Pith. sign in

REVIEW 4 major objections 3 minor 3 cited by

NGC 891 hosts a thick molecular disk (FWHM ~1.1 kpc) that can contain up to 27% of the galaxy’s total molecular gas, lifted by ordinary star-formation feedback.

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-4.5

2026-07-15 13:35 UTC pith:JGIV626Q

load-bearing objection Only the abstract is real for NGC 891; the body is the wrong paper, so the thick-disk CO claim is interesting but still unauditable. the 4 major comments →

arxiv 2603.06913 v2 pith:JGIV626Q submitted 2026-03-06 astro-ph.GA

Imaging the disk-halo interface of NGC 891: a 2.7 kpc-thick molecular gas disk

D. Jim\'enez-L\'opez , S. Garc\'ia-Burillo , M. Querejeta , A. Usero , P. Tarr\'io This is my paper
classification astro-ph.GA
keywords NGC 891molecular gasextraplanar gasthick diskgalactic fountainCO emissionstar-formation feedbackdisk-halo interface
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.

Edge-on spiral NGC 891 is used as a laboratory to measure how far molecular gas extends above the midplane. New IRAM 30 m CO(2-1) maps, after careful subtraction of residual error-beam sidelobes, show that the vertical distribution is best fit by two Gaussians: a bright thin disk (FWHM ~360 pc) plus a fainter thick component (FWHM ~1.1 kpc). Statistically significant emission reaches 1.3–1.4 kpc, and the thick component may hold as much as 27% of the total molecular mass. The result implies that ordinary (non-starburst) star-formation feedback can loft large amounts of molecular gas into the disk–halo interface, consistent with a galactic-fountain cycle that also feeds the multiphase halo.

Core claim

After residual error-beam removal, the vertical CO(2-1) distribution of NGC 891 is a two-component Gaussian: thin-disk FWHM ≃ 360 pc and thick-disk FWHM ≃ 1.1 kpc, with emission detected to 1.3–1.4 kpc and the thick component contributing up to 27% of the galaxy’s molecular gas mass; this shows that SF-driven feedback in a normal spiral can lift significant molecular material into the halo.

What carries the argument

Two-component Gaussian decomposition of the error-beam-cleaned IRAM 30 m CO(2-1) cube, which isolates the faint high-|z| thick-disk emission from residual single-dish sidelobes.

Load-bearing premise

The residual error-beam contribution from the IRAM 30 m telescope has been estimated and subtracted well enough that the faint high-|z| CO and the derived 27% thick-disk mass fraction are real, not leftover sidelobe artifacts.

What would settle it

An independent interferometric CO map (e.g., ALMA or NOEMA) of the same high-|z| regions that recovers the same thick-disk flux and kinematics without single-dish error beams would confirm or refute the 1.1 kpc component and its mass fraction.

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

If this is right

  • Ordinary star-forming spirals, not only starbursts, can supply a substantial molecular reservoir to the multiphase disk–halo interface.
  • Galactic-fountain models must include a cold molecular phase that can survive or reform at |z| ~1 kpc.
  • Extraplanar molecular gas can contribute a non-negligible fraction of a galaxy’s total H2 budget and therefore of its baryonic mass inventory.
  • Comparisons with HI, DIG and dust maps become more meaningful once the molecular thick-disk mass and kinematics are known.

Where Pith is reading between the lines

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

  • If the thick molecular layer is common, edge-on surveys with modern single-dish or interferometric CO mapping may systematically raise the cosmic molecular-gas mass density once high-|z| emission is properly recovered.
  • The survival of CO at 1 kpc height implies either rapid reformation of molecules in fountain gas or shielding by dust that also reaches those altitudes, both of which can be tested with multi-line and continuum follow-up.
  • The same error-beam-cleaning approach applied to other nearby edge-ons could reveal whether the ~25% thick-disk fraction is typical or special to NGC 891.

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

4 major / 3 minor

Summary. The abstract claims that new IRAM 30 m CO(2-1) maps of the edge-on spiral NGC 891 (D = 9.5 Mpc), after careful residual error-beam subtraction, reveal a two-component vertical molecular structure: a bright thin disk (deconvolved FWHM ≃ 360 pc) and a fainter thick disk (deconvolved FWHM ≃ 1.1 kpc), with statistically significant emission to |z| ≃ 1.3–1.4 kpc and up to ~27% of the galaxy’s molecular mass in the thick component. The authors interpret this as SF-driven galactic-fountain transport of molecular gas in a non-starburst system, and compare with literature HI, DIG, and dust maps. The title phrases the result as a “2.7 kpc-thick molecular gas disk.”

Significance. If the high-|z| CO is intrinsic and the mass fraction robust, the result would be an important empirical constraint on multiphase disk–halo exchange: it would show that ordinary star-formation feedback can loft a substantial molecular reservoir well above the midplane, not only in starbursts. That would tighten links among molecular, atomic, ionized, and dust extraplanar components in a classic edge-on laboratory and would matter for baryon-cycle and fountain models. The abstract’s emphasis on error-beam control is the right observational priority for single-dish high-|z| work. Those strengths cannot be credited as demonstrated until the actual methods, residual maps, and mass conversion are available for audit.

major comments (4)
  1. The package under review does not contain the NGC 891 manuscript. The supplied FULL MANUSCRIPT TEXT is the unrelated robotics paper SysNav (arXiv:2603.06914). Residual error-beam model, subtraction residuals, vertical profiles, kinematic analysis, X_CO / line-ratio assumptions, and mass-fraction derivation for NGC 891 are therefore unavailable. No load-bearing claim of the abstract can be checked against methods, figures, or tables. Proper refereeing of 2603.06913 is not possible from this package.
  2. Abstract (error-beam paragraph and two-component fit): the central empirical claim—that residual high-|z| CO is intrinsic extraplanar gas, not sidelobe artifact—rests entirely on “a careful method to estimate and remove the residual contribution of the error beam.” Single-dish error beams routinely produce faint high-|z| wings. Without residual maps, the error-beam model, pre-/post-subtraction profiles, and a quantitative residual budget relative to the thick-disk amplitude, the deconvolved FWHM ≃ 1.1 kpc, the 1.3–1.4 kpc detection, and the ≤27% thick-disk mass fraction remain unauditable and could be instrumental.
  3. Title vs abstract thickness: the title states a “2.7 kpc-thick molecular gas disk,” while the abstract reports a thick-component deconvolved FWHM ≃ 1.1 kpc and significant emission only to 1.3–1.4 kpc. These numbers are not interchangeable without an explicit definition (FWHM, 2σ full width, full vertical extent, or something else). Until the correct manuscript defines “2.7 kpc-thick,” the headline claim is inconsistent with the abstract’s quantitative results and risks overstating the vertical scale.
  4. Abstract (mass fraction “up to 27%”): the thick-disk mass fraction depends on the two-Gaussian decomposition, the adopted α_CO (or X_CO), the CO(2-1)/CO(1-0) ratio, and how the two mapped 6×6 kpc NE patches plus center are extrapolated to the whole galaxy. None of these choices, uncertainties, or extrapolation steps are inspectable in the supplied text. The “up to 27%” figure is therefore not yet a demonstrated galaxy-wide result.
minor comments (3)
  1. Abstract: “statistically significant CO(2-1) emission” at 1.3–1.4 kpc needs an explicit significance criterion (e.g., per-channel S/N, integrated intensity threshold, or false-detection rate) once the correct methods section is available.
  2. Abstract: comparison to HI, Hα DIG, and dust is promised but not quantified here; the full paper should state which literature maps, resolutions, and vertical scales are used so the multiphase comparison is reproducible.
  3. Distance D = 9.5 Mpc is stated without reference; a citation and sensitivity of pc-scale FWHMs to D should appear in the methods of the correct manuscript.

Circularity Check

0 steps flagged

No circularity: purely empirical two-Gaussian fit and mass fraction from IRAM CO(2-1) data after error-beam subtraction; no derivation reduces to its own inputs by construction.

full rationale

The paper's central claims (thin/thick disk FWHMs after deconvolution, detection to 1.3-1.4 kpc, thick-component mass fraction up to 27 %, galactic-fountain interpretation) are obtained by mapping, residual error-beam removal, vertical profile fitting, and mass conversion applied to new IRAM 30 m CO(2-1) cubes, then compared with literature HI/Hα/dust maps. None of these steps is definitional of the result, none renames a fitted parameter as a prediction, and no uniqueness theorem or self-citation chain forces the outcome. Standard observational practice (Gaussian decomposition of a cleaned cube, X_CO conversion) does not constitute circular reasoning. The supplied full-text body is an unrelated robotics manuscript, so residual maps cannot be inspected, but that is an auditability/correctness issue, not circularity. Score 0 is therefore required under the hard rules.

Axiom & Free-Parameter Ledger

3 free parameters · 4 axioms · 0 invented entities

Abstract-only review of an observational CO study. Load-bearing items are standard astrophysical conversion/assumptions and the paper’s own error-beam treatment; no new physical entities are postulated. Free parameters are those that convert CO intensity into H2 mass and that define the two-Gaussian vertical model.

free parameters (3)
  • CO-to-H2 conversion factor (X_CO or α_CO) and CO(2-1)/CO(1-0) line ratio
    Thick-disk mass fraction (up to 27%) depends on how CO(2-1) intensity is converted to molecular mass; values are typically assumed or taken from literature, not measured in the abstract.
  • Two-component Gaussian vertical model parameters (amplitudes, FWHMs, possibly centers)
    The thin/thick decomposition and deconvolved FWHMs (~360 pc, ~1.1 kpc) are fitted to the observed vertical profile after beam deconvolution.
  • Error-beam residual model / subtraction amplitude
    Abstract emphasizes careful estimation and removal of residual error-beam contribution; any free scale in that model directly affects faint high-|z| flux.
axioms (4)
  • domain assumption CO(2-1) emission is a faithful tracer of molecular hydrogen in both the thin and thick disks of NGC 891.
    Standard ISM assumption; excitation or abundance changes at high |z| could bias extent and mass.
  • ad hoc to paper After error-beam removal, residual high-|z| CO is intrinsic extraplanar emission, not instrumental artifact.
    Central to the detection height and thick-disk mass; abstract presents this as a careful method but provides no validation details here.
  • domain assumption Distance D = 9.5 Mpc for NGC 891 is accurate enough for linear scales (pc/kpc).
    Stated in abstract; scales FWHM and height linearly with D.
  • domain assumption Extraplanar molecular gas is produced by SF-driven galactic fountain rather than accretion or other channels as the primary interpretation.
    Interpretive framework stated in abstract; comparison to HI/DIG/dust is cited as supporting context.

pith-pipeline@v1.1.0-grok45 · 18041 in / 3026 out tokens · 25983 ms · 2026-07-15T13:35:12.547383+00:00 · methodology

0 comments
read the original abstract

Halos surrounding spiral galaxies act as the bridges connecting the galactic disk and the intergalactic medium (IGM). They host a significant fraction of the baryonic mass in the Universe, and feedback from star formation (SF) or active galactic nuclei (AGN) likely plays an important role in regulating this vertical baryonic component. Despite its importance, the contribution of extraplanar molecular gas remains poorly understood. We aim to characterize the vertical extent and the kinematics of molecular gas traced by CO(2-1) emission in the nearby (D = 9.5 Mpc) spiral galaxy NGC 891, one of the best studied edge-on galaxies. We also compare our results with HI, H$\alpha$-traced DIG and dust maps from the literature. Our analysis is based on new CO(2-1) observations of NGC 891 obtained with the IRAM 30m telescope. We mapped two 6 kpc $\times$ 6 kpc regions on the northeastern side and the area surrounding the galactic center. We apply a careful method to estimate and remove the residual contribution of the error beam to the CO cube. The vertical extent of the molecular gas is best described by a two-component Gaussian fit, consisting of a bright thin disk component (deconvolved FWHM $\simeq$ 360 pc) and a fainter thick disk component (deconvolved FWHM $\simeq$ 1.1 kpc). Statistically significant CO(2-1) emission is detected up to 1.3-1.4 kpc above the disk midplane. We estimate that the thick molecular disk component contains up to 27% of the total molecular gas mass of the galaxy. Our results demonstrate that SF-driven feedback in a non-starburst galaxy can lift significant amounts of molecular gas to large vertical distances. We interpret the presence of extraplanar molecular gas in NGC 891 in the framework of a galactic fountain scenario, in which material is expelled from star-forming regions and transported toward the outer halo.

discussion (0)

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

Forward citations

Cited by 3 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. The Edge-on Galaxies in the DESI survey (EGIDE): sample building and photometry

    astro-ph.GA 2026-06 unverdicted novelty 6.0

    The EGIDE project releases a tenfold larger catalogue of edge-on galaxies with griz photometry, stellar masses, redshifts and star formation rates, finding that red-sequence galaxies are thicker than blue-cloud ones a...

  2. The Edge-on Galaxies in the DESI survey (EGIDE): sample building and photometry

    astro-ph.GA 2026-06 conditional novelty 6.0

    EGIDE provides 149,215 edge-on galaxy candidates from DESI DR10 with homogeneous photometry, masses, and redshifts, ten times larger than EGIPS.

  3. A CO detection in the off-plane region of the edge-on galaxy NGC 4565 with the Nobeyama 45-m telescope

    astro-ph.GA 2026-07 conditional novelty 5.0

    CO emission is detected ~0.85 kpc above the disk of NGC 4565 at three positions, with a high-velocity component whose kinetic energy (~10^54 erg) is difficult to explain by stellar feedback alone, suggesting possible ...

Reference graph

Works this paper leans on

44 extracted references · 3 linked inside Pith · cited by 2 Pith papers

  1. [1]

    Spatialvlm: Endowing vision-language models with spatial reasoning capabilities,

    B. Chen, Z. Xu, S. Kirmani, B. Ichter, D. Sadigh, L. Guibas, and F. Xia, “Spatialvlm: Endowing vision-language models with spatial reasoning capabilities,” in����������� �� ��� �������� ���������� �� �������� ������ ��� ������� �����������, 2024, pp. 14 455–14 465

    2024

  2. [2]

    Gpt4scene: Understand 3d scenes from videos with vision-language models,

    Z. Qi, Z. Zhang, Y . Fang, J. Wang, and H. Zhao, “Gpt4scene: Understand 3d scenes from videos with vision-language models,” ����� �������� ����������������, 2025

    2025

  3. [3]

    Dd-ppo: Learning near-perfect pointgoal navigators from 2.5 billion frames,

    E. Wijmans, A. Kadian, A. Morcos, S. Lee, I. Essa, D. Parikh, M. Savva, and D. Batra, “Dd-ppo: Learning near-perfect pointgoal navigators from 2.5 billion frames,”����� �������� ����������������, 2019

    2019

  4. [4]

    Auxiliary tasks and explo- ration enable objectgoal navigation,

    J. Ye, D. Batra, A. Das, and E. Wijmans, “Auxiliary tasks and explo- ration enable objectgoal navigation,” in����������� �� ��� �������� ������������� ���������� �� �������� ������, 2021, pp. 16 117–16 126

    2021

  5. [5]

    Habitat- web: Learning embodied object-search strategies from human demon- strations at scale,

    R. Ramrakhya, E. Undersander, D. Batra, and A. Das, “Habitat- web: Learning embodied object-search strategies from human demon- strations at scale,” in����������� �� ��� �������� ���������� �� �������� ������ ��� ������� �����������, 2022, pp. 5173–5183

    2022

  6. [6]

    Pirlnav: Pretraining with imitation and rl finetuning for objectnav,

    R. Ramrakhya, D. Batra, E. Wijmans, and A. Das, “Pirlnav: Pretraining with imitation and rl finetuning for objectnav,” in����������� �� ��� �������� ���������� �� �������� ������ ��� ������� �����������, 2023, pp. 17 896–17 906

    2023

  7. [7]

    Uni-navid: A video-based vision-language- action model for unifying embodied navigation tasks,

    J. Zhang, K. Wang, S. Wang, M. Li, H. Liu, S. Wei, Z. Wang, Z. Zhang, and H. Wang, “Uni-navid: A video-based vision-language- action model for unifying embodied navigation tasks,”����� �������� ����������������, 2024

    2024

  8. [8]

    Object goal navigation using goal-oriented semantic exploration,

    D. S. Chaplot, D. P. Gandhi, A. Gupta, and R. R. Salakhutdinov, “Object goal navigation using goal-oriented semantic exploration,” �������� �� ������ ����������� ���������� �������, vol. 33, pp. 4247–4258, 2020

    2020

  9. [9]

    Poni: Potential functions for objectgoal navigation with interaction-free learning,

    S. K. Ramakrishnan, D. S. Chaplot, Z. Al-Halah, J. Malik, and K. Grauman, “Poni: Potential functions for objectgoal navigation with interaction-free learning,” in����������� �� ��� �������� ���������� �� �������� ������ ��� ������� �����������, 2022, pp. 18 890–18 900

    2022

  10. [10]

    3d-aware object goal navigation via simultaneous exploration and identification,

    J. Zhang, L. Dai, F. Meng, Q. Fan, X. Chen, K. Xu, and H. Wang, “3d-aware object goal navigation via simultaneous exploration and identification,” 2023. [Online]. Available: https: //arxiv.org/abs/2212.00338

    Pith/arXiv arXiv 2023

  11. [11]

    L3mvn: Leveraging large language models for visual target navigation,

    B. Yu, H. Kasaei, and M. Cao, “L3mvn: Leveraging large language models for visual target navigation,” in���� �������� ������������� ���������� �� ����������� ������ ��� ������� ������. IEEE, Oct. 2023, p. 3554–3560. [Online]. Available: http://dx.doi.org/10.1109/ IROS55552.2023.10342512

    arXiv 2023

  12. [12]

    Esc: Exploration with soft commonsense constraints for zero-shot object navigation,

    K. Zhou, K. Zheng, C. Pryor, Y . Shen, H. Jin, L. Getoor, and X. E. Wang, “Esc: Exploration with soft commonsense constraints for zero-shot object navigation,” 2023. [Online]. Available: https://arxiv.org/abs/2301.13166

    Pith/arXiv arXiv 2023

  13. [13]

    Vlfm: Vision- language frontier maps for zero-shot semantic navigation,

    N. Yokoyama, S. Ha, D. Batra, J. Wang, and B. Bucher, “Vlfm: Vision- language frontier maps for zero-shot semantic navigation,” in���� ���� ������������� ���������� �� �������� ��� ���������� ������. IEEE, 2024, pp. 42–48

    2024

  14. [14]

    Sg-nav: Online 3d scene graph prompting for llm-based zero-shot object navigation,

    H. Yin, X. Xu, Z. Wu, J. Zhou, and J. Lu, “Sg-nav: Online 3d scene graph prompting for llm-based zero-shot object navigation,”�������� �� ������ ����������� ���������� �������, vol. 37, pp. 5285–5307, 2024

    2024

  15. [15]

    Move to understand a 3d scene: Bridging visual grounding and exploration for efficient and versatile embodied navigation,

    Z. Zhu, X. Wang, Y . Li, Z. Zhang, X. Ma, Y . Chen, B. Jia, W. Liang, Q. Yu, Z. Deng�� ���, “Move to understand a 3d scene: Bridging visual grounding and exploration for efficient and versatile embodied navigation,” in����������� �� ��� �������� ������������� ���������� �� �������� ������, 2025, pp. 8120–8132

    2025

  16. [16]

    Apex- nav: An adaptive exploration strategy for zero-shot object navigation with target-centric semantic fusion,

    M. Zhang, Y . Du, C. Wu, J. Zhou, Z. Qi, J. Ma, and B. Zhou, “Apex- nav: An adaptive exploration strategy for zero-shot object navigation with target-centric semantic fusion,”����� �������� ����������������, 2025

    2025

  17. [17]

    Blip: Bootstrapping language- image pre-training for unified vision-language understanding and generation,

    J. Li, D. Li, C. Xiong, and S. Hoi, “Blip: Bootstrapping language- image pre-training for unified vision-language understanding and generation,” 2022. [Online]. Available: https://arxiv.org/abs/2201. 12086

    2022

  18. [18]

    Learning transferable visual models from natural language supervision,

    A. Radford, J. W. Kim, C. Hallacy, A. Ramesh, G. Goh, S. Agarwal, G. Sastry, A. Askell, P. Mishkin, J. Clark�� ���, “Learning transferable visual models from natural language supervision,” in������������� ���������� �� ������� ��������. PmLR, 2021, pp. 8748–8763

    2021

  19. [19]

    Gpt-4v (ision) is a human-aligned evaluator for text- to-3d generation,

    T. Wu, G. Yang, Z. Li, K. Zhang, Z. Liu, L. Guibas, D. Lin, and G. Wetzstein, “Gpt-4v (ision) is a human-aligned evaluator for text- to-3d generation,” in����������� �� ��� �������� ���������� �� �������� ������ ��� ������� �����������, 2024, pp. 22 227–22 238

    2024

  20. [20]

    Gem- ini: a family of highly capable multimodal models,

    G. Team, R. Anil, S. Borgeaud, J.-B. Alayrac, J. Yu, R. Soricut, J. Schalkwyk, A. M. Dai, A. Hauth, K. Millican�� ���, “Gem- ini: a family of highly capable multimodal models,”����� �������� ����������������, 2023

    2023

  21. [21]

    Instructnav: Zero-shot system for generic instruction navigation in unexplored environment,

    Y . Long, W. Cai, H. Wang, G. Zhan, and H. Dong, “Instructnav: Zero-shot system for generic instruction navigation in unexplored environment,” in��� ������ ���������� �� ����� ��������

  22. [22]

    Do vision-language models represent space and how? eval- uating spatial frame of reference under ambiguities,

    Z. Zhang, F. Hu, J. Lee, F. Shi, P. Kordjamshidi, J. Chai, and Z. Ma, “Do vision-language models represent space and how? eval- uating spatial frame of reference under ambiguities,”����� �������� ����������������, 2024

    2024

  23. [23]

    Hydra: A real-time spatial perception system for 3D scene graph construction and optimization,

    N. Hughes, Y . Chang, and L. Carlone, “Hydra: A real-time spatial perception system for 3D scene graph construction and optimization,” 2022

    2022

  24. [24]

    How to not train your dragon: Training-free embodied object goal navigation with semantic frontiers,

    J. Chen, G. Li, S. Kumar, B. Ghanem, and F. Yu, “How to not train your dragon: Training-free embodied object goal navigation with semantic frontiers,”����� �������� ����������������, 2023

    2023

  25. [25]

    Cows on pasture: Baselines and benchmarks for language-driven zero-shot object navigation,

    S. Y . Gadre, M. Wortsman, G. Ilharco, L. Schmidt, and S. Song, “Cows on pasture: Baselines and benchmarks for language-driven zero-shot object navigation,” in����������� �� ��� �������� ���������� �� �������� ������ ��� ������� �����������, 2023, pp. 23 171–23 181

    2023

  26. [26]

    Navigating to objects in the real world,

    T. Gervet, S. Chintala, D. Batra, J. Malik, and D. S. Chaplot, “Navigating to objects in the real world,”������� ��������, vol. 8, no. 79, p. eadf6991, 2023

    2023

  27. [27]

    Open scene graphs for open-world object- goal navigation,

    J. Loo, Z. Wu, and D. Hsu, “Open scene graphs for open-world object- goal navigation,” in����� �������� �� ��������������� ������ ��� ���������� ��� ������������ �� ���� ����

  28. [28]

    Etpnav: Evolving topological planning for vision-language navigation in continuous environments,

    D. An, H. Wang, W. Wang, Z. Wang, Y . Huang, K. He, and L. Wang, “Etpnav: Evolving topological planning for vision-language navigation in continuous environments,”���� ������������ �� ������� �������� ��� ������� ������������, 2024

    2024

  29. [29]

    Strive: Structured representation integrating vlm reasoning for efficient object navigation,

    H. Zhu, Z. Li, Z. Liu, W. Wang, J. Zhang, J. Francis, and J. Oh, “Strive: Structured representation integrating vlm reasoning for efficient object navigation,”����� �������� ����������������, 2025

    2025

  30. [30]

    V oronav: V oronoi-based zero-shot object navigation with large language model,

    P. Wu, Y . Mu, B. Wu, Y . Hou, J. Ma, S. Zhang, and C. Liu, “V oronav: V oronoi-based zero-shot object navigation with large language model,” ����� �������� ����������������, 2024

    2024

  31. [31]

    Hierarchical open-vocabulary 3d scene graphs for language-grounded robot navigation,

    A. Werby, C. Huang, M. B ¨uchner, A. Valada, and W. Burgard, “Hierarchical open-vocabulary 3d scene graphs for language-grounded robot navigation,” in����� �������� �� ��������������� ������ ��� ���������� ��� ������������ �� ���� ����, 2024

    2024

  32. [32]

    Yolo- world: Real-time open-vocabulary object detection,

    T. Cheng, L. Song, Y . Ge, W. Liu, X. Wang, and Y . Shan, “Yolo- world: Real-time open-vocabulary object detection,” in����������� �� ��� �������� ���������� �� �������� ������ ��� ������� �����������, 2024, pp. 16 901–16 911

    2024

  33. [33]

    Yoloe: Real-time seeing anything,

    A. Wang, L. Liu, H. Chen, Z. Lin, J. Han, and G. Ding, “Yoloe: Real-time seeing anything,”����� �������� ����������������, 2025

    2025

  34. [34]

    Sam 2: Segment anything in images and videos,

    N. Ravi, V . Gabeur, Y .-T. Hu, R. Hu, C. Ryali, T. Ma, H. Khedr, R. R ¨adle, C. Rolland, L. Gustafson, E. Mintun, J. Pan, K. V . Alwala, N. Carion, C.-Y . Wu, R. Girshick, P. Doll ´ar, and C. Feichtenhofer, “Sam 2: Segment anything in images and videos,”����� �������� ����������������, 2024. [Online]. Available: https://arxiv.org/abs/2408.00714

    Pith/arXiv arXiv 2024

  35. [35]

    Tare: A hierarchical framework for efficiently exploring complex 3d environments

    C. Cao, H. Zhu, H. Choset, and J. Zhang, “Tare: A hierarchical framework for efficiently exploring complex 3d environments.” in ��������� ������� ��� �������, vol. 5, 2021, p. 2

    2021

  36. [36]

    Autonomous exploration development environment and the planning algorithms,

    C. Cao, H. Zhu, F. Yang, Y . Xia, H. Choset, J. Oh, and J. Zhang, “Autonomous exploration development environment and the planning algorithms,” in���� ������������� ���������� �� �������� ��� ��� �������� ������. IEEE, 2022, pp. 8921–8928

    2022

  37. [37]

    Openfmnav: Towards open-set zero- shot object navigation via vision-language foundation models,

    Y . Kuang, H. Lin, and M. Jiang, “Openfmnav: Towards open-set zero- shot object navigation via vision-language foundation models,”����� �������� ����������������, 2024

    2024

  38. [38]

    Trihelper: Zero-shot object navigation with dynamic assistance,

    L. Zhang, Q. Zhang, H. Wang, E. Xiao, Z. Jiang, H. Chen, and R. Xu, “Trihelper: Zero-shot object navigation with dynamic assistance,” in ���� �������� ������������� ���������� �� ����������� ������ ��� ������� ������. IEEE, 2024, pp. 10 035–10 042

    2024

  39. [39]

    Autonomy stack for mecanum wheel platform,

    J. Zhang, “Autonomy stack for mecanum wheel platform,” https: //github.com/jizhang-cmu/autonomy stack mecanum wheel platform, 2024, accessed: 2025-04-29

    2024

  40. [40]

    Habitat challenge 2023,

    K. Yadav, J. Krantz, R. Ramrakhya, S. K. Ramakrishnan, J. Yang, A. Wang, J. Turner, A. Gokaslan, V .-P. Berges, R. Mootaghi, O. Maksymets, A. X. Chang, M. Savva, A. Clegg, D. S. Chaplot, and D. Batra, “Habitat challenge 2023,” https://aihabitat.org/challenge/ 2023/, 2023

    2023

  41. [41]

    Habitat challenge 2022,

    K. Yadav, S. K. Ramakrishnan, A. Gokaslan, O. Maksymets, R. Jain, R. Ramrakhya, A. X. Chang, A. Clegg, M. Savva, E. Undersander, D. S. Chaplot, and D. Batra, “Habitat challenge 2022,” https://aihabitat. org/challenge/2022/, 2022

    2022

  42. [42]

    Habitat 3.0: A co-habitat for humans, avatars and robots,

    X. Puig, E. Undersander, A. Szot, M. D. Cote, R. Partsey, J. Yang, R. Desai, A. W. Clegg, M. Hlavac, T. Min, T. Gervet, V . V ondruˇs, V .-P. Berges, J. Turner, O. Maksymets, Z. Kira, M. Kalakrishnan, J. Malik, D. S. Chaplot, U. Jain, D. Batra, A. Rai, and R. Mottaghi, “Habitat 3.0: A co-habitat for humans, avatars and robots,” 2023

    2023

  43. [43]

    Matterport3d: Learning from rgb-d data in indoor environments,

    A. Chang, A. Dai, T. Funkhouser, M. Halber, M. Niessner, M. Savva, S. Song, A. Zeng, and Y . Zhang, “Matterport3d: Learning from rgb-d data in indoor environments,”������������� ���������� �� �� ������ �����, 2017

    2017

  44. [44]

    Hm3d- ovon: A dataset and benchmark for open-vocabulary object goal navigation,

    N. Yokoyama, R. Ramrakhya, A. Das, D. Batra, and S. Ha, “Hm3d- ovon: A dataset and benchmark for open-vocabulary object goal navigation,” in���� �������� ������������� ���������� �� ����������� ������ ��� ������� ������. IEEE, 2024, pp. 5543–5550

    2024