Next Issue
Volume 26, July-2
Previous Issue
Volume 26, June-2
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
9.4
4.0
Journals
Sensors
Volume 26
Issue 13
sensors-logo

Journal Menu

► Journal Menu

Journal Browser

► Journal Browser
share announcement

Sensors, Volume 26, Issue 13 (July-1 2026) – 356 articles

Cover Story (view full-size image): In the field of shock physics, Fiber Bragg Gratings (FBGs) are used to measure shock velocity, detonation velocity and shock pressure levels. They are also used to measure strain. FBG sensors are known to be light and small, with a fast response. A high-speed dispersive spectrometer was developed using a mode-locked fiber laser with a wide spectrum and a long fiber for chromatic dispersion. It creates a ramp of wavelengths at 100 MHz to interrogate the FBG and observe complex phenomena. The experiments presented are a low-pressure shock velocity measurement in epoxy, a deflagration-to-detonation transition in a porous energetic material, a shock-to-detonation transition in a dense energetic material, a tentative-to-measure pressure level in epoxy from an FBG made in a sapphire fiber and multi-point strain measurements up to eight FBGs. View this paper
  • Issues are regarded as officially published after their release is announced to the table of contents alert mailing list.
  • You may sign up for e-mail alerts to receive table of contents of newly released issues.
  • PDF is the official format for papers published in both, html and pdf forms. To view the papers in pdf format, click on the "PDF Full-text" link, and use the free Adobe Reader to open them.
Order results
Result details
Section
Show export options expand_more Show export options expand_less
Select all
Export citation of selected articles as:
22 pages, 16089 KB  
Article
Real-Time Detection System for Road Roughness Based on Ultrasonic Technology
by Hongjia Zhao, Libo Wang, Yimin Zhao and Xiaodong Sun
Sensors 2026, 26(13), 4324; https://doi.org/10.3390/s26134324 - 7 Jul 2026
Viewed by 660
Abstract
With the rapid development of intelligent connected vehicles and autonomous driving, real-time and accurate road condition perception has become increasingly critical. Aiming at the limitations of traditional direct and indirect detection methods, this paper proposes an ultrasonic-based real-time detection system for road roughness. [...] Read more.
With the rapid development of intelligent connected vehicles and autonomous driving, real-time and accurate road condition perception has become increasingly critical. Aiming at the limitations of traditional direct and indirect detection methods, this paper proposes an ultrasonic-based real-time detection system for road roughness. Most urban roads today feature asphalt pavements; therefore, this system focuses its research on asphalt pavements. Under the same pavement type (asphalt roads), there is a strong correlation between pavement roughness and the friction coefficient. By measuring the roughness of different pavements, the friction coefficient is estimated using the fuzzy processing method. Then the system through measuring ultrasonic echo amplitude and sensor–road distance, combined with software digital filtering, dual-parameter compensation (distance and temperature–humidity), probabilistic statistical analysis, and fuzzy inference, the mapping relationship among echo signals, road roughness and friction coefficient is established. The system mainly includes an ultrasonic transceiver module, a hardware signal conditioning module, and an MCU-based data processing, display and transmission module. Both simulated experiments and real asphalt pavement tests are carried out for verification. The results show that the system can effectively suppress noise, compensate distance attenuation and environmental interference, and achieve accurate real-time detection of road roughness, with a relative error less than 10% compared with the reference value. The proposed system can provide reliable data support for vehicle active safety systems and autonomous driving applications. Full article
(This article belongs to the Section Physical Sensors)
►▼ Show Figures

Graphical abstract

26 pages, 1587 KB  
Article
Vibration-Based Machine Learning Model Training for Railway Bridge Health Monitoring
by Rocco Alaggio, Muhammad Asad, Riccardo Cirella, Stefania Costantini and Giovanni De Gasperis
Sensors 2026, 26(13), 4323; https://doi.org/10.3390/s26134323 - 7 Jul 2026
Cited by 1 | Viewed by 1054
Abstract
Bridge health monitoring and machine learning are increasingly intertwined for civil engineers and artificial intelligence experts. Bridges’ poor health can result in severe outcomes if not addressed in time. Therefore, continuous monitoring is required to detect any anomaly or damage. Sensors, such as [...] Read more.
Bridge health monitoring and machine learning are increasingly intertwined for civil engineers and artificial intelligence experts. Bridges’ poor health can result in severe outcomes if not addressed in time. Therefore, continuous monitoring is required to detect any anomaly or damage. Sensors, such as accelerometers, inclinometers, thermistors, etc., can help actively monitor these bridges. The signals from these sensors help record physiological activities. Such activities are helpful for anomaly detection, damage localization, and bridge health predictions with the help of machine learning algorithms. The proposed method extracts features from the dynamic response of a bridge to ambient excitation. It focuses on processing the signal received from different accelerometers installed on a steel railway bridge to determine the location of the damage and the level of the damage predictions. Initially, features are extracted from time-series data; then, they are fed to a deep neural network after some pre-processing. Normal and augmented data are used with different parameter tuning for results. Original data is also subdivided, and the effect of data slicing on the predictions is investigated. The results show that one-fourth of the slicing of the original data gives the best results for training and testing accuracy with a deep neural network. The results show that the reduced matrix representation, particularly the 40 × 40 feature slicing, improved the classification performance for the predefined bridge scenario classes under the considered experimental settings. For bridge scenario classification, the best reported accuracy was 93.54%, while for damage intensity classification the best reported accuracy was 98.21%. In the DNN-based optimizer comparison, the Adam optimizer achieved higher and more stable performance than Stochastic Gradient Descent (SGD), with test accuracies of 92.3% and 93.7% compared with 75.2% and 86.4%, respectively. It is also observed that the Adam optimizer outperformed Stochastic Gradient Descent (SGD) in terms of both damage localization and damage intensity estimation. Full article
(This article belongs to the Section Fault Diagnosis & Sensors)
►▼ Show Figures

Figure 1

23 pages, 2514 KB  
Article
FedHSFV: Federated Learning for Finger Vein Recognition via Hierarchical Decoupling and Subspace Metric
by Ximing Zhou, Yuhan Wang, Jiajun Cui, Jian Guo and Hengyi Ren
Sensors 2026, 26(13), 4322; https://doi.org/10.3390/s26134322 - 7 Jul 2026
Viewed by 659
Abstract
Finger vein recognition (FVR) has significant potential in biometrics due to its high accuracy and intrinsic liveness detection capabilities. However, the increasingly stringent privacy regulations have presented severe data security challenges for traditional centralized training. While federated learning (FL) mitigates these privacy concerns [...] Read more.
Finger vein recognition (FVR) has significant potential in biometrics due to its high accuracy and intrinsic liveness detection capabilities. However, the increasingly stringent privacy regulations have presented severe data security challenges for traditional centralized training. While federated learning (FL) mitigates these privacy concerns through a decentralized training paradigm, conventional FL algorithms that seek a single global model experience significant performance degradation on non-independent and identically distributed (Non-IID) data in real-world cross-institutional deployments. This degradation stems primarily from a dual-heterogeneity issue that involves domain shift caused by hardware discrepancies across acquisition devices, and label skew resulting from nonoverlapping user identities. To address this dual-heterogeneity challenge, we propose a personalized federated learning framework driven by hierarchical parameter decoupling and subspace metric. First, we designed a hierarchical parameter decoupling architecture. Macroscopically, the architecture retains the classifier locally to isolate label heterogeneity; microscopically, it introduces an additive parameter decomposition that decouples the feature extractor on a global full-rank basis (to capture domain-invariant semantics, namely, the shared physiological vein topologies) and a local low-rank adapter (that accommodates device-specific characteristics, such as hardware-induced noise and illumination discrepancies). Furthermore, we propose a subspace similarity matching strategy based on principal angles on the Grassmann manifold. By exploiting the geometric properties of low-rank projection matrices, this strategy accurately quantifies the underlying distribution discrepancies among clients to guide personalized weighted aggregation. Extensive experiments on six public finger vein datasets demonstrate that the proposed framework significantly improves the overall recognition performance and mitigates performance degradation caused by data heterogeneity. Full article
(This article belongs to the Section Intelligent Sensors)
►▼ Show Figures

Figure 1

30 pages, 21473 KB  
Article
Measuring Methane Emissions in Ambient Air with a Low-Cost, Portable Sensor System: Focus on Scalability and Transferability of the Model
by Lorenzo Bertin, Matteo Mentasti, Fabrizio Pittorino, Veronica Villa, Emanuele Zanni, Gabriele Viscardi, Yuri Ponzani, Andrea Massara, Manuel Roveri, Raffaele Dellaca’ and Laura Capelli
Sensors 2026, 26(13), 4321; https://doi.org/10.3390/s26134321 - 7 Jul 2026
Viewed by 679
Abstract
Landfills represent a significant source of methane emissions, with important environmental, climatic and safety impacts due to the widespread and variable nature of these emissions. Traditional monitoring methods, such as flow chambers coupled with flame ionisation detectors (FIDs), provide high accuracy but are [...] Read more.
Landfills represent a significant source of methane emissions, with important environmental, climatic and safety impacts due to the widespread and variable nature of these emissions. Traditional monitoring methods, such as flow chambers coupled with flame ionisation detectors (FIDs), provide high accuracy but are limited in terms of spatial representativeness, operational flexibility and cost, especially during large-scale or continuous monitoring campaigns. Within this context, the European ESCAPE project aims to develop a low-cost, portable and modular platform for the detection and quantification of low methane concentrations in ambient air at complex environmental sites. The system is based on commercial MOX and NDIR sensors integrated into portable toolboxes equipped with dedicated chambers, regulated suction systems and autonomous data acquisition units with real-time transmission. This work describes the development and testing of two identical toolboxes to assess system reproducibility and the transferability of predictive models between devices. Laboratory and field tests were carried out under controlled and real landfill conditions, with comparisons against portable FID measurements. Results showed good agreement between predicted methane concentrations and reference data, with correlation indexes up to 0.77. Moreover, transferring the machine learning model between toolboxes did not produce statistically significant performance reductions, demonstrating promising robustness and generalizability of the proposed calibration strategy. Full article
(This article belongs to the Special Issue Recent Advances in Sensors for Chemical Detection Applications (2nd Edition))
►▼ Show Figures

Figure 1

36 pages, 3485 KB  
Article
Auditing Road-Segment Speed Forecasting Under Sparse Mobile Probe Sensing: A Mask-Consistent Support-Chain Analysis
by Dingxin Wu, Zheng Xu, Zhiyuan Wang, Kai Huang, Hong Ki An and Dewen Kong
Sensors 2026, 26(13), 4320; https://doi.org/10.3390/s26134320 - 7 Jul 2026
Viewed by 475
Abstract
Ride-hailing global positioning system (GPS) mobile probe data provide flexible urban traffic observations, but their sparse and uneven coverage makes model evaluation difficult because observed targets, valid predictions, and historical input support do not always coincide. This study audits ultra-short-term road-segment speed forecasting [...] Read more.
Ride-hailing global positioning system (GPS) mobile probe data provide flexible urban traffic observations, but their sparse and uneven coverage makes model evaluation difficult because observed targets, valid predictions, and historical input support do not always coincide. This study audits ultra-short-term road-segment speed forecasting under sparse mobile sensing using a mask-consistent support-chain framework. A three-day GPS dataset is aggregated into 5 min speed observations over 1970 road segments and used as a controlled sparse-sensing case study rather than a general-purpose long-term forecasting benchmark. The evaluation protocol distinguishes the full test grid, the set of directly observed target speeds, model-valid prediction support, strict complete-history support, and common-support subsets for coverage-limited baselines. The directly observed target set is used as the primary relaxed support because it retains all verifiable ground-truth targets, while strict and common-support subsets are reported as sensitivity checks. Under this support-conditioned evaluation, the adaptive graph convolutional recurrent network (AGCRN) is associated with lower mean absolute error (MAE) among full-coverage models, the historical mean (HIST_MEAN) baseline is associated with lower root mean squared error (RMSE), and congestion recall remains below 0.24 for all full-coverage deep models. These complementary results indicate conditional and metric-dependent strengths rather than universal model superiority. Because the dataset covers only three consecutive days, weekday/weekend variation, incident-specific fluctuations, seasonal effects, and spatial transferability cannot be fully examined and are treated as limitations. Overall, the findings show that evaluation support should be reported as a first-order experimental factor alongside model accuracy under sparse mobile probe sensing. Full article
(This article belongs to the Special Issue Smart Traffic Control Based on Sensor Technology)
►▼ Show Figures

Figure 1

24 pages, 1226 KB  
Article
Generative Adversarial Network-Based Joint Mapping and Localization for Millimeter-Wave Communication Systems
by Zexu Zhao, Zhigang Chen and Lu Chen
Sensors 2026, 26(13), 4319; https://doi.org/10.3390/s26134319 - 7 Jul 2026
Viewed by 358
Abstract
In this paper, we propose a novel generative adversarial network (GAN)-based joint localization and mapping (JLAM) method using angle difference of arrival (ADOA) measurements for millimeter-wave (mmWave) communication systems. The proposed method adopts a deep auto-encoder neural network as the discriminator of the [...] Read more.
In this paper, we propose a novel generative adversarial network (GAN)-based joint localization and mapping (JLAM) method using angle difference of arrival (ADOA) measurements for millimeter-wave (mmWave) communication systems. The proposed method adopts a deep auto-encoder neural network as the discriminator of the GAN and models the generator as an explicit geometric ADOA function of the access point (AP) positions and the mobile terminal (MT) position, rather than as a conventional black-box neural network. By exploiting the two-dimensional distribution characteristics of high-dimensional ADOA vectors collected at a large number of random and unknown MT positions, the proposed method learns the ADOA data distribution and transforms it into the AP geometric topology. Then, the MT positions and the indoor map are estimated based on the recovered physical and virtual AP topology. The simulation results show that, under the representative setting with N=2000 measured ADOA vectors and σ=2° AOA measurement noise, the proposed method achieves an average localization error of about 0.25 m, compared with about 0.60 m for the JADE algorithm, corresponding to an error reduction of approximately 58%. The proposed method also provides more accurate room boundary estimation than JADE, confirming its effectiveness for mmWave JLAM. Full article
(This article belongs to the Special Issue 5G/6G Networks for Wireless Communication and IoT—2nd Edition)
►▼ Show Figures

Figure 1

19 pages, 7632 KB  
Article
BSEF-Stereo: A Stereo Matching Model Based on Branching Strategy and Error Feedback
by Kaicheng Li, Jinlong Yang and Chin Chi Choi
Sensors 2026, 26(13), 4318; https://doi.org/10.3390/s26134318 - 7 Jul 2026
Cited by 1 | Viewed by 455
Abstract
Iterative stereo matching remains challenging in weakly textured regions, repetitive patterns, occlusions, and object boundaries, where ambiguous correspondence cues require broad contextual reasoning while accurate reconstruction depends on preserving local structural details. Existing recurrent updaters with a fixed receptive field struggle to balance [...] Read more.
Iterative stereo matching remains challenging in weakly textured regions, repetitive patterns, occlusions, and object boundaries, where ambiguous correspondence cues require broad contextual reasoning while accurate reconstruction depends on preserving local structural details. Existing recurrent updaters with a fixed receptive field struggle to balance these requirements, and their initial disparity estimates may retain local geometric inconsistencies. To address these limitations, we propose BSEF-Stereo, an iterative framework that combines adaptive recurrent updating with explicit error-feedback refinement. A channel–position attention module strengthens discriminative channel and spatial cues, while a branch-strategy gated recurrent unit uses complementary small- and large-kernel branches to preserve boundary details and aggregate context in ambiguous regions. An error-aware refinement module subsequently exploits reprojection error and image guidance to correct the initial disparity map. Experiments on Scene Flow, KITTI 2012, KITTI 2015, and Middlebury demonstrate competitive performance across synthetic, outdoor, and indoor scenes. BSEF-Stereo achieves 0.41 px EPE and 2.27% D1 on Scene Flow and a D1-all error of 1.48% on KITTI 2015. Ablation and sensitivity studies verify the complementary contributions of the three modules and support the selected design settings. Full article
(This article belongs to the Special Issue AI-Based Visual Sensing for Object Detection)
►▼ Show Figures

Figure 1

20 pages, 584 KB  
Article
Cost-Aware Scheduling Under Latency Constraints for Multi-View 3D Reconstruction Across the Edge–Cloud Continuum
by Ivan Čilić, Ivana Podnar Žarko, Mario Kušek and Josip Štajdohar
Sensors 2026, 26(13), 4317; https://doi.org/10.3390/s26134317 - 7 Jul 2026
Viewed by 818
Abstract
Learning-based multi-view 3D reconstruction pipelines, such as transformer-based approaches, enable the accurate reconstruction of 3D scenes from multiple images, but their deployment across the edge–cloud continuum is challenging due to high computational demands and large intermediate data transfers. Effective pipeline scheduling in the [...] Read more.
Learning-based multi-view 3D reconstruction pipelines, such as transformer-based approaches, enable the accurate reconstruction of 3D scenes from multiple images, but their deployment across the edge–cloud continuum is challenging due to high computational demands and large intermediate data transfers. Effective pipeline scheduling in the continuum must therefore balance latency constraints with the cost of cloud resource usage. In this work, we address cost-aware scheduling under latency constraints for a multi-stage 3D reconstruction pipeline consisting of depth estimation, transformer-based multi-view fusion, and point cloud merging with export to a rendering-ready representation. We implement a service-oriented pipeline where each stage can be executed either on edge or cloud nodes, and we experimentally characterize its performance on representative hardware platforms. The results show a strong imbalance between the computational time and communication latency across platforms, mainly due to large intermediate data. Based on these insights, we propose an online scheduler that dynamically selects stage placements to minimize the cloud cost while satisfying latency constraints. The scheduler incorporates a top-K edge selection mechanism that reduces the decision complexity by jointly considering the network conditions and node utilization. Simulation results parameterized with real-system measurements show that the proposed approach effectively reduces cloud usage while meeting latency constraints, outperforming the baseline strategies based on single-node pipeline execution. Full article
(This article belongs to the Special Issue Optimizing the Future: Securing and Streamlining IoT Resource Allocation in the 6G Era with Edge Computing and UAV Integration)
►▼ Show Figures

Figure 1

25 pages, 15162 KB  
Article
A Novel Hybrid Chaotic Map and Cryptographic Hash Whitening for Optimized S-Box Design: Construction, Cryptanalysis, and Lightweight IoT Sensor Evaluation
by Nahar F. Alshammari, Faraj H. Alyami, Abdullah G. Alharbi, Saleh Al Dawsari, Yousaf Hameed Khattak and Faisal Baig
Sensors 2026, 26(13), 4316; https://doi.org/10.3390/s26134316 - 7 Jul 2026
Cited by 1 | Viewed by 564
Abstract
In this work, a novel method was introduced for the construction of the S-box based on delayed nonlinear chaotic systems coupled with a hybrid hash whitening mechanism. The evidence for strong nonlinear dynamical behavior is the strongly positive Lyapunov exponent with uniform statistical [...] Read more.
In this work, a novel method was introduced for the construction of the S-box based on delayed nonlinear chaotic systems coupled with a hybrid hash whitening mechanism. The evidence for strong nonlinear dynamical behavior is the strongly positive Lyapunov exponent with uniform statistical distribution of the generated binary sequence. Optimized 8-bit → 8-bit S-boxes, displayed in 16 × 16 hexadecimal format, were achieved by applying Fisher–Yates permutation to the chaotic sequences and further enhancing them with affine transformations over GF(2). Nonlinearity, differential uniformity, avalanche effect, strict avalanche criterion, linear approximation table, difference distribution table, and algebraic degree are cryptographic metrics that reveal very strong resistance against linear and differential attacks. Image encryption with the generated S-box further validates the confusion properties. The results confirm that the hybrid approach achieves high-quality S-boxes suitable for symmetric cryptography. Full article
(This article belongs to the Section Communications)
►▼ Show Figures

Figure 1

42 pages, 5892 KB  
Article
PM2.5/PM10 Forecasting System with Benchmarking of 44 Machine Learning Algorithms and Ensemble Learning Approaches
by Pedro Mamani-Suclla, Sharon Villavicencio-Siu and Antonio Arroyo-Paz
Sensors 2026, 26(13), 4315; https://doi.org/10.3390/s26134315 - 7 Jul 2026
Viewed by 1042
Abstract
Air pollution from particulate matter (PM2.5 and PM10) poses a serious public health risk in urban environments, particularly in areas with heavy vehicular traffic. Against this backdrop, the present study proposes an Internet of Things (IoT)-based system designed to support air quality monitoring [...] Read more.
Air pollution from particulate matter (PM2.5 and PM10) poses a serious public health risk in urban environments, particularly in areas with heavy vehicular traffic. Against this backdrop, the present study proposes an Internet of Things (IoT)-based system designed to support air quality monitoring and evidence-based decision-making regarding PM2.5 and PM10 concentrations, integrating low-cost sensors with a machine learning prediction module. The study follows an experimental-applied design with a quantitative–comparative approach. Its scientific contribution is organized around an integrated IoT-ML framework addressing a concrete gap in the literature: the lack of local empirical evidence regarding which family of machine learning algorithms delivers the greatest accuracy, stability, and computational efficiency for particulate matter forecasting in mid-altitude urban environments using low-cost sensors. On one hand, the framework proposes and deploys a four-node IoT network for continuous PM2.5 and PM10 monitoring in high-traffic urban microenvironments—representing one of the first sustained deployments with low-cost, high-temporal-resolution sensors (10-min intervals) in Arequipa, Peru. On the other hand, the study presents the most extensive benchmarking reported in the local literature: a systematic evaluation of 44 machine learning algorithms under homogeneous experimental conditions, covering classical statistical models, traditional machine learning techniques, deep learning architectures, and hybrid approaches, along with an analysis of ensemble learning strategies using Ridge stacking and K-Fold cross-validation. This unified comparative analysis—applying consistent metrics (MAE, RMSE, R2, and MAPE), the same prediction horizon, and a shared dataset—provides replicable empirical evidence that had not previously been reported for the urban context of Arequipa. The results show that traditional statistical models perform poorly overall, while tree-based and boosting algorithms consistently achieve R2 values above 0.90 for both pollutants. Ensemble models, particularly stacking with Ridge regression and cross-validation, yielded the strongest overall performance, demonstrating greater robustness and prediction stability. Explainability criteria were also incorporated, enabling an assessment of each base model’s individual contribution and identifying the variables most relevant to the prediction process. The methodological contribution provides future researchers with a rigorous reference framework for algorithm selection in environmental IoT systems. Taken together, the findings demonstrate that combining low-cost IoT networks with advanced machine learning and ensemble learning techniques constitutes an effective, scalable, and cost-efficient alternative for air quality monitoring, predictive analysis, and the support of informed mitigation strategies in urban environments. Full article
(This article belongs to the Section Environmental Sensing)
►▼ Show Figures

Figure 1

15 pages, 30250 KB  
Article
Learning Fine-Grained Video Anomaly Detection from Normal Videos
by Ruqin Wang, Yasumasa Tamura and Masahito Yamamoto
Sensors 2026, 26(13), 4314; https://doi.org/10.3390/s26134314 - 7 Jul 2026
Viewed by 671
Abstract
Video anomaly detection (VAD) aims to identify abnormal events in videos. Due to the lack of high-quality training data with detailed annotations, current VAD methods can only produce video-level predictions. To remedy this, several methods attempt to synthesize pseudo video anomalies. However, these [...] Read more.
Video anomaly detection (VAD) aims to identify abnormal events in videos. Due to the lack of high-quality training data with detailed annotations, current VAD methods can only produce video-level predictions. To remedy this, several methods attempt to synthesize pseudo video anomalies. However, these methods suffer from low realism and coarse annotations, which limits their performance in real-world scenarios. In this paper, we propose a framework for unsupervised anomaly video generation from solely normal videos, leveraging VLMs to generate structured textual descriptions of anomalies conditioned on the perception of this video. Then, abnormal segments are synthesized using VLMs based on the synthetic textual descriptions. As our framework is highly controllable, video-level and region-level labels can be obtained to provide fine-grained annotations. On top of the synthetic data, we develop a fine-grained VAD network to simultaneously produce video-level, frame-level, and region-level predictions. Experiments show that our method achieves remarkable fine-grained VAD performance. Full article
(This article belongs to the Special Issue Machine Learning and Deep Learning in Image/Video Processing and Sensing: 2nd Edition)
►▼ Show Figures

Figure 1

21 pages, 2365 KB  
Article
Q-GrAM: Fine-Grained Image–Text Retrieval via Grouped Query Routing and Conditional Query Modulation
by Guihe Gu, Huawei Li and Hong Qin
Sensors 2026, 26(13), 4313; https://doi.org/10.3390/s26134313 - 7 Jul 2026
Viewed by 554
Abstract
Existing image–text retrieval methods often compute cross-modal similarity using global single-vector representations. Although efficient for coarse semantic alignment, such compressed representations are limited when textual queries involve fine-grained semantics, including objects, attributes, relations, and their compositional structures. This paper focuses on fine-grained text-to-image [...] Read more.
Existing image–text retrieval methods often compute cross-modal similarity using global single-vector representations. Although efficient for coarse semantic alignment, such compressed representations are limited when textual queries involve fine-grained semantics, including objects, attributes, relations, and their compositional structures. This paper focuses on fine-grained text-to-image retrieval and proposes Q-GrAM, a retrieval-oriented adaptation of the BLIP-2 Q-Former. Instead of treating Q-Former queries as a homogeneous set, Q-GrAM partitions a fixed query budget into semantically differentiated groups. A text-guided router assigns token-level semantic demands to query groups, while query conditional initialization modulates each group according to group-level textual summaries. The resulting grouped visual query features are matched with text tokens through a group-aware late interaction scorer, and auxiliary routing balance and inter-group diversity regularization are introduced to stabilize semantic specialization. Experiments on MS-COCO 5K, Flickr30K, and Flickr30K-CFQ show that Q-GrAM achieves strong text-to-image retrieval performance against both global embedding baselines and representative fine-grained image–text matching methods, while maintaining competitive bidirectional retrieval performance. These results demonstrate the effectiveness of structured, text-conditioned Q-Former query specialization for fine-grained text-driven image search. Full article
(This article belongs to the Special Issue Artificial Intelligence in Computer Vision: Methods and Applications—2nd Edition)
►▼ Show Figures

Figure 1

30 pages, 11826 KB  
Article
Eddy-Current-Induced Waveform Reconstruction by Metallic Probe Carriers in Magnetic Flux Leakage Inspection
by Xiaoyuan Jiang, Bohan Jia and Yanhua Sun
Sensors 2026, 26(13), 4312; https://doi.org/10.3390/s26134312 - 7 Jul 2026
Viewed by 509
Abstract
Metallic probe carriers are commonly used in magnetic flux leakage (MFL) inspection to support sensing elements and maintain lift-off, but a conductive carrier located near the sensor can act as an active electromagnetic boundary. This study investigates the carrier-induced waveform reconstruction caused by [...] Read more.
Metallic probe carriers are commonly used in magnetic flux leakage (MFL) inspection to support sensing elements and maintain lift-off, but a conductive carrier located near the sensor can act as an active electromagnetic boundary. This study investigates the carrier-induced waveform reconstruction caused by such a conductive near-field boundary. A theoretical model is developed to describe the induced current, secondary magnetic field, and relaxation-related downstream memory generated when the carrier moves through a non-uniform leakage field. Transient finite-element simulations are used to examine the effects of carrier material, scanning speed, and concave carrier geometry. Compared with the air reference, aluminum and copper carriers produce stage-dependent waveform reconstruction, including valley modification, peak modulation, feature-position shift, and trailing-side extension. The quantitative waveform-deviation indicators increase with increasing speed and are further regulated by carrier geometry. Experimental results based on repeated magnetic response events confirm amplitude suppression, non-zero residual after amplitude matching, response broadening, and enhanced trailing asymmetry. These results demonstrate that the metallic probe carrier is not an electromagnetically transparent holder but an active near-field conductive boundary that should be considered in probe-carrier design and MFL signal interpretation. Full article
(This article belongs to the Section Physical Sensors)
►▼ Show Figures

Figure 1

14 pages, 4093 KB  
Article
Femtosecond Laser-Induced Graphene Modified with Platinum Nanoparticles for Advanced Multifunctional Sensing
by Jie Zhan, Mingle Guan, Zi Wang, Xiaolin Qi and Sumei Wang
Sensors 2026, 26(13), 4311; https://doi.org/10.3390/s26134311 - 7 Jul 2026
Viewed by 573
Abstract
Flexible sensors are important for wearable health monitoring, strain detection, and temperature sensing because of their mechanical flexibility and functional versatility. Here, a femtosecond laser direct scanning method was used to fabricate porous laser-induced graphene (LIG) and further modify it with platinum nanoparticles [...] Read more.
Flexible sensors are important for wearable health monitoring, strain detection, and temperature sensing because of their mechanical flexibility and functional versatility. Here, a femtosecond laser direct scanning method was used to fabricate porous laser-induced graphene (LIG) and further modify it with platinum nanoparticles (PtNPs), forming Pt/LIG. This mask-free and rapid process enables simultaneous patterning and functionalization of flexible sensors. The introduction of PtNPs improves the electron transport and surface adsorption properties of LIG. As a result, the sheet resistance of Pt/LIG is reduced to 2.41 Ω/sq, enhancing electrical conductivity and suitability for sensing applications. Based on this method, highly sensitive strain and temperature sensors were fabricated. The Pt/LIG strain sensor shows a ΔR/R0 of 1141.8 at a bending angle of 90°, about 213% higher than that of pristine LIG, with fast response and recovery times of 36 and 56 ms, respectively. The temperature sensitivity also improved by about 650%, with a temperature coefficient of resistance of 0.240%/°C, compared with −0.032%/°C for pristine LIG. Overall, this work provides a fast and precise strategy for fabricating nanoparticle–graphene composites for flexible electronics, wearable health monitoring, and environmental sensing. Full article
(This article belongs to the Section Nanosensors)
►▼ Show Figures

Figure 1

21 pages, 1453 KB  
Article
Characterizing Human Visual Performance in Dim Industrial Environments: An Eye-Tracking Sensor-Based Study on Mine Monitoring Interface Layouts
by Junqing Hao, Rui Chen and Wei Zong
Sensors 2026, 26(13), 4310; https://doi.org/10.3390/s26134310 - 7 Jul 2026
Viewed by 545
Abstract
In underground mining operations, intelligent electronic monitoring and control systems have gradually become essential tools for practitioners to obtain operational information. Because underground mines are characterized by low-light environments that differ substantially from above-ground natural lighting, screen-related factors can strongly affect visual search [...] Read more.
In underground mining operations, intelligent electronic monitoring and control systems have gradually become essential tools for practitioners to obtain operational information. Because underground mines are characterized by low-light environments that differ substantially from above-ground natural lighting, screen-related factors can strongly affect visual search tasks. It is therefore important to examine how interface layout influences visual search efficiency under low-illuminance mining conditions. This study utilized eye-tracking sensing technology to evaluate user performance within a simulated underground electronic monitoring and control system. Interface layout and ambient illuminance were set as experimental variables to investigate their effects on users’ search efficiency, quantitative eye-movement metrics, and usability satisfaction. The results showed that interface layout and ambient illuminance had significant main effects on task completion time, fixation count, saccade count, and subjective usability score. Among the tested layouts, the Three-Column Layout (THCL) showed the most favorable performance in task completion time, fixation count, and subjective usability score, while visual search efficiency was generally higher under the 50-lx condition. Full article
(This article belongs to the Section Intelligent Sensors)
►▼ Show Figures

Figure 1

31 pages, 6385 KB  
Article
Unsupervised Identification of Driving Styles from Naturalistic Driving Data Through a Context-Normalized Framework
by Cunzhi Xu, Reuben S.K. Agbozo, Liang Huang, Zheng Zhang, Tao Peng and Renzhong Tang
Sensors 2026, 26(13), 4309; https://doi.org/10.3390/s26134309 - 7 Jul 2026
Viewed by 653
Abstract
Identifying driving styles is essential for personalizing driving assistance systems and enhancing intelligent transportation services. However, existing approaches predominantly rely on experience-driven feature engineering and annotated data, limiting objectivity and hindering the exploitation of unlabeled naturalistic driving datasets. To address these limitations, this [...] Read more.
Identifying driving styles is essential for personalizing driving assistance systems and enhancing intelligent transportation services. However, existing approaches predominantly rely on experience-driven feature engineering and annotated data, limiting objectivity and hindering the exploitation of unlabeled naturalistic driving datasets. To address these limitations, this paper proposes an unsupervised framework for driving style identification from naturalistic driving data through context normalization. A Constrained Convolutional Autoencoder (CCAE) integrated with a global self-attention mechanism is developed to map unlabeled driving sequences onto a standardized dynamic reference defined by the Worldwide Harmonized Light Vehicles Test Cycle (WLTC). This process extracts Driving Adaptability Characteristics (DACs) as WLTC-anchored latent representations that characterize normalized driver-specific response patterns across heterogeneous naturalistic contexts. To ensure feature robustness, frequency-domain refinement is applied to eliminate high-frequency noise. The extracted DAC sequences are subsequently partitioned into distinct driving styles using a kernel-mapped clustering algorithm. To evaluate the external relevance and physical interpretability of the identified styles, actual vehicle accident records and raw CAN-bus feature backtracking are introduced as validation evidence. The results show that the identified driving styles exhibit different historical accident probabilities. The proposed CCAE model achieves clearer cluster-level differentiation than traditional feature engineering and unconstrained deep learning models, and the ablation analysis confirms the contribution of the WLTC-based constraint. These findings indicate that the context-normalization framework can extract interpretable and externally relevant driving style representations from unlabeled naturalistic data. Full article
(This article belongs to the Special Issue Feature Papers in "Industrial Sensors" Section 2026–2027)
►▼ Show Figures

Figure 1

12 pages, 735 KB  
Review
Transoral Robotic Cleft Palate Surgery: Communication-Related Outcomes and Feasibility
by Tim Frederik Peter Ritzen, Jill Goris, Lisa E. Ramaut, Darren I. Booi, René R. W. J. van der Hulst, Moustapha Hamdi, Nasser Nadjmi and Rutger M. Schols
Sensors 2026, 26(13), 4308; https://doi.org/10.3390/s26134308 - 7 Jul 2026
Viewed by 1154
Abstract
Treatment of cleft lip alveolus and/or palate includes surgical repair to treat communication-related outcomes such as velopharyngeal insufficiency and otological dysfunction. Robot-assisted surgery has recently evolved into a promising adjunct to conventional surgery, particularly for complex procedures such as transoral (reconstructive) surgery. This [...] Read more.
Treatment of cleft lip alveolus and/or palate includes surgical repair to treat communication-related outcomes such as velopharyngeal insufficiency and otological dysfunction. Robot-assisted surgery has recently evolved into a promising adjunct to conventional surgery, particularly for complex procedures such as transoral (reconstructive) surgery. This structured literature review aims to investigate whether robot-assisted transoral cleft palate repair enhances communication (i.e., speech and otological) outcomes compared to conventional manual cleft palate surgery. A literature search was performed using PubMed, Embase, the Cochrane Library and Google Scholar. Primary outcomes were the change in cleft speech characteristics and otological disease after robot-assisted cleft palate surgery versus manual cleft palate surgery. The available evidence on communication-related outcomes remains sparse. Six relevant articles were included. In only one study, transoral robotic cleft surgery (TORCS) significantly reduced otitis media with effusion (OME), need for ventilation tubes and hearing threshold within 2 years post-surgery. No postoperative speech or velopharyngeal outcomes were reported. In conclusion, transoral robotic cleft surgery (TORCS) appears safe and feasible for repair of a cleft palate. It provides superior intraoral view and improved surgeon ergonomics. Current drawbacks are the costs and the available tools, the extended surgical duration and the lack of haptic feedback, which limit the current clinical applicability of TORCS. Based on limited clinical evidence, TORCS may support faster recovery of Eustachian tube function and hearing, but no conclusions on speech outcomes can yet be drawn. Full article
(This article belongs to the Special Issue Feature Review Papers in Sensors and Robotics)
►▼ Show Figures

Figure 1

18 pages, 16960 KB  
Article
Application of Stationary IoT Sensor Networks to Reduce Exposure to Particulate Matter During Local Hotspot Events in a Megacity: A Fireworks Case Study
by Sangwon Bang, Kyung Hwan Kim, Kwang-Rae Kim, Hyun-Ju Ha, Jong-Cheol Yoon, Heejung Jung, Meehye Lee and Seung-Bok Lee
Sensors 2026, 26(13), 4307; https://doi.org/10.3390/s26134307 - 7 Jul 2026
Cited by 1 | Viewed by 596
Abstract
Growing public concern over the health impacts of particulate matter (PM2.5 and PM10) has highlighted the need for air quality monitoring networks that can raise an alert when high PM2.5 and PM10 concentrations are detected. However, existing networks [...] Read more.
Growing public concern over the health impacts of particulate matter (PM2.5 and PM10) has highlighted the need for air quality monitoring networks that can raise an alert when high PM2.5 and PM10 concentrations are detected. However, existing networks often have insufficient spatiotemporal resolution to detect localized events with high PM2.5 and PM10 concentrations. In this study, the existing air quality monitoring network in Seoul was complemented by Smart Seoul Data of Things (S-Dot), which is an Internet of Things (IoT)-based sensor network that collects various types of urban data. S-Dot sensors were used to track the spatial distribution and temporal evolution of high-concentration plumes generated by a fireworks festival, and the PM2.5 and PM10 concentrations measured by the S-Dot sensors were corrected with data measured by precision instruments at the closest air quality monitoring station. While the existing high-concentration alert system is designed to issue warnings when averaged PM2.5 concentrations measured at twenty-five air quality monitoring stations across the entire urban area exceed a threshold for a specified duration, the high spatial resolution of S-Dot was leveraged to provide localized alerts in near real time on plume location and movement. A new alert protocol is suggested to help reduce public exposure to pollutants, mitigate associated health risks, and encourage behavioral changes to improve air quality. Full article
(This article belongs to the Section Sensor Networks)
►▼ Show Figures

Figure 1

19 pages, 4219 KB  
Article
A Sensor to Analyze Fish Freshness: A Virtual Sensor Array Based on an Electrochemical Chemotransistor
by Yulia Efremenko, Eya Boughanmi and Vladimir M. Mirsky
Sensors 2026, 26(13), 4306; https://doi.org/10.3390/s26134306 - 7 Jul 2026
Viewed by 595
Abstract
The introduction of a quantitative definition of fish freshness enables the determination of the remaining storage time of raw fish materials. To measure this value, a virtual array of electrochemical chemotransistor-based chemical sensors was developed. The electrolyte used to electrically connect the four [...] Read more.
The introduction of a quantitative definition of fish freshness enables the determination of the remaining storage time of raw fish materials. To measure this value, a virtual array of electrochemical chemotransistor-based chemical sensors was developed. The electrolyte used to electrically connect the four measurement electrodes and the reference electrode was optimized. To achieve the high stability, high electrochemical activity of chemosensitive material, and reversible potential of the silver/silver chloride reference electrode, a chloride-containing ionic liquid and polymeric acid mixture was used as the electrolyte. Polyaniline in different redox states was applied as the chemosensitive material with electrically controlled affinity. First, the sensor was evaluated for trimethylammonium detection, and then it was applied to fish samples. Unlike the response observed for trimethylammonium, the sensor’s response to fish samples exhibited complex, non-exponential kinetics and a non-monotonic dependence on the storage duration of fish samples. To characterize these responses, a set of descriptors was introduced. The storage time was estimated by minimizing the Euclidean distance between the descriptors values obtained from fish samples and those determined during calibration. Based on the quantitative definition of freshness, this approach categorizes the current stage of fish products and predicts the remaining storage duration quantitatively. Full article
(This article belongs to the Special Issue Feature Papers in Biosensors Section 2026)
►▼ Show Figures

Figure 1

17 pages, 12133 KB  
Article
Simple, Fast and Sensitive Voltammetric Procedure for Copper Ion Determination Using a Solid Gold Microelectrode Array
by Mieczyslaw Korolczuk, Mateusz Ochab and Iwona Gęca
Sensors 2026, 26(13), 4305; https://doi.org/10.3390/s26134305 - 7 Jul 2026
Viewed by 584
Abstract
The present study reports the application of a gold microelectrode array to determine copper(II) ions by anodic stripping voltammetry (ASV). The microelectrode characterization of the presented working electrode was investigated. Moreover, the way of its preparation ensures its reusability and eco-friendly character, thanks [...] Read more.
The present study reports the application of a gold microelectrode array to determine copper(II) ions by anodic stripping voltammetry (ASV). The microelectrode characterization of the presented working electrode was investigated. Moreover, the way of its preparation ensures its reusability and eco-friendly character, thanks to the use of environmentally benign electrode material. The procedure does not require modification of the surface of the working electrode. Main experimental parameters were optimized, including pH and a concentration of the supporting electrolyte, activation and deposition conditions, and square wave parameters. The calibration graph was linear in the range of Cu(II) concentrations from 2 × 10−9 to 2 × 10−7 mol L−1 (with a deposition time of 30 s) and from 5 × 10−10 to 5 × 10−8 mol L−1 (with a deposition time of 90 s; RSD was 4.7% (n = 3) for a 1 × 10−8 mol L−1 of Cu(II)). The limit of detection was equal to 1.93 × 10−10 mol L−1 (tacc = 90 s). The correctness of the developed procedure was successfully checked by analysis of certified reference material and a tap water sample, confirming the possibility of its practical application. Satisfactory recovery values were also obtained during the analysis of an environmental water sample. Full article
(This article belongs to the Section Environmental Sensing)
►▼ Show Figures

Figure 1

36 pages, 6277 KB  
Review
A Survey on Security Threats and Mitigation Mechanisms for Smart Hospitals in the 6G Era
by Orestis Maraziotis, Georgios Mantas, Jonathan Rodriguez and Felipe Gil-Castiñeira
Sensors 2026, 26(13), 4304; https://doi.org/10.3390/s26134304 - 7 Jul 2026
Viewed by 653
Abstract
Smart Hospitals integrated within 6G edge networks aim to enhance hospital connectivity and operational efficiency by enabling intelligent and personalized e-health services and applications while optimizing resource utilization and maintaining a high degree of autonomy. Nevertheless, the interconnectivity and 6G integration, which comprise [...] Read more.
Smart Hospitals integrated within 6G edge networks aim to enhance hospital connectivity and operational efficiency by enabling intelligent and personalized e-health services and applications while optimizing resource utilization and maintaining a high degree of autonomy. Nevertheless, the interconnectivity and 6G integration, which comprise core components of Smart Hospitals, are susceptible to a wide range of security threats, posing significant risks to the confidentiality, integrity, and availability of hospital data and operations. Given that security is a critical concern for Smart Hospitals, there is an urgent need to develop novel security mechanisms to safeguard these environments within 6G edge networks. In particular, this work highlights how defining 6G characteristics, such as Ultra-Reliable Low-Latency Communications, massive IoMT connectivity, distributed edge intelligence, and AI-native network operation, not only enable next-generation hospital services but also reshape the security and privacy threat landscape and the requirements of mitigation mechanisms. In this context, the first essential step is to comprehensively understand both existing and emerging threats targeting Smart Hospitals in the 6G edge network ecosystem. Therefore, this article provides a categorization of security and privacy attacks based on their primary targets. Moreover, it presents a survey of mitigation techniques derived from recent literature, specifically designed to counter threats facing Smart Hospitals in 6G edge networks. The intent is to establish a foundation that supports ongoing research towards the development of effective, 6G-aware security countermeasures capable of protecting Smart Hospitals under the stringent latency, scalability, and reliability requirements of future healthcare environments. Full article
(This article belongs to the Section Internet of Things)
►▼ Show Figures

Graphical abstract

15 pages, 829 KB  
Article
One-Dimensional Position Detection Using a Cable Piezoelectric Sensor
by Yusuke Yamazoe, Kento Ise, Sayaka Kohno, Junei Kobayashi and Takashi Nakajima
Sensors 2026, 26(13), 4303; https://doi.org/10.3390/s26134303 - 7 Jul 2026
Viewed by 609
Abstract
This study established a sensor structure and readout circuit for identifying the point of force application within a single cable piezoelectric sensor. The sensor is highly flexible and enables wide-area sensing because it can be fabricated in long, continuous form. To determine the [...] Read more.
This study established a sensor structure and readout circuit for identifying the point of force application within a single cable piezoelectric sensor. The sensor is highly flexible and enables wide-area sensing because it can be fabricated in long, continuous form. To determine the point of force application, the central electrode of the cable sensor was designed to have high resistance, and charge amplifiers were connected to both ends of the cable. The generated charge was divided and measured by the two amplifiers in a proportion corresponding to the point of force application. The position was predicted using a normalized ratio calculated from the charge quantities measured by the two amplifiers. The principle was first verified by accurately identifying the loaded section in a five-segment cable connected through discrete resistors. For cable sensors with a high-resistance central electrode, we predicted the point of force application with a root mean square error of 25 mm by interpolating the relationship between the point of force application and the normalized charge ratio with a smoothing-spline calibration model. Full article
(This article belongs to the Section Electronic Sensors)
►▼ Show Figures

Graphical abstract

14 pages, 2703 KB  
Article
Decoding Multidimensional Machining Loads: iKIT Wireless Extrasensory Toolholder and Parametric Analysis in Aluminum Cutting
by Qian Qiao, Dawei Guo, Chi-Tat Kwok and Lap Mou Tam
Sensors 2026, 26(13), 4302; https://doi.org/10.3390/s26134302 - 7 Jul 2026
Viewed by 505
Abstract
Smart manufacturing requires real-time monitoring of multidimensional forces at the interface between the tool and workpiece in computer numerical control (CNC) machining. In this study, an innovative iKIT wireless extrasensory toolholder is introduced that is capable of high-fidelity, in situ, high-frequency sensing and [...] Read more.
Smart manufacturing requires real-time monitoring of multidimensional forces at the interface between the tool and workpiece in computer numerical control (CNC) machining. In this study, an innovative iKIT wireless extrasensory toolholder is introduced that is capable of high-fidelity, in situ, high-frequency sensing and monitoring of the cutting force, torque, and two-way bending moments. The hardware design of the system is outlined, highlighting a high-bandwidth miniature wireless transmission method and noncontact power supply and energy storage solution suitable for rotating machining environments. To assess the system performance, comprehensive milling tests were performed on aluminum alloy materials, and the relationship between the process parameters and changes in multidimensional mechanical loads was thoroughly examined. The experimental findings demonstrate that the smart toolholder detects precisely how parameter variations affect the loads. Multidimensional mechanical signals (torque and two-way bending moments) show a strong positive correlation with the feed rate and axial depth of cut, confirming the impact of the material removal rate on the system loads. Conversely, these signals are negatively correlated with spindle speed, accurately reflecting the effects of thermal softening and a reduced friction coefficient in aluminum alloys during high-speed cutting. This study not only offers a dependable hardware framework for integrating miniaturized sensors into toolholders, but also delivers accurate data to support digital twin models and adaptive control in machining processes. Full article
(This article belongs to the Special Issue AI-Enhanced Sensor Data Integration and Processing)
►▼ Show Figures

Figure 1

34 pages, 8018 KB  
Article
A Two-Stage GMFAMM Approximation for Joint Bias Correction of NASA POWER Hydroclimatic Data: The ColClim Web Application
by David Arango-Londoño, Delia Ortega-Lenis, Mauricio A. Mazo-Lopera and Paula Moraga
Sensors 2026, 26(13), 4301; https://doi.org/10.3390/s26134301 - 7 Jul 2026
Viewed by 431
Abstract
We propose and empirically evaluate a two-stage approximation to a Generalized Multivariate Functional Additive Mixed Model (GMFAMM) for the joint bias correction of five NASA POWER reanalysis variables: minimum and maximum temperature (Tmin, Tmax), relative humidity (RH), solar [...] Read more.
We propose and empirically evaluate a two-stage approximation to a Generalized Multivariate Functional Additive Mixed Model (GMFAMM) for the joint bias correction of five NASA POWER reanalysis variables: minimum and maximum temperature (Tmin, Tmax), relative humidity (RH), solar radiation (Rad), and precipitation occurrence (Pbin). Our primary contribution is the first operational-scale evaluation of such a framework (≈200,000 station–day observations, two orders of magnitude beyond previous studies) together with its deployment in an open-access web application. A systematic grid of more than 200 marginal configurations is evaluated on a strict chronological 70/30 hold-out (training 2016–2022; testing 2023–2025) to identify the optimal marginal specification per variable. Against a correctly specified marginal baseline, station-level linear calibration combined with the marginal GAMM removes the bulk of the systematic bias (RMSE reductions of ≈80%, 82% and 30% for Tmin, Tmax and RH). A shared latent step, using the first principal component of the marginal residual matrix as a scalar proxy for Λ0(t), yields additional but conditional out-of-sample reductions (≈17% Tmax, 10% RH, 9% Rad; negligible for Tmin, with precipitation occurrence retained in the shared representation but its joint gain treated as exploratory); because it requires co-located donor observations, at ungauged locations the deployed pipeline applies the marginal correction only, whose spatial transfer is confirmed by leave-one-station-out cross-validation. The residual cross-correlation structure is consistent with, though not in itself proof of, Clausius–Clapeyron coupling. The trained artefacts are deployed in ColClim, an open-access R Shiny application that queries the NASA POWER API and the Open-Meteo forecast service for any location in Colombia and delivers historical bias-corrected series and short-range (1–16 day) forecasts. Full article
(This article belongs to the Section Remote Sensors)
►▼ Show Figures

Figure 1

23 pages, 68438 KB  
Article
Infrared and Visible Image Fusion via Lightweight Semantic Prior Encoding and Cross-Attention Fusion
by Xun Zhang, Di Wu, Jianqi Li and Na Cui
Sensors 2026, 26(13), 4300; https://doi.org/10.3390/s26134300 - 6 Jul 2026
Viewed by 812
Abstract
Infrared (IR) and visible image fusion aims to synthesize a composite representation that integrates the thermal target saliency of IR imagery with the textural richness of visible imagery. Existing deep learning-based methods have achieved promising progress in this field. However, they either operate [...] Read more.
Infrared (IR) and visible image fusion aims to synthesize a composite representation that integrates the thermal target saliency of IR imagery with the textural richness of visible imagery. Existing deep learning-based methods have achieved promising progress in this field. However, they either operate at the pixel level without semantic priors, or rely on segmentation supervision to obtain such priors. Both approaches limit their practicality and performance in complex scenes. To design a lightweight fusion network that leverages semantic priors without segmentation supervision, we propose SPE2Fusion, a semantic prior-driven fusion network that operates through a dual-stage semantic injection paradigm. Specifically, a lightweight semantic encoder is designed to extract multi-scale scene priors in an end-to-end manner optimized solely by the fusion loss, without requiring segmentation mask annotations. Then, these priors are injected at two complementary stages: the Efficient Semantic Feature Awareness (ESFA) module applies spatially adaptive attention at the encoding stage to amplify semantically salient regions, while the Efficient Semantic Feature Embedding (ESFE) module applies semantically conditioned spatial normalization at the decoding stage to ensure coherent texture reconstruction. Finally, a bidirectional cross-attention fusion block is introduced to integrate complementary cross-modal features under this dual semantic guidance. The network is supervised by a multi-constraint loss combining gradient fidelity, intensity preservation, and structural similarity terms. Comprehensive experiments on the MSRS, LLVIP, and RoadScene benchmarks demonstrate that SPE2Fusion achieves state-of-the-art performance against representative methods (e.g., CrossFuse and DDBFusion), ranking first on four of six metrics on the MSRS test set, specifically EN (6.70), QAB/F (0.86), AG (6.06), and SD (43.44), while maintaining strong generalization on unseen datasets without domain adaptation. Full article
(This article belongs to the Section Sensing and Imaging)
►▼ Show Figures

Figure 1

31 pages, 14595 KB  
Article
A YOLOv8-Based Real-Time Road Congestion Decision-Making Approach Fused with Channel–Spatial Attention and Dynamic Weighted Loss
by Wei Huang, Heyang Xu, Hao Bai and Le Yu
Sensors 2026, 26(13), 4299; https://doi.org/10.3390/s26134299 - 6 Jul 2026
Viewed by 750
Abstract
Conventional object detection models suffer from significant performance degradation in dense urban traffic scenarios. To address these critical limitations and enable accurate real-time road congestion decision making, this study proposes an optimized YOLOv8-based detection paradigm that decouples multi-scale feature enhancement from dynamic focused [...] Read more.
Conventional object detection models suffer from significant performance degradation in dense urban traffic scenarios. To address these critical limitations and enable accurate real-time road congestion decision making, this study proposes an optimized YOLOv8-based detection paradigm that decouples multi-scale feature enhancement from dynamic focused bounding box regression. Specifically, a multi-scale feature enhancement (MFE) module is designed to extract high-resolution shallow features directly from the P2 layer of the YOLOv8 backbone. Then, a convolutional block attention module (CBAM) is embedded into the feature fusion neck to adaptively filter complex urban background noise and recalibrate channel–spatial feature responses for vehicle target saliency. Furthermore, the standard CIoU loss is replaced with the Wise-IoU (WIoU) dynamic focusing loss function, which suppresses gradient interference from low-quality, occluded samples and stabilizes bounding box regression for dense vehicle targets. The high-precision vehicle detection outputs are fed into a quantitative congestion index (CI) model, which fuses vehicle density and average speed to realize real-time congestion-level classification. Extensive experiments on the public UAVDT benchmark dataset demonstrate that the proposed model achieves an mAP@0.5 of 83.1% (3.8 percentage points higher than the YOLOv8 baseline), an mAP_S (small target) of 23.2% (a 4.3 percentage point improvement), and a real-time congestion decision accuracy of 83.8%. Ablation studies verify the independent and synergistic effectiveness of the MFE, CBAM, and WIoU modules, with the MFE module making the greatest contribution to small-target detection performance (+1.7% mAP@0.5). The proposed model maintains a real-time inference speed of 86 FPS (frames per second) on an NVIDIA RTX 3090 GPU, far exceeding the 30 FPS threshold for real-time traffic monitoring. Full article
(This article belongs to the Special Issue Multimodal Sensing Fusion-Based LLM Agent Methods, System, and Applications)
►▼ Show Figures

Figure 1

15 pages, 6647 KB  
Article
Adaptive Multi-Temporal Fusion and Cross-Modal Adversarial Alignment for Robust Driver Fatigue Detection
by Yanqiao Feng, Yong Peng and Dennis Z. Yu
Sensors 2026, 26(13), 4298; https://doi.org/10.3390/s26134298 - 6 Jul 2026
Viewed by 719
Abstract
To address the critical challenges of multi-scale temporal dynamics and sensor-intrusiveness in driver fatigue detection, this paper proposes the Multi-Temporal Fusion Attention Network (MTFA-Net). The framework integrates two core innovations: a Multi-scale Temporal Adaptive Fusion (MTAF) module that dynamically weights short-, mid-, and [...] Read more.
To address the critical challenges of multi-scale temporal dynamics and sensor-intrusiveness in driver fatigue detection, this paper proposes the Multi-Temporal Fusion Attention Network (MTFA-Net). The framework integrates two core innovations: a Multi-scale Temporal Adaptive Fusion (MTAF) module that dynamically weights short-, mid-, and long-term behavioral features via a scene-aware modulator, and a Physiological–Behavioral Cross-modal Adversarial Alignment (PBCAA) network that implicitly infers latent physiological states (e.g., HRV) from facial videos using adversarial learning and mutual information maximization. Experimental results on RLDD and NTHU-DDD datasets demonstrate that MTFA-Net achieves state-of-the-art accuracy (92.8%) while maintaining high interpretability and real-time efficiency, providing a robust, non-intrusive solution for intelligent cockpit safety. Full article
(This article belongs to the Special Issue AI and Smart Sensors for Intelligent Transportation Systems—2nd Edition)
►▼ Show Figures

Figure 1

29 pages, 795 KB  
Article
A Measurement-Supported Extrapolation Framework for Lowband MIMO Coverage and Capacity Enhancement in Future AAS-Assisted Wireless Networks
by Kornél Merkli, Szilvia Nagy and Péter Prukner
Sensors 2026, 26(13), 4297; https://doi.org/10.3390/s26134297 - 6 Jul 2026
Viewed by 481
Abstract
Low-frequency mobile bands remain essential for wide-area and penetration-limited wireless coverage, but their limited channel bandwidth constrains the achievable capacity. This paper presents a measurement-supported extrapolation framework for assessing how lowband MIMO and future AAS-assisted operation can enhance coverage and single-user throughput-oriented capacity [...] Read more.
Low-frequency mobile bands remain essential for wide-area and penetration-limited wireless coverage, but their limited channel bandwidth constrains the achievable capacity. This paper presents a measurement-supported extrapolation framework for assessing how lowband MIMO and future AAS-assisted operation can enhance coverage and single-user throughput-oriented capacity in wireless networks. The motivation is to evaluate whether such deployments can strengthen the lower-frequency layer as a robust coverage-and-capacity support layer for general traffic and reduce the load on midband and higher-frequency resources. Controlled radiated SISO and 2×2 MIMO measurements were performed with a base-station simulator and commercial user equipment in representative lowband and midband frequency bands. Measured RSRP, CQI, BLER, MAC-layer throughput, and IP-layer throughput thresholds for a 25 Mbit/s downlink target were used for coverage estimation and conditional extrapolation. Under the Extended Hata model, the measured 2×2 MIMO thresholds yielded a 43% larger estimated radius at 800 MHz than at 1800 MHz, while the same model indicated a 93% radius increase for a representative 10 dB AAS-related beamforming gain scenario. Conditional 4×4 MIMO extrapolations indicated data rates above 100 Mbit/s in 10 MHz and above 200 Mbit/s with 10 MHz two-component-carrier aggregation under ideal high-CQI conditions. The results support the potential of future lowband AAS deployments. The AAS and higher-order MIMO results are scenario-based estimates rather than direct field validation. Full article
(This article belongs to the Special Issue MIMO Antenna Design and Performance Enhancement for Wireless Applications)
►▼ Show Figures

Figure 1

24 pages, 12346 KB  
Article
Jamming Recognition Based on Adaptive Feature-Focusing Convolutional Neural Network for Agile Cognitive Radar
by Jialei Liu, Jiazhi Ma, Longfei Shi, Zhikang Lin, Yukai Kong and Junxian Chen
Sensors 2026, 26(13), 4296; https://doi.org/10.3390/s26134296 - 6 Jul 2026
Cited by 1 | Viewed by 506
Abstract
With the advancement of cognitive radar, applying deep neural networks to radar jamming recognition has become an indispensable research direction. However, as a common anti-jamming measure, the agility of radar waveform parameters degrades the effectiveness of jamming recognition, creating a trade-off between jamming [...] Read more.
With the advancement of cognitive radar, applying deep neural networks to radar jamming recognition has become an indispensable research direction. However, as a common anti-jamming measure, the agility of radar waveform parameters degrades the effectiveness of jamming recognition, creating a trade-off between jamming recognition and anti-jamming agility. Specifically, for the same type of jamming, radar agility in frequency, pulse width, and bandwidth alters the profile and scale features of the jamming, posing challenges to conventional CNN-based jamming recognition. To address this challenge, this paper proposes an Adaptive Feature-Focusing CNN (AFF-CNN). A pre-trained AFF module is designed to establish a mapping between agile parameters and adaptive feature scales. Operating on time-domain high-resolution range profiles (HRRP) and time–frequency domain short-time Fourier transform (STFT) data, this module calibrates deviations induced by radar inter-pulse parameter agility and enhances the capability of salient signal feature-focusing. Furthermore, a lightweight 1D-2D feature fusion CNN is designed to process these adaptive features and recognize jamming using single-pulse signals, thereby enhancing the network’s adaptability to inter-pulse parameter agility in radar systems. Simulation results demonstrate superior recognition accuracy and generalization capability compared to five comparative approaches, confirming effective adaptation to inter-pulse agility scenarios. Full article
(This article belongs to the Topic Radar Signal and Data Processing with Applications, 2nd Edition)
►▼ Show Figures

Figure 1

23 pages, 8813 KB  
Article
GNSS/eLoran Fusion-Based RAIM for Satellite-Deficient Environments
by Jingling Li and Huabing Wu
Sensors 2026, 26(13), 4295; https://doi.org/10.3390/s26134295 - 6 Jul 2026
Viewed by 476
Abstract
Global Navigation Satellite Systems (GNSS) provide essential positioning, navigation, and timing (PNT) services for a wide range of safety-critical applications. However, GNSS performance degrades significantly in satellite-deficient or interference-prone environments. To address this limitation, this study proposes a hybrid GNSS/eLoran integrity monitoring framework [...] Read more.
Global Navigation Satellite Systems (GNSS) provide essential positioning, navigation, and timing (PNT) services for a wide range of safety-critical applications. However, GNSS performance degrades significantly in satellite-deficient or interference-prone environments. To address this limitation, this study proposes a hybrid GNSS/eLoran integrity monitoring framework based on a simplified Receiver Autonomous Integrity Monitoring (RAIM) architecture. In the proposed method, GNSS observations from satellite constellations and range-equivalent measurements from the enhanced Loran (eLoran) terrestrial system are jointly processed using a weighted least-squares estimator. Integrity monitoring is performed through a global chi-square consistency test combined with a solution separation strategy for fault identification and exclusion. Horizontal Protection Level (HPL) is derived from the covariance of the estimation process to ensure bounded positioning error under nominal and fault conditions. Unlike conventional GNSS-only RAIM, the proposed framework enables improved redundancy and fault observability in satellite-deficient scenarios by incorporating heterogeneous terrestrial measurements. Simulation experiments consider satellite faults, eLoran measurement disturbances, and inter-system clock bias effects. Results demonstrate that the proposed method maintains reliable fault detection capability and ensures that positioning errors remain consistently bounded by the protection level under all tested scenarios. Full article
(This article belongs to the Section Navigation and Positioning)
►▼ Show Figures

Figure 1

Show export options expand_more Show export options expand_less
Select all
Export citation of selected articles as:
 
Displaying articles 1-356
Previous Issue
Next Issue
Back to TopTop