pith. sign in

arxiv: 2604.26759 · v1 · submitted 2026-04-29 · 📡 eess.SP

A New Location Estimator for Mixed LOS & NLOS scenarios

Gaurav Duggal , Richard M. Buehrer , Harpreet S. Dhillon , Jeffrey H. Reed This is my paper

Pith reviewed 2026-05-07 11:20 UTC · model grok-4.3

classification 📡 eess.SP
keywords TOA localizationmixed LOS NLOSdiffraction path modelmaximum likelihood estimationsemidefinite relaxationGauss-NewtonCramér-Rao boundvirtual anchor
0
0 comments X

The pith

A diffraction path-length model unifies mixed LOS/NLOS TOA localization and supports efficient estimators that reach near-Cramér-Rao performance.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper shows that replacing the usual straight-line distance with a diffraction path-length model produces a single geometric description that covers both clear line-of-sight paths and obstructed paths without any separate classification step. This unified model turns localization into a non-convex maximum-likelihood problem whose direct solution by Gauss-Newton often lands in poor local minima. The authors therefore build specialized estimators that exploit the model's structure: a virtual-anchor reduction for the two-dimensional case with known height, surrogate and semidefinite-relaxation variants that trade speed against fidelity, and a sample-polish-select procedure that collapses the three-dimensional search to one dimension before local refinement. These estimators deliver accuracy close to the theoretical Cramér-Rao lower bound while using far less computation than multistart Gauss-Newton and remaining reliable even with bad initial guesses.

Core claim

By adopting the diffraction path-length model the localization task becomes a single non-convex maximum-likelihood problem that admits structure-exploiting solvers; for known target height the model yields a virtual-anchor geometry, enabling surrogate and semidefinite-relaxation estimators, while the full three-dimensional case is handled by sample-polish-select procedures that reduce the global search to one dimension, solve fixed-height subproblems, and apply local nonlinear refinement; the resulting estimators attain performance near the Cramér-Rao lower bound with substantially lower complexity than multistart Gauss-Newton and greater robustness to initialization than a direct single-run

What carries the argument

The diffraction path-length model, which replaces Euclidean distance with a continuous path length that accounts for diffraction around obstacles and transitions smoothly between LOS and NLOS regimes; it induces a virtual-anchor representation of the reduced two-dimensional problem.

If this is right

  • For known target height the virtual-anchor geometry yields surrogate and semidefinite-relaxation formulations that occupy different points on the complexity-performance curve.
  • The sample-polish-select procedure reduces the full three-dimensional search to a one-dimensional line search followed by fixed-height two-dimensional solves and local polishing.
  • The estimators remain close to the Cramér-Rao lower bound across varying fractions of LOS and NLOS paths.
  • Robustness to poor initialization exceeds that of a direct single-start Gauss-Newton solver.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Receivers could perform TOA localization without a separate LOS/NLOS classifier or extra hardware for path identification.
  • The same structural reduction may apply to other ranging modalities such as time-difference-of-arrival when diffraction dominates.
  • Field tests in environments with known obstacle geometry would directly compare the model's predicted path lengths against measured delays.

Load-bearing premise

The diffraction path-length model accurately predicts the actual propagation distances that occur in both clear and obstructed conditions.

What would settle it

In a simulation or measurement campaign where the true time-of-arrival values are generated from the diffraction model, compute the root-mean-square position error of the proposed estimators and verify whether it remains within a small multiple of the Cramér-Rao bound while a single-start Gauss-Newton run from the same initialization produces visibly larger errors.

Figures

Figures reproduced from arXiv: 2604.26759 by Gaurav Duggal, Harpreet S. Dhillon, Jeffrey H. Reed, Richard M. Buehrer.

Figure 1
Figure 1. Figure 1: In the O2I scenario, we have K anchors transmitting orthog￾onal signals which are received by the n th target inside the building. The received signal includes several MPCs, from which the ranging measurement corresponding to the diffraction path length AkQeNn is extracted. A bandwidth of 200 MHz is assumed, ensuring that all MPCs are resolvable. path-length model varies continuously as the target transiti… view at source ↗
Figure 2
Figure 2. Figure 2: Monte Carlo simulation results for 2D localization with known target height z, showing the average 2D-RMSE/2D-PEB vs SNR over anchor sets with K = 6 anchors, 64 indoor target locations, and 100 noise realizations. (a) The 2D-GTRS estimator asymptotically approaches the 2D-PEB for SNR > 5 dB. (b) The 2D-SDP estimator provides the next best performance, while (c) The 2D-USR formulation performs the worst. C.… view at source ↗
Figure 3
Figure 3. Figure 3: Maximum-likelihood R-LS objective as a function of the target height z, referred to as the z-profile. The approximate z￾profiles are obtained by replacing the underlying 2D solve with the corresponding GTRS-, USR-, and SDP-based solutions. At high SNR (25 dB), the GTRS solution is exact for the surrogate SR-LS objective, which closely matches the R-LS objective, and therefore closely tracks the true z-prof… view at source ↗
Figure 4
Figure 4. Figure 4: 3D Localization Results: The single-start D-NLS baseline [37] does not attain the CRLB because its single arbitrary 3D initialization can lead to convergence to suboptimal local minima. The proposed sample–polish–select procedures, 3D-USR and 3D-GTRS, both approach the CRLB using only 8 seeds along the height dimension. By comparison, 3D-MS-GN with the same total seed count (2 3 = 8 seeds in 3D) fails to a… view at source ↗
Figure 5
Figure 5. Figure 5: Computation time comparison of 3D-MS-GN, 3D-USR, and 3D-GTRS. For 3D-MS-GN, results are shown for 8 and 27 seeds generated over the full 3D search space. While the 8-seed 3D-MS￾GN configuration has the lowest cost among the multi-start GN baselines, it does not attain the CRLB. Increasing the number of seeds to 27 improves estimation performance, but at a substantially higher computational cost. In contras… view at source ↗
read the original abstract

Time-of-arrival (TOA)-based localization in mixed line-of-sight (LOS) and non-line-of-sight (NLOS) environments is challenging because conventional Euclidean range models do not capture diffraction-dominated propagation. We show that the diffraction path-length model smoothly transitions between LOS and diffraction-dominated NLOS conditions, eliminating the need for explicit path classification. Although this model provides a unified geometric description of mixed LOS/NLOS propagation, the resulting 3D maximum-likelihood problem is nonconvex, and a direct Gauss--Newton estimator based on this model can converge to suboptimal local minima. This motivates the development of a class of structure-exploiting estimators. For known target height, the model induces a virtual-anchor representation of the reduced 2D problem, enabling estimators that exhibit a clear complexity--performance tradeoff: surrogate formulations provide structure and computational efficiency, while a semidefinite-relaxation formulation more faithfully preserves the original likelihood at higher cost. Building on this same structure, we develop 3D sample--polish--select estimators that reduce the global search to one dimension, solve the associated fixed-height 2D subproblems, and then apply local nonlinear refinement in 3D. The proposed estimators achieve near-Cram\'er--Rao lower bound (CRLB) performance with substantially lower complexity than multistart Gauss--Newton, while also being far more robust to initialization than a direct single-start Gauss--Newton estimator.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The manuscript introduces a diffraction path-length model for TOA-based localization in mixed LOS/NLOS environments. This model is claimed to provide a unified geometric description that smoothly transitions between LOS and diffraction-dominated NLOS without explicit path classification or extra parameters. The resulting 3D ML problem is nonconvex, motivating structure-exploiting estimators: virtual-anchor reductions for known-height 2D cases (with surrogate and SDR formulations offering a complexity-performance tradeoff) and 3D sample-polish-select procedures that reduce global search to one dimension before local 3D refinement. The proposed methods are stated to achieve near-CRLB performance with substantially lower complexity than multistart Gauss-Newton and greater robustness to initialization than single-start Gauss-Newton.

Significance. If the central claims hold under the assumed model, the work offers a parameter-light unified framework for mixed-propagation localization that avoids separate LOS/NLOS classification steps, together with computationally attractive estimators that approach theoretical bounds. The virtual-anchor reduction and 1D search reduction for the 3D problem are clear structural strengths that could be useful in wireless positioning applications. The explicit complexity-performance tradeoff between surrogate and SDR formulations is also a positive feature.

major comments (2)
  1. [Abstract] Abstract: The claim that the diffraction path-length model 'smoothly transitions between LOS and diffraction-dominated NLOS conditions, eliminating the need for explicit path classification' and requires 'no additional parameters beyond the model itself' is load-bearing for all subsequent estimator derivations and performance claims. If real TOA measurements contain unmodeled effects such as reflections or scattering, the likelihood function changes, the CRLB no longer bounds the true error, and the reported near-CRLB performance plus complexity/robustness advantages become model-specific rather than general; the manuscript must therefore include explicit discussion or simulations demonstrating behavior under such mismatches.
  2. [Abstract] Abstract and estimator sections: The near-CRLB performance, lower complexity than multistart Gauss-Newton, and improved initialization robustness are central claims, yet the abstract provides no numerical results, error bars, Monte Carlo trial counts, data-generation details, or CRLB derivation under the nonconvex model. Without these, the strength of the evidence supporting the claims cannot be assessed from the given text.
minor comments (1)
  1. [Abstract] Abstract: The description of the 3D sample-polish-select procedure is compressed; separating the sampling, polishing, and selection steps into distinct sentences would improve clarity for readers unfamiliar with the approach.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful and constructive review. We address each major comment below and indicate the revisions planned for the manuscript.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The claim that the diffraction path-length model 'smoothly transitions between LOS and diffraction-dominated NLOS conditions, eliminating the need for explicit path classification' and requires 'no additional parameters beyond the model itself' is load-bearing for all subsequent estimator derivations and performance claims. If real TOA measurements contain unmodeled effects such as reflections or scattering, the likelihood function changes, the CRLB no longer bounds the true error, and the reported near-CRLB performance plus complexity/robustness advantages become model-specific rather than general; the manuscript must therefore include explicit discussion or simulations demonstrating behavior under such mismatches.

    Authors: We agree that the performance guarantees and near-CRLB claims hold under the assumed diffraction path-length model. In the revised manuscript we will add a dedicated subsection on model mismatch, including analytical discussion of how reflections and scattering alter the likelihood and a set of Monte Carlo simulations that quantify the resulting performance degradation relative to the CRLB. This will explicitly delineate the scope of the reported advantages. revision: yes

  2. Referee: [Abstract] Abstract and estimator sections: The near-CRLB performance, lower complexity than multistart Gauss-Newton, and improved initialization robustness are central claims, yet the abstract provides no numerical results, error bars, Monte Carlo trial counts, data-generation details, or CRLB derivation under the nonconvex model. Without these, the strength of the evidence supporting the claims cannot be assessed from the given text.

    Authors: The abstract is a concise summary; the full manuscript already contains the requested details (1000-trial Monte Carlo results with error bars, explicit data-generation parameters, and the CRLB derivation for the nonconvex likelihood). To make the strength of evidence immediately visible, we will revise the abstract to include concise quantitative statements (e.g., “within 5 % of the CRLB with 10× lower complexity than multistart Gauss–Newton across 1000 trials”). We will also add a brief parenthetical reference to the CRLB derivation in the abstract. revision: yes

Circularity Check

0 steps flagged

No significant circularity in the derivation chain

full rationale

The paper's derivation begins with the diffraction path-length model as the foundation for a unified likelihood function in mixed LOS/NLOS TOA localization. From this, the 3D ML problem is formulated, recognized as nonconvex, and addressed via structure-exploiting estimators (virtual-anchor reduction for known height, surrogate/SDR formulations, and 1D sample-polish-select procedures). These are constructed directly from the model's geometry to approximate the global optimum. The near-CRLB performance, complexity reduction relative to multistart Gauss-Newton, and initialization robustness follow as consequences of better optimization on the same likelihood; the CRLB itself is the independent theoretical bound from the Fisher information matrix of the model, not a fitted or self-referential input. No steps reduce by construction to prior outputs, self-citations, or ansatzes; the claims remain self-contained against the stated model and standard ML theory.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The central claims rest on the diffraction path-length model being an accurate unified description and on standard maximum-likelihood estimation being solvable via the proposed structure-exploiting methods; no free parameters or invented entities are stated in the abstract.

axioms (2)
  • domain assumption The diffraction path-length model provides a valid geometric description of mixed LOS/NLOS propagation
    Invoked in the opening sentences to eliminate explicit path classification
  • domain assumption The resulting 3D ML problem can be reduced to 2D fixed-height subproblems whose solutions can be polished to 3D optimality
    Basis for the sample-polish-select estimator class

pith-pipeline@v0.9.0 · 5570 in / 1298 out tokens · 52651 ms · 2026-05-07T11:20:41.946582+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

43 extracted references · 43 canonical work pages

  1. [1]

    Zekavat and R

    R. Zekavat and R. M. Buehrer,Handbook of Position Location: Theory, Practice, and Advances. John Wiley & Sons, 2019

    work page 2019

  2. [2]

    Line-of-Sight Probability for Outdoor-to-Indoor UA V-Assisted Emergency Networks,

    G. Duggal, R. M. Buehrer, J. H. Reed, and N. Tripathi, “Line-of-Sight Probability for Outdoor-to-Indoor UA V-Assisted Emergency Networks,” IEEE Intl. Conf. on Communications (ICC), pp. 5004–5009, 2023

    work page 2023

  3. [3]

    The Non-Line-of-Sight Problem in Mobile Location Estimation,

    M. P. Wylie and J. Holtzman, “The Non-Line-of-Sight Problem in Mobile Location Estimation,” inProc. 5th IEEE Int. Conf. Universal Pers. Commun. (ICUPC), vol. 2, Cambridge, MA, USA, 1996, pp. 827– 831

    work page 1996

  4. [4]

    Decision Theoretic Framework for NLOS Identification,

    J. Borras, P. Hatrack, and N. B. Mandayam, “Decision Theoretic Framework for NLOS Identification,” inProc. 48th IEEE Veh. Technol. Conf. (VTC), vol. 2, Ottawa, ON, Canada, 1998, pp. 1583–1587

    work page 1998

  5. [5]

    Time-of-Arrival Based Localization Under NLOS Conditions,

    Y .-T. Chan, W.-Y . Tsui, H.-C. So, and P.-C. Ching, “Time-of-Arrival Based Localization Under NLOS Conditions,”IEEE Trans. on Vehicular Technology, vol. 55, no. 1, pp. 17–24, Jan. 2006

    work page 2006

  6. [6]

    A UWB Positioning Method Based on Clustering for NLOS Identification and Mitigation,

    Z. Li, Y . Huan, X. Zhang, Z. Li, and X. Chen, “A UWB Positioning Method Based on Clustering for NLOS Identification and Mitigation,” IEEE Trans. Ind. Electron., 2026

    work page 2026

  7. [7]

    On Geolocation Accuracy with Prior Infor- mation in Non-Line-of-Sight Environment,

    Y . Qi and H. Kobayashi, “On Geolocation Accuracy with Prior Infor- mation in Non-Line-of-Sight Environment,” inProc. IEEE 56th Veh. Technol. Conf. (VTC-Fall), Vancouver, BC, Canada, Sep. 2002, pp. 285– 288

    work page 2002

  8. [8]

    Analysis of Wireless Geolocation in a Non-Line-of-Sight Environment,

    Y . Qi, H. Kobayashi, and H. Suda, “Analysis of Wireless Geolocation in a Non-Line-of-Sight Environment,”IEEE Trans. on Wireless Communi- cations, vol. 5, no. 3, pp. 672–681, Mar. 2006

    work page 2006

  9. [9]

    On Time-of-Arrival Positioning in a Multipath Environment,

    Y . Qi, H. Suda, and H. Kobayashi, “On Time-of-Arrival Positioning in a Multipath Environment,”IEEE Trans. on Vehicular Technology, vol. 55, no. 5, pp. 1516–1526, Sep. 2006

    work page 2006

  10. [10]

    Position Error Bound for UWB Localization in Dense Cluttered Environments,

    D. B. Jourdan, D. Dardari, and M. Z. Win, “Position Error Bound for UWB Localization in Dense Cluttered Environments,”IEEE Trans. Aerosp. Electron. Syst., vol. 44, no. 2, pp. 613–628, Apr. 2008

    work page 2008

  11. [11]

    Fundamental Limits of Wideband Localization—Part I: A General Framework,

    Y . Shen and M. Z. Win, “Fundamental Limits of Wideband Localization—Part I: A General Framework,”IEEE Trans. on Infor- mation Theory, vol. 56, no. 10, pp. 4956–4980, Oct. 2010

    work page 2010

  12. [12]

    A Mathematical Justification for Exponentially Distributed NLOS Bias,

    C. E. O’Lone, H. S. Dhillon, and R. M. Buehrer, “A Mathematical Justification for Exponentially Distributed NLOS Bias,” inProc. IEEE Global Commun. Conf. (GLOBECOM), 2019, pp. 1–6

    work page 2019

  13. [13]

    Characterizing the First-Arriving Multipath Component in 5G Millimeter Wave Networks: TOA, AOA, and Non-Line-of-Sight Bias,

    ——, “Characterizing the First-Arriving Multipath Component in 5G Millimeter Wave Networks: TOA, AOA, and Non-Line-of-Sight Bias,” IEEE Trans. on Wireless Communications, vol. 21, no. 3, pp. 1602–1620, Mar. 2022

    work page 2022

  14. [14]

    A Survey on TOA Based Wireless Localization and NLOS Mitigation Techniques,

    I. Guvenc and C.-C. Chong, “A Survey on TOA Based Wireless Localization and NLOS Mitigation Techniques,”IEEE Commun. Surveys Tuts., vol. 11, no. 3, pp. 107–124, 2009

    work page 2009

  15. [15]

    NLOS Mitigation Using Linear Pro- gramming in Ultrawideband Location-Aware Networks,

    S. Venkatesh and R. M. Buehrer, “NLOS Mitigation Using Linear Pro- gramming in Ultrawideband Location-Aware Networks,”IEEE Trans. on Vehicular Technology, vol. 56, no. 5, pp. 3182–3198, Sep. 2007

    work page 2007

  16. [16]

    Cooperative Sensor Localization with NLOS Mitigation Using Semidefinite Programming,

    R. M. Vaghefi and R. M. Buehrer, “Cooperative Sensor Localization with NLOS Mitigation Using Semidefinite Programming,” inProc. 9th Workshop Positioning, Navigat. Commun. (WPNC), 2012, pp. 13–18

    work page 2012

  17. [17]

    NLOS Mitigation in TOA-Based Localization Using Semidefinite Programming,

    R. M. Vaghefi, J. Schloemann, and R. M. Buehrer, “NLOS Mitigation in TOA-Based Localization Using Semidefinite Programming,” inProc. 10th Workshop Positioning, Navigat. Commun. (WPNC), Dresden, Ger- many, 2013, pp. 1–6

    work page 2013

  18. [18]

    Non-Line- of-Sight Node Localization Based on Semi-Definite Programming in Wireless Sensor Networks,

    H. Chen, G. Wang, Z. Wang, H.-C. So, and H. V . Poor, “Non-Line- of-Sight Node Localization Based on Semi-Definite Programming in Wireless Sensor Networks,”IEEE Trans. on Wireless Communications, vol. 11, no. 1, pp. 108–116, Jan. 2012

    work page 2012

  19. [19]

    Collaborative Position Location with NLOS Mitigation,

    T. Jia and R. M. Buehrer, “Collaborative Position Location with NLOS Mitigation,” inProc., IEEE PIMRC Workshops, 2010, pp. 267–271

    work page 2010

  20. [20]

    Two-Stage Weighted Projection for Reliable Low-Complexity Cooperative and Non-Cooperative Localization,

    H. K. Dureppagari, R. M. Buehrer, and H. S. Dhillon, “Two-Stage Weighted Projection for Reliable Low-Complexity Cooperative and Non-Cooperative Localization,”IEEE Trans. on Wireless Communica- tions, vol. 25, pp. 5029–5045, 2026

    work page 2026

  21. [21]

    Ray Tracing for Radio Propagation Modeling: Principles and Applications,

    Z. Yun and M. F. Iskander, “Ray Tracing for Radio Propagation Modeling: Principles and Applications,”IEEE Access, vol. 3, pp. 1089– 1100, 2015

    work page 2015

  22. [22]

    UWB Positioning with Virtual Anchors and Floor Plan Information,

    P. Meissner, C. Steiner, and K. Witrisal, “UWB Positioning with Virtual Anchors and Floor Plan Information,” inProc. 7th Workshop Positioning, Navigat. Commun. (WPNC), Dresden, Germany, 2010, pp. 150–156

    work page 2010

  23. [23]

    Multipath- Assisted Maximum-Likelihood Indoor Positioning Using UWB Signals,

    E. Leitinger, M. Fr ¨ohle, P. Meissner, and K. Witrisal, “Multipath- Assisted Maximum-Likelihood Indoor Positioning Using UWB Signals,” inProc. IEEE Int. Conf. Commun. Workshops (ICC Workshops), 2014, pp. 170–175

    work page 2014

  24. [24]

    Multipath Assisted Positioning with Simultaneous Localization and Mapping,

    C. Gentner, T. Jost, W. Wang, S. Zhang, A. Dammann, and U.-C. Fiebig, “Multipath Assisted Positioning with Simultaneous Localization and Mapping,”IEEE Trans. on Wireless Communications, vol. 15, no. 9, pp. 6104–6117, Sep. 2016

    work page 2016

  25. [25]

    Evaluation of Position-Related Information in Multipath Components for Indoor Positioning,

    E. Leitinger, P. Meissner, C. R ¨udisser, G. Dumphart, and K. Witrisal, “Evaluation of Position-Related Information in Multipath Components for Indoor Positioning,”IEEE J. Sel. Areas Commun., vol. 33, no. 11, pp. 2313–2328, Nov. 2015

    work page 2015

  26. [26]

    UWB for Robust Indoor Tracking: Weighting of Multipath Components for Efficient Estimation,

    P. Meissner, E. Leitinger, and K. Witrisal, “UWB for Robust Indoor Tracking: Weighting of Multipath Components for Efficient Estimation,” IEEE Wireless Communications Letters, vol. 3, no. 5, pp. 501–504, Oct. 2014

    work page 2014

  27. [27]

    High-Accuracy Localization for Assisted Living: 5G Systems Will Turn Multipath Channels From Foe to Friend,

    K. Witrisal, P. Meissner, E. Leitinger, Y . Shen, C. Gustafson, F. Tufves- son, K. Haneda, D. Dardari, A. F. Molisch, A. Conti, and M. Z. Win, “High-Accuracy Localization for Assisted Living: 5G Systems Will Turn Multipath Channels From Foe to Friend,”IEEE Signal Processing Magazine, vol. 33, no. 2, pp. 59–70, Mar. 2016

    work page 2016

  28. [28]

    Impact of Frequency on Diffraction-Aided Wireless Positioning,

    G. Duggal, A. M. Kumar, R. M. Buehrer, H. S. Dhillon, N. Kumar, and J. H. Reed, “Impact of Frequency on Diffraction-Aided Wireless Positioning,”IEEE Intl. Conf. on Communications (ICC), 2025

    work page 2025

  29. [29]

    (2024, July) Wireless InSite MIMO: Version 3.4.4

    Remcom Inc . (2024, July) Wireless InSite MIMO: Version 3.4.4. [Online]. Available: https://www.remcom.com

    work page 2024

  30. [30]

    Estimation of Electrical Characteristics of Inhomogeneous Walls Using Generative Adversarial Networks,

    K. Yasmeen and S. S. Ram, “Estimation of Electrical Characteristics of Inhomogeneous Walls Using Generative Adversarial Networks,”IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, vol. 16, pp. 7009–7023, 2023

    work page 2023

  31. [31]

    3D Positioning Using a New Diffraction Path Model,

    G. Duggal, R. M. Buehrer, H. S. Dhillon, and J. H. Reed, “3D Positioning Using a New Diffraction Path Model,”IEEE Intl. Conf. on Communications (ICC), 2024

    work page 2024

  32. [32]

    A Novel NLOS Correction Approach for Harsh Indoor Settings,

    Y . Zhou, W. Jiang, B. Cai, X. Hu, J. Wang, C. Deng, J. Liu, and B. Li, “A Novel NLOS Correction Approach for Harsh Indoor Settings,”IEEE Internet of Things Journal, 2026

    work page 2026

  33. [33]

    Measurements of 28 GHz Diffraction Loss by Building Corners,

    P. A. Tenerelli and C. W. Bostian, “Measurements of 28 GHz Diffraction Loss by Building Corners,” inProc. 9th IEEE Int. Symp. Pers., Indoor Mobile Radio Commun. (PIMRC), vol. 3. IEEE, 1998, pp. 1166–1169

    work page 1998

  34. [34]

    Embracing Diffraction: A Paradigm Shift in Wireless Sensing and Communication,

    A. Pallaprolu, W. Hurst, and Y . Mostofi, “Embracing Diffraction: A Paradigm Shift in Wireless Sensing and Communication,”IEEE J. Sel. Topics Electromagn., Antennas Propag., vol. 1, no. 1, pp. 150–164, 2025

    work page 2025

  35. [35]

    Analytical Methods to Determine Diffraction Points on Multiple Edges and Cylindrical Scatterers in UTD Ray Tracing,

    T. W. Ang, S. Y . Tan, and H. S. Tan, “Analytical Methods to Determine Diffraction Points on Multiple Edges and Cylindrical Scatterers in UTD Ray Tracing,”Microw. Opt. Technol. Lett., vol. 22, no. 5, pp. 304–309, Sep. 1999

    work page 1999

  36. [36]

    Optimal Anchor Placement for Wireless Localization in Mixed LOS and NLOS Scenarios,

    G. Duggal, R. M. Buehrer, H. S. Dhillon, and J. H. Reed, “Optimal Anchor Placement for Wireless Localization in Mixed LOS and NLOS Scenarios,” 2026. [Online]. Available: https://arxiv.org/abs/2604.00863

    work page arXiv 2026

  37. [37]

    Diffraction-Aided Wireless Positioning,

    ——, “Diffraction-Aided Wireless Positioning,”IEEE Trans. on Wireless Communications, 2025

    work page 2025

  38. [38]

    Non-Line-of-Sight Localization in Multipath Environments,

    C. K. Seow and S. Y . Tan, “Non-Line-of-Sight Localization in Multipath Environments,”IEEE Trans. on Mobile Computing, vol. 7, no. 5, pp. 647–660, 2008

    work page 2008

  39. [39]

    Exact and Approximate Solutions of Source Localization Problems,

    A. Beck, P. Stoica, and J. Li, “Exact and Approximate Solutions of Source Localization Problems,”IEEE Trans. on Signal Processing, vol. 56, no. 5, pp. 1770–1778, 2008

    work page 2008

  40. [40]

    G. H. Golub and C. F. Van Loan,Matrix Computations, 4th ed. Johns Hopkins University Press, 2013

    work page 2013

  41. [41]

    Fundamental Limits of Wideband Localization — Part I: A General Framework,

    Y . Shen and M. Z. Win, “Fundamental Limits of Wideband Localization — Part I: A General Framework,”IEEE Trans. on Information Theory, Oct. 2010

    work page 2010

  42. [42]

    [Online]

    MOSEK,Optimization Toolbox for MATLAB 11.1.10, 2026. [Online]. Available: https://docs.mosek.com/latest/toolbox/index.html

    work page 2026

  43. [43]

    S. M. Kay,Fundamentals of Statistical Signal Processing: Estimation Theory. Prentice-Hall, Inc., 1993

    work page 1993