Search Results (5,416)

Induction motors play a vital role in industrial operations; however, stator inter-turn short-circuits faults remain a common and critical source of failure. This paper presents a machine learning-based diagnostic approach for detecting stator inter-turn short-circuit faults in three-phase induction motors operating at 50 Hz. Secondary data from a controlled test bench were processed using Power Spectral Density to determine energy distribution and guide the design of a Butterworth bandpass filter (20–350 Hz). The filtered signals were then analyzed using the Hilbert Transform to extract statistical features, which were ranked using the Minimum Redundancy Maximum Relevance algorithm to identify the most discriminative parameters. Two supervised classifiers, Support Vector Machine and Random Forest, were developed and validated using MATLAB’s Classification learner app with 5-fold cross-validation. The Support Vector Machine achieved an accuracy of 94.19%, while the Random Forest model achieved 99.51% with macro and F1-scores of 0.9951 and near-perfect area under the curve values. The results confirm that the Random Forest classifier provides superior generalization, sensitivity, and robustness in fault detection compared to Support Vector Machine. This study successfully demonstrates that combining Power Spectral Density, Hilbert Transform, Minimum Redundancy Maximum Relevance, and ensemble learning yields a highly effective framework for predictive maintenance and reliable fault diagnosis in industrial motor applications. Full article

Surface deformation triggered by coal extraction activities, together with the consequent development of unstable slopes within rugged mountainous landscapes, constitutes a critical focus for geological risk assessment and mitigation strategies. Conventional SBAS-InSAR processing pipelines suffer from inadequate tropospheric phase mitigation in topographically complex environments, while existing clustering-based recognition approaches fail to incorporate sufficient geophysical constraints. To overcome these deficiencies, the present investigation introduces a refined methodology that synergizes InSAR measurements with an enhanced clustering scheme for the automated screening of potentially unstable slope units. First, a two-stage coupled atmospheric correction framework is constructed within the SBAS-InSAR processing chain, comprising spatially varying stratified atmosphere estimation based on geographically weighted robust regression (GWRR-M) and turbulent atmosphere compensation based on structure-guided deformation-preserving interpolation (SGDPI); both stages require no external meteorological data and effectively protect deformation signals from overcorrection. Second, a spatiotemporally constrained density peak clustering algorithm (STC-DPC) is developed, which constructs a multi-dimensional feature space integrating spatial location, deformation rate, temporal evolution characteristics, and topographic-geological background, and introduces a spatiotemporally constrained distance metric together with an Unstable Slope Index (USI) to achieve automatic identification and quantitative discrimination of unstable slopes. The proposed method was evaluated using 120 ascending-track Sentinel-1A SAR images acquired from 2019 to 2023 over the coal-mining mountainous areas of Mentougou and Fangshan districts in western Beijing, China. The results show that the improved atmospheric correction reduces the phase standard deviation of a representative interferogram from 1.6 rad to 0.6 rad, with an average reduction of 42.3% across all interferograms. A total of 187 unstable slopes were identified by the STC-DPC algorithm, mainly distributed in abandoned mining areas and steep terrain with gradients of 10–35°, with a mean deformation rate of −25.3 mm/a; field investigations at representative sites confirmed significant deformation evidence (e.g., tension cracks and bulging), providing qualitative support for the identification results. Compared with the identification results obtained without atmospheric correction (79 unstable slopes), the improved method improves the detectability of weak deformation signals in areas with strong topographic relief and diverse deformation patterns. This study provides a practical technical pathway for the early screening and monitoring of geological hazards in coal-mining mountainous areas and holds great significance for mine ecological restoration and regional disaster prevention and mitigation. Full article

This work presents an incremental zoom analog-to-digital converter (ADC) readout integrated circuit (IC) and field-programmable gate array (FPGA)-based digital calibration system for high-precision capacitive micro electro-mechanical system (MEMS) pressure sensors, aimed at enhancing pressure readout resolution, pressure measurement accuracy, and long-term output stability. In the readout IC, the zoom ADC employs coarse-fine quantization, achieving high readout accuracy while relaxing requirements on integrator output swing and front-end linearity. On the digital side, a sparrow-search-algorithm-optimized Gaussian process regression (SSA-GPR) model constructs a nonlinear pressure–temperature mapping with limited calibration samples, while a segmented aging-compensation scheme based on dual-channel periodic self-test dynamically corrects aging-induced offset and sensitivity drifts. Together, these three techniques form a complete signal-conditioning chain addressing weak capacitance variation readout, pressure–temperature coupling, and long-term drift. Measured results show 0.11 Pa_RMS output root mean square (RMS) noise under a 205.2 ms conversion time, 37.44 Pa mean absolute error over −40 °C to 85 °C and 30 kPa to 120 kPa, and ±30 Pa residual error after 85 °C/1000 h aging, confirming the effectiveness of the proposed readout IC and digital calibration system. Full article

Conventional full-waveform centroid algorithms for pulsed LiDAR rely on rectangular integration and fixed windows, causing timing errors for asymmetric echo pulses and poor adaptability to varying pulse shapes. This paper proposes an adaptive window-width trapezoidal centroid algorithm (AWTWCA) that replaces rectangular with trapezoidal integration and dynamically determines the optimal window based on the rising-edge half-amplitude interval. Simulations under diverse echo conditions demonstrate that AWTWCA consistently achieves lower mean absolute error than rectangular and fixed-window trapezoidal methods. Experiments on a MEMS LiDAR system further validate its robustness: timing errors remain below 0.2 ns (3 cm ranging) and the detection success rate exceeds 99% under low-reflectivity and APD saturation conditions, significantly outperforming fixed-window algorithms. The algorithm simultaneously generates range and intensity point clouds, enabling clear discrimination of targets with different reflectivities at the same distance. With computational complexity virtually identical to conventional methods, AWTWCA offers a hardware-friendly solution for high-precision LiDAR signal processing in complex environments. Full article

Multiple-source DOA (Direction of Arrival) estimation is vital for array processing, radar, and integrated sensing and communications, yet classical subspace methods degrade under low signal-to-noise ratios and snapshot-starved conditions due to inaccurate sample covariance matrices. To address this, we propose a Toeplitz-enhanced neural network (TENN-DOA) for DOA estimation, a hybrid physics-informed framework for uniform linear arrays that combines the array signal processing prior with a lightweight learning-based regressor. The front end explicitly enforces the Hermitian–Toeplitz structure and fuses the projected matrix with the sample covariance via an analytically derived optimal shrinkage coefficient, yielding a robust covariance estimate. This enhanced representation is mapped onto an overcomplete angular dictionary, producing a feature sequence structurally coupled with the array manifold. A pooling-free one-dimensional convolutional neural network with decreasing kernel sizes starts with large kernels to capture the broad spectral envelope from grid mismatch, and then regresses to a pseudo spatial spectrum under multi-hot supervision for grid-point estimates. The Monte Carlo simulation results show that under the conditions of low signal-to-noise ratio, limited snapshots and the simulated ideal uniform linear array scenario, TENN-DOA achieves a higher resolution probability and lower root mean square error compared with MUSIC, TLS-ESPRIT and the deep learning-based baseline algorithm DA-MUSIC. Full article

To address the challenge of estimating the depth of a single-frequency (tonal) source using a vertical line array (VLA) without prior knowledge of seabed parameters, a method called OCMS-D (orthogonality-constrained modal search-based depth estimation) is presented. The method exploits the orthogonality of mode depth functions as a physical constraint and leverages the sparsity of propagating normal modes in the received field. A convex optimization model is formulated to jointly estimate modal wavenumbers, depth functions, and complex mode amplitudes. Furthermore, a depth-sign search (DSS) is introduced to compensate for mode phase signs, which effectively suppresses sidelobes in the depth ambiguity function and enables high-accuracy depth estimation. Numerical simulations analyze the influence of signal-to-noise ratio, array parameters and sound speed profile (SSP) uncertainty on the algorithm’s performance. Comparisons with matched-mode processing (MMP) show that OCMS-D exhibits superior robustness to SSP uncertainty and requires a smaller array aperture. Experimental data from the SWellEx-96 campaign demonstrate that the method achieves depth estimation errors of less than 5.4 m for the shallows and less than 10.8 m for the deep spurce. Notably, it requires neither seabed prior information nor source motion, offering a robust array signal processing framework for tonal source localization. Full article

Background/Objectives: Lateral flow assays (LFAs) are widely used point-of-care diagnostic devices due to their low cost, portability, and ease of use. However, most LFAs provide only qualitative results, limiting their utility for applications requiring continuous biomarker monitoring. This study introduces DeepBand, a deep learning-enabled multi-stage framework designed to automate the continuous quantification of analyte concentrations from unstandardized smartphone-captured lateral flow assay (LFA) images. Methods: A publicly available dataset containing 672 COVID-19 LFA images corresponding to four analyte concentrations (0.0, 1.8, 3.7, and 7.4 ng) was analyzed. A multi-stage pipeline was developed consisting of: (1) YOLOv11-based object detection to isolate the LFA cartridge from background artifacts, (2) a custom computer vision algorithm to identify and crop the test and control bands, and (3) a custom convolutional neural network (CNN) trained as a supervised regression model to predict continuous analyte concentrations. Data augmentation, hyperparameter optimization, and 5-fold cross-validation were used to improve model robustness. Results: The YOLOv11 model achieved approximately 99% mAP50 and 93.96% mAP95 for cartridge detection. Initial CNN models exhibited systematic underprediction of higher concentrations due to target imbalance; replacing mean squared error with Huber loss substantially improved performance, resulting in a final 20% held-out test set RMSE of 0.0292 ng. Analysis of HSV image channels demonstrated that the saturation-channel test-to-control intensity ratio was strongly correlated with analyte concentration (r = 0.94), consistent with the Beer–Lambert law governing LFA signal formation. Channel ablation studies confirmed the saturation channel as the most informative feature, while saliency mapping showed that the model primarily focused on biologically relevant test and control line regions. Conclusions: The proposed deep learning-enabled workflow, DeepBand, successfully integrates object detection, image processing, and CNN-based regression to provide automated quantitative interpretation of LFA results from smartphone images. Furthermore, the observed agreement between model behavior and Beer–Lambert theory suggests that the network learns biologically meaningful signal characteristics, supporting its potential for quantitative point-of-care diagnostics and longitudinal disease monitoring. Full article

Accurate interwell connectivity characterization is critical for fine-grained waterflood optimization and reservoir management, often requiring integrated analysis across multiple disciplines and methods. Among these, injection–production response analysis stands as the most cost-effective and widely adopted approach. However, it remains highly subjective, heavily reliant on senior engineers’ decades of accumulated experience, and prohibitively labor-intensive for large-scale oilfields with hundreds of wells. With the exponential growth of production data in modern oilfields, manual analysis has become the bottleneck restricting the timeliness of reservoir management decisions. While signal processing techniques offer a promising path to automation, general-purpose algorithms fail to incorporate fundamental reservoir fluid flow laws, resulting in insufficient accuracy for practical engineering applications. To address this gap, we propose a novel connectivity identification method that mimics expert analysis logic by focusing on large-amplitude fluctuation segments rather than full-curve matching. Using curve slope as the core metric, cosine similarity quantifies trend consistency, while Root Mean Square Error (RMSE) measures amplitude proximity. Three targeted strategies enhance accuracy: key region screening with segmented matching, outlier removal accounting for time-varying lags, and multi-index weighted fusion. Validated on synthetic and mature carbonate waterflood field cases, the method improves the identification performance over benchmark Normalized Cross-Correlation (NCC) and Capacitance-Resistance Model (CRM) methods by more than 12% in both cases. It retains the reliability of traditional response analysis while achieving full automation, and can help estimate the timing of preferential flow path formation, requiring only routine production data to provide valuable reference for timely field development decision-making. Full article

This paper describes a polarimetric, frequency-modulated continuous wave (FMCW) synthetic aperture radar (SAR) with an ultrawide bandwidth that spans 2–18 GHz. It is being developed as an airborne sensor called SKuSAR. The intent is to generate a set of sub-band images, thereby creating a multispectral SAR. This opens the prospect for new remote sensing algorithms that exploit variations in the scene’s polarization/frequency response occurring over the system’s three octaves of bandwidth. We describe the SKuSAR hardware and the processing steps applied to the FMCW data to create a multispectral SAR. The approach is practically demonstrated using data collected against calibration targets deployed in a field with the system mounted on a truck. This ground-based arrangement provided an inexpensive solution to test and fine-tune the system hardware and processing algorithms. A few complicating factors specific to the ground-based geometry were encountered, such as multipath signal contamination due to ground reflection and rather short data collections that affected attainable azimuth resolutions. Both these factors were identified and analyzed. The measured characteristics follow theoretical predictions rather well, giving confidence that the system meets its expected nominal performance once airborne, with the mentioned limiting factors absent or of reduced significance. Full article

Background/Objectives: Continuous injection-pressure monitoring may provide quantitative insight into fascial plane hydrodissection, but comparative data remain limited. This study compared baseline-corrected mean pressure and linear pressure trends during PECS I block (interpectoral plane block) versus superficial parasternal intercostal plane (SPIP) block and rectus sheath block (RSB) versus transversus abdominis plane block (TAPB). Methods: Each ultrasound-guided technique was performed once in 10 fresh, unfrozen human cadavers. Saline was infused at 10 mL/min, and pressure was recorded continuously. A deterministic signal-processing algorithm estimated the onset of the post-puncture pressure plateau; the subsequent 40 s segment was analysed after subtraction of the pre-insertion baseline. Paired t-tests were used, with exact sign-flip permutation tests as sensitivity analyses and with Holm adjustment for multiplicity. Results: Mean pressure was 4.99 psi for PECS I and 5.09 psi for SPIP, with no significant difference between techniques (mean paired difference, 0.09 psi; 95% confidence interval [CI], −0.67 to 0.85; p = 0.790). TAPB showed a higher mean pressure than RSB (5.50 versus 4.47 psi, respectively; mean paired difference, 1.03 psi; 95% CI, 0.36 to 1.69; p = 0.0134; paired effect size, d = 1.11). Permutation analyses were concordant. Linear pressure trends did not differ for SPIP versus PECS I (p = 0.484) or TAPB versus RSB (p = 0.742). Conclusions: Continuous pressure monitoring was used to characterize the hydrodissection pressure of four fascial plane block techniques. PECS I and SPIP yielded closely aligned mean values, whereas TAPB was associated with higher pressure than RSB. These findings provide quantitative reference data and support further clinical evaluation of continuous injection-pressure monitoring. Full article

Background: Associative learning, the process of binding stimuli, responses, and outcomes, is critical for behavioral adaptation. While event-related potentials (ERPs) like the P300 effectively index the cognitive effort and context updating required during trial-and-error learning, the precise cortical generators of these signals and their shift across development remain difficult to isolate due to the electroencephalography (EEG) inverse problem. The aim of the study was to examine the differences in P300 localization depending on the associative learning task stage and participants’ age. Methods: A total of 148 participants (aged 7–21) completed a visual-motor associative learning task while undergoing 64-channel EEG recording. Source localization was performed over the 300–500 ms (P300) time window as a conjunction of results of four inversion algorithms at FWE-corrected p < 0.05: multiple sparse priors with greedy search (GS), independent and identically distributed (IID), low-resolution electromagnetic tomography (LOR), and empirical Bayes beamformer (EBB). Exploratory three-algorithm (IID, LOR, and EBB) conjunction analysis results were also reported. Results: A spatial convergence analysis of Early > Late-stage differential maps revealed a consensus across all four models for both cue Onset-locked (right inferior parietal lobule/angular gyrus) and Feedback-locked (occipital, temporal (including bilateral middle temporal gyrus), and orbitofrontal regions) activity. The exploratory three-algorithm analysis additionally implicated left middle temporal gyrus and angular gyrus for cue Onset-locked rule acquisition. No results were produced for age-related effect surviving either the four-algorithm or three-algorithm consensus criterion. Conclusions: This multi-algorithm consensus links active rule search to distributed cortical pattern, involving angular (for cue stimulus processing) and middle temporal gyri (for feedback processing), while no age-related changes in these areas have been demonstrated. Full article

Online social networks generate large volumes of textual data that reflect users’ opinions, affective expressions, and broader patterns of engagement and social behavior. However, natural language processing approaches frequently examine sentiment, trust-related signals, and behavioral indicators independently, limiting their ability to represent the multidimensional nature of online interaction. This study conducts a systematic comparative evaluation of lexicon-enhanced fine-grained sentiment classification using linguistic, message-level statistical, and lexicon-derived affective information across a common experimental framework. The empirical analysis combines TF–IDF features, word-count information, and sentiment indicators derived from TextBlob, SentiStrength, and VADER, while the broader multi-level organization is used to relate the resulting affective evidence to online social-behavior analysis. Fifteen classical machine learning algorithms and seven deep learning architectures are evaluated on a real-world Twitter dataset containing 41,157 COVID-19-related tweets labeled across five sentiment-intensity classes. The experimental evaluation considers four feature configurations and seven performance metrics, complemented by Friedman and post hoc Wilcoxon signed-rank tests. The results show that TextBlob provides modest improvements, SentiStrength produces broader and more consistent gains, and VADER yields the strongest overall performance. AdaBoost combined with VADER achieves the best results, with 93.16% accuracy, 93.20% macro F1, 93.27% balanced accuracy, and an MCC of 0.913, while the Dense Neural Network is the strongest deep learning model. These results demonstrate that lexicon-derived affective features can substantially strengthen fine-grained sentiment classification, although their effectiveness depends strongly on the learning algorithm used to exploit them. The empirical contribution of this study is confined to fine-grained sentiment classification, while trust-related and attachment-related dimensions are retained as higher-order interpretive constructs rather than directly predicted or empirically validated outcomes. Full article

The development of accurate, robust and adaptive methods for monitoring the optical parameters of fiber-optic sensors (FOS) is one of the priority tasks in the field of precision measurements, especially in the context of rapidly growing requirements for intelligent monitoring systems. This paper presents a comprehensive approach to the use of modern deep learning algorithms for analyzing and processing spectral data coming from FOS. The proposed solution is based on the use of convolutional neural networks (CNN) for the automatic extraction of informative features, as well as autoencoders for noise suppression and signal restoration. A hybrid architecture combining CNN and recurrent neural networks (RNN) was developed. The experiments conducted confirmed the effectiveness of the evaluated models. On the independent regression test set, the CNN-only model achieved a macro-averaged R²-based prediction score of 98.2% without added noise and 89.4% under high-noise conditions; on a separate temporal test sequence, the hybrid CNN + RNN model achieved 95.0% compared with 88.0% for CNN alone. The presented approach has high resistance to noise and the ability to scale to various types of FOS. At the conclusion, the prospects for the practical applications of the proposed system are discussed: the structural monitoring of buildings and structures and the automation of processes in industry and energy, with an emphasis on reliability, autonomy and integration with existing platforms. Full article

Water treatment plants are essential urban infrastructure with direct implications for public health and everyday life. Data-driven management has received growing attention in drinking water treatment, particularly for optimizing chemical dosing to improve operational efficiency, reduce costs, and ease operator workload. AI-based prediction of coagulant dosage has therefore become an active research topic. Existing studies, however, have focused mainly on model architecture, with less attention to data validity and cost control. In practice, many plants face data-quality problems, including inconsistent dosing records under similar water-quality conditions. Conventional data cleaning may also remove large portions of the dataset, which can weaken model reliability. This study proposes an artificial intelligence (AI) modeling framework for coagulation dosing that handles anomalous data, emphasizes data quality assurance, and combines cost-oriented feedforward prediction with feedback control. A genetic algorithm-optimized backpropagation (GA-BP) neural network was first evaluated on controlled laboratory data and full-scale plant data using the same core model architecture, allowing the effects of model configuration to be separated from those of data quality. Historical plant records were subsequently cleaned through expert-guided validation, approximate time-delay alignment, and turbidity-based classification of operating conditions. Settled-water turbidity was then used as a feedback signal to dynamically adjust subsequent coagulant dosage and assess the resulting chemical savings. Changes in the input structure produced only modest improvements in full-scale prediction performance (R² = 0.53–0.72). In contrast, data cleaning and process-based data organization markedly improved predictive performance, with R² values increasing to 0.927–0.969. Standalone AI models achieved only moderate dosage reductions, while their integration with real-time turbidity feedback provided the best cost-control performance. The model-based control strategy reduced average coagulant consumption by 10.37%, with a maximum reduction of 21.33% at a settled-water turbidity target of 1.9 nephelometric turbidity units (NTU). Across the evaluated feedback-control scenarios, manual dosing was up to 32.83% higher than the corresponding feedback-controlled dosage. Overall, AI models can fit coagulation-dosing data and predict coagulant dosage with sufficient accuracy, but data quality assurance remains the main factor determining model performance. Effective cost control also requires real-time turbidity-based feedback regulation rather than model outputs alone. Full article

Reinforced concrete structures are critical components of modern infrastructure, and the detection of internal damage within these structures has long been an important research topic. These damages can be detected using the vibration response of the structures, and laser Doppler vibrometry (LDV) on a moving platform offers an efficient, long-range sensing approach for structural vibration monitoring. However, extending LDV-based damage detection from static to mobile measurement requires addressing the effects of measurement signal frequency shift, platform vibrations, and speckle noise. To address these issues, this paper proposes a mobile measurement damage detection framework that integrates multi-source information with artificial intelligence algorithms. First, theoretical derivation and numerical simulation demonstrate that the vibration frequency shift induced by moving speed is negligible, proving that mobile and static measurement signals are similar in both time and frequency domains. Then, a multi-sensor data processing framework is used to decouple the platform vibration and suppress speckle noise. Finally, a spatial-aware CNN network is employed to achieve damage detection under mobile measurement. The results reveal that the vibration signals for large-scale voids were effectively recovered, whereas signals for small-scale voids and healthy regions were only partially recovered. Voids with a tested size of 0.4 m and larger were successfully identified under the experimental conditions. The results demonstrate the feasibility of extending static LDV-based void detection to mobile measurement, providing a theoretical and technical basis for efficient, non-contact mobile inspection of infrastructure. Full article