–1560, 1988

    1988

  18. [26]

    Principles of modern radar: Basic principles,

    M. A. Richards, J. A. Scheer, and W. A. Holm, “Principles of modern radar: Basic principles,” 2013. [Online]. Available: https://api.semanticscholar.org/CorpusID:114114032

    2013

  19. [27]

    Floating point operations in matrix-vector calculus

    R. Hunger, “Floating point operations in matrix-vector calculus.” Tech. Rep., October 2005. [Online]. Available: https://mediatum.ub.tum.de/ doc/625604/625604

    2005

  20. [28]

    Golub and C

    G. Golub and C. Van Loan,Matrix Computations, ser. Johns Hopkins Studies in the Mathematical Sciences. Johns Hopkins University Press, 2013. [Online]. Available: https://books.google.be/books?id= X5YfsuCWpxMC

    2013

  21. [29]

    Direct FFT oversampling without zero-padding,

    C. Donciu, M. C. Temneanu, and E. Serea, “Direct FFT oversampling without zero-padding,”Scientific Reports, vol. 15, 2025

    2025

  22. [30]

    1–1020, 2025

    “IEEE standard for information technology–telecommunications and information exchange between systems local and metropolitan area networks–specific requirements - part 11: Wireless LAN medium access control (MAC) and physical layer (PHY) specifications amendment 2: Enhancement...

    2024

  23. [31]

    Detection of signals by information theoretic criteria,

    M. Wax and T. Kailath, “Detection of signals by information theoretic criteria,”IEEE Transactions on Acoustics, Speech, and Signal Process- ing, vol. 33, no. 2, pp. 387–392, 1985

    1985

  24. [32]

    Orthogonal least squares with integrated information theoretic criteria for joint number of targets and doa estimation,

    M. Willame, G. Monnoyer, F. Horlin, and J. Louveaux, “Orthogonal least squares with integrated information theoretic criteria for joint number of targets and doa estimation,” 2026. [Online]. Available: https://arxiv.org/abs/2605.06198

    2026 arXiv

  25. [33]

    Diagonality measures of hermitian positive-definite matrices with application to the approximate joint diagonalization problem,

    K. Alyani, M. Congedo, and M. Moakher, “Diagonality measures of hermitian positive-definite matrices with application to the approximate joint diagonalization problem,”Linear Algebra and its Applications, vol. 528, pp. 290–320, 2017, special Issue Dedicated to Rajendra Bhatia....

    2017

  26. [34]

    H. W. Kuhn,The Hungarian Method for the Assignment Problem. Berlin, Heidelberg: Springer Berlin Heidelberg, 2010, pp. 29–47

    2010

  27. [35]

    Index for rating diagnostic tests,

    W. J. Youden, “Index for rating diagnostic tests,”Cancer, vol. 3, no. 1, pp. 32–35, 1950. SUBMITTED TO IEEE TRANSACTIONS ON SIGNAL PROCESSING 13

    1950

  28. [36]

    The youden index in the generalized receiver operating characteristic curve context,

    P. Mart ´ınez-Camblor and J. C. Pardo-Fern ´andez, “The youden index in the generalized receiver operating characteristic curve context,”Interna- tional Journal of Biostatistics, vol. 15, 2019

    2019

Pith tools

Reviewed June 29, 2026 · model on record in the stance chip above.