Pith. sign in

REVIEW 3 major objections 2 minor 80 references

OmniLoc uses unified tokenization of wireless signals plus geometry-aware attention to localize devices without anchors across many different buildings.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · grok-4.3

2026-06-27 13:36 UTC pith:UG5ZV3TC

load-bearing objection OmniLoc claims to be the first foundation model for anchor-free wireless localization but the abstract supplies no equations, numbers, or mechanism details to back the generalization claims. the 3 major comments →

arxiv 2606.11490 v1 pith:UG5ZV3TC submitted 2026-06-09 cs.LG cs.SYeess.SY

OmniLoc: A Geometry-Aware Foundation Model for Anchor-Free UE Localization Across Diverse Indoor Environments

Lei Chu , Yuning Zhang , Omer Gokalp Serbetci , Anushka Katiyar , Bassel Abou Ali Modad , Andreas F. Molisch This is my paper
classification cs.LG cs.SYeess.SY
keywords indoor localizationanchor-freefoundation modelwireless measurementsgeometry-aware transformeruser equipmentcross-environment generalizationaccess point
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The paper presents OmniLoc as a foundation model that takes heterogeneous wireless measurements and produces location estimates for user equipment without relying on fixed anchors. It converts varied signal types into one common token form, runs a Transformer that weights dominant access points while pulling supporting information from others, and ties the final regression step to geometric embeddings so predictions stay consistent with building structure. The result is meant to work when the set of visible access points, the building layout, and the signal statistics all shift between training and test settings. A sympathetic reader would care because current methods typically need retraining or fail when moved to a new environment, so a single model that generalizes could simplify large-scale indoor deployments.

Core claim

OmniLoc is the first foundation-model approach built directly on wireless measurements for anchor-free UE localization. Its three core designs are a unified input tokenization module that turns heterogeneous measurements into a shared representation, a geometry-aware Transformer that extracts features by emphasizing dominant APs and aggregating from supporting APs, and a geometry-aware location estimation module that conditions regression on geometric embeddings. On a large in-house dataset and a public benchmark, the model outperforms prior methods, improves existing backbones when its components are added, and shows strong cross-environment generalization.

What carries the argument

Geometry-aware Transformer performing AP-aware feature extraction by emphasizing dominant APs while aggregating complementary evidence from supporting APs.

Load-bearing premise

A single unified tokenization of wireless measurements together with geometry-aware attention will keep location predictions accurate when the detectable access points, building geometry, and signal statistics all change substantially between training and test environments.

What would settle it

Measure localization error after training on one set of buildings and testing on a new building whose access-point placement, floor plan, and signal statistics were never seen during training; the claim fails if median error rises sharply compared with environment-specific baselines.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • Adding the three design components to existing localization backbones consistently raises their accuracy.
  • The same trained model maintains performance when the number and identity of detectable APs vary across buildings.
  • Location estimates remain geometrically consistent even though no explicit floor-plan input is required at inference time.
  • Cross-environment testing shows the model does not require per-building retraining to reach competitive error levels.

Where Pith is reading between the lines

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

  • The tokenization step could be reused for other wireless tasks that must combine RSSI, CSI, and time-of-flight data from changing transmitter sets.
  • If the geometric embeddings capture building structure well, training data requirements might drop because explicit maps are less critical.
  • Extending the same attention pattern to outdoor or hybrid indoor-outdoor settings would test whether the geometry conditioning generalizes beyond walls and ceilings.

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

3 major / 2 minor

Summary. The paper introduces OmniLoc, presented as the first foundation-model approach for anchor-free UE localization directly from wireless measurements across diverse indoor environments. It proposes three components: (1) a unified input tokenization module to convert heterogeneous measurements into a common representation, (2) a geometry-aware Transformer that performs AP-aware feature extraction by emphasizing dominant APs and aggregating from supporting APs, and (3) a geometry-aware location estimation module that conditions regression on geometric embeddings. Evaluations on a large-scale in-house dataset and a public benchmark claim significant outperformance over existing methods, consistent improvements when components are added to backbones, and strong generalization in cross-environment settings.

Significance. If the generalization results hold under the reported conditions, the work would address a central limitation in learning-based indoor localization by providing an architecture intended to remain accurate despite changes in detectable APs, building geometry, and signal statistics. The explicit incorporation of geometry-aware attention and embeddings is a constructive direction; however, the foundation-model framing would require the experiments to demonstrate that performance gains derive from the proposed mechanisms rather than environment-specific fitting.

major comments (3)
  1. [§3.1] §3.1 (Unified Input Tokenization): the description does not specify whether tokenization employs fixed AP indexing, learned per-AP embeddings, or fully permutation-invariant set processing. This detail is load-bearing for the cross-environment generalization claim, because any implicit dependence on the training-environment AP set would undermine the reported gains when the detectable AP support changes substantially between train and test environments.
  2. [§4.3] §4.3 (Cross-Environment Evaluations): the reported improvements are presented without ablations that isolate the contribution of the geometric embeddings versus environment-specific statistics learned during pre-training. Without such controls, it is not possible to confirm that the geometry-aware location estimation module produces predictions that remain consistent when the coordinate frame and signal distributions shift, as required by the central claim.
  3. [§3.2] §3.2 (Geometry-Aware Transformer): the AP-aware feature extraction is described as emphasizing dominant APs while aggregating complementary evidence, yet no concrete mechanism (e.g., attention masking, learned AP importance scores, or explicit geometric distance embeddings) is provided. This omission prevents assessment of whether the module can operate without retraining when both the set of APs and the building geometry change.
minor comments (2)
  1. The abstract and introduction would benefit from explicit statements of the number of environments, number of APs per environment, and the precise train/test split protocol used in the cross-environment experiments.
  2. Figure captions for the architecture diagram should clarify the dimensionality of the geometric embeddings and how they are concatenated with the tokenized measurements.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive and detailed comments, which have identified opportunities to improve the clarity of our descriptions and strengthen the supporting analyses. We address each major comment below and commit to revisions that will incorporate the requested details and ablations.

read point-by-point responses

  1. Referee: [§3.1] §3.1 (Unified Input Tokenization): the description does not specify whether tokenization employs fixed AP indexing, learned per-AP embeddings, or fully permutation-invariant set processing. This detail is load-bearing for the cross-environment generalization claim, because any implicit dependence on the training-environment AP set would undermine the reported gains when the detectable AP support changes substantially between train and test environments.

    Authors: We agree that the tokenization details are essential to substantiate the generalization claims. The unified input tokenization module processes measurements as a set without fixed AP indexing or per-AP embeddings tied to any training environment; each token is formed from signal features and a geometry-derived context vector that is independent of specific AP identities. This permutation-invariant design is what enables handling of varying AP supports. In the revised manuscript we will expand §3.1 with an explicit statement of this property, the tokenization equations, and pseudocode. revision: yes

  2. Referee: [§4.3] §4.3 (Cross-Environment Evaluations): the reported improvements are presented without ablations that isolate the contribution of the geometric embeddings versus environment-specific statistics learned during pre-training. Without such controls, it is not possible to confirm that the geometry-aware location estimation module produces predictions that remain consistent when the coordinate frame and signal distributions shift, as required by the central claim.

    Authors: The referee correctly notes the absence of targeted ablations. While the cross-environment results in §4.3 show overall gains, they do not yet isolate the geometry-aware location estimation module from possible environment-specific fitting. We will add the requested ablation studies in the revised §4.3, comparing the full model against a variant that removes the geometric conditioning, evaluated specifically on the cross-environment splits. These new results will be included to directly address whether the geometric embeddings support consistency under coordinate-frame and signal-distribution shifts. revision: yes

  3. Referee: [§3.2] §3.2 (Geometry-Aware Transformer): the AP-aware feature extraction is described as emphasizing dominant APs while aggregating complementary evidence, yet no concrete mechanism (e.g., attention masking, learned AP importance scores, or explicit geometric distance embeddings) is provided. This omission prevents assessment of whether the module can operate without retraining when both the set of APs and the building geometry change.

    Authors: We acknowledge that the current description in §3.2 is high-level and lacks the concrete implementation details needed for independent assessment. The AP-aware feature extraction modulates self-attention using learned importance scores (derived from signal strength and geometric priors) together with explicit pairwise geometric distance embeddings between APs; the attention is set-based and does not rely on a fixed AP vocabulary. In the revised manuscript we will supply the precise attention equations, the formulation of the importance scores, and the distance-embedding construction in §3.2 so that readers can verify the mechanism's invariance to AP-set and geometry changes. revision: yes

Circularity Check

0 steps flagged

No circularity: model described via architectural components with empirical claims only.

full rationale

The paper presents OmniLoc as a foundation model with three design elements (unified tokenization, geometry-aware Transformer, geometry-aware regression) but supplies no equations, parameter-fitting procedures, or derivations that could reduce to self-defined quantities. Claims rest on reported performance on in-house and public datasets plus cross-environment evaluations; no self-citation chains, uniqueness theorems, or ansatzes are invoked in the provided text. The derivation chain is therefore self-contained as an empirical architecture description rather than a closed mathematical reduction.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

Abstract-only review yields no visible free parameters, invented entities, or non-standard axioms; the model relies on standard Transformer assumptions and domain assumptions about wireless signal geometry.

axioms (2)
  • domain assumption Heterogeneous wireless measurements can be converted into a common token representation amenable to Transformer processing
    Described as the first key design in the abstract.
  • domain assumption Emphasizing dominant APs while aggregating supporting APs produces geometrically consistent location estimates
    Second and third design components stated in the abstract.

pith-pipeline@v0.9.1-grok · 5781 in / 1335 out tokens · 18912 ms · 2026-06-27T13:36:27.751370+00:00 · methodology

0 comments
read the original abstract

Indoor localization from wireless measurements remains challenging in large-scale deployments due to substantial variation in building geometry, the set of detectable access points (APs), and the heterogeneity of received signals. Existing learning-based methods often perform well only in limited settings and degrade under environmental shifts, making robust anchor-free localization across diverse indoor environments notoriously difficult. In this paper, we present OmniLoc, an environment-interactive foundation model for anchor-free user equipment localization across diverse indoor environments. To the best of our knowledge, OmniLoc is the first foundation-model-based approach built directly on wireless measurements for this task. OmniLoc is built on three key designs. First, a unified input tokenization module converts heterogeneous wireless measurements into a common representation that is more amenable to learning. Second, a geometry-aware Transformer performs AP-aware feature extraction by emphasizing dominant APs while aggregating complementary evidence from supporting APs. Third, a geometry-aware location estimation module conditions regression on geometric embeddings to produce geometrically consistent location predictions. We evaluate OmniLoc on both a large-scale in-house dataset and a public benchmark dataset. Results show that OmniLoc significantly outperforms existing methods, consistently improves existing backbones when its design components are integrated, and demonstrates strong generalization in cross-environment evaluations.

Figures

Figures reproduced from arXiv: 2606.11490 by Andreas F. Molisch, Anushka Katiyar, Bassel Abou Ali Modad, Lei Chu, Omer Gokalp Serbetci, Yuning Zhang.

Figure 1
Figure 1. Figure 1: Variable AP Length and Power Variations distinct blockage patterns and multipath effects, causing signal observations to vary significantly across buildings. As a re￾sult, representations learned in one building may not transfer reliably to another. Similar heterogeneity also exists within a single building, where floor-specific layouts and localized clutter further affect both AP visibility and received s… view at source ↗
Figure 2
Figure 2. Figure 2: Overview of our OmniLoc-based Anchor-free UE Localization framework. [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Qualitative comparisons with SOTA methods. [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Building-wise qualitative results [PITH_FULL_IMAGE:figures/full_fig_p008_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Floor-wise qualitative results. splitting. However, due to the high density of the fingerprinting (< 0.01 m spacing) and the Wi-Fi beacon interval of 102.4 ms, residual spatial and temporal correlations likely persist between the training and testing sets, yielding optimistic performance. For cross-environment evaluation, we consider three chal￾lenging settings: Scenario I: leave-one-floor-out, Scenario II… view at source ↗
Figure 7
Figure 7. Figure 7: Test RMSE vs. Label Ratio (Scenario I). cluding the original baselines and their enhanced “+” variants, which retain the original backbone while incorporating Om￾niLoc’s design components (unified tokenization, customized feature extraction, and hierarchical prediction). It is shown in [PITH_FULL_IMAGE:figures/full_fig_p009_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Test RMSE vs. Label Ratio (Scenario II). [PITH_FULL_IMAGE:figures/full_fig_p010_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Test RMSE versus the percentage of labeled data [PITH_FULL_IMAGE:figures/full_fig_p010_9.png] view at source ↗

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

80 extracted references · 4 canonical work pages · 1 internal anchor

  1. [1]

    Zekavat and R

    R. Zekavat and R. M. Buehrer,Handbook of position location: theory, practice, and advances. John Wiley & Sons, 2019

    2019

  2. [2]

    A. F. Molisch,Wireless communications, 3rd. ed. - From fundamentals to beyond 5G. IEEE Press - Wiley, 2022

    2022

  3. [3]

    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. Contiet al., “High- accuracy localization for assisted living: 5G systems will turn multipath channels from foe to friend,”IEEE Signal Process. Mag., vol. 33, no. 2, pp. 59–70, 2016

    2016

  4. [4]

    A survey of indoor localization systems and technologies,

    F. Zafari, A. Gkelias, and K. K. Leung, “A survey of indoor localization systems and technologies,”IEEE Commun. Surveys Tuts., vol. 21, no. 3, pp. 2568–2599, 2019

    2019

  5. [5]

    Context-aware telco outdoor localization,

    Y . Zhang, W. Rao, M. Yuan, J. Zeng, and P. Hui, “Context-aware telco outdoor localization,”IEEE Trans. Mobile Comput., vol. 21, no. 4, pp. 1211–1225, 2020

    2020

  6. [6]

    Self-sustaining ultrawideband positioning system for event-driven indoor localization,

    P. Mayer, M. Magno, and L. Benini, “Self-sustaining ultrawideband positioning system for event-driven indoor localization,”IEEE Internet Things J., vol. 11, no. 1, pp. 1272–1284, 2023

    2023

  7. [7]

    Localization and positioning systems for emergency responders: A survey,

    A. F. G. Ferreira, D. M. A. Fernandes, A. P. Catarino, and J. L. Monteiro, “Localization and positioning systems for emergency responders: A survey,”IEEE Commun. Surveys Tuts., vol. 19, no. 4, pp. 2836–2870, 2017

    2017

  8. [8]

    A platform for power management based on indoor localization in smart buildings using long short-term neural networks,

    M. Elbes, T. Alrawashdeh, E. Almaita, S. AlZu’bi, and Y . Jararweh, “A platform for power management based on indoor localization in smart buildings using long short-term neural networks,”Trans. Emerg. Telecommun. Technol., vol. 33, no. 3, p. e3867, 2022

    2022

  9. [9]

    A comprehensive survey of machine learning based localization with wireless signals,

    D. Burghal, A. T. Ravi, V . Rao, A. A. Alghafis, and A. F. Molisch, “A comprehensive survey of machine learning based localization with wireless signals,”arXiv preprint arXiv:2012.11171, 2020

    work page arXiv 2012

  10. [10]

    Machine learning based indoor localization using wi-fi rssi fingerprints: An overview,

    N. Singh, S. Choe, and R. Punmiya, “Machine learning based indoor localization using wi-fi rssi fingerprints: An overview,”IEEE access, vol. 9, pp. 127 150–127 174, 2021

    2021

  11. [11]

    An overview of indoor localization techniques,

    S. Tiku and S. Pasricha, “An overview of indoor localization techniques,” Machine Learning for Indoor Localization and Navigation, pp. 3–25, 2023

    2023

  12. [12]

    Uncovering the potential of indoor localization: Role of deep and transfer learning,

    O. Kerdjidj, Y . Himeur, S. S. Sohail, A. Amira, F. Fadli, S. Atalla, W. Mansoor, A. Copiaco, A. Gawanmeh, S. Miniaouiet al., “Uncovering the potential of indoor localization: Role of deep and transfer learning,” IEEE Access, vol. 12, pp. 73 980–74 010, 2024

    2024

  13. [13]

    A survey on the impact of multipath on wideband time-of-arrival based localization,

    S. Aditya, A. F. Molisch, and H. M. Behairy, “A survey on the impact of multipath on wideband time-of-arrival based localization,”Proc. IEEE, vol. 106, no. 7, pp. 1183–1203, 2018

    2018

  14. [14]

    Exploiting semantic local- ization in highly dynamic wireless networks using deep homoscedastic domain adaptation,

    L. Chu, A. Alghafis, and A. F. Molisch, “Exploiting semantic local- ization in highly dynamic wireless networks using deep homoscedastic domain adaptation,”IEEE Trans. Commun., vol. 73, no. 3, pp. 2032– 2046, 2024

    2032

  15. [15]

    Blockage intelligence in complex environments for beyond 5G localization,

    G. Torsoli, M. Z. Win, and A. Conti, “Blockage intelligence in complex environments for beyond 5G localization,”IEEE J. Sel. Areas Commun., vol. 41, no. 6, pp. 1688–1701, 2023

    2023

  16. [16]

    Precise power delay profiling with commodity WiFi,

    Y . Xie, Z. Li, and M. Li, “Precise power delay profiling with commodity WiFi,” inProc. ACM Int. Conf. Mobile Comput. Netw. (MobiCom), 2015, pp. 53–64

    2015

  17. [17]

    From error analysis to mitigation: A hybrid framework for enhancing Wi-Fi FTM positioning in multipath- prone indoor scenarios,

    W. Tong, B. Li, and J. Qiao, “From error analysis to mitigation: A hybrid framework for enhancing Wi-Fi FTM positioning in multipath- prone indoor scenarios,”IEEE Trans. Wireless Commun., vol. 25, pp. 9859–9874, 2026

    2026

  18. [18]

    Large multi-modal models (LMMs) as universal foundation models for AI-native wireless systems,

    S. Xu, C. K. Thomas, O. Hashash, N. Muralidhar, W. Saad, and N. Ramakrishnan, “Large multi-modal models (LMMs) as universal foundation models for AI-native wireless systems,”IEEE Netw., vol. 38, no. 5, pp. 10–20, 2024

    2024

  19. [19]

    Radio foundation models: Pre-training transformers for 5G-based indoor localization,

    J. Ott, J. Pirkl, M. Stahlke, T. Feigl, and C. Mutschler, “Radio foundation models: Pre-training transformers for 5G-based indoor localization,” in Proc. Int. Conf. Indoor Positioning Indoor Navig. (IPIN), 2024, pp. 1–6

    2024

  20. [20]

    Large wireless localization model (LWLM): A foundation model for positioning in 6G networks,

    G. Pan, K. Huang, H. Chen, S. Zhang, C. H ¨ager, and H. Wymeersch, “Large wireless localization model (LWLM): A foundation model for positioning in 6G networks,”arXiv preprint arXiv:2505.10134, 2025

    work page arXiv 2025

  21. [21]

    WiFo: Wireless foundation model for channel prediction,

    B. Liu, S. Gao, X. Liu, X. Cheng, and L. Yang, “WiFo: Wireless foundation model for channel prediction,”Sci. China Inf. Sci., vol. 68, no. 6, p. 162302, 2025

    2025

  22. [22]

    Cross-scenario foundation localization models: Architecture, key technologies, and challenges,

    H. Si, X. Guo, G. O. Boateng, and H. Xia, “Cross-scenario foundation localization models: Architecture, key technologies, and challenges,” IEEE Wireless Communications, 2025

    2025

  23. [23]

    6g wavesfm: A foundation model for sensing, communication, and localization,

    A. Aboulfotouh, E. Mohammed, and H. Abou-Zeid, “6g wavesfm: A foundation model for sensing, communication, and localization,”IEEE Open Journal of the Communications Society, 2025

    2025

  24. [24]

    A foundation model for wireless technology recognition and localiza- tion tasks,

    M. Cheraghinia, E. De Poorter, J. Fontaine, M. Debbah, and A. Shahid, “A foundation model for wireless technology recognition and localiza- tion tasks,”IEEE Open Journal of the Communications Society, vol. 6, pp. 9879–9896, 2025

    2025

  25. [25]

    Wireless AI evolution: From statistical learners to electromagnetic-guided foundation models,

    J. Xiao, J. Wang, K. Cao, X. Li, Z. Chen, and C. Yuen, “Wireless AI evolution: From statistical learners to electromagnetic-guided foundation models,”IEEE Commun. Mag., 2026

    2026

  26. [26]

    DeepMIMO: A Generic Deep Learning Dataset for Millimeter Wave and Massive MIMO Applications

    A. Alkhateeb, “Deepmimo: A generic deep learning dataset for millimeter wave and massive mimo applications,”arXiv preprint arXiv:1902.06435, 2019

    work page internal anchor Pith review Pith/arXiv arXiv 1902

  27. [27]

    WiLoc: Massive measured dataset of Wi-Fi channel state information with application to machine-learning based localization,

    Y . Zhang, L. Chu, O. G. Serbetci, J. Gomez-Ponce, and A. F. Molisch, “WiLoc: Massive measured dataset of Wi-Fi channel state information with application to machine-learning based localization,” inProc. IEEE INFOCOM, 2026

    2026

  28. [28]

    WiDeep: WiFi-based accurate and robust indoor localization system using deep learning,

    M. Abbas, M. Elhamshary, H. Rizk, M. Torki, and M. Youssef, “WiDeep: WiFi-based accurate and robust indoor localization system using deep learning,” inProc. IEEE Int. Conf. Pervasive Comput. Commun. (PerCom), 2019, pp. 1–10

    2019

  29. [29]

    Model-based localization and tracking using bluetooth low-energy beacons,

    F. S. Danis ¸ and A. T. Cemgil, “Model-based localization and tracking using bluetooth low-energy beacons,”Sensors, vol. 17, no. 11, p. 2484, 2017

    2017

  30. [30]

    Localization via ultra-wideband radios: A look at positioning aspects for future sensor networks,

    S. Gezici, Z. Tian, G. B. Giannakis, H. Kobayashi, A. F. Molisch, H. V . Poor, and Z. Sahinoglu, “Localization via ultra-wideband radios: A look at positioning aspects for future sensor networks,”IEEE Signal Process. Mag., vol. 22, no. 4, pp. 70–84, 2005

    2005

  31. [31]

    Anchor- free distributed localization in sensor networks,

    N. B. Priyantha, H. Balakrishnan, E. Demaine, and S. Teller, “Anchor- free distributed localization in sensor networks,” inProc. ACM Conf. Embedded Netw. Sensor Syst. (SenSys), 2003, pp. 340–341

    2003

  32. [32]

    Accurate anchor-free node localization in wireless sensor networks,

    A. Youssef, A. Agrawala, and M. Younis, “Accurate anchor-free node localization in wireless sensor networks,” inProc. IEEE Int. Perform. Comput. Commun. Conf. (IPCCC), 2005, pp. 465–470

    2005

  33. [33]

    Anchor-free localization: Estimation of relative locations of sensors,

    S. Shioda and K. Shimamura, “Anchor-free localization: Estimation of relative locations of sensors,” inProc. IEEE Int. Symp. Pers. Indoor Mobile Radio Commun. (PIMRC), 2013, pp. 2087–2092. JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2015 13

    2013

  34. [34]

    Connectivity- based and anchor-free localization in large-scale 2D/3D sensor net- works,

    G. Tan, H. Jiang, S. Zhang, Z. Yin, and A.-M. Kermarrec, “Connectivity- based and anchor-free localization in large-scale 2D/3D sensor net- works,”ACM Trans. Sensor Netw., vol. 10, no. 1, pp. 1–21, 2013

    2013

  35. [35]

    Cooperative indoor localization using mobile robot anchors via factor graph optimization,

    B. Zhou, M. Tang, C. Liu, X. Zhong, H. He, X. Chen, J. Song, Y . Wang, X. Zhang, and Q. Li, “Cooperative indoor localization using mobile robot anchors via factor graph optimization,”IEEE Internet Things J., 2025

    2025

  36. [36]

    SpringLoc: A device-free localization technique for indoor positioning and tracking using adaptive RSSI spring relaxation,

    D. Konings, F. Alam, F. Noble, and E. M. K. Lai, “SpringLoc: A device-free localization technique for indoor positioning and tracking using adaptive RSSI spring relaxation,”IEEE Access, vol. 7, pp. 56 960– 56 973, 2019

    2019

  37. [37]

    Anchor- free TDOA self-localization,

    J. Wendeberg, F. H ¨oflinger, C. Schindelhauer, and L. Reindl, “Anchor- free TDOA self-localization,” inProc. Int. Conf. Indoor Positioning Indoor Navig. (IPIN), 2011, pp. 1–10

    2011

  38. [38]

    Anchor-free localization method for mobile targets in coal mine wireless sensor networks,

    Z. Pei, Z. Deng, S. Xu, and X. Xu, “Anchor-free localization method for mobile targets in coal mine wireless sensor networks,”Sensors, vol. 9, no. 4, pp. 2836–2850, 2009

    2009

  39. [39]

    A state-of-the- art survey on multidimensional scaling-based localization techniques,

    N. Saeed, H. Nam, T. Y . Al-Naffouri, and M.-S. Alouini, “A state-of-the- art survey on multidimensional scaling-based localization techniques,” IEEE Commun. Surveys Tuts., vol. 21, no. 4, pp. 3565–3583, 2019

    2019

  40. [40]

    SmartPDR: Smartphone-based pedestrian dead reckoning for indoor localization,

    W. Kang and Y . Han, “SmartPDR: Smartphone-based pedestrian dead reckoning for indoor localization,”IEEE Sensors J., vol. 15, no. 5, pp. 2906–2916, 2014

    2014

  41. [41]

    UBiGTLoc: A unified BiLSTM-graph transformer localization framework for IoT sensor networks,

    A. A. Lehyeh, A. Gharib, T. Xia, D. Huston, and S. Wshah, “UBiGTLoc: A unified BiLSTM-graph transformer localization framework for IoT sensor networks,”IEEE Internet Things J., 2025

    2025

  42. [42]

    Anchor selection for SLAM based on graph topology and submodular optimiza- tion,

    Y . Chen, L. Zhao, Y . Zhang, S. Huang, and G. Dissanayake, “Anchor selection for SLAM based on graph topology and submodular optimiza- tion,”IEEE Trans. Robot., vol. 38, no. 1, pp. 329–350, 2021

    2021

  43. [43]

    CSI-based indoor localization,

    K. Wu, J. Xiao, Y . Yi, D. Chen, X. Luo, and L. M. Ni, “CSI-based indoor localization,”IEEE Trans. Parallel Distrib. Syst., vol. 24, no. 7, pp. 1300–1309, 2012

    2012

  44. [44]

    A robust vector matching localization approach based on multiple channels SSD fingerprinting of ZigBee networks,

    L. Chu, X.-S. Guo, B. Li, B. Xu, Q. Wan, and Y . Shen, “A robust vector matching localization approach based on multiple channels SSD fingerprinting of ZigBee networks,”Prog. Electromagn. Res., vol. 144, pp. 133–140, 2014

    2014

  45. [45]

    A survey on wireless indoor localization from the device perspective,

    J. Xiao, Z. Zhou, Y . Yi, and L. M. Ni, “A survey on wireless indoor localization from the device perspective,”ACM Comput. Surv., vol. 49, no. 2, pp. 1–31, 2016

    2016

  46. [46]

    Micro-doppler-based target detection and feature extraction in indoor and outdoor environ- ments,

    T. Thayaparan, L. J. Stankovi ´c, and I. Djurovi ´c, “Micro-doppler-based target detection and feature extraction in indoor and outdoor environ- ments,”J. Franklin Inst., vol. 345, no. 6, pp. 700–722, 2008

    2008

  47. [47]

    TDOA-based optical wireless indoor localization using LED ceiling lamps,

    S.-Y . Jung, S. Hann, and C.-S. Park, “TDOA-based optical wireless indoor localization using LED ceiling lamps,”IEEE Trans. Consum. Electron., vol. 57, no. 4, pp. 1592–1597, 2011

    2011

  48. [48]

    Recent advances in indoor localization: A survey on theoretical approaches and applications,

    A. Yassin, Y . Nasser, M. Awad, A. Al-Dubai, R. Liu, C. Yuen, R. Raulefs, and E. Aboutanios, “Recent advances in indoor localization: A survey on theoretical approaches and applications,”IEEE Commun. Surveys Tuts., vol. 19, no. 2, pp. 1327–1346, 2016

    2016

  49. [49]

    Joint azimuth, elevation, and delay estimation for 3-D indoor localization,

    F. Wen, P. Liu, H. Wei, Y . Zhang, and R. C. Qiu, “Joint azimuth, elevation, and delay estimation for 3-D indoor localization,”IEEE Trans. Veh. Technol., vol. 67, no. 5, pp. 4248–4261, 2018

    2018

  50. [50]

    PiLoc: A self-calibrating participa- tory indoor localization system,

    C. Luo, H. Hong, and M. C. Chan, “PiLoc: A self-calibrating participa- tory indoor localization system,” inProc. Int. Symp. Inf. Process. Sensor Netw. (IPSN), 2014, pp. 143–153

    2014

  51. [51]

    Large-scale WiFi indoor localization via extreme learning machine,

    J. Zhang, J. Sun, H. Wang, W. Xiao, and L. Tan, “Large-scale WiFi indoor localization via extreme learning machine,” inProc. Chin. Control Conf. (CCC), 2017, pp. 4115–4120

    2017

  52. [52]

    The microsoft indoor localization competition: Experiences and lessons learned,

    D. Lymberopoulos and J. Liu, “The microsoft indoor localization competition: Experiences and lessons learned,”IEEE Signal Process. Mag., vol. 34, no. 5, pp. 125–140, 2017

    2017

  53. [53]

    Physical- statistical modeling of dynamic indoor power delay profiles,

    E. Vinogradov, A. Bamba, W. Joseph, and C. Oestges, “Physical- statistical modeling of dynamic indoor power delay profiles,”IEEE Trans. Wireless Commun., vol. 16, no. 10, pp. 6493–6502, 2017

    2017

  54. [54]

    Learn- ing to localize: A 3D CNN approach to user positioning in massive MIMO-OFDM systems,

    C. Wu, X. Yi, W. Wang, L. You, Q. Huang, X. Gao, and Q. Liu, “Learn- ing to localize: A 3D CNN approach to user positioning in massive MIMO-OFDM systems,”IEEE Trans. Wireless Commun., vol. 20, no. 7, pp. 4556–4570, 2021

    2021

  55. [55]

    Channel charting: Locating users within the radio environment using channel state information,

    C. Studer, S. Medjkouh, E. Gonultas ¸, T. Goldstein, and O. Tirkkonen, “Channel charting: Locating users within the radio environment using channel state information,”IEEE Access, vol. 6, pp. 47 682–47 698, 2018

    2018

  56. [56]

    Channel charting in real-world coordinates with distributed MIMO,

    S. Taner, V . Palhares, and C. Studer, “Channel charting in real-world coordinates with distributed MIMO,”IEEE Trans. Wireless Commun., 2025

    2025

  57. [57]

    Indoor localization without the pain,

    K. Chintalapudi, A. Padmanabha Iyer, and V . N. Padmanabhan, “Indoor localization without the pain,” inProc. ACM Int. Conf. Mobile Comput. Netw. (MobiCom), 2010, pp. 173–184

    2010

  58. [58]

    Smartphones based crowdsourcing for indoor localization,

    C. Wu, Z. Yang, and Y . Liu, “Smartphones based crowdsourcing for indoor localization,”IEEE Trans. Mobile Comput., vol. 14, no. 2, pp. 444–457, 2014

    2014

  59. [59]

    Secure and reliable indoor localization based on multitask collaborative learning for large-scale buildings,

    C. Wang, J. Luo, X. Liu, and X. He, “Secure and reliable indoor localization based on multitask collaborative learning for large-scale buildings,”IEEE Internet Things J., vol. 9, no. 22, pp. 22 291–22 303, 2021

    2021

  60. [60]

    The wisdom of 1,170 teams: Lessons and experiences from a large indoor localization competition,

    Y . Hu, X. Fan, Z. Yin, F. Qian, Z. Ji, Y . Shu, Y . Han, Q. Xu, J. Liu, and P. Bahl, “The wisdom of 1,170 teams: Lessons and experiences from a large indoor localization competition,” inProc. ACM Int. Conf. Mobile Comput. Netw. (MobiCom), 2023, pp. 1–15

    2023

  61. [61]

    A structured self-attentive sentence embedding,

    Z. Lin, M. Feng, C. N. dos Santos, M. Yu, B. Xiang, B. Zhou, and Y . Bengio, “A structured self-attentive sentence embedding,” inProc. Int. Conf. Learn. Represent. (ICLR), 2017

    2017

  62. [62]

    CSI-based fingerprinting for indoor localization: A deep learning approach,

    X. Wang, L. Gao, S. Mao, and S. Pandey, “CSI-based fingerprinting for indoor localization: A deep learning approach,”IEEE Trans. Veh. Technol., vol. 66, no. 1, pp. 763–776, 2016

    2016

  63. [63]

    Deep convolutional neural networks for indoor localization with CSI images,

    X. Wang, X. Wang, and S. Mao, “Deep convolutional neural networks for indoor localization with CSI images,”IEEE Trans. Netw. Sci. Eng., vol. 7, no. 1, pp. 316–327, 2018

    2018

  64. [64]

    An improved k-nearest- neighbor indoor localization method based on spearman distance,

    Y . Xie, Y . Wang, A. Nallanathan, and L. Wang, “An improved k-nearest- neighbor indoor localization method based on spearman distance,”IEEE Signal Process. Lett., vol. 23, no. 3, pp. 351–355, 2016

    2016

  65. [65]

    Localization in wireless sensor networks based on support vector machines,

    D. A. Tran and T. Nguyen, “Localization in wireless sensor networks based on support vector machines,”IEEE Trans. Parallel Distrib. Syst., vol. 19, no. 7, pp. 981–994, 2008

    2008

  66. [66]

    Transform- ing indoor localization: Advanced transformer architecture for NLOS- dominated wireless environments with distributed sensors,

    S. Masrur, J.-F. Cheng, A. R. Khamesi, and ˙I. G ¨uvenc ¸, “Transform- ing indoor localization: Advanced transformer architecture for NLOS- dominated wireless environments with distributed sensors,”IEEE Trans. Mach. Learn. Commun. Netw., vol. 4, pp. 161–177, 2025

    2025

  67. [67]

    TransAoA: Transformer-based angle of arrival estimation for BLE indoor localiza- tion,

    W. Wu, D. Zhou, L. Shen, Z. Zhao, C. Li, and G. Q. Huang, “TransAoA: Transformer-based angle of arrival estimation for BLE indoor localiza- tion,”IEEE Trans. Instrum. Meas., vol. 74, pp. 1–12, 2025

    2025

  68. [68]

    Temporal self-supervised learning for RSSI-based indoor localization,

    J. Paulavi ˇcius, S. Jardak, R. McConville, R. Piechocki, and R. Santos- Rodriguez, “Temporal self-supervised learning for RSSI-based indoor localization,” inProc. IEEE Int. Conf. Commun. (ICC), 2022, pp. 3046– 3051

    2022

  69. [69]

    Self-supervised and invariant rep- resentations for wireless localization,

    A. Salihu, M. Rupp, and S. Schwarz, “Self-supervised and invariant rep- resentations for wireless localization,”IEEE Trans. Wireless Commun., vol. 23, no. 8, pp. 8281–8296, 2024

    2024

  70. [70]

    Sentence-BERT: Sentence embeddings using siamese BERT-networks,

    N. Reimers and I. Gurevych, “Sentence-BERT: Sentence embeddings using siamese BERT-networks,” inProc. Conf. Empir. Methods Nat. Lang. Process. Int. Joint Conf. Nat. Lang. Process. (EMNLP-IJCNLP), 2019, pp. 3982–3992

    2019

  71. [71]

    Attention is all you need,

    A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, Ł. Kaiser, and I. Polosukhin, “Attention is all you need,” inAdv. Neural Inf. Process. Syst. (NeurIPS), vol. 30, 2017

    2017

  72. [72]

    Self-attention graph pooling,

    J. Lee, I. Lee, and J. Kang, “Self-attention graph pooling,” inProc. Int. Conf. Mach. Learn. (ICML), 2019, pp. 3734–3743

    2019

  73. [73]

    Pooling Methods in Deep Neural Networks, a Review

    H. Gholamalinezhad and H. Khosravi, “Pooling methods in deep neural networks: A review,”arXiv preprint arXiv:2009.07485, 2020

    work page Pith review arXiv 2009

  74. [74]

    CSI-based positioning in massive MIMO systems using convolutional neural networks,

    S. De Bast, A. P. Guevara, and S. Pollin, “CSI-based positioning in massive MIMO systems using convolutional neural networks,” inProc. IEEE Veh. Technol. Conf. (VTC-Spring), 2020, pp. 1–5

    2020

  75. [75]

    DyLoc: Dynamic localiza- tion for massive MIMO using predictive recurrent neural networks,

    F. Hejazi, K. Vuckovic, and N. Rahnavard, “DyLoc: Dynamic localiza- tion for massive MIMO using predictive recurrent neural networks,” in Proc. IEEE INFOCOM, 2021, pp. 1–9

    2021

  76. [76]

    Finding the missing data: A BERT-inspired approach against package loss in wireless sensing,

    Z. Zhao, T. Chen, F. Meng, H. Li, X. Li, and G. Zhu, “Finding the missing data: A BERT-inspired approach against package loss in wireless sensing,” inProc. IEEE INFOCOM Workshops, 2024, pp. 1–6

    2024

  77. [77]

    LoRA: Low-rank adaptation of large language models,

    E. J. Hu, Y . Shen, P. Wallis, Z. Allen-Zhu, Y . Li, S. Wang, L. Wang, W. Chenet al., “LoRA: Low-rank adaptation of large language models,” inProc. Int. Conf. Learn. Represent. (ICLR), 2022

    2022

  78. [78]

    Overcoming catastrophic forgetting in neural networks,

    J. Kirkpatrick, R. Pascanu, N. Rabinowitz, J. Veness, G. Desjardins, A. A. Rusu, K. Milan, J. Quan, T. Ramalho, A. Grabska-Barwinska et al., “Overcoming catastrophic forgetting in neural networks,”Proc. Natl. Acad. Sci. U.S.A., vol. 114, no. 13, pp. 3521–3526, 2017

    2017

  79. [79]

    Verification: Accuracy evaluation of WiFi fine time measurements on an open platform,

    M. Ibrahim, H. Liu, M. Jawahar, V . Nguyen, M. Gruteser, R. Howard, B. Yu, and F. Bai, “Verification: Accuracy evaluation of WiFi fine time measurements on an open platform,” inProc. ACM Int. Conf. Mobile Comput. Netw. (MobiCom), 2018, pp. 417–427

    2018

  80. [80]

    Deep sets,

    M. Zaheer, S. Kottur, S. Ravanbakhsh, B. Poczos, R. R. Salakhutdinov, and A. J. Smola, “Deep sets,” inAdv. Neural Inf. Process. Syst. (NeurIPS), vol. 30, 2017

    2017