Pith. sign in

REVIEW 4 major objections 5 minor 23 references

This paper claims that a purely RGB aerial video, aligned to onboard egomotion, can yield metric-scale 3D terrain maps accurate enough to plan safe deployment of a ground robot into hidden spaces like culverts.

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 · deepseek-v4-flash

2026-08-02 18:07 UTC pith:JEEHCXWZ

load-bearing objection A sensible, well-scoped integration of feed-forward RGB reconstruction with egomotion scale grounding for UAV-to-UGV deployment-zone selection, but the metric scale that everything rests on is never actually validated. the 4 major comments →

arxiv 2603.14639 v2 pith:JEEHCXWZ submitted 2026-03-15 cs.RO cs.CV

Seeing Where to Deploy: Metric RGB-Based Traversability Analysis for Aerial-to-Ground Hidden Space Inspection

Seoyoung Lee , Shaekh Mohammad Shithil , Durgakant Pushp , Lantao Liu , Zhangyang Wang This is my paper
classification cs.RO cs.CV
keywords aerial-to-ground deploymenttraversability analysismetric scale recoveryRGB 3D reconstructionUGV deploymenttop-down aerial inspectionsemantic mappingegomotion prior
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 tries to establish that an aerial drone can decide where to safely drop a ground robot without any LiDAR or ground-level views, using only top-down RGB video plus its own motion estimate. The claim is that a feed-forward neural reconstruction (VGGT) produces geometry that is correct up to scale, and that this scale can be recovered by aligning the predicted camera trajectory to the drone's onboard egomotion. If true, this removes the need for heavy depth sensors on the drone and makes metric terrain reasoning—slope, roughness, clearance—possible from ordinary cameras. This matters because hidden infrastructure such as culverts is accessible only from above, and deployment decisions are safety-critical and hard to reverse.

Core claim

The central discovery is that a single global similarity transform, fit by comparing inter-frame camera displacements predicted by VGGT with those from drone egomotion (Umeyama alignment on relative motion), resolves the metric scale ambiguity of monocular RGB reconstruction. With this scale fixed, every point in the reconstructed point cloud, every depth value, and every candidate zone becomes metrically meaningful. The paper then shows that combining these metric geometric features (height, slope, roughness, clearance) with semantic compatibility scores in a confidence-aware BEV grid yields a traversability map that suppresses hazardous regions—such as rocks and structural obstacles—that p

What carries the argument

The load-bearing mechanism is the 'embodied motion prior': a closed-form Umeyama alignment (Equation 6) that fits scale s and rotation R to match predicted VGGT camera displacements ∆C_v against platform egomotion displacements ∆C_p, using multiple temporal strides for robustness. This single fitted (s, R) transforms all predicted geometry and camera poses into metric scale. The second mechanism is a BEV traversability model that fuses linear-penalty slope/roughness/clearance scores with semantic class compatibility, weighted by per-cell reconstruction confidence, yielding a continuous traversability map and a distance-aware zone selection objective.

Load-bearing premise

The core assumption is that VGGT's predicted camera trajectories and depth field, observed only from hovering top-down views, are accurate up to a single global scale—so that one fitted scale fixes the entire scene—when in reality depth errors under low parallax and high altitude may be non-uniform across regions.

What would settle it

Collect a hidden-space scene with both RGB video and dense LiDAR ground truth; run the proposed pipeline to produce a metric point cloud and traversability map; then compute the absolute per-point scale error (predicted depth vs LiDAR depth) and show whether a single global scale factor actually aligns the whole ground surface, or whether per-region residuals exceed the threshold that would flip a safe deployment zone into an unsafe one.

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

If this is right

  • If the metric-scale recovery works as claimed, drones can produce deployment-relevant metric maps (obstacle clearance, slope limits, energy margins) using only a lightweight RGB camera and standard VIO/GPS egomotion, replacing heavier LiDAR payloads.
  • Traversability maps that fuse semantics with geometry are shown to be substantially more accurate (mACC 0.958 vs 0.779 geometric-only) at identifying safe versus unsafe zones, suggesting that conservative, semantics-informed filtering is a reliable way to compensate for reconstruction smoothing.
  • The method generalizes across altitudes from 10 m to over 70 m even with purely top-down views and limited parallax, where classical SfM and depth sensors degrade, implying a broad operating envelope for aerial inspection.
  • Because the framework runs in a single forward pass (VGGT) plus lightweight alignment, it requires far less memory and computation than pairwise reconstruction baselines, making it feasible for onboard deployment on computation-limited UAVs.
  • If scale-grounded, this map can directly feed planners (e.g., the HALO trajectory generation demonstrated) to produce dynamically feasible deployment paths without additional sensing.

Where Pith is reading between the lines

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

  • Editorial inference: The paper does not report absolute metric scale error against LiDAR or surveyed ground truth, only relative trajectory errors after Sim(3) alignment. A reader should suspect that the single global similarity fit may be over-optimistic; translating this to deployment safety would require reporting per-region scale error, especially for ground-level discontinuities seen at obliq
  • Editorial inference: This approach could be transferred beyond hidden-space deployment to general UAV landing-zone selection, perching, or aerial drop of sensor packages, since it only assumes an egomotion signal that most drones already possess.
  • Editorial inference: A natural testable extension is to compute traversability scores from the same RGB pipeline and compare them to scores computed from onboard LiDAR in the same scene; differences would directly quantify the safety margin lost by avoiding dense depth sensing.
  • Editorial inference: The paper's adaptive keyframe strategy for semantic discovery is a practical engineering choice, but its coverage-trigger threshold (ρ_t) is an added parameter; sensitivity to this threshold is not reported, so reproducibility on longer or more complex flights is uncertain.

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 / 5 minor

Summary. The paper proposes a UAV-UGV deployment framework that, from purely top-down RGB video, builds a metric 3D geometric-semantic reconstruction and a confidence-aware BEV traversability map for selecting safe deployment zones. The reconstruction uses VGGT as a feed-forward multi-view backbone; instance-consistent masks are obtained via SAM2; metric scale is recovered by aligning VGGT-predicted camera displacements with onboard egomotion through a closed-form Sim(3) fit (Eqs. (4)-(8)). The traversability score fuses slope, roughness, clearance, semantic compatibility, and reconstruction confidence (Eqs. (16)-(19)), and deployment candidates are selected under a distance-aware objective (Eq. (20)). Experiments compare pose estimation and efficiency against several RGB reconstruction baselines and report traversability accuracy on field-collected hidden-space scenes. The central claim is that metric, deployment-relevant terrain reasoning can be obtained from RGB plus egomotion only, without LiDAR.

Significance. If the metric-grounding claim holds, the framework is practically significant: it would enable lightweight aerial platforms to produce physically meaningful obstacle clearance, slope, and roughness estimates for UGV deployment without dense LiDAR. The paper also contributes an efficient feed-forward pipeline, real-world top-down field data at multiple altitudes, and a semantic-geometric fusion mechanism that visibly suppresses hazardous regions. The efficiency gains in Table II and the qualitative altitude robustness in Figure 3 are strengths. However, the core contribution—metric reliability of the reconstructed terrain—is not directly validated, and the semantic label source is under-specified, so the significance currently rests on assumptions rather than demonstrated measurements.

major comments (4)
  1. [§IV-A.2, Eq. (6); §V-B, Table I] The load-bearing metric claim is not validated. Eq. (6) fits a single global scale s from relative camera displacements under the assumption that VGGT's reconstruction is shape-correct up to one similarity transform. Under top-down limited-parallax views at altitudes up to 70 m, depth errors can be spatially non-rigid (per-region flattening or drift), and a single fitted scale will not correct them. Table I reports ATE/RPE after Sim(3) alignment of the whole trajectory, which absorbs global scale error and does not test dense geometry. Please report absolute scale error against ground truth (e.g., per-scene fitted s vs. true scale, or dense point-to-LiDAR distances), at each altitude, and check whether residuals are consistent with a global similarity model. Without this, obstacle clearance and slope figures cannot be claimed metrically reliable.
  2. [§IV-A.1, §IV-A.2; Eq. (18)] The semantic class labels used in Eq. (18) are never defined. The method section describes SAM2 as producing temporally consistent instance masks, but traversability requires semantic classes c and a compatibility map τ(c). There is no description of a classifier that maps instances to classes, nor of the class set, nor of how 'per-pixel semantic probabilities' in §IV-A.2 are obtained. This is not a presentation detail: the fusion map in Eq. (19) and the quantitative gains in Table III depend on these labels, and the experiments cannot be reproduced or assessed without specifying the semantic source.
  3. [§V-D, Table III; §VI] The traversability evaluation is against manually labeled ground truth, but the annotation protocol is not described (who labeled, at what resolution, with what class definitions, how boundary and ambiguous cells were treated), and no inter-annotator agreement or per-scene confidence intervals are given. More importantly, no physical UGV traversal is used to confirm that cells labeled safe are actually traversable. The conclusion and future-work section explicitly state that full end-to-end field validation with physical UGV traversal remains future work. Thus the claim of 'reliable deployment-zone identification' is currently supported only by proxy labels. Please add per-scene results, annotation details, and ideally a physical traversal check or at least clearly state the proxy nature of the labels.
  4. [§V-C, Fig. 3] The altitude-robustness discussion is qualitative. Figure 3 shows reconstructions at increasing elevations, but there is no quantitative error metric as a function of altitude. Since the paper argues that RGB-based reconstruction remains stable at high altitudes and that LiDAR degrades, this claim should be supported by numbers (e.g., pose/depth error per altitude bin, or scale-error per scene). Without such data, it is hard to judge whether the method degrades gracefully or substantially at 70 m.
minor comments (5)
  1. [§IV-B, Eq. (19)] The reconstruction confidence π_i is said to be per-point, but its derivation from VGGT's depth confidence C_i is not explained. Please specify how per-pixel confidences are normalized, aggregated, and associated with unprojected 3D points.
  2. [§IV-B, Eqs. (16)-(20)] The thresholds s_soft/s_hard, σ_soft/σ_hard, d_soft/d_hard, and the weights α, λ, w_s/w_r/w_c, along with T_th and r_max, appear hand-selected. A sensitivity analysis (e.g., varying T_th or d_res) would help establish that the reported accuracy is not an artifact of a particular threshold choice.
  3. [Tables I and II] No standard deviations or number of evaluation sequences are reported for the pose and efficiency metrics. Given the small field dataset and the variability across scenes, error bars or per-sequence results are needed to compare methods meaningfully.
  4. [§V-A] The paper states that observations are exclusively top-down, but Figure 3 shows 'ground-level views' for qualitative visualization. It would help to clarify whether these views are rendered from the reconstruction or captured separately, since the latter would conflict with the stated data constraint.
  5. [§III-B and §IV-A.1] The relationship between SAM instance masks and semantic class labels should be clarified even in the related-work/preliminaries; currently the reader must infer that an additional classification step exists, but none is described.

Circularity Check

0 steps flagged

No significant circularity: metric-scale recovery is an explicit fit to an external onboard-egomotion signal, not a disguised reuse of the predicted quantity.

full rationale

The paper's central derivation is self-contained in the relevant sense. VGGT provides an up-to-scale reconstruction from RGB, and onboard egomotion (FAST-LIO) provides an independent metric signal. Equations (4)-(6) explicitly fit a similarity transform (s,R) by aligning predicted relative camera displacements to platform displacement; Eq. (8) then propagates that transform to the reconstruction. This is a stated calibration/fusion step, not a hidden reuse of the output as input: the metric scale is not claimed to be predicted from RGB alone, and the paper explicitly credits the embodied motion prior. Downstream traversability quantities (slope, roughness, clearance) are computed from the scaled geometry, but they are not presented as independent predictions that avoid the egomotion input; they inherit the scale by design, which is disclosed in Sec. IV-A.2. Pose evaluation in Table I is checked against ground-truth trajectories with ATE/RPE, an external benchmark, and no load-bearing result is justified solely by self-citation. References [2], [3], and [22] overlap with the authors, but they are used for baselines, platform context, and evaluation protocol, not as uniqueness theorems or as substitutes for the derivation. The absence of an absolute metric-scale error against LiDAR or surveyed coordinates is a genuine validation gap and a correctness risk, but it is not circular reasoning: the method's equations do not reduce to their own inputs by construction. Therefore no specific circular step can be exhibited, and the appropriate score is 0.

Axiom & Free-Parameter Ledger

4 free parameters · 5 axioms · 0 invented entities

No new physical entities are introduced. The main free quantity is the metric scale estimated from egomotion, plus a set of hand-set traversability thresholds and weights whose values are not disclosed.

free parameters (4)
  • metric scale s (and similarity transform R,t) = estimated via Eq. (6)-(7) from egomotion
    Central to the metric claim; fitted by minimizing displacement mismatch between VGGT camera centers and onboard egomotion.
  • traversability thresholds s_soft, s_hard, sigma_soft, sigma_hard, d_soft, d_hard = not stated
    Appear in Eq. (17); hand-chosen soft/hard bounds that directly determine which cells are safe.
  • fusion weight alpha and semantic compatibility map tau(c) = not stated
    Eqs. (18)-(19); alpha balances geometry vs semantics, tau encodes expert terrain-suitability labels.
  • selection parameters T_th, lambda, r_max, min separation, keyframe trigger delta, BEV resolution d_res = not stated
    Control deployment-zone ranking in Eq. (20) and the adaptive segmentation keyframeing in Eq. (2).
axioms (5)
  • domain assumption VGGT predicts shape and camera trajectories up to one global similarity transform under purely top-down views
    Sec. IV-A.2; if depth errors are non-rigid under limited parallax, a single scale fit is insufficient.
  • domain assumption Onboard egomotion (FAST-LIO) is metric and accurate enough to anchor scale
    Sec. V-A; all metric quantities inherit this signal's accuracy.
  • domain assumption SAM2 yields temporally consistent masks and the coverage-trigger keyframeing discovers new regions
    Sec. IV-A.1; semantic 3D labels depend on tracker continuity.
  • domain assumption Terrain safety can be summarized by slope, roughness, clearance, and a hand-assigned semantic score
    Sec. IV-B; traction and softness are not measured directly.
  • standard math Closed-form Umeyama alignment gives the optimal similarity transform for noise-free displacements
    Sec. IV-A.2 Eq. (6); standard result.

pith-pipeline@v1.3.0-alltime-deepseek · 10480 in / 12051 out tokens · 115920 ms · 2026-08-02T18:07:42.781921+00:00 · methodology

0 comments
read the original abstract

Inspection of confined infrastructure such as culverts often requires accessing hidden spaces whose entrances are reachable primarily from elevated viewpoints. Aerial-ground cooperation enables a UAV to deploy a compact UGV for interior exploration, but selecting a suitable deployment region from aerial observations requires metric terrain reasoning involving scale ambiguity, reconstruction uncertainty, and terrain semantics. We present a metric RGB-based geometric-semantic reconstruction and traversability analysis framework for aerial-to-ground hidden space inspection. A feed-forward multi-view RGB reconstruction backbone produces dense geometry, while temporally consistent semantic segmentation yields a 3D semantic map. To enable deployment-relevant measurements without requiring LiDAR-based dense mapping, we introduce an embodied motion prior that recovers metric scale by aligning predicted camera motion with onboard platform egomotion. From the metrically grounded reconstruction, we construct a confidence-aware geometric-semantic traversability map and evaluate candidate deployment zones under reachability constraints. Experiments on a tethered UAV-UGV platform demonstrate reliable deployment-zone identification in hidden space scenarios.

Figures

Figures reproduced from arXiv: 2603.14639 by Durgakant Pushp, Lantao Liu, Seoyoung Lee, Shaekh Mohammad Shithil, Zhangyang Wang.

Figure 1
Figure 1. Figure 1: Overview of the proposed RGB-based geometric and semantic reconstruction pipeline with embodied metric grounding. Given a top-down RGB video captured by a UAV, VGGT performs feed-forward multi-view geometric reconstruction to predict camera parameters and depth, from which dense 3D geometry is obtained via depth unprojection. In parallel, an adaptive segmentation module generates temporally consistent inst… view at source ↗
Figure 2
Figure 2. Figure 2: Experimental Setup: An Integrated aerial–ground robotic system. The DJI Matrice M600 hexacopter carries a compact tracked ground robot and integrates a Livox Mid-360 LiDAR, RealSense camera, and Jetson Xavier NX compute module. Dense LiDAR data are not used in the proposed method. hard thresholds controlling the transition between safe and unsafe terrain. All component scores are clipped to the interval [0… view at source ↗
Figure 3
Figure 3. Figure 3: RGB-based geometric and semantic 3D reconstruction from purely top-down aerial views at increasing UAV elevations for two hidden-space inspection scenes: a culvert entrance (left) and a vent system occluded by surrounding structures (right). For each elevation, RGB reconstructions are shown above their corresponding semantic reconstructions for both top-down and ground-level views. Ground-level views also … view at source ↗
Figure 4
Figure 4. Figure 4: Visual Results of Traversability mapping. From left to right: RGB-based 3D reconstruction, semantic segmentation, geometric traversability map, and geometric–semantic fusion result. The geometric map assigns high scores to planar regions but partially misclassifies rocky areas due to smoothed geometry. The fusion map assigns low traversability to terrain classes such as rocks and structural obstacles, prod… view at source ↗
Figure 5
Figure 5. Figure 5: Deployment location selection and trajectory generation. High traversability regions (green) indicate safe deployment areas, while red regions denote unsafe terrain. Candidate deployment locations (blue) are identified and feasible trajectories are generated from the UAV to the deployment zone. illustrates regions with high traversability (green), candidate deployment locations (blue), and the resulting tr… 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

23 extracted references · 2 linked inside Pith

  1. [1]

    Trial by fire [rescue robots],

    R. R. Murphy, “Trial by fire [rescue robots],”IEEE Robotics & Automation Magazine, vol. 11, no. 3, pp. 50–61, 2004

    2004

  2. [2]

    Cipher: Culvert inspection through pairwise frame selection and high-efficiency reconstruction,

    S. Lee and Z. Wang, “Cipher: Culvert inspection through pairwise frame selection and high-efficiency reconstruction,”IEEE/CVF Con- ference on Computer Vision and Pattern Recognition (CVPR) End-to- End 3D Learning (E2E3D) Workshop, 2025

    2025

  3. [3]

    Miniugv2: A compact uav-deployable tracked ground vehicle with manipulation capabilities,

    D. Pushp, S. Kalhapure, S. M. Shithil, and L. Liu, “Miniugv2: A compact uav-deployable tracked ground vehicle with manipulation capabilities,”arXiv preprint arXiv:2603.00972, 2026

    arXiv 2026

  4. [4]

    Uav-miniugv hybrid sys- tem for hidden area exploration and manipulation,

    D. Pushp, S. Kalhapure, K. Das, and L. Liu, “Uav-miniugv hybrid sys- tem for hidden area exploration and manipulation,” in2022 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2022, pp. 1297–1304

    2022

  5. [5]

    Vggt: Visual geometry grounded transformer,

    J. Wang, M. Chen, N. Karaev, A. Vedaldi, C. Rupprecht, and D. Novotny, “Vggt: Visual geometry grounded transformer,” inPro- ceedings of the Computer Vision and Pattern Recognition Conference, 2025, pp. 5294–5306

    2025

  6. [6]

    Fast3r: Towards 3d reconstruction of 1000+ images in one forward pass,

    J. Yang, A. Sax, K. J. Liang, M. Henaff, H. Tang, A. Cao, J. Chai, F. Meier, and M. Feiszli, “Fast3r: Towards 3d reconstruction of 1000+ images in one forward pass,” inProceedings of the Computer Vision and Pattern Recognition Conference, 2025, pp. 21 924–21 935

    2025

  7. [7]

    Nondestructive evaluation sensor fusion with autonomous robotic system for civil infrastructure inspection,

    S. Gibb, H. M. La, T. Le, L. Nguyen, R. Schmid, and H. Pham, “Nondestructive evaluation sensor fusion with autonomous robotic system for civil infrastructure inspection,”Journal of Field Robotics, vol. 35, no. 6, pp. 988–1004, 2018

    2018

  8. [8]

    Inspection of an underwater structure using point-cloud slam with an auv and a laser scanner,

    A. Palomer, P. Ridao, and D. Ribas, “Inspection of an underwater structure using point-cloud slam with an auv and a laser scanner,” Journal of field robotics, vol. 36, no. 8, pp. 1333–1344, 2019

    2019

  9. [9]

    Robotic system for inspection by contact of bridge beams using uavs,

    P. J. Sanchez-Cuevas, P. Ramon-Soria, B. Arrue, A. Ollero, and G. Heredia, “Robotic system for inspection by contact of bridge beams using uavs,”Sensors, vol. 19, no. 2, p. 305, 2019

    2019

  10. [10]

    Automated robotic monitoring and inspection of steel structures and bridges,

    H. M. La, T. H. Dinh, N. H. Pham, Q. P. Ha, and A. Q. Pham, “Automated robotic monitoring and inspection of steel structures and bridges,”Robotica, vol. 37, no. 5, pp. 947–967, 2019

    2019

  11. [11]

    Cable surface damage detection in cable-stayed bridges using optical techniques and image mosaicking,

    X. Li, C. Gao, Y . Guo, F. He, and Y . Shao, “Cable surface damage detection in cable-stayed bridges using optical techniques and image mosaicking,”Optics & Laser Technology, vol. 110, pp. 36–43, 2019

    2019

  12. [12]

    Structure-from-Motion Revis- ited,

    J. L. Sch ¨onberger and J.-M. Frahm, “Structure-from-Motion Revis- ited,” inConference on Computer Vision and Pattern Recognition (CVPR), 2016

    2016

  13. [13]

    Grounding image matching in 3d with mast3r,

    V . Leroy, Y . Cabon, and J. Revaud, “Grounding image matching in 3d with mast3r,” inEuropean Conference on Computer Vision. Springer, 2024, pp. 71–91

    2024

  14. [14]

    3d gaussian splatting for real-time radiance field rendering,

    B. Kerbl, G. Kopanas, T. Leimk ¨uhler, and G. Drettakis, “3d gaussian splatting for real-time radiance field rendering,”ACM Transactions on Graphics, vol. 42, no. 4, July 2023. [Online]. Available: https://repo-sam.inria.fr/fungraph/3d-gaussian-splatting/

    2023

  15. [15]

    Large spatial model: End-to- end unposed images to semantic 3d,

    Z. Fan, J. Zhang, W. Cong, P. Wang, R. Li, K. Wen, S. Zhou, A. Kadambi, Z. Wang, D. Xuet al., “Large spatial model: End-to- end unposed images to semantic 3d,”Advances in Neural Information Processing Systems, vol. 37, pp. 40 212–40 229, 2024

    2024

  16. [16]

    Easi3r: Estimating disentangled motion from dust3r without training,

    X. Chen, Y . Chen, Y . Xiu, A. Geiger, and A. Chen, “Easi3r: Estimating disentangled motion from dust3r without training,” inProceedings of the IEEE/CVF International Conference on Computer Vision, 2025, pp. 9158–9168

    2025

  17. [17]

    Probabilistic terrain mapping for mobile robots with uncertain localization,

    P. Fankhauser and M. Hutter, “Probabilistic terrain mapping for mobile robots with uncertain localization,”IEEE Robotics and Automation Letters, vol. 3, no. 4, pp. 3019–3026, 2018

    2018

  18. [18]

    Where should i walk? predicting terrain properties from images via self-supervised learning,

    L. Wellhausen, A. Dosovitskiy, R. Ranftl, K. Walas, C. Cadena, and M. Hutter, “Where should i walk? predicting terrain properties from images via self-supervised learning,”IEEE Robotics and Automation Letters, vol. 4, no. 2, pp. 1509–1516, 2019

    2019

  19. [19]

    Wild visual navigation: Fast traversability learning via pre-trained models and online self-supervision,

    M. Mattamala, J. Frey, P. Libera, N. Chebrolu, G. Martius, C. Cadena, M. Hutter, and M. Fallon, “Wild visual navigation: Fast traversability learning via pre-trained models and online self-supervision,”Au- tonomous Robots, vol. 49, no. 3, pp. 1–19, 2025

    2025

  20. [20]

    Afrda: Attentive feature re- finement for domain adaptive semantic segmentation,

    M. A.-M. Khan, D. Pushp, and L. Liu, “Afrda: Attentive feature re- finement for domain adaptive semantic segmentation,”IEEE Robotics and Automation Letters, vol. 10, no. 9, pp. 9573–9580, 2025

    2025

  21. [21]

    Tns: Terrain traversability mapping and navigation system for autonomous excava- tors,

    T. Guan, Z. He, R. Song, D. Manocha, and L. Zhang, “Tns: Terrain traversability mapping and navigation system for autonomous excava- tors,” inProceedings of Robotics: Science and Systems (RSS), New York City, NY , USA, Jun. 2022

    2022

  22. [22]

    E3d-bench: A bench- mark for end-to-end 3d geometric foundation models,

    W. Cong, Y . Liang, Y . Zhang, Z. Yang, Y . Wang, B. Ivanovic, M. Pavone, C. Chen, Z. Wang, and Z. Fan, “E3d-bench: A bench- mark for end-to-end 3d geometric foundation models,”arXiv preprint arXiv:2506.01933, 2025

    Pith/arXiv arXiv 2025

  23. [23]

    Halo: Hazard-aware landing optimization for autonomous systems,

    C. R. Hayner, S. C. Buckner, D. Broyles, E. Madewell, K. Leung, and B. Acikmese, “Halo: Hazard-aware landing optimization for autonomous systems,” 2023. [Online]. Available: https://arxiv.org/abs/ 2304.01583

    Pith/arXiv arXiv 2023