pith. sign in

arxiv: 2605.08159 · v1 · submitted 2026-05-04 · 📡 eess.SP

Sensor-Outage-Aware Spatio-Temporal Graph Reconstruction of High-Rise Facade Pressure Fields

Seyedeh Fatemeh Mirfakhar , Reda Snaiki This is my paper

Pith reviewed 2026-05-12 02:26 UTC · model grok-4.3

classification 📡 eess.SP
keywords sensor outagespatio-temporal graphfacade pressure fieldhigh-rise buildingpressure reconstructionwind tunnel datagraph propagationtime series imputation
0
0 comments X

The pith

Spatio-temporal graph reconstruction completes high-rise facade pressure fields from sparse outage-prone sensors.

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

This paper develops a graph-based method to reconstruct time-varying pressure distributions across building facades using only limited sensor readings that may drop out temporarily. Such reconstructions matter for designing tall buildings to withstand wind loads without needing prohibitively expensive full instrumentation. The approach builds a single representation of the facade surface, pulls out time patterns from whatever data is currently available, and spreads that information across connected locations to fill gaps both at broken sensors and at completely unmeasured spots. Validation against wind-tunnel measurements under varying wind angles confirms that the main time changes, frequency content, and overall spatial structures are retained, although side and rear faces prove harder to recover than the front face. An extension shows how the same model can support forecasting future pressure fields by first predicting at available sensors and then completing the map.

Core claim

The framework represents the building facade as a graph domain and combines a temporal feature extractor with spatial message passing, guided by a binary indicator that marks which sensor channels are currently observed. This allows simultaneous recovery of dropped signals at instrumented locations and estimation of pressure at all other facade points. On wind-tunnel pressure coefficient records for a high-rise model, the reconstructions match the recorded data closely at instrumented sites and produce plausible full-field maps elsewhere, maintaining the dominant temporal behavior, key spectral peaks, and coherent spatial organization while leaving the largest errors in localized high-freq u

What carries the argument

Unified facade graph with temporal feature extraction and spatial propagation driven by an explicit sensor-availability mask.

Load-bearing premise

The specific graph connectivity and the temporal feature extractor tuned on this wind-tunnel dataset will transfer directly to different building shapes and real outdoor sensor outage patterns.

What would settle it

If the model is applied to pressure measurements from a second, geometrically distinct building or from actual field sensors and the reconstructed fields deviate substantially from direct measurements in both time and space, the claim of reliable outage-tolerant reconstruction would be falsified.

Figures

Figures reproduced from arXiv: 2605.08159 by Reda Snaiki, Seyedeh Fatemeh Mirfakhar.

Figure 2
Figure 2. Figure 2 [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
read the original abstract

Time-resolved facade pressure fields are essential for the wind-resistant design and aerodynamic assessment of high-rise buildings. However, dense instrumentation is costly and often impractical, and sensor outages can further reduce data availability. This study proposes a sensor-outage-aware spatio-temporal graph reconstruction framework for completing facade pressure fields from sparse measurements. The method couples temporal feature extraction with graph-based spatial propagation on a unified facade-domain representation and uses an explicit observation-availability indicator to handle temporarily unavailable sensor signals while reconstructing both missing instrumented channels and non-instrumented locations. The framework is evaluated using wind-tunnel pressure coefficient data for a high-rise building across windward, lateral, and leeward facades under multiple wind directions. The results show reliable outage-tolerant reconstruction at instrumented sensors and accurate full-field completion at non-instrumented nodes, with reconstruction generally most accurate on the windward facade and more challenging on the lateral and leeward facades. Time-domain, spectral, and spatial validations further show that the framework preserves the dominant temporal evolution, principal dynamic content, and coherent large-scale pressure-field organization, while the largest residual discrepancies remain localized in higher-frequency or intermittent components. A two-stage predictive extension is also outlined, in which future sensor signals are forecast at available instrumented locations and then mapped to future full-field pressure estimates through the proposed reconstruction model.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 2 minor

Summary. The manuscript proposes a sensor-outage-aware spatio-temporal graph reconstruction framework that couples temporal feature extraction with graph-based spatial propagation over a unified facade representation, using an explicit observation-availability indicator to reconstruct both missing instrumented channels and non-instrumented locations from sparse wind-tunnel pressure-coefficient measurements on a single high-rise building across multiple wind directions. The central claims are that the method delivers reliable outage-tolerant reconstruction at instrumented sensors, accurate full-field completion at non-instrumented nodes, and preservation of dominant temporal evolution, principal dynamic content, and large-scale spatial organization, with a two-stage predictive extension also outlined.

Significance. If the quantitative claims are substantiated, the work would offer a practical advance in wind-engineering signal processing by enabling robust completion of facade pressure fields under realistic sensor failures, reducing the need for dense instrumentation. The explicit outage handling and graph formulation for irregular physical domains represent a relevant contribution to spatio-temporal modeling; however, the single-geometry evaluation limits broader significance until generalization is demonstrated.

major comments (3)
  1. Evaluation/results: no quantitative metrics (RMSE, correlation, spectral error norms), error bars, or baseline comparisons (e.g., against kriging, PCA, or standard GNNs) are supplied to support the assertions of 'reliable' and 'accurate' reconstruction or the preservation of temporal/spectral content; the abstract alone supplies none of these, and the full text must provide them to make the central claim verifiable.
  2. Experimental setup and generalization: all reported results use a single high-rise geometry and sensor layout from wind-tunnel tests under a limited set of wind directions; no ablation on graph connectivity, no tests on altered aspect ratios or sensor densities, and no real-building data are described, leaving the weakest assumption (generalization of the learned graph and temporal extractor) untested and load-bearing for the reliability claim.
  3. Method description: the precise definition of the facade graph (adjacency construction, node features) and the architecture/details of the temporal feature extractor are insufficiently specified to allow reproduction or to confirm that the observation-availability indicator is integrated without introducing circular dependence on the training data.
minor comments (2)
  1. Abstract: the phrase 'time-domain, spectral, and spatial validations' is used without naming the concrete metrics or figures that implement them.
  2. The two-stage predictive extension is only 'outlined'; if intended as a contribution, it requires at least a brief quantitative demonstration or explicit statement that it is future work.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive and detailed comments. We address each major point below, indicating where revisions will be made to strengthen the manuscript.

read point-by-point responses

  1. Referee: Evaluation/results: no quantitative metrics (RMSE, correlation, spectral error norms), error bars, or baseline comparisons (e.g., against kriging, PCA, or standard GNNs) are supplied to support the assertions of 'reliable' and 'accurate' reconstruction or the preservation of temporal/spectral content; the abstract alone supplies none of these, and the full text must provide them to make the central claim verifiable.

    Authors: We agree that the current presentation relies primarily on qualitative descriptions of performance. The revised manuscript will incorporate quantitative metrics, including RMSE, correlation coefficients, and spectral error norms, with error bars derived from repeated trials or cross-validation folds. Baseline comparisons against kriging, PCA-based methods, and standard GNNs will also be added to allow direct assessment of the proposed approach. revision: yes

  2. Referee: Experimental setup and generalization: all reported results use a single high-rise geometry and sensor layout from wind-tunnel tests under a limited set of wind directions; no ablation on graph connectivity, no tests on altered aspect ratios or sensor densities, and no real-building data are described, leaving the weakest assumption (generalization of the learned graph and temporal extractor) untested and load-bearing for the reliability claim.

    Authors: The evaluation is performed on wind-tunnel data from one high-rise geometry, which is standard for initial validation studies in wind engineering. We will add an ablation study on graph connectivity in the revision and expand the discussion to explicitly address the scope and limitations of generalization. New experiments on varied aspect ratios, sensor densities, or real-building data lie outside the current scope and are identified as future work. revision: partial

  3. Referee: Method description: the precise definition of the facade graph (adjacency construction, node features) and the architecture/details of the temporal feature extractor are insufficiently specified to allow reproduction or to confirm that the observation-availability indicator is integrated without introducing circular dependence on the training data.

    Authors: We acknowledge that the method section requires greater specificity. The revised manuscript will provide: explicit construction details for the facade graph (adjacency via spatial proximity with distance thresholds and node features combining pressure coefficients with normalized coordinates); full architectural specifications for the temporal feature extractor (layer types, dimensions, and hyperparameters); and a clear description, including pseudocode, of how the observation-availability indicator is applied to prevent any circular dependence on training data. revision: yes

Circularity Check

0 steps flagged

No circularity: empirical framework evaluated on external wind-tunnel benchmarks

full rationale

The paper describes a sensor-outage-aware spatio-temporal graph reconstruction method that couples temporal feature extraction, graph-based spatial propagation, and an observation-availability indicator. All reported results are obtained by applying this model to independent wind-tunnel pressure-coefficient measurements for a single high-rise geometry. No equations, derivations, or self-citations are shown that reduce the reconstructed fields or performance metrics to quantities defined from the same data by construction. The evaluation metrics (time-domain, spectral, spatial) are computed against held-out sensor readings and full-field references that are not used in fitting the graph connectivity or feature extractor. This keeps the derivation chain self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the domain assumption that facade pressure can be usefully represented as a graph whose edges capture spatial correlations and that temporal features can be extracted independently before spatial propagation.

axioms (1)
  • domain assumption Facade pressure fields admit a graph representation in which nodes are spatial locations and edges encode relevant spatial correlations.
    Invoked by the choice of graph-based spatial propagation on a unified facade-domain representation.

pith-pipeline@v0.9.0 · 5547 in / 1242 out tokens · 34462 ms · 2026-05-12T02:26:44.887876+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

41 extracted references · 41 canonical work pages

  1. [1]

    Modal Analysis of Offshore Monopile Wind Turbine: An Analytical Solution,

    H. Pezeshki, D. Pavlou, H. Adeli, and S. C. Siriwardane, “Modal Analysis of Offshore Monopile Wind Turbine: An Analytical Solution,” Journal of Offshore Mechanics and Arctic Engineering, vol. 145, p. 010907, 2022

    work page 2022

  2. [2]

    Review of Wind -Induced Effects Estimation through Nonlinear Analysis of Tall Buildings, High -Rise Structures, Flexible Bridges and Transmission Lines,

    S. Zhao, C. Zhang, X. Dai, and Z. Yan, “Review of Wind -Induced Effects Estimation through Nonlinear Analysis of Tall Buildings, High -Rise Structures, Flexible Bridges and Transmission Lines,” Buildings, vol. 13, no. 8, p. 2033, 2023

    work page 2033

  3. [3]

    A fully-coupled generalized model for multi -directional wind loads on tall buildings: A development of the quasi-steady theory,

    W. Cui and L. Cara coglia, “A fully-coupled generalized model for multi -directional wind loads on tall buildings: A development of the quasi-steady theory,” Journal of Fluids and Structures, vol. 78, pp. 52–68, 2018

    work page 2018

  4. [4]

    The effects of a double -skin façade on the cladding pressure around a tall building,

    G. Hu, J. Song, S. Hassanli, R. Ong, and K. C. Kwok, “ The effects of a double -skin façade on the cladding pressure around a tall building,” Journal of Wind Engineering and Industrial Aerodynamics, vol. 191, pp. 239–251, 2019

    work page 2019

  5. [5]

    Ensemble flow reconstruction in the atmospheric boundary layer from spatially limited measurements through latent diffusion models,

    A. Rybchuk, M. Hassanaly, N. Hamilton, P. Doubrawa, M. J. Fulton, and L. A. Mart ínez-Tossas, “Ensemble flow reconstruction in the atmospheric boundary layer from spatially limited measurements through latent diffusion models,” Physics of Fluids, vol. 35, no. 12, p. 126604, 2023

    work page 2023

  6. [6]

    High -Performance Tall Buildings: An Overview of Recent Developments,

    K. Al -Kodmany and M. M. Ali, “High -Performance Tall Buildings: An Overview of Recent Developments,” Encyclopedia, vol. 5, no. 2, p. 53, 2025

    work page 2025

  7. [7]

    Dynamic performance of façade systems during lifting and transportation in prefabricated modules,

    R. Hajirezaei and P. Sharafi, “Dynamic performance of façade systems during lifting and transportation in prefabricated modules,” Structures, vol. 85, p. 111089, 2026

    work page 2026

  8. [8]

    Comparative study between field measurements of wind pressures on a 600 -m-high skyscraper during Super Typhoon Mangkhut and wind tunnel test,

    X.-L. Han, Q.-S. Li, K. Zhou, and X. Li, “Comparative study between field measurements of wind pressures on a 600 -m-high skyscraper during Super Typhoon Mangkhut and wind tunnel test,” Engineering Structures, vol. 272, p. 114958, 2022

    work page 2022

  9. [9]

    CFD evaluation of mean and turbulent wind characteristics around a high-rise building affected by its surroundings,

    M. Shirzadi and Y. Tominaga, “CFD evaluation of mean and turbulent wind characteristics around a high-rise building affected by its surroundings,” Building and Environment , vol. 225, p. 109637, 2022

    work page 2022

  10. [10]

    Influence of geometric parameters of façade protruding ribs on turbulent flow statistics in street canyons: A large-eddy simulation study,

    Y. Zhang and X. Zheng, “Influence of geometric parameters of façade protruding ribs on turbulent flow statistics in street canyons: A large-eddy simulation study,” Building and Environment, vol. 287, p. 113845, 2026

    work page 2026

  11. [11]

    Dynamic mode decomposition of random pressure fields over bluff bodies,

    X. Luo and A. Kareem, “Dynamic mode decomposition of random pressure fields over bluff bodies,” Journal of Engineering Mechanics, vol. 147, no. 4, p. 04021007, 2021

    work page 2021

  12. [12]

    Wind load impact on tall building facades: Damage observations during severe wind events and wind tunnel testing,

    O. Metwally, H. A. Ibrahim, A. Elawady, I. Zisis, and A. G. Chowdhury, “Wind load impact on tall building facades: Damage observations during severe wind events and wind tunnel testing,” Frontiers in Built Environment, vol. 10, p. 1514523, 2025

    work page 2025

  13. [13]

    Deep reinforcement learning-based active flow control for a tall building,

    L. Yan, H. Cai, Q. Wang, L. Chen, C. Li, and G. Hu, “Deep reinforcement learning-based active flow control for a tall building,” Physics of Fluids, vol. 37, no. 4, 2025

    work page 2025

  14. [14]

    Unsteady RANS simulation of wind flow around a building shape obstacle,

    S. Hassan, Md. M. Molla, P. Nag, N. Akhter, and A. Khan, “Unsteady RANS simulation of wind flow around a building shape obstacle,” Building Simulation, vol. 15, no. 2, pp. 291–312, 2022

    work page 2022

  15. [15]

    Numerical simulation of turbulent airflow around a tall telecommunication tower model,

    A. Moradi, E. Salajegheh, M. M. Tavakol, A. Heidari, and G. Ahmadi, “Numerical simulation of turbulent airflow around a tall telecommunication tower model,” Advances in Aerodynamics, vol. 6, no. 1, p. 26, 2024

    work page 2024

  16. [16]

    Aerodynamic performance opti mization of high -rise structures under extreme wind,

    H. Yadav and A. K. Roy, “Aerodynamic performance opti mization of high -rise structures under extreme wind,” Journal of Building Engineering, vol. 114, p. 114364, 2025. 25

    work page 2025

  17. [17]

    Aerodynamic shape optimization of civil structures: A CFD -enabled Kriging-based approach,

    E. Bernardini, S. M. Spence, D. Wei, and A. Kareem, “Aerodynamic shape optimization of civil structures: A CFD -enabled Kriging-based approach,” Journal of Wind Engineering and Industrial Aerodynamics, vol. 144, pp. 154–164, 2015

    work page 2015

  18. [18]

    CFD simulation advances in urban aerodynamics: Accuracy, validation, and high-rise building applications,

    A. Idrissi, H. El Mghari, and R. El Amraoui, “CFD simulation advances in urban aerodynamics: Accuracy, validation, and high-rise building applications,” Results in Engineering, vol. 26, p. 105148, 2025

    work page 2025

  19. [19]

    Sensitivity analysis of wind pressure coefficients on CAARC standard tall buildings in CFD simulations,

    F.-Q. Meng, B.-J. He, J. Zhu, D. -X. Zhao, A. Darko, and Z. -Q. Zhao, “Sensitivity analysis of wind pressure coefficients on CAARC standard tall buildings in CFD simulations,” Journal of Building Engineering, vol. 16, pp. 146–158, 2018

    work page 2018

  20. [20]

    Active Simulation of Transient Wind Field in a Multiple -Fan Wind Tunnel via Deep Reinforcement Learning,

    S. Li, R. Snaiki, and T. Wu, “Active Simulation of Transient Wind Field in a Multiple -Fan Wind Tunnel via Deep Reinforcement Learning,” Journal of Engineering Mechanics , vol. 147, no. 9, p. 04021056, 2021

    work page 2021

  21. [21]

    Cyber-physical aerodynamic shape optimization of a tall building in a wind tunnel using an active fin system,

    M. L. Whiteman, P. L. Fernández-Cabán, B. M. Phillips, F. J. Masters, J. R. Davis, and J. A. Bridge, “Cyber-physical aerodynamic shape optimization of a tall building in a wind tunnel using an active fin system,” Journal of Wind Engineering and Industrial Aerodynamics, vol. 220, p. 104835, 2022

    work page 2022

  22. [22]

    Accuracy of CFD simulations in urban aerodynamics and microclimate: Progress and challenges,

    Y. Tominaga, L. Wang, Z. Zhai, and T. Stathopoulos, “Accuracy of CFD simulations in urban aerodynamics and microclimate: Progress and challenges,” Building and Environment, vol. 243, p. 110723, 2023

    work page 2023

  23. [23]

    Design and demonstration of a sensing network for full -scale wind pressure measurements on buildings,

    J. Hochschild an d C. Gorlé, “Design and demonstration of a sensing network for full -scale wind pressure measurements on buildings,” Journal of Wind Engineering and Industrial Aerodynamics , vol. 250, p. 105760, 2024

    work page 2024

  24. [24]

    Sensor se lection with cost constraints for dynamically relevant bases,

    E. Clark, J. N. Kutz, and S. L. Brunton, “Sensor se lection with cost constraints for dynamically relevant bases,” IEEE Sensors Journal, vol. 20, no. 19, pp. 11674–11687, 2020

    work page 2020

  25. [25]

    Computational methodologies for optimal sensor placement in structural health monitoring: A review,

    Y. Tan and L. Zhang, “Computational methodologies for optimal sensor placement in structural health monitoring: A review,” Structural Health Monitoring, vol. 19, no. 4, pp. 1287–1308, 2020

    work page 2020

  26. [26]

    Optimal sensor placement for reconstructing wind pressure field around buildings using compressed sensing,

    X. Luo, A. Kareem, and S. Yoo, “Optimal sensor placement for reconstructing wind pressure field around buildings using compressed sensing,” Journal of Building Engineering , vol. 75, p. 106855, 2023

    work page 2023

  27. [27]

    Reliability in wireless sensor networks: A survey and challenges ahead,

    M. A. Mahmood, W. K. G. Seah, and I. Welch, “Reliability in wireless sensor networks: A survey and challenges ahead,” Computer Networks, vol. 79, pp. 166–187, 2015

    work page 2015

  28. [28]

    Degradation Detection in a Redundant Sensor Architecture,

    A. Kajmakovic, K. Diwold, K. Römer, J. Pestana, and N. Kajtazovic, “Degradation Detection in a Redundant Sensor Architecture,” Sensors, vol. 22, no. 12, p. 4649, 2022

    work page 2022

  29. [29]

    A comprehensive review of sensor technologies, instrumentation, and signal processing solutions for low -power Internet of Things systems with mini-computing devices,

    A. Gazis, I. Papadongonas, A. Andriopoulos, C. Zioudas, and T. Vavouras, “A comprehensive review of sensor technologies, instrumentation, and signal processing solutions for low -power Internet of Things systems with mini-computing devices,” arXiv preprint arXiv:2503.13466, 2025

    work page arXiv 2025

  30. [30]

    Higher order dynamic mode decomposition of wind pressures on square buildings,

    L. Zhou, K. T. Tse, G. Hu, and Y. Li, “Higher order dynamic mode decomposition of wind pressures on square buildings,” Journal of Wind Engineering and Industrial Aerodynamics, vol. 211, p. 104545, 2021

    work page 2021

  31. [31]

    Machi ne learning enhanced real -time aerodynamic forces prediction based on sparse pressure sensor inputs,

    J. Duan, Q. Wang, and J. S. Hesthaven, “Machi ne learning enhanced real -time aerodynamic forces prediction based on sparse pressure sensor inputs,” AIAA Journal, vol. 62, no. 7, pp. 2601 –2621, 2024

    work page 2024

  32. [32]

    A hybrid machine learning framework for wind pressure prediction on buildings with constrained sensor networks,

    F. M. Nav, S. F. Mirfakhar, and R. Snaiki, “A hybrid machine learning framework for wind pressure prediction on buildings with constrained sensor networks,” Computer-Aided Civil and Infrastructure Engineering, vol. 40, no. 19, pp. 2816–2832, 2025

    work page 2025

  33. [33]

    Discovering governing equations from data by sparse identification of nonlinear dynamical systems,

    S. L. Brunton, J. L. Proctor, and J. N. Kutz, “Discovering governing equations from data by sparse identification of nonlinear dynamical systems,” Proceedings of the National Academy of Sciences , vol. 113, no. 15, pp. 3932–3937, 2016

    work page 2016

  34. [34]

    Data -Driven Sparse Sensor Placement for Reconstruction: De monstrating the Benefits of Exploiting Known Patterns,

    K. Manohar, B. W. Brunton, J. N. Kutz, and S. L. Brunton, “Data -Driven Sparse Sensor Placement for Reconstruction: De monstrating the Benefits of Exploiting Known Patterns,” IEEE Control Systems Magazine, vol. 38, no. 3, pp. 63–86, 2018. 26

    work page 2018

  35. [35]

    Global field reconstruction from sparse sensors with Voronoi tessellation-assisted deep learning,

    K. Fukami, R. Maulik, N. Ramachandra, K. Fukagata, and K. Taira, “Global field reconstruction from sparse sensors with Voronoi tessellation-assisted deep learning,” Nature Machine Intelligence, vol. 3, no. 11, pp. 945–951, 2021

    work page 2021

  36. [36]

    Machine learning based algorithms for wind pressure prediction of high-rise buildings,

    Y. Li, X. Huang, Y. -G. Li, F.-B. Chen, and Q.-S. Li, “Machine learning based algorithms for wind pressure prediction of high-rise buildings,” Advances in Structural Engineering, vol. 25, no. 10, pp. 2222–2233, 2022

    work page 2022

  37. [37]

    Forecasting of pressure coefficient for wind interference due to surrounding tall building on a tall rectangular building using CFD data trained machine learning models,

    H. Verma and R. Sonparote, “Forecasting of pressure coefficient for wind interference due to surrounding tall building on a tall rectangular building using CFD data trained machine learning models,” Structures, vol. 75, p. 108705, 2025

    work page 2025

  38. [38]

    Mitigation of wind load on tall buildings through aerodynamic modifications: Review,

    A. Sharma, H. Mittal, and A. Gairola, “Mitigation of wind load on tall buildings through aerodynamic modifications: Review,” Journal of Building Engineering, vol. 18, pp. 180–194, 2018

    work page 2018

  39. [39]

    Advances in Computer Vision -Based Civil Infrastructure Inspection and Monitoring,

    B. F . Spencer, V. Hoskere, and Y. Narazaki, “Advances in Computer Vision -Based Civil Infrastructure Inspection and Monitoring,” Engineering, vol. 5, no. 2, pp. 199–222, 2019

    work page 2019

  40. [40]

    Machine learning paradigm for structural health monitoring,

    Y. Bao and H. Li, “Machine learning paradigm for structural health monitoring,” Structural Health Monitoring, vol. 20, no. 4, pp. 1353–1372, 2021

    work page 2021

  41. [41]

    Deep Learning for Structural Health Monitoring: Data, Algorithms, Applications, Challenges, and Trends,

    J. Jia and Y. Li, “Deep Learning for Structural Health Monitoring: Data, Algorithms, Applications, Challenges, and Trends,” Sensors, vol. 23, no. 21, p. 8824, 2023

    work page 2023