Pith. sign in

REVIEW 4 major objections 6 minor 44 references

A Transformer that fuses mismatched ultrasonic guided-wave and FBG strain streams predicts composite health indicators with under-0.1 error and localizes damage under fatigue loading.

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

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

T0 review · grok-4.5

2026-07-12 12:10 UTC pith:Q3LSEZP3

load-bearing objection Solid applied multi-modal Transformer for asynchronous PZT/FBG fusion on real ReMAP panels; gains are real relative to single-sensor and DNN baselines, but rest on AE-derived labels without independent NDT ground truth. the 4 major comments →

arxiv 2607.02545 v1 pith:Q3LSEZP3 submitted 2026-06-24 eess.SP cs.LG

Transformer-based Multisensor Data Fusion of Ultrasonic Guided Wave and FBG-based Strain Measurements for Multitask Aerospace Structural Health Monitoring

Xin Yang , Morteza Moradi , Tongtong Yan , Jinbo Du , Yunlai Liao , Dimitrios Zarouchas , Dimitrios Chronopoulos This is my paper
classification eess.SP cs.LG
keywords Structural health monitoringMultisensor data fusionTransformer networkMultitask learningUltrasonic guided wavesFiber Bragg grating sensorHealth indicatorDamage localization
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.

Aircraft composite panels fail gradually under cyclic loads, and no single sensor sees the full picture: ultrasonic guided waves from PZTs catch discontinuities at high frequency but only every few thousand cycles, while FBG strain gauges run more often yet lack the same spatial sensitivity. This paper claims that tokenizing both streams, padding the gaps, and letting a Transformer attend across them yields a single model that both tracks a continuous health indicator and produces a damage heatmap. On four stiffened composite skins run to failure under compression-compression fatigue, the fused model keeps health-indicator MAE and RMSE below 0.1—roughly 60 percent better than either sensor alone or ordinary deep-learning baselines—and keeps localization MAE/RMSE below 0.0465/0.1571. Attention maps show the network actually trading information between the two modalities rather than simply concatenating them. If the claim holds, condition-based maintenance of composite airframes can rest on fewer, cheaper sensor suites that still deliver both prognosis and location.

Core claim

When PZT spectrogram patches and multi-channel FBG time-series segments are turned into a shared token sequence, masked for missing acquisition intervals, and fused by multi-head cross-modal attention inside a Transformer, the resulting multitask network simultaneously predicts an acoustic-emission-derived health indicator and a pixel-wise damage localization map with substantially lower error than single-modality or conventional DNN baselines on the same fatigue-tested composite panels.

What carries the argument

Cycle-aware masked token fusion: STFT patches of guided-wave signals and windowed FBG strain vectors are linearly projected to a common dimension, concatenated, zero-padded to fixed length, and tagged with fatigue-cycle and modality embeddings; multi-head self-attention then performs both intra- and cross-modal fusion before shared multitask heads (health-indicator regression plus localization heatmap plus a reconstruction regularizer).

Load-bearing premise

The paper treats averaged acoustic-emission cumulative energy and Gaussian-KDE maps built from AE time-of-arrival triangulation as the true ground-truth labels for both health-indicator regression and damage location.

What would settle it

Repeat the identical fatigue campaign on a fifth panel instrumented with an independent imaging modality (for example, full-field digital image correlation or ultrasonic C-scan at every PZT acquisition step) and check whether the Transformer’s predicted heatmaps and health-indicator curves still match the new, non-AE labels within the reported error bounds.

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

If this is right

  • SHM systems can drop the requirement that every sensor fire at the same rate; missing intervals are simply masked and still contribute to both prognosis and localization.
  • A single trained network replaces separate pipelines for health-indicator tracking and damage imaging on stiffened composite skins.
  • Attention weights between PZT spectral patches and FBG channels become an inspectable diagnostic of which sensor events drive each prediction.
  • Cross-validation on an out-of-distribution panel (roughly double the fatigue life and a different impact site) still yields R^{2} > 0.8, suggesting the fusion transfers beyond the training lifetime distribution.

Where Pith is reading between the lines

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

  • The same tokenization-plus-masking pattern could absorb a third sparse modality (for example, acoustic emission raw waveforms or vibration spectra) without redesigning the architecture.
  • Because the reconstruction head is discarded at inference, the model could be distilled into a lighter student network for on-board avionics once the fused representation is learned.
  • If AE labels prove biased, the attention maps themselves might be re-purposed as an unsupervised anomaly detector that does not rely on AE at all.

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 / 6 minor

Summary. The paper proposes a Transformer-based early-fusion framework that tokenizes heterogeneous PZT ultrasonic guided-wave spectrograms and multi-channel FBG strain series, aligns them with cycle-aware temporal and modality embeddings plus binary padding masks, and jointly performs health-indicator (HI) regression and pixel-wise damage localization. Labels are derived from continuous AE: cumulative energy (selected via Mo–Pr–Tr) for HI and Gaussian-KDE maps from anisotropic TDoA triangulation for localization. On four ReMAP CFRP skin-stiffener panels under compression–compression fatigue, leave-one-panel-out results show fusion MAE/RMSE below ~0.08 for HI (claimed ~60% better than PZT-only, FBG-only, and 1D-CNN/Bi-LSTM/CNN-LSTM baselines) and localization MAE/RMSE below 0.0465/0.1571 with improved SSIM. Attention maps are used to illustrate cross-modal focus.

Significance. If the reported gains hold under stronger validation, the work is a useful contribution to aerospace SHM: it addresses a practical multi-rate fusion problem (PZT every 5000 cycles vs FBG every 500), provides an end-to-end multitask Transformer with explicit missing-modality handling, and supplies attention visualizations that aid interpretability. The experimental setting (real Embraer-design stiffened panels, fatigue to failure, multi-sensor ReMAP data) is more realistic than many synthetic SHM studies. Strengths include clear single-sensor and DNN baselines (Tables 3–5), quantitative localization metrics including SSIM, and an explicit reconstruction regularizer for sparse tokens. The main significance is methodological (heterogeneous token fusion under fatigue acquisition mismatch) rather than a definitive new physical damage metric.

major comments (4)
  1. [§3.1–3.2, Tables 3–4, Fig. 12] §3.1–3.2 and Tables 3–4: All supervised targets are AE-derived (channel-averaged cumulative energy; Gaussian-KDE of TDoA sources under the anisotropic velocity model in Eqs. 17–24). PZT/FBG never see independent damage references (C-scan, residual stiffness/strength, post-test NDT, or even direct comparison of localization peaks to the known 10 J impact sites beyond the qualitative red boxes in Fig. 12). The headline “nearly 60% improvement” and “highest accuracy” therefore measure agreement with AE proxies, not verified internal damage. AE can miss quiescent growth and is sensitive to placement and velocity assumptions. Reframe claims as “prediction of AE-based HI/localization from PZT+FBG,” quantify agreement of AE maps with known impact locations, and discuss bias risk; ideally add at least one AE-independent check on a subset of cycles or panels.
  2. [§4.3.2–4.3.3, Tables 3–4, Abstract] §4.3.2–4.3.3: Evaluation uses only four panels in leave-one-out; L04 is a clear OOD case (roughly double life, edge impact) and already shows the worst HI and localization errors. No confidence intervals, repeated splits, or significance tests accompany MAE/RMSE/R²/SSIM. With N=4 the ~60% improvement claim is fragile. Report variability (e.g., bootstrap or multi-seed runs), temper generalization language in the abstract/conclusion, and expand the limitations discussion on specimen count and OOD behavior.
  3. [§4.3.2, Eq. (9)] §4.3.2: OOD handling for L04 relies on an impact-region spatial embedding described only narratively (“plate partitioned into predefined spatial regions”). There is no ablation removing this prior, no formal definition of the embedding, and no sensitivity study. Because the paper attributes L04 generalization largely to this physics-informed input, the embedding must be specified (region map, encoding, how it is injected into Eq. 9 / H^(0)) and ablated against a no-region baseline so that fusion gains are not confounded with location side-information.
  4. [§2.2.4, Eq. (12)–(13), §4.3.1] §2.2.4, Eq. (12)–(13): Multitask weights λ_det=1.0, λ_loc=0.5, λ_rec=0.1 and the two-stage freeze of prediction/localization heads for 10 epochs are stated without sensitivity analysis. Given sparse, asynchronous tokens, results may depend strongly on these choices and on T_max / STFT–window tokenization. Provide at least a limited ablation on λ’s and on reconstruction loss on/off, and state how T_max and token lengths were chosen.
minor comments (6)
  1. [Abstract, §5] Abstract and §5: “nearly 60% performance improvement” should cite the exact baseline average (which metric, which folds) so the percentage is reproducible from Table 3.
  2. [§4.3.3] Table 4 text: “0.0465 and 1571” appears to omit the decimal for RMSE (should be 0.1571); fix typographical error.
  3. [Figures 9–13, Tables 3–4] Fig. 9–13 and several figure captions have spacing/OCR artifacts (“T rain”, “Data F usion”, “/s8722”); clean for production.
  4. [§2.1–2.2] §2.1 Eq. (2): notation mixes Y_det and G_loc; later sections use detection/prediction interchangeably for HI. Unify terminology (HI prediction vs detection) throughout.
  5. [§1] Related work claims no prior PZT+FBG fusion for aircraft composites; a short check against multi-sensor ReMAP/companion papers (e.g., Broer et al. fusion diagnostics) would strengthen novelty positioning.
  6. [Table 5, §4.3.4] Inference-time comparison (Table 5) is useful; note hardware and batch size explicitly and whether times are per sample or per panel trajectory.

Circularity Check

1 steps flagged

No definitional or fitted-input circularity; AE labels are an independent sensor modality used as supervised targets, and the fusion gains are empirical leave-one-out comparisons against single-modality baselines on the same targets.

specific steps
  1. self citation load bearing [§3.1 (HI labeling) and references [38–41]]
    "Following the evaluation of multiple AE features using the Mo–Pr–Tr criteria, cumulative energy was selected as the most effective HI for fatigue loading. ... Recent works [38–40] have suggested three criteria ..."

    The choice of cumulative AE energy as the regression target rests on Mo–Pr–Tr metrics and prior HI-construction papers whose author lists heavily overlap with the present paper. This is ordinary self-citation of methodology, not a uniqueness theorem that forbids alternatives, and the subsequent fusion performance numbers are still measured against held-out data rather than being forced by the citation itself. Mild elevation of score only.

full rationale

The paper is an empirical multi-task Transformer fusion study. Inputs are PZT (STFT tokens) and FBG (segmented strain tokens) with zero-padding for asynchronous acquisition (Eqs. 1–9); targets are AE cumulative energy (selected by Mo–Pr–Tr, §3.1) and Gaussian-KDE heatmaps from AE TDoA triangulation (Eqs. 17–24, §3.2). The model is trained to map the former onto the latter; evaluation (Tables 3–4, Figs. 11–13) is leave-one-panel-out MAE/RMSE/R²/SSIM against held-out AE-derived labels, plus comparison to PZT-only, FBG-only and 1D-CNN/Bi-LSTM/CNN-LSTM baselines that share the identical tokenization and labels. Nothing in the architecture or loss (Eqs. 12–13) makes the predicted HI or localization map equal by construction to any fitted input parameter or to the PZT/FBG tokens themselves. The reconstruction term is an ordinary self-supervised regularizer that is disabled at inference. Self-citations (ReMAP dataset papers, prior AE-HI work by overlapping authors) supply the experimental campaign and the Mo–Pr–Tr selection criteria; they do not supply a uniqueness theorem or an ansatz that forces the reported fusion numbers. The residual concern that AE may be an imperfect proxy for true internal damage is a label-validity / correctness issue, not a circular reduction of the claimed derivation. Hence the score is 1 (minor self-citation present but non-load-bearing) rather than 0 only out of caution; the central performance claims remain independent empirical results.

Axiom & Free-Parameter Ledger

4 free parameters · 4 axioms · 2 invented entities

The central performance claim rests on a handful of free hyper-parameters, standard ML and SHM modeling assumptions, and the treatment of AE-derived quantities as ground truth. No new physical entities are postulated; the novelty is architectural.

free parameters (4)
  • λ_det, λ_loc, λ_rec = 1.0, 0.5, 0.1
    Multi-task loss weights fixed at 1.0 / 0.5 / 0.1 with no sensitivity study; they directly control the reported MAE/RMSE trade-off.
  • learning rate & warm-up = 0.001
    Adam lr=0.001 with 100-step warm-up chosen by hand; affects convergence of the attention layers.
  • T_max (max token sequence length)
    Fixed maximum sequence length used for padding/truncation; determines how much temporal context the Transformer sees.
  • STFT patch size / FBG window length
    Tokenization hyper-parameters that convert raw signals into the latent sequences; not ablated.
axioms (4)
  • domain assumption AE cumulative energy (averaged over four channels) is a monotonic, prognosable and trendable health indicator that can serve as regression target.
    Section 3.1; Mo/Pr/Tr criteria used to select the label, but the criteria themselves are taken from prior SHM literature without re-validation on these panels.
  • domain assumption Gaussian KDE of AE TDoA triangulations yields a pixel-wise ground-truth damage map comparable to true internal damage.
    Section 3.2; orthotropic velocity model and Levenberg-Marquardt localization assumed accurate enough for SSIM/MAE evaluation.
  • ad hoc to paper Zero-padding of missing modalities plus binary attention masks preserves architectural consistency without introducing systematic bias.
    Eqs. (6)–(7); necessary for batching asynchronous streams but never stress-tested with controlled missingness rates.
  • ad hoc to paper Impact-region spatial embedding supplies sufficient physics-informed prior for OOD generalization (L04).
    Section 4.3.2; hand-crafted regional encoding used to explain why L04 still works despite doubled lifetime.
invented entities (2)
  • Cycle-aware temporal + modality embedding for mixed PZT/FBG tokens no independent evidence
    purpose: Aligns asynchronous sensor streams inside the Transformer so that self-attention can perform cross-modal fusion.
    Defined in Eq. (9) and Fig. 1; no independent physical measurement validates the embedding beyond the downstream task metrics.
  • Content-reconstruction regularizer on sparse multi-modal tokens no independent evidence
    purpose: Prevents representation collapse when PZT and FBG tokens are highly unbalanced.
    Eq. (12)–(13); self-supervised term whose necessity is asserted but not quantified by ablation.

pith-pipeline@v1.1.0-grok45 · 23941 in / 3401 out tokens · 37158 ms · 2026-07-12T12:10:04.012463+00:00 · methodology

0 comments
read the original abstract

Structural health monitoring (SHM) has emerged as an essential tool for ensuring the integrity and reliability of critical engineering structures, particularly in aerospace applications. Since each sensing technology has its limitations, the fusion of different modalities enables capturing a more complete picture of inhomogeneous materials, like composites. However, effective multisensor data fusion in SHM is often hindered by heterogeneous sensing modalities that operate at disparate sampling frequencies and acquisition intervals. To address these challenges, this paper proposes a Transformer-based data fusion framework that integrates multisensor data streams from piezoelectric transducer (PZT) capturing ultrasonic guided wave signals and fiber Bragg grating (FBG) sensors for strain measurements. By incorporating an attention-mechanism visualization, the proposed framework enables transparent, multitask learning for both health indicator (HI) prediction and damage localization. The framework was experimentally validated using aircraft composite structures subjected to compression-compression fatigue cyclic loading. For HI prediction, the framework consistently achieved a mean absolute error (MAE) and root mean squared error (RMSE) below 0.1, representing a nearly 60% performance improvement over single-sensor approaches (PZT or FBG alone) and baseline deep learning models. For damage localization, the model demonstrated the highest accuracy, maintaining an MAE and RMSE below 0.0465 and 0.1571, respectively. These results demonstrate that the proposed Transformer-based data fusion framework significantly outperforms single-source models and state-of-the-art deep learning models in both HI prediction and damage localization accuracy.

Figures

Figures reproduced from arXiv: 2607.02545 by Dimitrios Chronopoulos, Dimitrios Zarouchas, Jinbo Du, Morteza Moradi, Tongtong Yan, Xin Yang, Yunlai Liao.

Figure 1
Figure 1. Figure 1: Multisensor data fusion based on the Transformer architecture. PZT and FBG sensor data are tokenized and fused within a Transformer network, accounting for their differing acquisition intervals under specific fatigue cycles. The model then performs two tasks: HI prediction and damage localization. 2.2.1. Multisensor data tokenization The framework utilizes a unified tokenization and padding mechanism to tr… view at source ↗
Figure 3
Figure 3. Figure 3: The initial impact locations for the ReMAP composite panels [ Disbond location [PITH_FULL_IMAGE:figures/full_fig_p011_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Normalized AE cumulative energy considered as HI labels. Subfigures (a-d) represent the normalized [PITH_FULL_IMAGE:figures/full_fig_p014_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Gaussian KDE-based damage localization by clustering triggered AE hits. L03, L04, L05, L09 AE [PITH_FULL_IMAGE:figures/full_fig_p017_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: 10-channel FBG raw data visualization for L04 composite panel in ReMAP dataset. For each channel, [PITH_FULL_IMAGE:figures/full_fig_p018_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: The 10-channel FBG segmentation results (at 500th cycle) of the L04 composite panel in the ReMAP [PITH_FULL_IMAGE:figures/full_fig_p019_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: The multichannel UGW data tokenization results. From the raw multichannel PZT series, the STFT [PITH_FULL_IMAGE:figures/full_fig_p020_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: The training loss curves using data from L03, L04 and L05 specimens that consist of total training [PITH_FULL_IMAGE:figures/full_fig_p022_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Attention layer visualization during the Transformer model training process. a) Cross-modal [PITH_FULL_IMAGE:figures/full_fig_p024_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: HI prediction results on L03, L04, L05 and L09 specimens with AE cumulative energy as health [PITH_FULL_IMAGE:figures/full_fig_p026_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: The damage localization results for composite panels L03, L04, L05, and L09 are presented from [PITH_FULL_IMAGE:figures/full_fig_p029_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: A comparison of HI prediction results between the proposed Transformer-based data fusion model [PITH_FULL_IMAGE:figures/full_fig_p033_13.png] view at source ↗

discussion (0)

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

Reference graph

Works this paper leans on

44 extracted references · 31 canonical work pages

  1. [1]

    The need for multi-sensor data fusion in structural health monitoring of composite aircraft structures.Aerospace2022;9(4):183

    Broer, A.A., Benedictus, R., Zarouchas, D.. The need for multi-sensor data fusion in structural health monitoring of composite aircraft structures.Aerospace2022;9(4):183. doi:https://doi.org/10.3390/aerospace9040183

    work page doi:10.3390/aerospace9040183

  2. [2]

    Aggelis, D., Barkoula, N.M., Matikas, T., Paipetis, A.. Acoustic structural health monitoring of composite materials : Damage identification and evaluation in cross ply laminates using acoustic emission and ultrasonics.Composites Science and Tech- nology2012;72(10):1127–1133. doi:https://doi.org/10.1016/j.compscitech.2011. 10.011

    work page doi:10.1016/j.compscitech.2011 2011

  3. [3]

    Health monitoring of FRP using acoustic emission and artificial neural networks.Computers & Structures2008;86(3):367–373

    de Oliveira, R., Marques, A.. Health monitoring of FRP using acoustic emission and artificial neural networks.Computers & Structures2008;86(3):367–373. doi:https: //doi.org/10.1016/j.compstruc.2007.02.015

    work page doi:10.1016/j.compstruc.2007.02.015 2007

  4. [4]

    3 - materials and material requirements for aerospace structures and engines

    Mouritz, A.P.. 3 - materials and material requirements for aerospace structures and engines. In:Introduction to Aerospace Materials. Woodhead Publishing. ISBN 978-1- 85573-946-8; 2012:39–56. doi:https://doi.org/10.1533/9780857095152.39

    work page doi:10.1533/9780857095152.39 2012

  5. [5]

    Sensing Techniques for Structural Health Monitoring: A State-of-the-Art Review on Performance Criteria and New-Generation Technologies.Sensors2025;25(5):1424

    Mardanshahi, A., Sreekumar, A., Yang, X., Barman, S.K., Chronopoulos, D.. Sensing Techniques for Structural Health Monitoring: A State-of-the-Art Review on Performance Criteria and New-Generation Technologies.Sensors2025;25(5):1424. doi:https://doi.org/10.3390/s25051424

    work page doi:10.3390/s25051424

  6. [6]

    Deep generative models in condition and structural health monitoring: Opportunities, limi- tations and future outlook.Mechanical Systems and Signal Processing2026;253:114277

    Yang, X., Fang, C., Liao, Y., Yang, J., Gryllias, K., Chronopoulos, D.. Deep generative models in condition and structural health monitoring: Opportunities, limi- tations and future outlook.Mechanical Systems and Signal Processing2026;253:114277. doi:https://doi.org/10.1016/j.ymssp.2026.114277

    work page doi:10.1016/j.ymssp.2026.114277 2026

  7. [8]

    Fang, C., Yang, X., Gryllias, K., Vandepitte, D., Liu, X., Zhang, L., Chronopoulos, D.. Using removable sensors in structural health monitoring: A Bayesian methodol- ogy for attachment-to-attachment uncertainty quantification.Mechanical Systems and Signal Processing2025;224:111973. doi:https://doi.org/10.1016/j.ymssp.2024. 111973. 35

    doi:10.1016/j.ymssp.2024 2024

  8. [10]

    Bayesian waveform inversion for structural parameter estimation using narrow spectral characteristics of guided waves.Mechanical Systems and Signal Processing2026;244:113827

    Fang, C., Yang, X., Shi, S., Gryllias, K., Vandepitte, D., Liu, X., Zhang, L., Chronopoulos, D.. Bayesian waveform inversion for structural parameter estimation using narrow spectral characteristics of guided waves.Mechanical Systems and Signal Processing2026;244:113827. doi:https://doi.org/10.1016/j.ymssp.2025.113827

    work page doi:10.1016/j.ymssp.2025.113827 2025

  9. [11]

    Yang, X., Fang, C., Kundu, P., Yang, J., Chronopoulos, D.. A decision-level sensor fusion scheme integrating ultrasonic guided wave and vibration measurements for damage identification.Mechanical Systems and Signal Processing2024;219:111597. doi:https://doi.org/10.1016/j.ymssp.2024.111597

    work page doi:10.1016/j.ymssp.2024.111597 2024

  10. [13]

    Acoustic emission monitoring for damage diagnosis in composite laminates based on deep learning with attention mechanism.Mechanical Systems and Signal Processing 2025;222:111770

    Du, J., Zeng, J., Chen, C., Ni, M., Guo, C., Zhang, S., Wang, H., Ding, H.. Acoustic emission monitoring for damage diagnosis in composite laminates based on deep learning with attention mechanism.Mechanical Systems and Signal Processing 2025;222:111770. doi:https://doi.org/10.1016/j.ymssp.2024.111770

    work page doi:10.1016/j.ymssp.2024.111770 2025

  11. [14]

    Embedding eddy current sensors into LPBF components for structural health monitoring.Progress in Additive Manu- facturing2021;6(3):445–453

    Stoll, P., Gasparin, E., Spierings, A., Wegener, K.. Embedding eddy current sensors into LPBF components for structural health monitoring.Progress in Additive Manu- facturing2021;6(3):445–453. doi:https://doi.org/10.1007/s40964-021-00204-3

    work page doi:10.1007/s40964-021-00204-3

  12. [15]

    Estimating ultrasonic time of flight using envelope and quasi maximum likelihood method for damage detection and assessment

    Hoseini, M.R., Wang, X., Zuo, M.J.. Estimating ultrasonic time of flight using envelope and quasi maximum likelihood method for damage detection and assessment. Measurement2012;45(8):2072–2080. doi:https://doi.org/10.1016/j.measurement. 2012.05.008

    doi:10.1016/j.measurement 2072

  13. [16]

    Damage localization method for plates based on the time reversal of the mode-converted Lamb waves.Ultrasonics2019;91:19–29

    Mori, N., Biwa, S., Kusaka, T.. Damage localization method for plates based on the time reversal of the mode-converted Lamb waves.Ultrasonics2019;91:19–29. doi:https://doi.org/10.1016/j.ultras.2018.07.007. 36

    work page doi:10.1016/j.ultras.2018.07.007 2018

  14. [17]

    Guided wave phased array beamforming and imaging in composite plates.Ultrasonics2016;68:43–53

    Yu, L., Tian, Z.. Guided wave phased array beamforming and imaging in composite plates.Ultrasonics2016;68:43–53. doi:https://doi.org/10.1016/j.ultras.2016. 02.001

    work page doi:10.1016/j.ultras.2016 2016

  15. [18]

    Yassin, M.H., Farhat, M.H., Soleimanpour, R., Nahas, M.. Fiber Bragg grat- ing (FBG)-based sensors: a review of technology and recent applications in structural health monitoring (SHM) of civil engineering structures.Discover Civil Engineering 2024;1(1):151. doi:https://doi.org/10.1007/s44290-024-00141-4

    work page doi:10.1007/s44290-024-00141-4 2024

  16. [19]

    Optical fiber Bragg grating (FBG)-based strain sensor embedded in different 3D-printed materials: A comparison of performance.Measurement2024;225:114060

    Alias, M.A., Ahmad, H., Zaini, M.K.A., Samion, M.Z., Sa’ad, M.S.M., Sing, L.K., Grattan, K.T., Rahman, B.A., Brambilla, G., Reduan, S.A., et al. Optical fiber Bragg grating (FBG)-based strain sensor embedded in different 3D-printed materials: A comparison of performance.Measurement2024;225:114060. doi:https://doi.org/ 10.1016/j.measurement.2023.114060

    work page doi:10.1016/j.measurement.2023.114060 2023

  17. [20]

    Structural Health Monitoring of Civil Infrastructure Using Optical Fiber Sensing Technology: A Comprehensive Review.The Scientific World Journal2014;2014(1):652329

    Ye, X.W., Su, Y.H., Han, J.P.. Structural Health Monitoring of Civil Infrastructure Using Optical Fiber Sensing Technology: A Comprehensive Review.The Scientific World Journal2014;2014(1):652329. doi:https://doi.org/10.1155/2014/652329

    work page doi:10.1155/2014/652329 2014

  18. [21]

    Properties of optical fibres and fbg sensors and strain monitoring of composite aircraft structure

    Ruzicka, M., Dvorak, M., Sasek, L.. Properties of optical fibres and fbg sensors and strain monitoring of composite aircraft structure. 2016

    2016

  19. [22]

    Damage de- tection in Aerospace structures using Chirped fiber Bragg grating.Materials Today: Proceedings2022;64:632–636

    Dinesh kumar, D., Sridhar raj, S., Sivananth, V., Ramkumar, V.. Damage de- tection in Aerospace structures using Chirped fiber Bragg grating.Materials Today: Proceedings2022;64:632–636. doi:https://doi.org/10.1016/j.matpr.2022.05.149

    work page doi:10.1016/j.matpr.2022.05.149 2022

  20. [23]

    Development of a Fault Detection Instrument for Fiber Bragg Grating Sensing System on Airplane.Micromachines2022;13(6):882

    Du, C., Kong, D., Xu, C.. Development of a Fault Detection Instrument for Fiber Bragg Grating Sensing System on Airplane.Micromachines2022;13(6):882. doi:https: //doi.org/10.3390/mi13060882

    work page doi:10.3390/mi13060882

  21. [25]

    Hakimi, O., Liu, H., Abudayyeh, O.. Deep learning-driven multi-level data fusion framework for predictive maintenance and structural health monitoring of concrete bridge decks.Automation in Construction2025;175:106180. doi:https://doi.org/ 10.1016/j.autcon.2025.106180. 37

    work page doi:10.1016/j.autcon.2025.106180 2025

  22. [26]

    Li, H., Wang, J., Wu, S., Gao, Y., Wang, Y., Nie, G.. Transforming Structural Health Monitoring: Leveraging Multi-Source Data Fusion with Two Stage Encoder TransformerforBridgeDeformationPrediction.IEEE Transactions on Instrumentation and Measurement2025;doi:https://doi.org/10.1109/TIM.2025.3550241

    work page doi:10.1109/tim.2025.3550241 2025

  23. [27]

    Data anomaly segmentation for structural healthmonitoringusingdenoisingdiffusionprobabilisticmodels.Engineering Structures 2026;354:122376

    Zheng, W., Li, J., Hao, H., Ni, P.. Data anomaly segmentation for structural healthmonitoringusingdenoisingdiffusionprobabilisticmodels.Engineering Structures 2026;354:122376. doi:https://doi.org/10.1016/j.engstruct.2026.122376

    work page doi:10.1016/j.engstruct.2026.122376 2026

  24. [28]

    Multimodal deep neural network- based sensor data anomaly diagnosis method for structural health monitoring.Buildings 2023;13(8):1976

    Nong, X., Luo, X., Lin, S., Ruan, Y., Ye, X.. Multimodal deep neural network- based sensor data anomaly diagnosis method for structural health monitoring.Buildings 2023;13(8):1976. doi:https://doi.org/10.3390/buildings13081976

    work page doi:10.3390/buildings13081976 2023

  25. [29]

    U-net transformer: Self and cross attention for medical image segmentation

    Petit, O., Thome, N., Rambour, C., Themyr, L., Collins, T., Soler, L.. U-net transformer: Self and cross attention for medical image segmentation. In:Interna- tional Workshop on Machine Learning in Medical Imaging. Springer; 2021:267–276. doi:https://doi.org/10.1007/978-3-030-87589-3_28

    work page doi:10.1007/978-3-030-87589-3_28 2021

  26. [30]

    Attention is all you need.Advances in neural information processing systems2017;30

    Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A.N., Kaiser, Ł., Polosukhin, I.. Attention is all you need.Advances in neural information processing systems2017;30

  27. [31]

    Damage imaging in structural health monitoring with fine-tuned conditional diffusion model.Mechanical Systems and Signal Processing2025;236:112996

    Yang, X., Cantero-Chinchilla, S., Moradi, M., Komninos, P., Fang, C., Liao, Y., Kundu, P., Zarouchas, D., Chronopoulos, D.. Damage imaging in structural health monitoring with fine-tuned conditional diffusion model.Mechanical Systems and Signal Processing2025;236:112996. doi:https://doi.org/10.1016/j.ymssp.2025.112996

    work page doi:10.1016/j.ymssp.2025.112996 2025

  28. [32]

    Towards good practices for missing modality robust action recognition

    Woo, S., Lee, S., Park, Y., Nugroho, M.A., Kim, C.. Towards good practices for missing modality robust action recognition. 2023. URL:https://arxiv.org/abs/ 2211.13916.arXiv:2211.13916

    Pith/arXiv arXiv 2023

  29. [33]

    Deep multimodal learning with missing modality: A survey

    Wu, R., Wang, H., Chen, H.T., Carneiro, G.. Deep multimodal learning with missing modality: A survey. 2026. URL:https://arxiv.org/abs/2409.07825. arXiv:2409.07825

    Pith/arXiv arXiv 2026

  30. [34]

    Fusion- based damage diagnostics for stiffened composite panels.Structural Health Monitoring 2022;21(2):613–639

    Broer, A., Galanopoulos, G., Benedictus, R., Loutas, T., Zarouchas, D.. Fusion- based damage diagnostics for stiffened composite panels.Structural Health Monitoring 2022;21(2):613–639. doi:10.1177/14759217211007127. 38

    work page doi:10.1177/14759217211007127 2022

  31. [35]

    Anovelstrain-basedhealthindicatorfortheremainingusefullifeestimation of degrading composite structures.Composite Structures2023;306:116579

    Galanopoulos, G., Eleftheroglou, N., Milanoski, D., Broer, A., Zarouchas, D., Loutas, T.. Anovelstrain-basedhealthindicatorfortheremainingusefullifeestimation of degrading composite structures.Composite Structures2023;306:116579. doi:https: //doi.org/10.1016/j.compstruct.2022.116579

    work page doi:10.1016/j.compstruct.2022.116579 2022

  32. [36]

    Health Monitoring of Aerospace Structures Utilizing Novel Health Indicators Extracted from Complex Strain and Acoustic Emission Data.Sensors2021;21(17)

    Galanopoulos, G., Milanoski, D., Broer, A., Zarouchas, D., Loutas, T.. Health Monitoring of Aerospace Structures Utilizing Novel Health Indicators Extracted from Complex Strain and Acoustic Emission Data.Sensors2021;21(17). doi:https://doi. org/10.3390/s21175701

    work page doi:10.3390/s21175701

  33. [37]

    Assess- ing stiffness degradation of stiffened composite panels in post-buckling compression- compression fatigue using guided waves.Composite Structures2022;293:115751

    Yue, N., Broer, A., Briand, W., Rébillat, M., Loutas, T., Zarouchas, D.. Assess- ing stiffness degradation of stiffened composite panels in post-buckling compression- compression fatigue using guided waves.Composite Structures2022;293:115751. doi:https://doi.org/10.1016/j.compstruct.2022.115751

    work page doi:10.1016/j.compstruct.2022.115751 2022

  34. [39]

    A novel intelligent health indicator using acoustic waves: CEEMDAN-driven semi-supervised ensemble deep learning.Mechanical Systems and Signal Processing2025;224:112156

    Moradi, M., Galanopoulos, G., Kuiters, T., Zarouchas, D.. A novel intelligent health indicator using acoustic waves: CEEMDAN-driven semi-supervised ensemble deep learning.Mechanical Systems and Signal Processing2025;224:112156. doi:https: //doi.org/10.1016/j.ymssp.2024.112156

    work page doi:10.1016/j.ymssp.2024.112156 2024

  35. [40]

    A novel machine learning model to design historical-independent health indicators for composite structures.Composites Part B: Engineering2024;275:111328

    Moradi, M., Gul, F.C., Zarouchas, D.. A novel machine learning model to design historical-independent health indicators for composite structures.Composites Part B: Engineering2024;275:111328. doi:https://doi.org/10.1016/j.compositesb.2024. 111328

    work page doi:10.1016/j.compositesb.2024 2024

  36. [41]

    Moradi, M., Chiachío, J., Zarouchas, D.. Health indicator modeling leveraging time- independent and time-dependent subtasks with adaptive standardization and physics- based Bayesian optimization for aeronautical structures.Engineering Applications of Artificial Intelligence2026;163:112767. doi:https://doi.org/10.1016/j.engappai. 2025.112767

    doi:10.1016/j.engappai 2025

  37. [42]

    Damage classification and evolution in composite under low-velocity 39 impact using acoustic emission, machine learning and wavelet packet decomposition

    Du, J., Wang, H., Chen, C., Ni, M., Guo, C., Zhang, S., Ding, H., Wang, H., Bi, Y.. Damage classification and evolution in composite under low-velocity 39 impact using acoustic emission, machine learning and wavelet packet decomposition. Engineering Fracture Mechanics2024;306:110238. doi:https://doi.org/10.1016/j. engfracmech.2024.110238

    doi:10.1016/j 2024

  38. [43]

    Acoustic emission-based remaining useful life prognosis of aeronautical structures subjected to compressive fatigue loading.Engineering Structures2023;290:116391

    Galanopoulos, G., Milanoski, D., Eleftheroglou, N., Broer, A., Zarouchas, D., Loutas, T.. Acoustic emission-based remaining useful life prognosis of aeronautical structures subjected to compressive fatigue loading.Engineering Structures2023;290:116391. doi:https://doi.org/10.1016/j.engstruct.2023.116391

    work page doi:10.1016/j.engstruct.2023.116391 2023

  39. [44]

    Acoustic emission source localization in arbitrary anisotropic plates: Equation error and Bayesian inference approaches.Me- chanical Systems and Signal Processing2025;236:112960

    Raorane, S., Papadimitriou, C., Packo, P.. Acoustic emission source localization in arbitrary anisotropic plates: Equation error and Bayesian inference approaches.Me- chanical Systems and Signal Processing2025;236:112960. doi:https://doi.org/10. 1016/j.ymssp.2025.112960

    arXiv 2025

  40. [45]

    Impact localization in composite structures of arbitrary cross section.Structural Health Monitoring2012;11(6):643–655

    Ciampa, F., Meo, M., Barbieri, E.. Impact localization in composite structures of arbitrary cross section.Structural Health Monitoring2012;11(6):643–655. doi:https: //doi.org/10.1177/1475921712451

    work page doi:10.1177/1475921712451

  41. [46]

    The Levenberg-Marquardt algorithm for nonlinear least squares curve- fitting problems.Department of Civil and Environmental Engineering Duke University August2019;3:1–23

    Gavin, H.P.. The Levenberg-Marquardt algorithm for nonlinear least squares curve- fitting problems.Department of Civil and Environmental Engineering Duke University August2019;3:1–23

  42. [47]

    Multivariate density estimation: theory, practice, and visualization

    Scott, D.W.. Multivariate density estimation: theory, practice, and visualization. John Wiley & Sons; 2015

    2015

  43. [48]

    Understanding SSIM

    Nilsson, J., Akenine-Möller, T.. Understanding SSIM. 2020. URL:https://arxiv. org/abs/2006.13846.arXiv:2006.13846

    Pith/arXiv arXiv 2020

  44. [49]

    Image quality assessment: from error visibility to structural similarity.IEEE Transactions on Image Processing 2004;13(4):600–612

    Wang, Z., Bovik, A., Sheikh, H., Simoncelli, E.. Image quality assessment: from error visibility to structural similarity.IEEE Transactions on Image Processing 2004;13(4):600–612. doi:https://doi.org/10.1109/TIP.2003.819861. 40

    doi:10.1109/tip.2003.819861 2004