Search Results (217)

Drivetrain systems operate under varying operating conditions (OCs), which often obscure early-stage fault signatures and hinder robust condition monitoring (CM). This work introduces an AI digital platform developed during the EEDRIVEN project, featuring a holistic CM framework that integrates statistical time series methods—using Generalized AutoRegressive (GAR) models in a multiple model fault diagnosis scheme—with deep learning approaches, including autoencoders and convolutional neural networks, enhanced through a dedicated decision fusion methodology. The platform addresses all key CM tasks, including fault detection, fault type identification, fault severity characterization, and remaining useful life (RUL) estimation, which is performed using a dynamics-informed health indicator derived from GAR parameters and a simple linear Wiener process model. Training for the platform relies on a limited set of experimental vibration signals from the physical drivetrain, augmented with high-fidelity multibody dynamics simulations and surrogate-model realizations to ensure coverage of the full space of OCs and fault scenarios. Its performance is validated on hundreds of inspection experiments using confusion matrices, ROC curves, and metric-based plots, while the decision fusion scheme significantly strengthens diagnostic reliability across the CM stages. The results demonstrate near-perfect fault detection (99.8%), 97.8% accuracy in fault type identification, and over 96% in severity characterization. Moreover, the method yields reliable early-stage RUL estimates for the outer gear of the drivetrain, with normalized errors < 20% and consistently narrow confidence bounds, which confirms the platform’s robustness and practicality for real-world drivetrain systems monitoring. Full article

The expansion force generated by lithium-ion batteries during charge–discharge cycles is a key indicator of their structural safety and health. Recently, flexible pressure-sensing technologies have emerged as promising solutions for in situ swelling monitoring, owing to their high flexibility, sensitivity and integration capability. This review provides a systematic summary of progress in this field. Firstly, we discuss the mechanisms of battery swelling and the principles of conventional measurement methods. It then compares their accuracy, dynamic response and environmental adaptability. Subsequently, the main flexible pressure-sensing mechanisms are categorized, including piezoresistive, capacitive, piezoelectric and triboelectric types, and their material designs, structural configurations and sensing behaviors are discussed. Building on this, we examine integration strategies for flexible pressure sensors in battery systems. It covers surface-mounted and embedded approaches at the cell level, as well as array-based and distributed schemes at the module level. A comparative analysis highlights the differences in installation constraints and monitoring capabilities between these approaches. Additionally, this section also summarizes the characteristics of swelling signals and recent advances in data processing techniques, including AI-assisted feature extraction, fault detection and health state correlation. Despite their promise, challenges such as long-term material stability and signal interference remain. Future research is expected to focus on high-performance sensing materials, multimodal sensing fusion and intelligent data processing, with the aim of further advancing the integration of flexible sensing technologies into battery management systems and enhancing early warning and safety protection capabilities. Full article

In recent years, financial statement fraud schemes have evolved to become markedly more sophisticated and concealed, thereby posing severe threats to both social stability and economic health. Traditional detection methods, which rely primarily on fragmented corporate data, exhibit significant limitations in capturing the dynamic evolution and spatial diffusion characteristics of fraudulent behaviors over time and space. To address this issue, in this study, we undertake a thorough analysis of the intrinsic nature of fraud risk from a sociotechnical systems perspective and construct a multi-level indicator system to comprehensively quantify risk elements. Furthermore, recognizing the dynamic evolution nature and propagating characteristics of fraud risk, we propose a novel financial statement fraud detection framework to capture behavior patterns in temporal and spatial dimensions. Experiments on A-share-listed companies of high-risk industries in China demonstrate that the proposed framework significantly outperforms other mainstream machine learning and deep learning techniques. In addition, we open the “black box” of the detection framework and empirically validate fraud risk patterns with respect to social–technical elements by leveraging explainable AI techniques. Practically, the proposed framework and interpretable analysis are capable of providing precise early warnings and supervision. Full article

Arc fault detection is a key technology for preventing electrical fires. However, existing research has primarily focused on series connections, with insufficient attention paid to parallel load conditions, which are prevalent in real-world residential electricity usage. In accordance with the UL 1699 and IEC 62606 standards, this study established an experimental platform for arc faults, incorporating seven single loads (categorized into four types) and nine multi-load combinations. A systematic analysis of the differences in time–frequency characteristics under different connection modes was conducted. Time-domain and frequency-domain analyses revealed that under parallel connection the dispersion of arc fault time-domain characteristics decreases by more than 50% and the fundamental frequency component increases significantly. For parallel multi-load scenarios, the fundamental component of resistive combinations can reach 90%, while the frequency variance of inductive combinations can be as high as 400,000. By elucidating the time–frequency domain characteristics of parallel arc faults, this study proposes an optimized feature parameter analysis scheme for electrical fire monitoring systems. Based on this, this paper proposes an arc fault detection method using the Dual-Channel Convolutional Neural Network (DCNN). The method achieves 97.09% recognition accuracy for arc faults with different connection modes. Comparative experiments with other models and ablation studies show that the model attains 98.52% detection accuracy, verifying the effectiveness of the proposed method. This approach can significantly improve the accuracy of arc fault detection in multi-load environments, thereby enabling early warning of electrical circuit faults and potential fire hazards. Full article

Atrial fibrillation (AF) is the most common type of heart rhythm disorder worldwide. Early recognition of brief episodes of atrial fibrillation can provide important diagnostic information and lead to prompt treatment. AF is mainly characterized by an irregular heartbeat. Today, many personal devices such as smartphones, smartwatches, smart rings, or small wearable medical devices can detect heart rhythm. Sensors can acquire different types of heart-related signals and extract the sequence of inter-beat intervals, i.e., the instantaneous heart rate. Various algorithms, some of which are very complex and require significant computational resources, are used to recognize AF based on inter-beat intervals (RR). This study aims to verify the possibility of using neural networks algorithms directly on a microcontroller connected to sensors for AF detection. Sequences of 25, 50, and 100 RR were extracted from a public database of electrocardiographic signals with annotated episodes of atrial fibrillation. A custom 1D convolutional neural network (1D-CNN) was designed and then validated via a 5-fold subject-wise split cross-validation scheme. In each fold, the model was tested on a set of 3 randomly selected subjects, which had not previously been used for training, to ensure a subject-independent evaluation of model performance. Across all folds, all models achieved high and stable performance, with test accuracies of 0.963 ± 0.031, 0.976 ± 0.022, and 0.980 ± 0.023, respectively, for models using 25 RR, 50 RR, and 100 RR sequences. Precision, recall, F1-score, and AUC-ROC exhibited similarly high performance, confirming robust generalization across unseen subjects. Performance systematically improved with longer RR windows, indicating that richer temporal context enhances discrimination of AF rhythm irregularities. A complete Edge AI prototype integrating a low-power ECG analog front-end, an ARM Cortex M7 microcontroller and an IoT transmitting module was utilized for realistic tests. Inferencing time, peak RAM usage, flash usage and current absorption were measured. The results obtained show the possibility of using neural network algorithms directly on microcontrollers for real-time AF recognition with very low power consumption. The prototype is also capable of sending the suspicious ECG trace to the cloud for final validation by a physician. The proposed methodology can be used for personal screening not only with ECG signals but with any other signal that reproduces the sequence of heartbeats (e.g., photoplethysmographic, pulse oximetric, pressure, accelerometric, etc.). Full article

Gas disasters in coal mines are the principal constraint on safe operations; accordingly, accurate gas time-series forecasting and real-time fluctuation monitoring are essential for prevention and early warning. A method termed Decomposition-Enhanced Cross-Graph Forecasting and Anomaly Finding is proposed. The Multi-Variate Variational Mode Decomposition (MVMD) algorithm is refined by integrating wavelet denoising with an Entropy Weight Method (EWM) multi-index scheme (seven indicators, including SNR and PSNR; weight-solver error ≤ 0.001, defined as the maximum absolute change between successive weight vectors in the entropy-weight iteration). Through this optimisation, the decomposition parameters are selected as K = 4 (modes) and $\alpha$ = 1000, yielding effective noise reduction on 83,970 multi-channel records from longwall faces; after joint denoising, SSIM reaches 0.9849, representing an improvement of 0.5%–18.7% over standalone wavelet denoising. An interpretable Cross Interaction Refinement Graph Neural Network (CrossGNN) is then constructed. Shapley analysis is employed to quantify feature contributions; the m1t2 gas component attains a SHAP value of 0.025, which is 5.8× that of the wind-speed sensor. For multi-timestep prediction (T0–T2), the model achieves MAE = 0.008705754 and MSE = 0.000242083, which are 8.7% and 12.7% lower, respectively, than those of STGNN and MTGNN. For fluctuation detection, Pruned Exact Linear Time (PELT) with minimum segment length L_min = 58 is combined with a circular block bootstrap test to identify sudden-growth and high-fluctuation segments while controlling FDR = 0.10. Hasse diagrams are further used to elucidate dominance relations among components (e.g., m3t3, the third decomposed component of the T2 gas sensor). Field data analyses substantiate the effectiveness of the approach and provide technical guidance for the intellectualisation of coal-mine safety management. Full article

In the wild, snow is not merely additive noise; it is a non-stationary, semi-transparent veil whose spatial statistics vary with depth, illumination, and wind. Because conventional two-stage pipelines first detect a binary mask and then inpaint the occluded regions, any early mis-classification is irreversibly baked into the final result, leading to over-smoothed textures or ghosting artifacts. We propose BAT-Net, a Bidirectional Attention Transformer Network that frames desnowing as a coupled representation learning problem, jointly disentangling snow appearance and scene radiance in a single forward pass. Our core contributions are as follows: (1) A novel dual-decoder architecture where a background decoder and a snow decoder are coupled via a Bidirectional Attention Module (BAM). The BAM implements a continuous predict–verify–correct mechanism, allowing the background branch to dynamically accept, reject, or refine the snow branch’s occlusion hypotheses, dramatically reducing error accumulation. (2) A lightweight yet effective multi-scale feature fusion scheme comprising a Scale Conversion Module (SCM) and a Feature Aggregation Module (FAM), enabling the model to handle the large scale variance among snowflakes without a prohibitive computational cost. (3) The introduction of the FallingSnow dataset, curated to eliminate the label noise caused by irremovable ground snow in existing benchmarks, providing a cleaner benchmark for evaluating dynamic snow removal. Extensive experiments on synthetic and real-world datasets demonstrate that BAT-Net sets a new state of the art. It achieves a PSNR of 35.78 dB on the CSD dataset, outperforming the best prior model by 1.37 dB, and also achieves top results on SRRS (32.13 dB) and Snow100K (34.62 dB) datasets. The proposed method has significant practical applications in autonomous driving and surveillance systems, where accurate snow removal is crucial for maintaining visual clarity. Full article

Drowsy driving is a major cause of traffic accidents worldwide, and its early detection remains essential for road safety. Conventional driver monitoring systems (DMS) primarily rely on behavioral indicators such as eye closure, gaze, or head pose, which typically appear only after a significant decline in alertness. This study explores the potential of facial near-infrared (NIR) imaging as a hypothetical physiological indicator of drowsiness. Because NIR light penetrates more deeply into biological tissue than visible light, it may capture subtle variations in blood flow and oxygenation near superficial vessels. Based on this hypothesis, we conducted a pilot feasibility study involving young adult participants to investigate whether drowsiness levels could be estimated from single-frame NIR facial images acquired at 940 nm—a wavelength already used in commercial DMS and suitable for both physiological sensitivity and practical feasibility. A convolutional neural network (CNN) was trained to classify multiple levels of drowsiness, and Gradient-weighted Class Activation Mapping (Grad-CAM) was applied to interpret the discriminative regions. The results showed that classification based on 940 nm NIR images is feasible, achieving an optimal accuracy of approximately 90% under the binary classification scheme (Pattern A). Grad-CAM revealed that regions around the nasal dorsum contributed to this, consistent with known physiological signs of drowsiness. These findings support the feasibility of NIR-based drowsiness classification in young drivers and provide a foundation for future studies with larger and more diverse populations. Full article

Recent centimeter-to-millimeter monitoring of nearby gamma-ray bursts (GRBs) has revealed late-time ($\gtrsim {10}^2$–${10}^4$ days) radio rebrightenings and spectral turnovers not explained by standard forward-shock scenarios with steady microphysics. We attribute these features to a buried millisecond magnetar whose surface dipole, initially submerged by early fallback (hours after birth), re-emerges via Hall–Ohmic diffusion on year–to–decade timescales, partially re-energizing the external shock. We combine a minimally parametric analytic framework with axisymmetric magnetohydrodynamic simulations of the hypercritical fallback phase to characterize burial depths and the initial conditions for reemergence. The growth of the external dipole is modeled as $\dot{E}\left(t\right)\propto {\dot{E}}_0{f}_G{\left(t\right)}^{\sigma }$ and calibrated against physically plausible diffusion timescales ${\tau }_m\sim \mathrm{years}$–$\mathrm{decades}$. Spin-down power couples to the afterglow through the surrounding ejecta via a single effective coupling factor and a causal delay kernel, encapsulating mediation by supernova ejecta/pulsar-wind nebulae in collapsars and by merger ejecta/winds in compact-object mergers. Applied to a representative set of events with late-time radio detections and upper limits, our scheme reproduces the observed rebrightenings and turnovers with modest coupling efficiencies. Within this picture, late-time centimeter–millimeter afterglows provide a practical diagnostic of magnetic-burial depth and crustal conductivity in newborn magnetars powering GRB afterglows, and motivate systematic radio follow-up hundreds to thousands of days after the trigger. Full article

Depression is among the most prevalent mental disorders, and its early detection is essential to improving therapeutic outcomes in psychotherapy. This systematic review and meta-analysis evaluated the accuracy, interpretability, and generalizability of supervised algorithms (SVM, Random Forest, XGBoost, and GCN) for clinical detection of depression using real-world data. Following PRISMA guidelines, 20 studies published between 2014 and 2025 were analyzed across major scientific databases. Extracted metrics included F1-Score, AUC-ROC, interpretability methods (SHAP/LIME), and cross-validation strategies, with statistical analyses using ANOVA and Pearson correlations. Results showed that XGBoost achieved the best average performance (F1-Score: 0.86; AUC-ROC: 0.84), although differences across algorithms were not statistically significant (p > 0.05), challenging claims of algorithmic superiority. SHAP was the predominant interpretability approach (70% of studies). Studies implementing combined SHAP+LIME showed higher F1-Score values (F(1,7) = 8.71, p = 0.021), although this association likely reflects greater overall methodological rigor rather than a direct causal effect of interpretability on predictive performance. Clinical surveys and electronic health records demonstrated the most stable predictive outputs across validation schemes, whereas neurophysiological data achieved the highest point estimates but with limited sample representation. F1-Score strongly correlated with AUC-ROC (r = 0.950, p < 0.001). Considerable heterogeneity was observed for both metrics (I² = 74.37% for F1; I² = 71.49% for AUC), and Egger’s test indicated a publication bias for AUC (p = 0.0048). Overall, findings suggest that algorithmic performance depends more on data quality, context, and interpretability than on the choice of model, with explainable approaches offering practical value for personalized and collaborative clinical decision-making. Full article

We propose a GPU-parallelized collision-detection and response framework for rigid-body dynamics, designed to efficiently handle densely populated 3D simulations in real time. The method combines explicit Euler time integration with a hierarchical Octree–AABB collision-detection scheme, enabling early pruning and localized refinement of contact checks. To resolve collisions, we employ a two-step response algorithm that integrates non-penetration correction and impulse-based velocity updates, stabilized through smoothing, clamping, and bias mechanisms. The framework is fully implemented within Unity3D using compute shaders and optimized GPU kernels. Experiments across multiple mesh models and increasing object counts demonstrate that the proposed hierarchical configuration significantly improves scalability and frame stability compared to conventional flat AABB methods. In particular, a two-level hierarchy achieves the best trade-off between spatial resolution and computational cost, maintaining interactive frame rates (≥30 fps) under high-density scenarios. These results suggest the practical applicability of our method to real-time simulation systems involving complex collision dynamics. Full article

This paper proposes two process monitoring schemes, namely the Shiryaev–Roberts (SR) procedure and the cumulative sum (CUSUM) procedure, to detect shifts in the shape and scale parameters of Type I right-censored Gamma-distributed lifetime data. The performance of the proposed schemes is compared with that of an exponentially weighted moving average (EWMA) control chart based on deep learning networks. The performance of the proposed schemes is evaluated under various censoring rates using Monte Carlo simulations, with the average run length (ARL) as the primary metric. Furthermore, the SR and CUSUM schemes are compared for both zero-state and steady-state shifts. Simulation results indicate that the SR and CUSUM procedures exhibit superior performance, with the SR scheme showing particular advantages when the actual shift is small, while the CUSUM chart proves more effective for identifying larger shifts. The shape parameter has a significant effect on the performance of the control charts such that a reduction in the shape parameter effectively improves the ability to capture early offsets. Increased censoring rates reduce detection sensitivity. To maintain ${\mathrm{ARL}}_0$= 370, control limits h adapt differentially. The SR and CUSUM charts with different censoring rates need to recalibrate the parameter to mitigate performance losses under higher censoring conditions. The monitoring performance of the SR and CUSUM chart is enhanced by an increase in sample size. Finally, a practical example is provided to illustrate the application of the proposed monitoring schemes. Full article

This paper presents a compact dynamic optical frequency domain reflectometry (D-OFDR) platform enabling millimeter-scale, distributed strain sensing for real-time structural health monitoring (SHM) of reinforced concrete subjected to seismic loading. The proposed D-OFDR interrogator employs a dual-interferometer architecture: a main interferometer for strain sensing and an auxiliary interferometer for nonlinear frequency sweep compensation. Both signals are detected by photodetectors and digitized via a dual-channel FPGA-based DAQ board, enabling high-speed embedded signal processing. A dual-edge triggering scheme exploits both the up-chirp and down-chirp of a 50 Hz bidirectional sweep to achieve a 100 Hz interrogation rate without increasing the sweep speed. Laboratory validation tests on stainless steel cantilever beams showed sub-hertz frequency fidelity (an error of 0.09 Hz) relative to conventional strain gauges. Shake-table tests on a 2 m RC column under incremental seismic excitations (scaled 10–130%, peak acceleration 0.864 g) revealed distinct damage regimes. Distributed strain data and frequency-domain analysis revealed a clear frequency reduction from approximately 3.82 Hz to 1.48 Hz, signifying progressive stiffness degradation and structural yielding prior to visible cracking. These findings demonstrate that the bidirectional sweep-triggered D-OFDR method offers enhanced real-time monitoring capabilities, substantially outperforming traditional point sensors in the early and precise detection of seismic-induced structural damage. Full article

This paper proposes a hybrid predictive maintenance framework that combines the discriminative power of XGBoost with the interpretability of a Bayesian Network automatically learned from sensor data. Targeted at textile manufacturing equipment operating under Industry 4.0 conditions, the system addresses the trade-off between early fault detection and decision transparency. Sensor data, including vibration, temperature, and electric current, were collected from a multi-needle quilting machine using a custom IoT-based platform. A degradation-aware labeling scheme was implemented using historical maintenance logs to assign semantic labels to sensor readings. A Bayesian Network structure was learned from this data via a Hill Climbing algorithm optimized with the Bayesian Information Criterion, capturing interpretable causal dependencies. In parallel, an XGBoost model was trained to improve classification accuracy for incipient faults. Experimental results demonstrate that XGBoost achieved an F1-score of 0.967 on the high-degradation class, outperforming the Bayesian model in raw accuracy. However, the Bayesian Network provided transparent probabilistic reasoning and root cause explanation capabilities—essential for operator trust and human-in-the-loop diagnostics. The integration of both models yields a robust and interpretable solution for predictive maintenance, enabling early alerts, visual diagnostics, and scalable deployment. The proposed architecture is validated in a real production line and demonstrates the practical value of hybrid AI systems in bridging performance and interpretability for predictive maintenance in Industry 4.0 environments. Full article

Following the COVID-19 pandemic, public interest in healthcare has significantly in-creased, emphasizing the importance of early disease detection through lung sound analysis. Lung sounds serve as a critical biomarker in the diagnosis of pulmonary diseases, and numerous deep learning-based approaches have been actively explored for this purpose. Existing lung sound classification models have demonstrated high accuracy, benefiting from recent advances in artificial intelligence (AI) technologies. However, these models often rely on transmitting data to computationally intensive servers for processing, introducing potential security risks due to the transfer of sensitive medical information over networks. To mitigate these concerns, on-device AI has garnered growing attention as a promising solution for protecting healthcare data. On-device AI enables local data processing and inference directly on the device, thereby enhancing data security compared to server-based schemes. Despite these advantages, on-device AI is inherently limited by computational constraints, while conventional models typically require substantial processing power to maintain high performance. In this study, we propose a lightweight lung sound classification model designed specifically for on-device environments. The proposed scheme extracts audio features using Mel spectrograms, chromagrams, and Mel-Frequency Cepstral Coefficients (MFCC), which are converted into image representations and stacked to form the model input. The lightweight model performs convolution operations tailored to both temporal and frequency–domain characteristics of lung sounds. Comparative experimental results demonstrate that the proposed model achieves superior inference performance while maintaining a significantly smaller model size than conventional classification schemes, making it well-suited for deployment on resource-constrained devices. Full article