Sensors

18 pages, 9404 KB

Open AccessArticle

A Distance-Adaptive Method for Three-Axis Angle Measurement Based on an Optical Wedge

by Jinkai Wang, Chengping Ran, Lihui Wang, Lizhong Zhang and Yangyang Bai

Sensors 2026, 26(14), 4430; https://doi.org/10.3390/s26144430 (registering DOI) - 12 Jul 2026

Existing non-contact three-axis angle-measurement methods are unsuitable for measuring the relative three-axis angles between the inner and outer ring frames of fifth-generation airborne optoelectronic gimbal platforms. Our previous study proposed a compact three-axis angle-measurement method based on an optical wedge for this application. [...] Read more.

Existing non-contact three-axis angle-measurement methods are unsuitable for measuring the relative three-axis angles between the inner and outer ring frames of fifth-generation airborne optoelectronic gimbal platforms. Our previous study proposed a compact three-axis angle-measurement method based on an optical wedge for this application. However, the fixed-coefficient angle-solving model used in that method does not account for variations in the measurement distance L, which can produce distance-dependent nonlinear errors when axial displacement of the inner ring frame occurs. To address this limitation, the present study proposes a distance-adaptive method for three-axis angle measurements. An analytical measurement-distance model is established using ABCD ray-transfer-matrix theory, and L is incorporated into the fixed-coefficient angle-solving model. The fixed coefficients in the polynomial error-compensation model are thereby expressed as functions of L and updated according to the estimated measurement distance. Experimental results demonstrate that the proposed method significantly improves the measurement accuracy under varying-distance conditions. Taking the measurement-distance condition with the largest error, L = −2 mm, as an example, the RMS values of the measurement errors for the pitch, yaw, and roll angles are reduced from 13.0″, 8.4″, and 31.1″ to 5.6″, 3.8″, and 18.6″, respectively. Full article

(This article belongs to the Section Optical Sensors)

► Show Figures

Figure 1

19 pages, 4587 KB

Open AccessArticle

Low-Temperature Ammonia Sensing Using Sulfurized WS₂ Thin Films

by Naman Kumar, Ivan Hotový, Magdaléna Kadlečíková, Michaela Sojková, Ivan Kostič, Dagmar Gregušová, Ondrej Pohorelec, Viera Stopjaková and Edmund Dobročka

Sensors 2026, 26(14), 4429; https://doi.org/10.3390/s26144429 (registering DOI) - 12 Jul 2026

Abstract

This study addresses key challenges in gas sensing, particularly achieving high sensitivity at low operating temperatures. Tungsten disulfide (WS₂) thin films were deposited by sputtering and sulfurized at 800 °C for 30 and 60 min on quartz and sapphire substrates. Raman [...] Read more.

This study addresses key challenges in gas sensing, particularly achieving high sensitivity at low operating temperatures. Tungsten disulfide (WS₂) thin films were deposited by sputtering and sulfurized at 800 °C for 30 and 60 min on quartz and sapphire substrates. Raman spectroscopy confirmed multilayer WS₂ formation, with characteristic modes at 349 and 417 cm⁻¹. X-ray diffraction revealed that the films are polycrystalline with the formation of the hexagonal 2H-WS₂ phase on both substrates. FESEM images showed clearly visible nanoparticles, but they were not evenly distributed and were differently oriented, with a hint of texturing. The WS₂ films exhibited stable sensing performance at low operating temperatures from 30 to 150 °C and high sensitivity toward NH₃ in the 10–200 ppm range. Experimental results indicate a limit of detection, highlighting the potential of WS₂ films for low operating temperature ammonia sensing. A minimum value of the limit of detection of 0.6 ppm for an operating temperature of 150 °C was achieved for the WS₂ sample sulfurized at 800 °C for 60 min on a quartz substrate. Full article

(This article belongs to the Section Sensor Materials)

25 pages, 4257 KB

Open AccessArticle

High-Sensitivity Identification of Micro-Voids at Thick Steel Shell–Concrete Interfaces Using Elastic Wave Analysis and Feature Attention Mechanisms

by Yan Zhang, Siying Qu, Songhui Li, Yi Liu and Xunnan Liu

Sensors 2026, 26(14), 4428; https://doi.org/10.3390/s26144428 (registering DOI) - 12 Jul 2026

Abstract

The steel–concrete interface in steel–concrete composite structures is susceptible to interfacial void defects during both casting and service, posing a significant threat to structural load-bearing capacity. For early-stage micro-voids exceeding 2 mm in height, signal variations are weak and exhibit response characteristics similar [...] Read more.

The steel–concrete interface in steel–concrete composite structures is susceptible to interfacial void defects during both casting and service, posing a significant threat to structural load-bearing capacity. For early-stage micro-voids exceeding 2 mm in height, signal variations are weak and exhibit response characteristics similar to dense states, leading to feature ambiguity when using conventional criteria based on time-domain amplitude and attenuation or frequency-domain peak values and resulting in a high risk of missed detections. To address this limitation for early warning purposes, this study proposes a high-sensitivity identification method integrating an impact elastic wave response feature system with a feature-attention gated multi-layer perceptron (Feature-attention MLP). Based on full-scale model experiments from an engineering project, the temporal and spectral evolution patterns of impact elastic wave responses under varying dense conditions were analyzed. A comprehensive feature system, including time-domain statistical descriptors, spectral peaks, and sub-band energy distributions, was constructed, with Random Forest used for feature importance ranking and Top-K selection. An MLP classifier was then developed for automatic discrimination of dense states. A feature-level attention gating mechanism was introduced to enable adaptive weighting across feature dimensions, enhancing sensitive features while suppressing noise and structural variability. The final lightweight classifier contains 4052 trainable parameters, enabling rapid execution with an average CPU inference time of approximately 1.24 ms per sample. The average CPU inference time was approximately 1.24 ms per sample. Under the original train–validation split, the recall-prioritized operating point achieved a Void recall of 0.978 and a weighted F1-score of 0.780, accompanied by a non-negligible false-positive screening burden. Stratified five-fold internal validation yielded a balanced accuracy of 0.682 ± 0.021 and a Void recall of 0.845 ± 0.035 under the inner-validation-optimized threshold. These results demonstrate the preliminary potential of the proposed lightweight framework for engineering-oriented micro-void screening under the investigated full-scale conditions. Full article

(This article belongs to the Special Issue Sensing Techniques for Intelligent Tunnel Construction)

► Show Figures

Figure 1

28 pages, 5431 KB

Open AccessArticle

ECG-Only Cognitive Workload State Classification in Laparoscopic Training Using Raw and Recurrence-Plot Representations

by Kaizhe Jin, Adrian Rubio-Solis, Ravi Naik and George Mylonas

Sensors 2026, 26(14), 4427; https://doi.org/10.3390/s26144427 (registering DOI) - 12 Jul 2026

Abstract

Electrocardiography (ECG)-only workload-state classification offers a lower-burden physiological sensing route than denser multimodal, multi-sensor physiological, or neuroimaging setups for controlled laparoscopic training research. This study evaluated whether ECG-only representations can classify condition-derived cognitive workload states during a controlled laparoscopic peg transfer task performed [...] Read more.

Electrocardiography (ECG)-only workload-state classification offers a lower-burden physiological sensing route than denser multimodal, multi-sensor physiological, or neuroimaging setups for controlled laparoscopic training research. This study evaluated whether ECG-only representations can classify condition-derived cognitive workload states during a controlled laparoscopic peg transfer task performed under Control and auditory N-back conditions (N0, N1, and N2). Twenty surgical trainees from an advanced surgical-skills course completed the task protocol, and 17 participants entered ECG modelling after ECG quality-control exclusions. The retained ECG modelling dataset comprised 268 task blocks (Control/N0/N1/N2: 68/68/68/64), evaluated as held-out task-block predictions in a known-participant four-fold leave-one-round-out (LOTO) evaluation. Branch-specific raw ECG windows, recurrence-plot sequences, and heart-rate/time-domain heart-rate-variability inputs are detailed in the Methods, and model metrics were computed after reduction to task-block predictions. The primary endpoint was Surgery Task Load Index (SURG-TLX)-aligned low/high workload, defined as Control plus N0 versus N1 plus N2. Four-class and three-level endpoints were retained as secondary views. Raw ECG, recurrence-plot (RP)-derived, hybrid score-level fusion, and conventional heart-rate/time-domain heart-rate-variability Random Forest (HRV-RF) models were compared using a locked evaluation protocol, leakage-aware train-fold-only preprocessing, participant-clustered confidence intervals, and planned paired tests for the primary endpoint. On the primary low/high endpoint, raw ECG achieved the highest macro-F1/balanced accuracy (0.865/0.865), followed by the hybrid branch (0.847/0.847) and HRV-RF (0.648/0.649). Raw ECG and hybrid were supported over RP-derived and HRV-RF under the planned paired tests. On selected secondary endpoints, hybrid achieved higher macro-F1 values than raw ECG, consistent with possible endpoint-dependent RP-derived complementarity rather than a general hybrid advantage. These findings support ECG-only block-level workload-state classification in this controlled training setting. The evidence is retrospective and based on held-out rounds from known participants rather than subject-independent or deployment validation. Full article

(This article belongs to the Section Biomedical Sensors)

20 pages, 29336 KB

Open AccessArticle

Acoustic Emission Characteristics During Shear Failure of Active Waveguide Structure for Rock Slope Monitoring

by Zhihui Wu, Lingjun Zhang, Jianjun Yang, Jie Dong, Yongxin Yu and Yunlong Sun

Sensors 2026, 26(14), 4426; https://doi.org/10.3390/s26144426 (registering DOI) - 12 Jul 2026

Abstract

This study investigates the acoustic emission (AE) characteristics associated with the shear failure mode based on the principles of active waveguide monitoring for the bedding rock slopes. Physical simulation experiments were conducted to assess the AE response during the shear-induced failure process of [...] Read more.

This study investigates the acoustic emission (AE) characteristics associated with the shear failure mode based on the principles of active waveguide monitoring for the bedding rock slopes. Physical simulation experiments were conducted to assess the AE response during the shear-induced failure process of active waveguide structures. The findings indicate that during the initial loading phase, the scatter points of the signals are concentrated within a relatively narrow range. As the shear stress exceeds 90% of the peak stress and approaches the failure stage, there is a significant increase in the AE count and a rise in the high-frequency signals. Additionally, the distribution range of signals in the parameter correlation plot expands progressively. With increasing shear stress, the AE count, amplitude, and energy also rise gradually. And the emergence of continuous high-frequency signals is noted. During the failure stage, numerous microcracks initiate and propagate within the specimen, with signal amplitudes ranging between 40 and 90 dB. The peak frequency range of the AE signals broadens, with high-frequency components mainly concentrated between 350 and 450 kHz. Loading tests conducted at shear displacement rates of 0.25–1.5 mm/min reveal a strong correlation between the AE count and the shear displacement rate. Furthermore, prior to the shear failure of the active waveguide structures, the AE count shows a positive correlation with shear displacement. After the shear failure of the waveguide structure specimens, the AE count gradually decreases from a higher level to a lower level, demonstrating a negative correlation with shear displacement. The active waveguide structure can monitor the internal deformation conditions of the bedding rock slope so as to provide some reference for the early warning research. In addition, quantitative statistical analysis and curve fitting are conducted on the relationship between AE statistical count and shear displacement under different loading rates. The measured data show good agreement with the fitted curves, and a distinct two-stage evolutionary pattern (positive correlation before peak and negative correlation after peak) is quantitatively identified. These results further enhance the reliability of using AE parameters for quantitative evaluation of shear failure characteristics and displacement rate effects in bedding rock slopes. Full article

(This article belongs to the Special Issue Advanced Sensors for Monitoring and Detection in Geotechnical Engineering)

► Show Figures

Figure 1

16 pages, 8344 KB

Open AccessArticle

A Surface Thermal Sensing Framework for Internal Winding Temperature Estimation in Oil-Immersed Converter Transformers

by Sheng Han, Zhiqiang Wang, Yuchen Tang, Li Huang, Hui Jiang and Xing Li

Sensors 2026, 26(14), 4425; https://doi.org/10.3390/s26144425 (registering DOI) - 12 Jul 2026

Abstract

Accurate monitoring of internal winding temperature is essential for assessing the thermal state and operational reliability of oil-immersed transformers. However, direct deployment of distributed temperature sensors inside transformer windings is difficult because of insulation constraints, structural complexity, and potential reliability risks. To address [...] Read more.

Accurate monitoring of internal winding temperature is essential for assessing the thermal state and operational reliability of oil-immersed transformers. However, direct deployment of distributed temperature sensors inside transformer windings is difficult because of insulation constraints, structural complexity, and potential reliability risks. To address this problem, this paper proposes a non-invasive internal winding temperature estimation method based on surface temperature sensing and a hybrid deep learning model. In the proposed framework, external surface temperature measurements are used as sensor inputs to infer the internal transient thermal state of the transformer. First, an extreme gradient boosting (XGBoost) model is employed to evaluate the contribution of different surface temperature measurement points and select the sensing locations that are most strongly correlated with internal winding temperature variations. Then, the selected surface temperature time-series data are used to train a Long Short-Term Memory (LSTM) network, which captures the temporal evolution of the transformer temperature field under different operating conditions. The proposed method is verified through both numerical simulation and experimental testing on a scaled single-phase oil-immersed converter transformer model (D-800/35) developed in this study. The results show that the proposed XGBoost-LSTM model can estimate internal winding temperature with an error of less than 1.5 K. Compared with direct internal sensing, the proposed method provides a non-invasive and sensor-efficient solution for internal temperature monitoring. The results demonstrate its potential for real-time thermal state estimation, condition monitoring, and fault diagnosis of oil-immersed converter transformers. Full article

(This article belongs to the Section Fault Diagnosis & Sensors)

► Show Figures

Figure 1

19 pages, 10173 KB

Open AccessArticle

Cross-Verification and Consistency Evaluation of Monochromator-Based Radiance Calibration Systems Under Varying Bandwidths Against an SI-Traceable Reference

by Kaichao Lei, Hongjia Xu, Xin Ye, Jingying Yue, Ye Jiang, Shuqi Li and Yachao Zhang

Sensors 2026, 26(14), 4424; https://doi.org/10.3390/s26144424 (registering DOI) - 12 Jul 2026

Abstract

High-precision absolute radiometric calibration is the cornerstone of space-based quantitative remote sensing. Monochromator-based radiance calibration is a key approach for space-borne radiometric benchmark transfer; however, the influence of the monochromator bandwidth on calibration accuracy urgently requires validation via an independent, SI-traceable benchmark. Taking [...] Read more.

High-precision absolute radiometric calibration is the cornerstone of space-based quantitative remote sensing. Monochromator-based radiance calibration is a key approach for space-borne radiometric benchmark transfer; however, the influence of the monochromator bandwidth on calibration accuracy urgently requires validation via an independent, SI-traceable benchmark. Taking a spectrometer as a case study, this paper establishes an equivalent theoretical calibration model. The model reveals that discrepancies in the equivalent spectral distributions of the calibration sources primarily contribute to their calibration deviations. Within the 775–855 nm range, cross-comparison experiments were conducted between the monochromator facility and a reference standard lamp-plaque. Experimental results demonstrate that under bandwidth configurations of 2 nm, 4 nm, and 6 nm, relative deviations in calibrated radiance responsivity stabilize at approximately 2.3%, exhibiting no significant dependence on bandwidth variations. Crucially, the consistency between the two methodologies is rigorously verified, as the radiance responsivity ratios uniformly fall within the combined expanded measurement uncertainty bounds. By demonstrating that bandwidth expansion within the specified 2–6 nm range does not degrade calibration accuracy, this study provides a critical empirical basis for safely increasing solar monochromator bandwidths to improve the signal-to-noise ratio (SNR) in next-generation benchmark missions like the Space-based Radiometric Measurement Benchmark Mission (LIBRA) program. Full article

(This article belongs to the Section Optical Sensors)

► Show Figures

Figure 1

24 pages, 3952 KB

Open AccessArticle

LSTM–Transformer-Based Mine Pressure Prediction Using Hydraulic-Support Monitoring Data

by Ran Tao, Xiaowan Lei, Lirong Wan, Yan Wang and Nan Xu

Sensors 2026, 26(14), 4423; https://doi.org/10.3390/s26144423 (registering DOI) - 12 Jul 2026

Abstract

Accurate mine pressure prediction is essential for understanding roof–support interaction and supporting intelligent monitoring in fully mechanized longwall mining. In underground production, hydraulic-support pressure sensors provide continuous pressure sequences that reflect the mechanical response of the support–roof system. However, these sequences are affected [...] Read more.

Accurate mine pressure prediction is essential for understanding roof–support interaction and supporting intelligent monitoring in fully mechanized longwall mining. In underground production, hydraulic-support pressure sensors provide continuous pressure sequences that reflect the mechanical response of the support–roof system. However, these sequences are affected by local mining disturbances, missing records, abnormal zero-value segments, nonstationary fluctuations, and periodic weighting, which make future pressure forecasting challenging. To address this issue, an LSTM–Transformer hybrid model is proposed for hydraulic-support pressure forecasting. The LSTM module extracts local nonlinear pressure-evolution features from recent historical windows, whereas the Transformer module captures temporal dependencies and periodic pressure patterns through global sequence modeling. Support-wise experiments were conducted using field monitoring data from Yili No. 1 Mine, and the pressure sequence of each support was processed independently to avoid mixing information from different support locations. In the representative test case, the proposed model achieved an R² of 0.971 and reduced the MAE to 0.471 MPa, while improving the phase consistency of predicted pressure peaks. Further analysis indicates that sufficient historical data coverage is necessary to capture complete pressure-evolution cycles, and that the 25-step forecasting case maintains stable accuracy for short-term mine pressure estimation. These findings demonstrate the feasibility of the proposed approach for hydraulic-support pressure prediction under the monitored conditions of the studied working face. Full article

(This article belongs to the Section Electronic Sensors)

17 pages, 3301 KB

Open AccessArticle

Reconstruction-Driven Induction Thermography for AI-Assisted Surface Defect Detection in Welded Structures

by Xiang Zhang, Shenghao Huang, Xiaolu Cui, Yu Cao, Dong Wang and Guiyun Tian

Sensors 2026, 26(14), 4422; https://doi.org/10.3390/s26144422 (registering DOI) - 12 Jul 2026

Abstract

Welding is widely used to join steel components in energy storage systems, wind power generation systems, and transportation infrastructure. During service, surface defects in steel welds can promote rapid crack propagation and sudden fracture. To address this issue, this study presents an induction [...] Read more.

Welding is widely used to join steel components in energy storage systems, wind power generation systems, and transportation infrastructure. During service, surface defects in steel welds can promote rapid crack propagation and sudden fracture. To address this issue, this study presents an induction thermography framework for post-weld quality inspection of surface-breaking defects in welded steel structures. First, a yoke excitation structure was optimized and evaluated through numerical simulation and experimental validation, and its performance was compared with those of a straight coil and a three-loop coil. The results show that the yoke configuration maintains sufficient heating intensity in the weld region while improving the spatial uniformity of the thermal response. Furthermore, a speed-based reconstruction method was developed to reduce the effects of specimen motion and coil occlusion. By aligning thermal responses from the same physical locations in sequential thermograms, the method reconstructs a continuous temperature field over an extended inspection area. Consequently, the reconstructed images retain clear crack-related thermal features and provide stable inputs for automated analysis. In addition, DeepLabv3+, DSCA-UNet, and feature pyramid network (FPN) were used as representative segmentation models to evaluate the suitability of the reconstructed thermograms for automated defect extraction. On the current laboratory dataset, the three models achieved precision values of 90.4%, 88.6%, and 92.6%, respectively. These results indicate the potential of the proposed framework, while further validation with larger datasets and natural weld defects is still required. Full article

(This article belongs to the Special Issue Advanced Non-Contact Sensing Technologies for Structural Health Monitoring)

19 pages, 6694 KB

Open AccessArticle

A Single-IMU Wearable System with 1D U-Net for Knee Adduction Moment Waveform Reconstruction During Gait

by Hiroko Sakamoto, Hiroshi Yoshihara, Ayako Akiba, Kohei Nishizawa, Masaki Nagashima, Kengo Harato, Takeo Nagura and Masaya Nakamura

Sensors 2026, 26(14), 4421; https://doi.org/10.3390/s26144421 (registering DOI) - 12 Jul 2026

Abstract

The knee adduction moment (KAM) is an important biomechanical marker of medial knee joint loading, but conventional assessment requires laboratory-based motion analysis. This study aimed to develop and evaluate an improved inertial measurement unit (IMU)-based KAM estimation system that incorporates a one-dimensional U-Net [...] Read more.

The knee adduction moment (KAM) is an important biomechanical marker of medial knee joint loading, but conventional assessment requires laboratory-based motion analysis. This study aimed to develop and evaluate an improved inertial measurement unit (IMU)-based KAM estimation system that incorporates a one-dimensional U-Net (1D U-Net) with a gated recurrent unit (GRU) bottleneck. Gait data from 50 participants were used for model development and internal evaluation, and an independent holdout dataset collected using the same study protocol included 45 participants. Simultaneously recorded IMU signals and reference KAM obtained from motion capture-based inverse dynamics were used for training and evaluation. The model used a single IMU attached to the tibial tuberosity to reconstruct the KAM waveform during the stance phase. For peak KAM, Pearson’s r values were 0.819, 0.602, and 0.621 for the training, internal validation, and holdout datasets, respectively. For KAM impulse, the corresponding values were 0.941, 0.744, and 0.745. These findings indicate that a simplified wearable system combining a single tibial IMU with a 1D U-Net incorporating a GRU bottleneck can reconstruct the KAM waveform during walking and derive clinically relevant parameters from it. Although further improvement in prediction accuracy is required, this framework may provide a practical basis for simplified screening or monitoring of medial knee loading in settings closer to routine clinical practice. Full article

(This article belongs to the Special Issue AI-Enabled Sensor Technologies for Gait and Posture Recognition: Methods, Applications, and Future Directions)

► Show Figures

Figure 1

21 pages, 722 KB

Open AccessArticle

Game-Theoretic Optimized Federated Learning for Heterogeneous IoT Object Detection

by Zepeng Wang and Jie Chen

Sensors 2026, 26(14), 4420; https://doi.org/10.3390/s26144420 (registering DOI) - 12 Jul 2026

Abstract

Federated object detection allows distributed IoT cameras to learn a shared detector without exposing raw images. Its performance is limited by non-IID scenes, unequal device resources, intermittent links, and selfish participation. We propose GO-FedDet, a game-theoretic optimized federated detection framework for heterogeneous IoT [...] Read more.

Federated object detection allows distributed IoT cameras to learn a shared detector without exposing raw images. Its performance is limited by non-IID scenes, unequal device resources, intermittent links, and selfish participation. We propose GO-FedDet, a game-theoretic optimized federated detection framework for heterogeneous IoT application. Client participation is formulated as a Stackelberg game, selection is formulated as an exact-potential resource game, and the equilibrium is embedded into a proximal detection objective. A utility-aligned aggregation balances detection contribution, communication/energy cost, and long-term fairness. We prove the existence of a Stackelberg equilibrium and the finite-improvement convergence of the selection game, and we derive a non-convex convergence bound under client drift and partial participation. Experiments on heterogeneous edge object detection show that GO-FedDet improves accuracy, lowers communication cost, and stabilizes fairness compared with representative federated baselines. Full article

(This article belongs to the Special Issue Emerging Trends in Cybersecurity for Wireless Communication and IoT)

19 pages, 16632 KB

Open AccessArticle

Simultaneous Measurement of Peripheral Ocular Aberrations Using a Virtual Multi-Eccentric Hartmann–Shack Aberrometer

by Jennyfer Morales-Marín, Andrés Osorno-Quiroz, Walter Torres-Sepúlveda and Alejandro Mira-Agudelo

Sensors 2026, 26(14), 4419; https://doi.org/10.3390/s26144419 (registering DOI) - 12 Jul 2026

Abstract

We present a computational proof of concept, inspired by a recent patent developed by Nikolai Suchkov and Siegfried Wahl (European Patent Office EP4179957A1), of a virtual multi-eccentric Hartmann–Shack (HS) aberrometer for simultaneous measurement of peripheral ocular aberrations. The simulated system records nine wavefronts [...] Read more.

We present a computational proof of concept, inspired by a recent patent developed by Nikolai Suchkov and Siegfried Wahl (European Patent Office EP4179957A1), of a virtual multi-eccentric Hartmann–Shack (HS) aberrometer for simultaneous measurement of peripheral ocular aberrations. The simulated system records nine wavefronts distributed within a region of approximately 20° around the fovea and is evaluated under ideal and realistic conditions, including speckle and illumination inhomogeneities in HS sensor images. Aberrations are quantified by two approaches: a traditional centroid-based method and a custom-made ResNet convolutional neural network (CNN) trained on simulated HS images. Results show that speckle markedly degrades traditional reconstructions for eccentric points, whereas the CNN (trained with speckled data) exhibits greater robustness and accuracy across all nine wavefronts. This model serves as a proof of concept toward a physical multi-eccentric aberrometer that acquires nine wavefronts in a single shot, thereby eliminating sequential point-by-point acquisition, reducing measurement time, and limiting motion-induced variability between measurements. The approach has potential applications in evaluating ocular treatments and in longitudinal studies of ametropias where peripheral aberrations are implicated. Full article

(This article belongs to the Special Issue Advances in Optical Sensing, Instrumentation and Systems: 3rd Edition)

► Show Figures

Figure 1

21 pages, 6586 KB

Open AccessArticle

Investigation of Magnetoelectric Properties and Applications in Multiferroic Composite

by Tingyu Deng, Jinlou Gu, Dong Wang and Jie Jiao

Sensors 2026, 26(14), 4418; https://doi.org/10.3390/s26144418 (registering DOI) - 12 Jul 2026

Abstract

In this work, resonator applications based on the magnetoelectric coupling effect in multiferroic materials are systematically investigated, with particular emphasis on mechanically driven ME antennas. A finite-element model is established to analyze the electromechanical response and coupling behavior of the device. To better [...] Read more.

In this work, resonator applications based on the magnetoelectric coupling effect in multiferroic materials are systematically investigated, with particular emphasis on mechanically driven ME antennas. A finite-element model is established to analyze the electromechanical response and coupling behavior of the device. To better describe the converse ME process, a nonlinear magnetostrictive model is introduced to evaluate the influence of material properties, structural configuration, and DC bias magnetic field on resonance characteristics and radiation performance. The simulation results show that the radiation intensity of the ME antenna is strongly dependent on the applied bias magnetic field and can be significantly enhanced under the optimal operating condition. On this basis, key parameters are optimized to reduce the resonance frequency and improve the radiation response. Prototype devices are then fabricated and experimentally characterized. The measured results verify the predicted resonance behavior and demonstrate the feasibility of the proposed devices in low-frequency wireless communication and magnetic-anomaly sensing. This study provides theoretical guidance and experimental support for the design of portable low-frequency ME antenna systems and other resonator-based magnetoelectric devices. Full article

(This article belongs to the Section Physical Sensors)

► Show Figures

Figure 1

24 pages, 7130 KB

Open AccessArticle

S2-DyGNN: A Spectro-Spatial Dynamic Graph Neural Network for Acoustic Event Classification in Distributed Acoustic Sensing

by Seunghun Jeong, Huioon Kim, Young Ho Kim, Hyoyoung Jung and Hong Kook Kim

Sensors 2026, 26(14), 4417; https://doi.org/10.3390/s26144417 (registering DOI) - 12 Jul 2026

Abstract

Distributed acoustic sensing (DAS) systems capture complex, nonlinear wave propagation across fiber-optic cables. Conventional event classification architectures, constrained by static physical topologies or isolated spatial grids, fail to effectively adapt to the dynamic feature relationships associated with such events, particularly when modeling complex [...] Read more.

Distributed acoustic sensing (DAS) systems capture complex, nonlinear wave propagation across fiber-optic cables. Conventional event classification architectures, constrained by static physical topologies or isolated spatial grids, fail to effectively adapt to the dynamic feature relationships associated with such events, particularly when modeling complex spatiotemporal interactions across sensor arrays. To resolve these structural limitations, we introduce the Spectro-Spatial Dynamic Graph Neural Network (S2-DyGNN), whose architecture couples a two-dimensional frequency–time convolutional front-end with a dual-matrix graph neural network (GNN). First, the convolutional module extracts spectro-temporal features, explicitly capturing localized acoustic dynamics independent of inter-sensor interference. Subsequently, the graph module constructs a dual-matrix topology, fusing a static physical distance prior with a data-driven adjacency matrix that recalculates spatial connections frame by frame from input signals. When evaluated on a highly skewed nine-class DAS field dataset, S2-DyGNN outperformed other conventional models by achieving a peak macro-averaged F1-score of 86.6% and an overall accuracy of 94.0%. The dual-matrix graph topology prevented dominant background features from washing out sparse transient events, improving the minority “openclose” class F1-score to 55.7% compared to the 48.0% ceiling of a static graph topology. These results demonstrate that explicitly coupling localized spectro-temporal representations with physically anchored spatial topologies consistently outperforms models that process these domains in isolation, providing a highly robust and scalable solution for real-world continuous monitoring systems. Full article

(This article belongs to the Special Issue Distributed Acoustic Sensing and Applications)

► Show Figures

Figure 1

24 pages, 3474 KB

Open AccessArticle

ASBR-CL: Stage-Aware Balanced Replay for Memory-Limited Continual Fault Diagnosis

by Jiaxin Li and Guanghe Zhu

Sensors 2026, 26(14), 4416; https://doi.org/10.3390/s26144416 (registering DOI) - 11 Jul 2026

Abstract

Vibration-based sensing systems for deployed industrial fault diagnosis often face incremental fault classes, changing degradation stages, and limited permission to retain historical sensor streams. Static fault classifiers are therefore insufficient for online maintenance settings in which a model must learn new sensor-observed states [...] Read more.

Vibration-based sensing systems for deployed industrial fault diagnosis often face incremental fault classes, changing degradation stages, and limited permission to retain historical sensor streams. Static fault classifiers are therefore insufficient for online maintenance settings in which a model must learn new sensor-observed states while preserving previous diagnostic knowledge under a bounded memory budget. This paper proposes ASBR-CL, an adaptive stage-aware balanced replay framework for continual fault diagnosis under a fixed exemplar-memory budget. ASBR-CL combines dataset-adaptive exemplar memory, balanced replay between current data and retained exemplars, and a conditional validation checkpoint module that is enabled only when it improves balanced stage recognition. Experiments on SEU-enhanced92, the 92-dimensional feature construction for SEU, and XJTU-bearing23, the 23-dimensional feature construction for XJTU-SY, compare ASBR-CL with DGGN/MFF same-backbone continual-learning baselines and XJTU imbalance-aware variants under five random seeds and

K = 100

training exemplars. On SEU, the selected ASBR-CL setting reports

97.49 \pm 1.18

Average Accuracy,

90.37 \pm 4.53

Final Accuracy,

90.37 \pm 4.53

Macro Recall, and

11.50 \pm 5.25

Average Forgetting. On XJTU, the conservative ASBR-CL-BoundedVal-K100 setting is not a universal Final Accuracy winner: DGGN-ER reaches a slightly higher Final Accuracy (

89.52 \pm 2.68

versus

89.05 \pm 2.48

). The ASBR-CL evidence instead lies in balanced recognition and retention, with Macro Recall

75.00 \pm 2.00

, Macro-F1

67.66 \pm 3.83

, and Average Forgetting

20.39 \pm 4.89

, compared with DGGN-ER at

63.94 \pm 7.50

,

55.48 \pm 12.82

, and

41.11 \pm 13.61

. Additional validation-resource, RMS-derived stage-definition, and imbalance-aware baseline analyses show that the revised XJTU claim should be framed as more stable Macro Recall, Macro-F1, and forgetting control under memory-limited continual diagnosis, not as superiority on every accuracy metric. Full article

(This article belongs to the Section Fault Diagnosis & Sensors)

► Show Figures

Figure 1

27 pages, 14879 KB

Open AccessArticle

The Role of Testing and Characterization in the Methodical Development of Sensor-Integrating Bolts for Multi-Axial Force Measurement

by Julian Peters, Klaus Rappenecker, Fabian Deeg, Felix Herbst, David Riehl, Dirk Leiacker, Mario Kupnik, Klaus Hofmann and Sven Matthiesen

Sensors 2026, 26(14), 4415; https://doi.org/10.3390/s26144415 (registering DOI) - 11 Jul 2026

Abstract

Sensor-integrating bolts offer a promising solution for in situ data acquisition by integrating sensory functions into existing machine elements. The standardized mechanical interfaces and performance must be maintained in order to ensure widespread use and retrofitting, which leads to a conflict of objectives [...] Read more.

Sensor-integrating bolts offer a promising solution for in situ data acquisition by integrating sensory functions into existing machine elements. The standardized mechanical interfaces and performance must be maintained in order to ensure widespread use and retrofitting, which leads to a conflict of objectives between mechanical and sensory performance. The state of the art lacks multi-axial force-measuring bolts that meet these objectives. Therefore, we present a methodical approach that develops evolving prototypes of sensor-integrating bolts and applies testing activities to characterize the bolts’ performance and gain knowledge to further improve the prototypes. Strain gauges and a specifically developed electronics platform for data acquisition with a BLE interface are used. The prototypes are tested for axial and bending loads. Mechanical and sensory performance indicators are investigated, such as von Mises stress increase, linearity errors, hysteresis errors and sensitivity. The characterization results of the testing activities and the knowledge extracted from each iteration concerning aspects like strain gauge type and position or sensor body integration and their influence on the performance indicators are shown. Overall, the prototypes improved the linearity error from 3.3% to 1% (axial) and 3.7% to 0.1% (bending). The error is below 6% and sensitivities are around 10 µV/V/kN (axial) and 1 µV/V/Nm (bending). The integration of the measurement chain reduces the mechanical performance within a range of 1.3% to 24.6%, underlining the conflict of objectives. Measures for compensating the reduction need to be taken, such as increasing the bolt’s strength. Moreover, the measurements indicate that bending torques are superposed, even in axial load cases, making investigations of the effect on loadability necessary. Full article

(This article belongs to the Section Industrial Sensors)

► Show Figures

Figure 1

30 pages, 6223 KB

Open AccessArticle

Design and Testing of an Integrated Robot for Harvesting, Stipe-Cutting, and Grading of Agaricus bisporus

by Tianhang Ding, Yingying Zhou, Hao Ma, Yulong Ding and Hongwei Cui

Sensors 2026, 26(14), 4414; https://doi.org/10.3390/s26144414 (registering DOI) - 11 Jul 2026

Abstract

To address the limitations of current Agaricus bisporus harvesting robots, including low picking efficiency, susceptibility to mechanical damage, poor operational stability, and discontinuous harvesting processes, an integrated robotic system capable of mushroom detection and localization, picking sequence planning, stipe-cutting, and grading was developed. [...] Read more.

To address the limitations of current Agaricus bisporus harvesting robots, including low picking efficiency, susceptibility to mechanical damage, poor operational stability, and discontinuous harvesting processes, an integrated robotic system capable of mushroom detection and localization, picking sequence planning, stipe-cutting, and grading was developed. The robot adopts a left-right symmetrical configuration and operates along the side of the mushroom cultivation racks. It mainly consists of a mobile platform, a lifting mechanism, dual picking units, a receiving unit, a stipe-cutting device, a collection system, and an electronic control system. A picking sequence planning method based on YOLOv8n-USD and KD-tree nearest neighbor search was employed to determine the optimal harvesting order for densely clustered and adhered mushrooms. The dual robotic arms executed the picking operations, after which the mushrooms were transferred to the receiving unit for stipe-cutting and subsequently classified according to the detection results, thereby completing the integrated process of detection, picking, stipe-cutting, and grading. Experimental results show that the average detection accuracy reaches 96.64%, the picking success rate is 95.39%, and the harvesting damage rate is 1.63%, while the average interaction failure rate and grading error rate are 0.35% and 1.28%, respectively. These results indicate that the proposed robot can achieve accurate detection and efficient integrated operation for densely clustered Agaricus bisporus. The findings provide a reference for flexible and efficient harvesting of Agaricus bisporus with synchronized stipe-cutting and grading operations. Full article

(This article belongs to the Section Sensors and Robotics)

12 pages, 6826 KB

Open AccessArticle

Frequency-Dependent Responses of Extensometers to Atmospheric Loading: Evidence from Geodynamical Observatory Jingyuan in the NE Margin of the Tibetan Plateau

by Jinling Yang and Xiaolin Yang

Sensors 2026, 26(14), 4413; https://doi.org/10.3390/s26144413 (registering DOI) - 11 Jul 2026

Abstract

Extensometers housed in vaults are capable of resolving strain changes of less than a nanostrain; however, they are differently distorted by variations in barometric pressure at different frequencies, which may result in complicated strain noise. Therefore, how to quantify the barometric effects in [...] Read more.

Extensometers housed in vaults are capable of resolving strain changes of less than a nanostrain; however, they are differently distorted by variations in barometric pressure at different frequencies, which may result in complicated strain noise. Therefore, how to quantify the barometric effects in different frequency bands is a perennial challenge in extensometric observation and research. Since the first strain observations at the Geodynamical Observatory Jingyuan (NW China) in 2007, the barometric effect has proven significant. Nevertheless, the frequency dependence and underlying mechanism of this effect remain unknown. With targeted research lacking at present, this study attempts to adopt the transfer function method to provide a systematic diagnosis. The results indicate that (1) the North–South (NS) component responds significantly to low-frequency pressure waves, with a slight phase shift leading the phenomenon; (2) the East–West (EW) component responds well to pressure waves in high-, medium-, and low-frequency bands, and the frequency dependence of the barometric coefficient spectrum and phase shift spectrum is strong; (3) the frequency-dependent response effect of the EW component may be related to factors such as horizontal pressure gradient force, topography, and fracture medium. These findings not only contribute to sub-frequency band correction of the barometric effect at the Geodynamical Observatory Jingyuan but also deepen our understanding of the ground–atmosphere coupling mechanism in the fault zone. Full article

(This article belongs to the Section Fault Diagnosis & Sensors)

36 pages, 3243 KB

Open AccessArticle

WebView-Based Hybrid Analysis of Link and Event for On-Device QR Phishing Detection Framework

by Jian Woo, Seungmin Lee, Inseok Park and Sejong Lee

Sensors 2026, 26(14), 4412; https://doi.org/10.3390/s26144412 (registering DOI) - 11 Jul 2026

Abstract

Quishing, a form of phishing conducted through QR codes, has emerged as a critical threat to user information security in mobile environments. Quishing attacks exploit the QR scanning workflow by opening malicious URLs in WebView and impersonating legitimate services to steal user credentials. [...] Read more.

Quishing, a form of phishing conducted through QR codes, has emerged as a critical threat to user information security in mobile environments. Quishing attacks exploit the QR scanning workflow by opening malicious URLs in WebView and impersonating legitimate services to steal user credentials. Recent variants further evade static inspection by exposing credential-harvesting behavior only after user interaction, form submission, redirection, or page-state changes. In this paper, we propose WebView-Based Hybrid Analysis of Link and Event for On-Device QR Phishing Detection (WHALE), an On-Device multi-stage phishing detection framework based on an isolated Sandbox WebView. WHALE first loads the QR-decoded URL into the Sandbox WebView instead of directly delivering it to the User WebView, thereby separating the analysis process from the user session. In the static stage, WHALE extracts 54 features from the URL string, initial HTML, and DOM snapshot, and computes a static phishing risk score using a lightweight model. Inputs with uncertain static scores are forwarded to the dynamic stage. In the dynamic stage, WHALE inserts decoy credentials instead of real user credentials, triggers a controlled submit event, and analyzes 59 credential-flow state-transition features extracted before and after submission. The static model achieved an accuracy of 93.86%, precision of 93.08%, recall of 94.78%, and F1-score of 93.92%. The dynamic model achieved an accuracy of 0.915, precision of 0.895, recall of 0.946, specificity of 0.882, and F1-score of 0.920 on a source-group-disjoint independent test set. Real-device evaluation on a Samsung Galaxy S23 Ultra showed that WHALE maintains practical mobile overhead, with an average static internal runtime of 58.75 ms, dynamic internal runtime of 4165.42 ms, combined model inference time of 0.088 ms, and model asset size of 0.681 MB. These results demonstrate that WHALE can detect QR-based phishing threats On-Device while reducing user credential exposure through sandboxed credential-flow analysis. Full article

(This article belongs to the Special Issue Intelligence, Security, Trust and Privacy Advances in IoT, Bigdata and 5G Networks (3rd Edition))

Journal Description

Sensors

Latest Articles

Journal Menu

Journal Browser

Highly Accessed Articles

Latest Books

E-Mail Alert

News

Topics

Conferences

Special Issues

Topical Collections

Further Information

Guidelines

MDPI Initiatives

Follow MDPI