Pith. sign in

REVIEW 4 major objections 4 minor 68 references

BIM-Loc claims that a robot can localize accurately in feature-sparse indoor environments using only the as-designed Building Information Model, while simultaneously flagging which designed structures are missing or changed in reality.

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 11:28 UTC pith:EARQRE66

load-bearing objection A genuinely new BIM-to-LiDAR localization pipeline with strong independent ground-truth results in simulation and office settings, but the construction-site evaluation is partly circular and the discrepancy model only handles presence/absence — worth a serious referee. the 4 major comments →

arxiv 2606.14237 v1 pith:EARQRE66 submitted 2026-06-12 cs.RO

BIM-Loc: BIM-Integrated Discrepancy-Aware LiDAR-based Indoor Localization

Yinqiang Zhang , Liang Lu , Yipeng Pan , Maolin Lei , Yuhan Xie , Zhanteng Xie , Xiaowei Luo , Jia Pan This is my paper
classification cs.RO
keywords LiDAR localizationBuilding Information Modeldiscrepancy detectionpose graph optimizationray castingindoor roboticsmap priorsBayesian inference
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 show that a robot can localize indoors using only the design-phase BIM model, even when the real building differs from the model, by simultaneously estimating a BIM-aligned trajectory and detecting which BIM structures actually exist. It claims this works in feature-sparse corridors and construction sites where point-cloud maps are unavailable, and that the resulting accuracy and robustness exceed those of map-based methods. The key is to treat discrepancy detection and trajectory optimization as coupled problems solved iteratively, so unbuilt structures are removed as constraints instead of pulling the pose estimate astray. If right, service and inspection robots could navigate buildings from design files alone and monitor construction progress as a byproduct.

Core claim

On its own terms, the paper's discovery is that BIM-reality discrepancy does not have to be a nuisance to be filtered; it can be estimated online as a structured variable that controls which BIM surfaces are allowed to constrain the trajectory. The system alternates between optimizing a pose graph that aligns LiDAR scans to BIM surfaces and updating a hierarchical Bayesian map of pixel-level hit/miss evidence, so only structures confirmed as built constrain the pose. In simulation and real-world office and construction benchmarks, it reports trajectory errors of about a centimeter in simulation and worst-case 0.147 m / 2.91 degrees in an office benchmark, with scan-to-BIM distances consisten

What carries the argument

The load-bearing mechanism is multi-hit ray casting: each LiDAR ray is cast through the BIM model from the sensor to just past its measured endpoint, collecting every surface intersection rather than only the first. That single operation supplies three things: hit/miss associations linking scan points to specific BIM facets, synthetic surface points with normals used as plane constraints in pose-graph optimization, and barycentric UV coordinates that project sparse 3D observations onto 2D texture images for discrepancy detection. The optimized trajectory and the discrepancy indicators are then updated iteratively, with a Bayesian kernel updating pixel posteriors and propagating them through

Load-bearing premise

The system assumes every real BIM structure can be classified as simply present, absent, or unknown; if a real wall is moved by more than about 0.2 m it is treated as missing rather than as a usable surface, so localization depends on enough unchanged structures remaining.

What would settle it

Take a corridor where a single wall is shifted 0.3 m but everything else matches the BIM. If BIM-Loc classifies that wall as missing and still keeps trajectory error near its reported 0.01–0.15 m level, the discrepancy-robustness claim survives; if drift grows or the map mislabels nearby intact walls, the presence/absence-only discrepancy model is the cause.

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

If this is right

  • Using only as-designed BIM as a prior, the method claims centimeter-level indoor localization and drift bounding without a pre-built point-cloud map.
  • Discrepancy detection prevents unbuilt or modified structures from corrupting pose estimates; only structures classified as built contribute constraints.
  • Explicit BIM surfaces resolve thin walls that voxel-based bundle adjustment confuses, improving consistency in buildings with partitions as thin as 7.5–10 cm.
  • The online discrepancy output yields construction progress information (built vs unbuilt vs unknown) as a byproduct of localization.
  • The system runs on a CPU within a 1.5 s batch cycle, suggesting deployability on resource-constrained robot hardware.

Where Pith is reading between the lines

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

  • If the discrepancy model were extended from presence/absence to continuous deformation, moved walls could become usable constraints instead of being discarded; that would likely help renovated buildings where most surfaces are shifted slightly.
  • The texture-space representation suggests a compact memory scheme for lifelong mapping: revisiting a floor could refine the pixel-level Bayesian posteriors to detect new changes without storing dense point clouds, an extension the paper does not develop.
  • The same multi-hit ray casting could transfer to other CAD or mesh priors, such as photogrammetric meshes; whether accuracy degrades when the prior is itself noisy would be a straightforward test.

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 / 4 minor

Summary. The paper presents BIM-Loc, a LiDAR-based indoor localization system that uses as-designed BIM models as the sole prior map, without requiring pre-built point-cloud maps. Its contributions are: (1) a multi-hit ray-casting module that associates LiDAR rays with BIM surfaces and projects observations into 2D texture space; (2) a pose-graph optimization framework combining odometry, inter-scan, and BIM-scan residuals; and (3) a hierarchical Bayesian module that incrementally estimates pixel-, facet-, and structure-level BIM-reality discrepancies. The method is evaluated in a custom Gazebo simulation benchmark and on two real-world benchmarks (HKUST office, CityU construction), where it is reported to outperform map-based baselines such as PALoc, Fast-LIO-Localization, HBA, and BALM2 by large margins on ATE, MME, scan-to-BIM distance, and Wasserstein distance metrics.

Significance. If the claims hold, BIM-Loc would be a practically valuable contribution: it demonstrates that a lightweight, design-stage BIM can serve as a global reference for indoor localization, and it addresses the important but often ignored problem of BIM-reality discrepancies in an online manner. The manuscript is strong in its breadth of evaluation: it includes GT-bearing simulation and office benchmarks, a no-GT construction benchmark, a controlled discrepancy stress test, ablation of the discrepancy module, initialization sensitivity analysis, runtime profiling, and hyperparameter sensitivity. The reported GT-based ATE improvements (e.g., worst-case 0.147 m / 2.91 deg vs. 1.121 m / 12.99 deg for PALoc on HKUST) are substantial. However, the correctness of the discrepancy-detection mathematics and the independence of some evaluation metrics need to be addressed before the central claims can be fully accepted.

major comments (4)
  1. [§3.4.2–§3.4.3, Eqs. (11)–(15)] The meaning of the discrepancy indicator is internally inconsistent. In Eq. (3), I_n=0 means 'exists in reality' and the residual is activated by (1−I_n). In §3.4, however, I_S=1 is defined as 'a discrepancy exists', and Eq. (15) labels I=1 as 'consistent' and I=0 as 'discrepant'. At the same time, Eq. (11) increments α with misses and β with hits, so E[I_I]=α/(α+β) is large when many rays miss the BIM surface — i.e., when the structure is likely discrepant. Under Eq. (15)'s labeling, a large E would then mark the structure as 'consistent'. These conventions cannot all be correct. Because Eq. (15) feeds the (1−I_n) factor in Eq. (3), the sign convention is load-bearing. Please adopt a single convention and adjust Eqs. (10)–(15) accordingly, and confirm the implementation matches the corrected equations.
  2. [§3.4.3, Eqs. (13)–(14)] The variance formulas for the model-averaged indicators are incorrect. For I_F = (1/N)Σ I_I with independent Bernoulli pixel indicators, Var(I_F) = (1/N²)Σ Var(I_I). The printed expression Var(I_F)=E[I_F]−(1/N²)ΣVar(I_I)−(E[I_F])² is not equal to this except in special cases; for identical means p it gives (1−1/N)p(1−p) instead of (1/N)p(1−p). The analogous formula in Eq. (14) has the same issue. Since Eq. (15) uses the variance threshold γσ to separate 'consistent' from 'unknown', the incorrect variance directly affects the discrepancy classification and the F1 scores in Table 2. Please correct these formulas and rerun the sensitivity analysis of Table 7 with the corrected statistics.
  3. [§4.2.4 (construction benchmark), Tables 4–5, Figs. 14–16] For the CityU construction benchmark there is no ground truth, and the reported metrics (scan-to-BIM distance, MME ratio, and Wasserstein distance against a BIM-derived map) all measure, directly or indirectly, how well the estimated trajectory agrees with the BIM. This is the same objective that the BIM-scan residual in Eq. (3) minimizes, so the construction benchmark is partially circular: it can show consistency with the BIM prior but cannot by itself establish that BIM-Loc is more accurate than PALoc or Fast-Loc in that setting. The headline 'outperforms SOTA in localization accuracy and robustness' should be tied to the GT-bearing simulation and HKUST office benchmarks, and the construction results presented as supporting consistency evidence rather than as independent accuracy validation.
  4. [§3.2/§3.4, Appendix C.1, and §6] The discrepancy model only distinguishes present/absent (or unknown) structures; geometric modifications such as shifted, thickened, or partially deformed walls are not modeled. A structure shifted by more than the 0.2 m scan-to-BIM threshold will be classified as miss/unknown and removed from the optimization through the (1−I_n) factor, rather than used as a movable constraint. Appendix C.1 tests random shifts of 0–0.5 m, but only against DLO odometry, and it does not report the fraction of shifted structures or sweep against map-based baselines. The paper acknowledges this in §6 ('geometric modifications... are not considered'), yet the abstract claims robustness under 'significant discrepancies' without this qualification. Please narrow the claim to missing/extra structures and provide a stress test with controlled fractions of shifted/deformed elements against SOTA map-based methods,
minor comments (4)
  1. [§3.3.1, Eq. (5)] The notation 'λ1c', 'λ2c' is ambiguous; it should read λ_1^c, λ_2^c (eigenvalues of cluster c). Please also define e_x, e_y precisely, as they are used before Table 6.
  2. [§5 / Fig. 19 / Table 8] The ablation in Figure 19 shows only a small improvement from the discrepancy module (0.036 vs. 0.038 m on Floor 06, 0.038 vs. 0.039 m on Floor 08). Calling discrepancy handling 'essential' in §5 is overstated; 'mild but consistent benefit' is more accurate.
  3. [§3.4.3, Eq. (15)] The labels '(consistent)' and '(discrepant)' in Eq. (15) are inconsistent with the definition of I in §3.1. Even apart from the sign issue raised above, please unify the terminology so that the reader does not have to reverse-engineer the intended meaning.
  4. [§4.1.4 / Table 1] In the simulation table, BIM-Loc's RMSE of 0.007 m translation appears below the physical resolution of the VLP-16 at typical ranges. A brief note on how such sub-centimeter values arise (e.g., simulation ground truth, multi-scan averaging, or map alignment) would help the reader calibrate the claim.

Circularity Check

1 steps flagged

Partial evaluative circularity in no-ground-truth construction benchmark; independent ATE in office/simulation benchmarks keeps the central claim from reducing to its objective.

specific steps
  1. fitted input called prediction [Section 4.1.3 (Metrics), Section 4.2.3/4.2.4 (Construction benchmark); Eq. 3 and Eq. 6]
    "The scan-to-BIM distance metric assesses per-scan consistency by computing distances between points and BIM surfaces. To reduce noise impact, only distances below a predefined threshold (0.2 m in all experiments) are considered for RMSE computation."

    In the no-ground-truth construction benchmark, scan-to-BIM RMSE (and MME/WD against the BIM map) are reported as localization accuracy/consistency. But the trajectory optimization minimizes exactly this quantity: Eq. 3's E_consistency and Eq. 6's E_bim = Σ ||n'_i^T(R_k p_i + t_k − p'_i)||^2 are point-to-BIM-surface distances over the same thresholded associations. The §4.1.4 statement 'the minimal scan-to-BIM distances indicate optimal alignment between scans and the BIM model' is therefore a restatement of the objective, not an independent test. This is partial: the office and simulation benchmarks have independent ground-truth ATE, so the central accuracy claim does not reduce entirely to the objective.

full rationale

The derivation chain of BIM-Loc is not fundamentally circular: the MAP formulation (Eq. 1), the residual decomposition (Eq. 3), the multi-hit ray-casting association, the pose-graph factors (Eqs. 4-6), and the hierarchical Bayesian discrepancy inference (Eqs. 7-15) form a standard coupled estimation loop in which trajectory and discrepancy indicators are alternately refined. The discrepancy detector is validated against known removed structures in the CityU-02 simulation with F1=0.913, which is an external label set, and the real-world office benchmark provides independent ground-truth ATE where BIM-Loc's worst-case 0.147 m/2.91 deg clearly beats PALoc's 1.121 m/12.99 deg. These results give the central 'accuracy' claim independent support. The circular aspect is limited to the construction benchmark, where no ground-truth trajectory exists and the reported scan-to-BIM RMSE, MME ratio, and Wasserstein distance against the BIM map are essentially the same residuals being optimized in Eqs. 3 and 6; outperformance there is partly by construction. Self-citations such as Zhang et al. (2024) are used for initialization and as a baseline, not as load-bearing evidence for the core derivation. Section 6's explicit admission that geometric modifications (shape changes, boundary shifts, thickness variations) are not modeled limits the breadth of the 'robustness' claim but is a scope limitation, not circularity. Overall, the paper has one partial evaluative circularity, while its central claims retain independent content.

Axiom & Free-Parameter Ledger

2 free parameters · 5 axioms · 0 invented entities

The central claim rests on the threshold-based separation of BIM-reality discrepancy from sensor noise, the straight-ray surface assumption, and the front-end odometry prior. The two tuned hyperparameters (0.2 m threshold, (0.85,0.15,2.0)) are fitted to data. No new physical entity is introduced; the texture-space 'unknown' state is a representation choice, not a new entity.

free parameters (2)
  • scan-to-BIM distance threshold = 0.2 m
    Selected via sensitivity analysis in Figure B.1 (U-shaped RMSE curve); affects all association and discrepancy decisions.
  • discrepancy decision thresholds (γ+μ, γ−μ, γσ) = (0.85, 0.15, 2.0)
    Chosen in Table B.1 to maximize discrepancy-detection F1 (0.913) on the CityU-02 simulation benchmark; a mild parameter fit to the validation data.
axioms (5)
  • domain assumption LiDAR rays follow straight lines and reflect at surfaces; the sensor endpoint lies on or near the first unmet surface.
    Basis of the multi-hit ray casting module (Section 3.2); if surfaces are transparent or specular, or if rays are blocked by occluders, the association model breaks.
  • domain assumption The as-designed BIM is geometrically correct to within 0.2 m for all 'built' structures.
    The fixed association threshold assumes that deviations larger than threshold indicate absence/discrepancy rather than modeling error; introduced in Section 3.2 and Appendix B.
  • domain assumption The environment is static over the batch period (1.5 s), so multi-hit ray associations are consistent.
    Batch-wise processing in Section 5.4 and Appendix D assumes no significant motion of people/equipment within a batch.
  • domain assumption Front-end odometry (DLO) provides a locally consistent trajectory prior with bounded error.
    The pose graph in Section 3.3 uses odometry factors as priors; if the front end fails (e.g., fast rotation in featureless corridor), the BIM factors alone may not recover.
  • standard math Bayes' theorem and SVD/eigendecomposition properties used in Eqs. 5-14.
    Standard background for the probabilistic model and the inter-scan eigenvalue formulation.

pith-pipeline@v1.3.0-alltime-deepseek · 33069 in / 10595 out tokens · 108544 ms · 2026-08-02T11:28:56.323839+00:00 · methodology

0 comments
read the original abstract

Accurate and robust localization is a fundamental requirement for service and inspection robots, particularly in feature-sparse indoor environments where traditional systems struggle due to a lack of distinct landmarks. While prior maps can enhance robustness, precise and compact maps capturing real-world details are often unavailable for new or frequently changing environments. This paper presents BIM-Loc, a novel discrepancy-aware LiDAR-based localization method that directly integrates Building Information Models (BIM) from the design phase. BIM-Loc simultaneously estimates trajectories aligned with the BIM coordinate system and identifies discrepancies between real-world observations and the as-designed BIM in an online fashion. Our core contributions include: (1) a novel multi-hit ray casting strategy for efficient BIM-point data association and projection of 3D observations into 2D texture space; (2) a pose graph optimization framework with BIM-integrated factors that enforces consistency among odometry, sequential scans, and BIM structures; and (3) a hierarchical Bayesian inference module that incrementally updates a continuous 2D surface representation for discrepancy detection, propagating updates from the pixel to the structure level. Extensive evaluations in both simulation and real-world applications demonstrate that BIM-Loc significantly outperforms state-of-the-art map-based methods in localization accuracy and robustness.

Figures

Figures reproduced from arXiv: 2606.14237 by Jia Pan, Liang Lu, Maolin Lei, Xiaowei Luo, Yinqiang Zhang, Yipeng Pan, Yuhan Xie, Zhanteng Xie.

Figure 1
Figure 1. Figure 1: Visualization of BIM-Loc’s outputs. The right panel displays blue reconstructed points aligned with a projected 2D BIM model. The left panel and bottom 3D meshes show the discrepancy map generated from localization. Two snapshots demonstrate localization performance by comparing actual scenes with corresponding BIM rendering. consumption, making them ideal for edge computing devices. While 2D floor plan lo… view at source ↗
Figure 2
Figure 2. Figure 2: Framework of the proposed BIM-Loc method. The system processes as-designed BIM models and front-end odometry measurements through three core modules: (1) multi-hit ray casting for BIM-observation association, (2) BIM-aided trajectory optimization, and (3) BIM-observation discrepancy detection. Using an iterative strategy, the method outputs an optimized trajectory and discrepancy maps, relying solely on de… view at source ↗
Figure 3
Figure 3. Figure 3: , upper right). If the ray has not yet reached its endpoint neighborhood, the RayUpdate function sets the intersection point as a new origin and continues casting. This iterative process continues until all rays are processed, generating a set of synthetic counterpart points in the BIM coordinate system. For each intersection, the RayCasting function returns the following information: (1) association: the … view at source ↗
Figure 4
Figure 4. Figure 4: Schema of the factor graph for trajectory optimization. Scans are transformed into a common temporary state, then processed via batch multi-hit ray casting. Synthetic points are represented in the BIM coordinate system. The pose graph incorporates three types of constraint factors to balance inter-scan consistency and BIM-scan alignment during online optimization. BIM structure BIM structure Front-end odom… view at source ↗
Figure 5
Figure 5. Figure 5: Definition of constraints for BIM-aided trajectory optimization. Multi-hit ray casting associates scans with BIM structures. The optimization incorporates three constraints: front-end odometry, inter-scan consistency, and scan-BIM alignment, which jointly determine the trajectory estimation. The multi-hit ray casting operation establishes data association by assigning each point in the hit set H a tuple of… view at source ↗
Figure 6
Figure 6. Figure 6: Challenge of voxel-based data association with thin wall structures. Walls measuring 7.5 to 10cm in thickness are comparable to typical voxel resolutions used in large-scale buildings, leading to ambiguous data associations. 2023a) and BALM2 (Liu et al. 2023b) fail to distinctly separate the two sides of a thin wall. 3.4 Discrepancy Detection This section addresses the subproblem of detecting discrepancies… view at source ↗
Figure 7
Figure 7. Figure 7: Hierarchical Bayesian model for online discrepancy detection. The model infers structural discrepancies from sparse ray-casting observations using a multi-stage Bayesian inference pipeline. Sparse observations are first processed through kernelized Bayesian inference to predict pixel status in texture images, then propagated through pixel-facet-structure hierarchies to estimate BIM structure discrepancy in… view at source ↗
Figure 8
Figure 8. Figure 8: Texture mapping visualization for discrepancy detection. Cases (a) and (b) show wrapped BIM models aligned with texture maps, highlighting discrepancies from elevator and roller doors. Cases (c) and (d) show H-shape beam detection behind unfinished and finished walls. Color coding indicates alignment status: blue for consistent, orange for discrepant, and gray for unknown regions. nearby pixels contribute … view at source ↗
Figure 9
Figure 9. Figure 9: Simulation benchmark structure, comprising a Gazebo simulator with BIM models and an exploration algorithm. Five BIM model groups generate test data: four unmodified models (CaseStudy, CityU-01, Office-01, Office-02) evaluate localization without BIM-reality discrepancies, while a synthetic model (CityU-02) focuses explicitly on such discrepancies. between the MME of the aggregated map and that of the samp… view at source ↗
Figure 10
Figure 10. Figure 10: MME ratios for evaluated methods across data sequences in the simulation benchmark. Performance is grouped into three tiers, with BIM-Loc consistently ranking first. second best in the CaseStudy, Office-02, and CityU-02 sequences, while HDL performs second best in CityU￾01. This variation stems from the voxel-based association mechanisms in HBA and BALM2, which are misled by the narrow double sides of thi… view at source ↗
Figure 11
Figure 11. Figure 11: Wasserstein distance errors for evaluated algorithms in the simulation benchmark. Box plots show median WD values (logarithmic y-axis). Results for unmodified BIM models (left of dashed line) and synthetic BIM models (right) are grouped separately. CityU-02-F07 CityU-02-F08 CityU-02-F10 CityU-02-F12 Texture Map Discrepancy Map [PITH_FULL_IMAGE:figures/full_fig_p012_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Discrepancy detection results for CityU-02 data group. Top row: texture maps for Floors 07, 08, 10, and 12. Bottom row: corresponding discrepancy maps showing built (blue) and unbuilt (orange) structures. in above settings are mostly owning to optimizations with BIM priors. In [PITH_FULL_IMAGE:figures/full_fig_p012_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Real-world evaluation benchmarks representing different building lifecycle phases. The HKUST office benchmark assesses localization in a completed building using a handheld Livox Mid-360 sensor with built-in IMU. The CityU construction benchmark evaluates performance in an active construction environment with significant BIM-reality discrepancies, using both Ouster OS0-128 and Livox Mid-360 sensors to enh… view at source ↗
Figure 14
Figure 14. Figure 14: MME ratios for evaluated methods in two real-world benchmarks. BIM-Loc achieves the highest ratios, with a substantial performance gap in the HKUST office benchmark and a smaller but consistent lead in CityU construction benchmark. 3F-Region1 3F-Region2 3F-Region3 4F-Region1 4F-Region2 4F-Region3 5F-Region1 5F-Region2 5F-Region3 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Wasserstein Distance Error 0.275 0.153 0.… view at source ↗
Figure 15
Figure 15. Figure 15: Wasserstein distance errors for evaluated methods in the HKUST office benchmark. Red lines indicate median WD errors within each box plot. Floor-06 Ouster Floor-07 Ouster Floor-08 Ouster Floor-09 Ouster Floor-10 Ouster Floor-11 Ouster Floor-12 Ouster Floor-06 Mid-360 Floor-08 Mid-360 0.0 0.2 0.4 0.6 Wasserstein Distance Error 0.223 0.143 0.150 0.224 0.210 0.180 0.138 0.151 0.174 0.241 0.153 0.128 0.143 0.… view at source ↗
Figure 16
Figure 16. Figure 16 [PITH_FULL_IMAGE:figures/full_fig_p015_16.png] view at source ↗
Figure 17
Figure 17. Figure 17: Localization and discrepancy detection results of BIM-Loc in the CityU construction benchmark. Top row: estimated trajectories in BIM coordinates for Floors 06 and 08. Middle rows: real construction site images (A, C, E) and corresponding BIM-rendered overlays (B, D, F). Bottom row: texture maps and discrepancy maps generated during localization. nature of construction environments. Nonetheless, BIM￾Loc s… view at source ↗
Figure 18
Figure 18. Figure 18: Sensitivity analysis of scan-to-BIM distances on real-world CityU dataset [PITH_FULL_IMAGE:figures/full_fig_p022_18.png] view at source ↗
Figure 19
Figure 19. Figure 19: Scan-to-BIM RMSE over time with and without the discrepancy detection module on CityU construction dataset with Livox Mid360 (Floor 06 and Floor 08) [PITH_FULL_IMAGE:figures/full_fig_p023_19.png] view at source ↗
Figure 21
Figure 21. Figure 21: Initialization sensitivity on CityU dataset (Livox Mid360). (a)–(b) Scan-to-BIM distance RMSE for perturbed initializations on Floor 06 and Floor 08. (c)–(d) Range of RMSE across the 50 trials over time. The band narrows quickly, indicating low variance in final accuracy. indicating low variance in final accuracy and robustness to moderate initialization errors. D Runtime Analysis A runtime analysis is pr… 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

68 extracted references · 8 canonical work pages · 1 internal anchor

  1. [1]

    , " * write output.state after.block = add.period write newline

    ENTRY address author booktitle chapter doi edition editor eid howpublished institution isbn journal key month note number organization pages publisher school series title type url volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.all := #1 'mid...

  2. [2]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize ":" * " " *...

  3. [3]

    Automation Science and Engineering, IEEE Transactions on 12(2): 494--506

    Aguero C, Koenig N, Chen I, Boyer H, Peters S, Hsu J, Gerkey B, Paepcke S, Rivero J, Manzo J, Krotkov E and Pratt G (2015) Inside the Virtual Robotics Challenge: Simulating Real-Time Robotic Disaster Response . Automation Science and Engineering, IEEE Transactions on 12(2): 494--506. doi:10.1109/TASE.2014.2368997

    arXiv 2015

  4. [4]

    IEEE Transactions on Robotics 32(6): 1309--1332

    Cadena C, Carlone L, Carrillo H, Latif Y, Scaramuzza D, Neira J, Reid I and Leonard JJ (2016) Past, Present, and Future of Simultaneous Localization and Mapping: Toward the Robust-Perception Age . IEEE Transactions on Robotics 32(6): 1309--1332. doi:10.1109/TRO.2016.2624754

    arXiv 2016

  5. [5]

    In: Robotics: Science and Systems , volume 5

    Cao C, Zhu H, Choset H and Zhang J (2021) TARE: A Hierarchical Framework for Efficiently Exploring Complex 3D Environments . In: Robotics: Science and Systems , volume 5. p. 2. doi:10.15607/RSS.2021.XVII.018

    work page doi:10.15607/rss.2021.xvii.018 2021

  6. [7]

    IEEE Robotics and Automation Letters 10(4): 3414--3421

    Chen C, Lu L, Yang L, Zhang Y, Chen Y, Jia R and Pan J (2025) Signage-Aware Exploration in Open World Using Venue Maps . IEEE Robotics and Automation Letters 10(4): 3414--3421. doi:10.1109/LRA.2025.3540390

    arXiv 2025

  7. [8]

    IEEE Robotics and Automation Letters 7(2): 2000--2007

    Chen K, Lopez BT, Agha-mohammadi Aa and Mehta A (2022) Direct LiDAR Odometry: Fast Localization With Dense Point Clouds . IEEE Robotics and Automation Letters 7(2): 2000--2007. doi:10.1109/LRA.2022.3142739

    arXiv 2022

  8. [9]

    IEEE Robotics and Automation Letters 9(11): 9375--9382

    Chung D and Kim J (2024) NV-LIOM: LiDAR-Inertial Odometry and Mapping Using Normal Vectors Towards Robust SLAM in Multifloor Environments . IEEE Robotics and Automation Letters 9(11): 9375--9382. doi:10.1109/LRA.2024.3457373

    arXiv 2024

  9. [10]

    doi:10.5281/zenodo.5794541

    Dellaert F and Contributors G (2022) borglab/gtsam. doi:10.5281/zenodo.5794541

    doi:10.5281/zenodo.5794541 2022

  10. [11]

    IEEE Transactions on Robotics 35(4): 953--966

    Doherty K, Shan T, Wang J and Englot B (2019) Learning-Aided 3-D Occupancy Mapping With Bayesian Generalized Kernel Inference . IEEE Transactions on Robotics 35(4): 953--966. doi:10.1109/TRO.2019.2912487

    arXiv 2019

  11. [12]

    In: ISARC

    Dreher M, Blum H, Siegwart R and Gawel A (2021) Global localization in meshes . In: ISARC. Proceedings of the International Symposium on Automation and Robotics in Construction, volume 38. IAARC Publications, pp. 747--754. doi:https://doi.org/10.22260/ISARC2021/0101

    work page doi:10.22260/isarc2021/0101 2021

  12. [13]

    IEEE Transactions on Robotics 40: 936--959

    Ebadi K, Bernreiter L, Biggie H, Catt G, Chang Y, Chatterjee A, Denniston CE, Deschênes SP, Harlow K, Khattak S, Nogueira L, Palieri M, Petráček P, Petrlík M, Reinke A, Krátký V, Zhao S, Agha-mohammadi Aa, Alexis K, Heckman C, Khosoussi K, Kottege N, Morrell B, Hutter M, Pauling F, Pomerleau F, Saska M, Scherer S, Siegwart R, Williams JL and Carlone L (20...

    arXiv 2024

  13. [14]

    Robotics and Autonomous Systems 69: 28--39

    Einhorn E and Gross HM (2015) Generic NDT mapping in dynamic environments and its application for lifelong SLAM . Robotics and Autonomous Systems 69: 28--39. doi:https://doi.org/10.1016/j.robot.2014.08.008

    work page doi:10.1016/j.robot.2014.08.008 2015

  14. [16]

    Journal of Field Robotics doi:https://doi.org/10.1002/rob.70044

    Fan T, Shen B, Zhang Y, Zhang C, Yang L, Chen H, Zhang W and Pan J (2025) S2MAT: Simultaneous and Self-Reinforced Mapping and Tracking in Dynamic Urban Scenarios . Journal of Field Robotics doi:https://doi.org/10.1002/rob.70044

    work page doi:10.1002/rob.70044 2025

  15. [17]

    IEEE Robotics and Automation Letters 9(11): 9175--9182

    Ferrari S, Giammarino LD, Brizi L and Grisetti G (2024) MAD-ICP: It is All About Matching Data – Robust and Informed LiDAR Odometry . IEEE Robotics and Automation Letters 9(11): 9175--9182. doi:10.1109/LRA.2024.3456509

    arXiv 2024

  16. [19]

    IEEE Intelligent Transportation Systems Magazine 2(4): 31--43

    Grisetti G, Kümmerle R, Stachniss C and Burgard W (2010) A Tutorial on Graph-Based SLAM . IEEE Intelligent Transportation Systems Magazine 2(4): 31--43. doi:10.1109/MITS.2010.939925

    arXiv 2010

  17. [20]

    In: 2016 IEEE International Conference on Robotics and Automation (ICRA)

    Hess W, Kohler D, Rapp H and Andor D (2016) Real-time loop closure in 2D LIDAR SLAM . In: 2016 IEEE International Conference on Robotics and Automation (ICRA). pp. 1271--1278. doi:10.1109/ICRA.2016.7487258

    arXiv 2016

  18. [21]

    In: 2022 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

    Holoch M, Kurz G and Biber P (2022) Detecting Invalid Map Merges in Lifelong SLAM . In: 2022 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). pp. 11039--11046. doi:10.1109/IROS47612.2022.9981564

    arXiv 2022

  19. [22]

    IEEE Robotics and Automation Letters 10(5): 4228--4235

    Hu X, Wu J, Jia M, Yan H, Jiang Y, Jiang B, Zhang W, He W and Tan P (2025) MapEval: Towards Unified, Robust and Efficient SLAM Map Evaluation Framework . IEEE Robotics and Automation Letters 10(5): 4228--4235. doi:10.1109/LRA.2025.3548441

    arXiv 2025

  20. [23]

    IEEE/ASME Transactions on Mechatronics 29(6): 4297--4308

    Hu X, Zheng L, Wu J, Geng R, Yu Y, Wei H, Tang X, Wang L, Jiao J and Liu M (2024) PALoc: Advancing SLAM Benchmarking With Prior-Assisted 6-DoF Trajectory Generation and Uncertainty Estimation . IEEE/ASME Transactions on Mechatronics 29(6): 4297--4308. doi:10.1109/TMECH.2024.3362902

    arXiv 2024

  21. [24]

    In: 2025 IEEE International Conference on Robotics and Automation (ICRA)

    Huang H, Qiao Z, Yu Z, Liu C, Shen S, Zhang F and Yin H (2025) SLABIM: A SLAM-BIM Coupled Dataset in HKUST Main Building . In: 2025 IEEE International Conference on Robotics and Automation (ICRA). pp. 9032--9038. doi:10.1109/ICRA55743.2025.11127798

    arXiv 2025

  22. [25]

    IEEE Robotics and Automation Letters 6(4): 8269--8276

    Huang H, Sun Y, Wu J, Jiao J, Hu X, Zheng L, Wang L and Liu M (2021) On Bundle Adjustment for Multiview Point Cloud Registration . IEEE Robotics and Automation Letters 6(4): 8269--8276. doi:10.1109/LRA.2021.3105686

    arXiv 2021

  23. [26]

    Automation in Construction 134: 104057

    Huang R, Xu Y, Hoegner L and Stilla U (2022) Semantics-aided 3D change detection on construction sites using UAV-based photogrammetric point clouds . Automation in Construction 134: 104057. doi:https://doi.org/10.1016/j.autcon.2021.104057

    arXiv 2022

  24. [27]

    ://github.com/HViktorTsoi/FAST_LIO_LOCALIZATION

    HViktorTsoi (2021) FAST-LIO-LOCALIZATION: A simple localization framework that can re-localize in built maps based on FAST-LIO . ://github.com/HViktorTsoi/FAST_LIO_LOCALIZATION. Accessed: 2024-05-20

    2021

  25. [28]

    IEEE Transactions on Multimedia

    Jia Y, Cao F, Wang T, Tang Y, Shao S and Liu L (2025) CAD-Mesher: A Convenient, Accurate, Dense Mesh-based Mapping Module in SLAM for Dynamic Environments . IEEE Transactions on Multimedia

    2025

  26. [29]

    ://github.com/jpcy/xatlas

    Jonathan Young (2025) Xatlas library . ://github.com/jpcy/xatlas

    2025

  27. [30]

    International Journal of Advanced Robotic Systems 16(2): 1729881419841532

    Kenji K, Jun M and Emanuele M (2019) A portable three-dimensional LIDAR-based system for long-term and wide-area people behavior measurement . International Journal of Advanced Robotic Systems 16(2): 1729881419841532. doi:https://doi.org/10.1177/1729881419841532

    doi:10.1177/1729881419841532 2019

  28. [31]

    ISPRS Open Journal of Photogrammetry and Remote Sensing 2: pp

    Khoshelham K, Tran H, Acharya D, Vilari \ n o LD, Kang Z and Dalyot S (2021) Results of the ISPRS benchmark on indoor modelling . ISPRS Open Journal of Photogrammetry and Remote Sensing 2: pp. 100008. doi:https://doi.org/10.1016/j.ophoto.2021.100008

    arXiv 2021

  29. [32]

    IEEE Transactions on Robotics 38(3): 1856--1874

    Kim G, Choi S and Kim A (2022) Scan Context++: Structural Place Recognition Robust to Rotation and Lateral Variations in Urban Environments . IEEE Transactions on Robotics 38(3): 1856--1874. doi:10.1109/TRO.2021.3116424

    arXiv 2022

  30. [33]

    In: 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

    Kim G and Kim A (2020) Remove, then Revert: Static Point cloud Map Construction using Multiresolution Range Images . In: 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). pp. 10758--10765. doi:10.1109/IROS45743.2020.9340856

    arXiv 2020

  31. [34]

    In: 2022 International Conference on Robotics and Automation (ICRA)

    Kim G and Kim A (2022) LT-mapper: A Modular Framework for LiDAR-based Lifelong Mapping . In: 2022 International Conference on Robotics and Automation (ICRA). pp. 7995--8002. doi:10.1109/ICRA46639.2022.9811916

    arXiv 2022

  32. [35]

    In: 2024 IEEE International Conference on Robotics and Automation (ICRA)

    Koide K, Oishi S, Yokozuka M and Banno A (2024 a ) Tightly Coupled Range Inertial Localization on a 3D Prior Map Based on Sliding Window Factor Graph Optimization . In: 2024 IEEE International Conference on Robotics and Automation (ICRA). pp. 1745--1751. doi:10.1109/ICRA57147.2024.10611195

    arXiv 2024

  33. [36]

    In: 2021 IEEE International Conference on Robotics and Automation (ICRA)

    Koide K, Yokozuka M, Oishi S and Banno A (2021) Voxelized GICP for Fast and Accurate 3D Point Cloud Registration . In: 2021 IEEE International Conference on Robotics and Automation (ICRA). pp. 11054--11059. doi:10.1109/ICRA48506.2021.9560835

    arXiv 2021

  34. [37]

    Robotics and Autonomous Systems 179: 104750

    Koide K, Yokozuka M, Oishi S and Banno A (2024 b ) GLIM: 3D range-inertial localization and mapping with GPU-accelerated scan matching factors . Robotics and Autonomous Systems 179: 104750. doi:https://doi.org/10.1016/j.robot.2024.104750

    arXiv 2024

  35. [38]

    In: 2021 European Conference on Mobile Robots (ECMR)

    Kornilova A and Ferrer G (2021) Be your own Benchmark: No-Reference Trajectory Metric on Registered Point Clouds . In: 2021 European Conference on Mobile Robots (ECMR) . pp. 1--8. doi:10.1109/ECMR50962.2021.9568822

    arXiv 2021

  36. [39]

    Trust, but Verify: Cross-Modality Fusion for HD Map Change Detection

    Lambert JW and Hays J (2021) Trust, but Verify : Cross-modality fusion for hd map change detection. In: Advances in Neural Information Processing Systems Track on Datasets and Benchmarks. pp. 8578--8585. doi:https://doi.org/10.48550/arXiv.2212.07312

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2212.07312 2021

  37. [42]

    IEEE Robotics and Automation Letters 8(3): 1523--1530

    Liu X, Liu Z, Kong F and Zhang F (2023 a ) Large-Scale LiDAR Consistent Mapping Using Hierarchical LiDAR Bundle Adjustment . IEEE Robotics and Automation Letters 8(3): 1523--1530. doi:10.1109/LRA.2023.3238902

    arXiv 2023

  38. [43]

    IEEE Transactions on Robotics 39(6): 4366--4386

    Liu Z, Liu X and Zhang F (2023 b ) Efficient and Consistent Bundle Adjustment on Lidar Point Clouds . IEEE Transactions on Robotics 39(6): 4366--4386. doi:10.1109/TRO.2023.3311671

    arXiv 2023

  39. [44]

    Automation in Construction 141: 104442

    Meyer T, Brunn A and Stilla U (2022) Change detection for indoor construction progress monitoring based on BIM, point clouds and uncertainties . Automation in Construction 141: 104442. doi:https://doi.org/10.1016/j.autcon.2022.104442

    arXiv 2022

  40. [45]

    In: 2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

    Mock A, Wiemann T, P \"u tz S and Hertzberg J (2024) MICP-L: Mesh-based ICP for Robot Localization using Hardware-Accelerated Ray Casting . In: 2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, pp. 10664--10671. doi:10.1109/IROS58592.2024.10802360

    arXiv 2024

  41. [46]

    In: 2018 IEEE International Conference on Robotics and Automation (ICRA)

    Nobre F, Heckman C, Ozog P, Wolcott RW and Walls JM (2018) Online Probabilistic Change Detection in Feature-Based Maps . In: 2018 IEEE International Conference on Robotics and Automation (ICRA). pp. 3661--3668. doi:10.1109/ICRA.2018.8461111

    arXiv 2018

  42. [47]

    In: 2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

    Park JM, Jang JH, Yoo SM, Lee SK, Kim UH and Kim JH (2021) ChangeSim: Towards End-to-End Online Scene Change Detection in Industrial Indoor Environments . In: 2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). pp. 8578--8585. doi:10.1109/IROS51168.2021.9636350

    arXiv 2021

  43. [48]

    Computer Networks 270: 111527

    Roohi K and Fekr AR (2025) A comparative analysis of indoor localization technologies . Computer Networks 270: 111527. doi:https://doi.org/10.1016/j.comnet.2025.111527

    arXiv 2025

  44. [49]

    In: Advanced Concepts for Intelligent Vision Systems: 20th International Conference, ACIVS 2020, Auckland, New Zealand, February 10--14, 2020, Proceedings 20

    Rozsa Z, Golarits M and Sziranyi T (2020) Localization of map changes by exploiting SLAM residuals . In: Advanced Concepts for Intelligent Vision Systems: 20th International Conference, ACIVS 2020, Auckland, New Zealand, February 10--14, 2020, Proceedings 20. Springer, pp. 312--324. doi:https://doi.org/10.1007/978-3-030-40605-9_27

    work page doi:10.1007/978-3-030-40605-9_27 2020

  45. [50]

    IEEE Robotics and Automation Letters

    Shaheer M, Millan-Romera JA, Bavle H, Giberna M, Sanchez-Lopez JL, Civera J and Voos H (2025) Tightly Coupled SLAM with Imprecise Architectural Plans . IEEE Robotics and Automation Letters

    2025

  46. [51]

    In: 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

    Shan T and Englot B (2018) LeGO-LOAM: Lightweight and Ground-Optimized Lidar Odometry and Mapping on Variable Terrain . In: 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). pp. 4758--4765. doi:10.1109/IROS.2018.8594299

    arXiv 2018

  47. [52]

    In: IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

    Shan T, Englot B, Meyers D, Wang W, Ratti C and Daniela R (2020) LIO-SAM: Tightly-coupled Lidar Inertial Odometry via Smoothing and Mapping . In: IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, pp. 5135--5142

    2020

  48. [53]

    In: Billard A, Dragan A, Peters J and Morimoto J (eds.) Proceedings of The 2nd Conference on Robot Learning, Proceedings of Machine Learning Research, volume 87

    Shan T, Wang J, Englot B and Doherty K (2018) Bayesian Generalized Kernel Inference for Terrain Traversability Mapping . In: Billard A, Dragan A, Peters J and Morimoto J (eds.) Proceedings of The 2nd Conference on Robot Learning, Proceedings of Machine Learning Research, volume 87. PMLR, pp. 829--838

    2018

  49. [54]

    Advanced Engineering Informatics 66: 103477

    St \"u hrenberg J and Smarsly K (2025) LIO-BIM--Coupling Lidar Inertial Odometry with Building Information Modeling for Robot Localization and Mapping . Advanced Engineering Informatics 66: 103477

    2025

  50. [55]

    In: ECPPM 2022-eWork and eBusiness in Architecture, Engineering and Construction 2022

    Torres MV, Braun A and Borrmann A (2023) OGM2PGBM: Robust BIM-based 2D-LiDAR localization for lifelong indoor navigation . In: ECPPM 2022-eWork and eBusiness in Architecture, Engineering and Construction 2022. CRC Press, pp. 567--574. doi:https://doi.org/10.1201/9781003354222

    work page doi:10.1201/9781003354222 2023

  51. [56]

    Construction Robotics 8(2): 13

    Vega-Torres MA, Braun A and Borrmann A (2024) SLAM2REF: advancing long-term mapping with 3D LiDAR and reference map integration for precise 6-DoF trajectory estimation and map extension . Construction Robotics 8(2): 13. doi:https://doi.org/10.1007/s41693-024-00126-w

    work page doi:10.1007/s41693-024-00126-w 2024

  52. [57]

    IEEE Robotics and Automation Letters (RA-L) 8(2): 1029--1036

    Vizzo I, Guadagnino T, Mersch B, Wiesmann L, Behley J and Stachniss C (2023) KISS-ICP: In Defense of Point-to-Point ICP -- Simple, Accurate, and Robust Registration If Done the Right Way . IEEE Robotics and Automation Letters (RA-L) 8(2): 1029--1036. doi:10.1109/LRA.2023.3236571

    arXiv 2023

  53. [58]

    In: 2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

    Wang X, Marcotte RJ and Olson E (2019) GLFP: Global localization from a floor plan . In: 2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). pp. 1627--1632. doi:10.1109/IROS40897.2019.8968061

    arXiv 2019

  54. [59]

    IEEE Robotics and Automation Letters 7(4): 9818--9825

    Wang Y, Song W, Lou Y, Zhang Y, Huang F, Tu Z and Liang Q (2022) Rail Vehicle Localization and Mapping With LiDAR-Vision-Inertial-GNSS Fusion . IEEE Robotics and Automation Letters 7(4): 9818--9825. doi:10.1109/LRA.2022.3190093

    arXiv 2022

  55. [60]

    In: 2017 IEEE International Symposium on Safety, Security and Rescue Robotics (SSRR)

    Wellhausen L, Dubé R, Gawel A, Siegwart R and Cadena C (2017) Reliable real-time change detection and mapping for 3D LiDARs . In: 2017 IEEE International Symposium on Safety, Security and Rescue Robotics (SSRR). pp. 81--87. doi:10.1109/SSRR.2017.8088144

    arXiv 2017

  56. [61]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision

    Wild L, Valencia R and Jensfelt P (2025) ArgoTweak: Towards Self-Updating HD Maps through Structured Priors . In: Proceedings of the IEEE/CVF International Conference on Computer Vision. pp. 6091--6100

    2025

  57. [62]

    IEEE Robotics and Automation Letters 9(1): 531--538

    Xie F and Schwertfeger S (2023) Robust lifelong indoor lidar localization using the area graph . IEEE Robotics and Automation Letters 9(1): 531--538. doi:10.1109/LRA.2023.3334158

    arXiv 2023

  58. [63]

    IEEE Transactions on Robotics 38(4): 2053--2073

    Xu W, Cai Y, He D, Lin J and Zhang F (2022) FAST-LIO2: Fast Direct LiDAR-Inertial Odometry . IEEE Transactions on Robotics 38(4): 2053--2073. doi:10.1109/TRO.2022.3141876

    arXiv 2022

  59. [64]

    Automation in Construction 162: 105344

    Yarovoi A and Cho YK (2024) Review of simultaneous localization and mapping (SLAM) for construction robotics applications . Automation in Construction 162: 105344. doi:https://doi.org/10.1016/j.autcon.2024.105344

    arXiv 2024

  60. [65]

    Automation in Construction 146: 104641

    Yin H, Lin Z and Yeoh JK (2023) Semantic localization on BIM-generated maps using a 3D LiDAR sensor . Automation in Construction 146: 104641. doi:https://doi.org/10.1016/j.autcon.2022.104641

    arXiv 2023

  61. [66]

    International Journal of Computer Vision 132(8): 3139--3171

    Yin H, Xu X, Lu S, Chen X, Xiong R, Shen S, Stachniss C and Wang Y (2024) A survey on global lidar localization: Challenges, advances and open problems . International Journal of Computer Vision 132(8): 3139--3171. doi:https://doi.org/10.1007/s11263-024-02019-5

    doi:10.1007/s11263-024-02019-5 2024

  62. [67]

    In: 2023 IEEE International Conference on Robotics and Automation (ICRA)

    Yuan C, Lin J, Zou Z, Hong X and Zhang F (2023) STD: Stable Triangle Descriptor for 3D place recognition . In: 2023 IEEE International Conference on Robotics and Automation (ICRA). pp. 1897--1903. doi:10.1109/ICRA48891.2023.10160413

    arXiv 2023

  63. [68]

    Automation in Construction 178: 106415

    Zhang H, Yan J, Yang J, Meng W and Chen S (2025) Two-stage point cloud registration using multi-scale edge convolution for digital twin-based bridge construction progress monitoring . Automation in Construction 178: 106415. doi:https://doi.org/10.1016/j.autcon.2025.106415

    arXiv 2025

  64. [69]

    In: Proceedings of Robotics: Science and Systems

    Zhang J and Singh S (2014) LOAM: Lidar odometry and mapping in real-time . In: Proceedings of Robotics: Science and Systems. Berkeley, CA, pp. 1--9. doi:10.15607/RSS.2014.X.007

    doi:10.15607/rss.2014.x.007 2014

  65. [70]

    Automation in Construction 168: 105796

    Zhang Y, Lu L, Luo X and Pan J (2024) Global BIM-point cloud registration and association for construction progress monitoring . Automation in Construction 168: 105796. doi:https://doi.org/10.1016/j.autcon.2024.105796

    arXiv 2024

  66. [71]

    IEEE Robotics and Automation Letters 7(3): 7163--7170

    Zhou L, Huang G, Mao Y, Yu J, Wang S and Kaess M (2022) PLC -LiSLAM: LiDAR SLAM With Planes, Lines, and Cylinders . IEEE Robotics and Automation Letters 7(3): 7163--7170. doi:10.1109/LRA.2022.3180116

    arXiv 2022

  67. [72]

    IEEE Robotics and Automation Letters 6(4): 7073--7080

    Zhou L, Koppel D and Kaess M (2021) LiDAR SLAM With Plane Adjustment for Indoor Environment . IEEE Robotics and Automation Letters 6(4): 7073--7080. doi:10.1109/LRA.2021.3092274

    arXiv 2021

  68. [73]

    Journal of Field Robotics 41(7): 2455--2474

    Zou Z, Yuan C, Xu W, Li H, Zhou S, Xue K and Zhang F (2024) LTA-OM: Long-term association LiDAR--IMU odometry and mapping . Journal of Field Robotics 41(7): 2455--2474. doi:https://doi.org/10.1002/rob.22337

    doi:10.1002/rob.22337 2024