Search Results (43)

Progressive extrusion-screw wear alters screw–material interaction and may become observable through process signals acquired during operation. This study compared raw-signal and continuous wavelet transform (CWT)-based pipelines while explicitly controlling temporal information leakage. Synchronized 1 Hz measurements of three motor phase currents and four barrel temperatures were recorded during two soybean-extrusion campaigns representing independently verified new and worn screw states. Five pipelines were evaluated using identical leakage-controlled temporal folds: time-domain descriptors with RBF-SVM, handcrafted CWT descriptors with RBF-SVM, raw-current 1D-CNN, CWT-scalogram 2D-CNN, and CWT-CNN with thermal-feature fusion. The raw-current 1D-CNN achieved the highest mean balanced accuracy (0.700 ± 0.075), whereas the CWT-based 2D-CNN produced the highest mean ROC-AUC (0.839 ± 0.160); however, the paired M3–M2 ROC-AUC confidence interval included zero. Handcrafted CWT descriptors underperformed conventional time-domain descriptors, and thermal fusion did not improve temporal generalization. The results therefore do not establish an inherent advantage of CWT. Because M2 and M3 differ in both representation and network architecture, their contrast is interpreted at pipeline level rather than as an isolated CWT effect. The evidence is limited to within-campaign temporal discrimination of the two recorded screw conditions and does not establish transferable wear diagnostics across independent campaigns. Full article

Groundwater variability is recognized as a critical constraint on long-term water resource sustainability in South Korea under nonstationary climate forcing. This study developed a groundwater-level (GWL) projection framework based on CMIP6 simulations, validated against national monitoring data, screened for physical realism of underlying climate projections, and translated into station-scale climate risk metrics. GWL observations from 199 stations of the National Groundwater Monitoring Network (2009–2025) were related to CMIP6 precipitation and soil-moisture forcings using a hybrid deep learning architecture, the hybrid attention-based convolutional neural network–long short-term memory (HACL) model, combining multiscale temporal convolution, bidirectional memory, and self-attention. Of ten candidate GCMs, four (ACCESS, GISS, INM, and NorESM) met pre-defined performance criteria (R² > 0.70, NSE ≥ 0.70, KGE > 0.50), forming a filtered ensemble more internally consistent than the unfiltered multi-model mean. This ensemble projected national-average GWL increases of 16.60 m (SSP245), 15.92 m (SSP370), and 17.85 m (SSP585), the largest in the western alluvial lowlands—substantially exceeding historical observed rates and indicating sensitivity signals warranting further investigation rather than direct station-level forecasts. The results were incorporated into the Climate Groundwater Risk Index (CGRI) integrating the magnitude of change, ensemble spread, observed variability, and vulnerability, offering a replicable approach for monsoon-affected regions. Independent out-of-sample evaluation (2022–2025) confirmed robust generalization (R² = 0.965, NSE = 0.965, KGE = 0.974, RMSE = 14.26 m). Ablation benchmarking showed HACL substantially outperformed standalone BiLSTM (NSE = 0.861), 1D-CNN (NSE = 0.810), and linear regression (NSE = 0.628). Predictor sensitivity analysis revealed that year as a continuous covariate accounts for ~85% of the projected 16–18 m signal; constrained strictly to physical forcing, projected increases are +2.42 m (SSP245), +2.15 m (SSP370), and +3.78 m (SSP585) by 2081–2100, aligning with historical trends (~2–3 m) and restoring station-level hydrogeological sensitivity. Full article

The mooring system is the dominant single point of failure on a floating offshore wind turbine, and inspecting it drives much of its operations and maintenance cost. Data-driven damage classifiers could ease that burden, but they are usually trained and tested on one sea-state distribution, so accuracy drops for the severe conditions that lie outside it. This work presents a domain-adversarial gated recurrent unit (GRU) that identifies eleven mooring conditions from the platform’s six rigid-body motions and transfers, under a semi-supervised protocol, to sea states for which no damage labels are available in training. The conditions span a healthy baseline, single-line and multi-line stiffness loss, non-uniform degradation, and two biofouling severities; all are sub-failure damage cases whose motion signatures are weak and easily confused. The model is trained on operational sea states and tested on unseen extreme storm states (22–26 m/s wind, 8.0–9.5 m significant wave height), simulated in OpenFAST with MoorDyn on the IEA 15 MW UMaine VolturnUS-S platform. A three-phase schedule combining a labeled intermediate near-target domain with adversarial alignment of the training and test feature distributions raises per-simulation macro-F1 across three seeds to 0.87, against 0.29 for a source-only model, 0.47 for the near-target bridge alone and 0.73 for adversarial alignment alone. Neither component reaches this level by itself: alignment supplies the larger share of the gain and the bridge stabilizes it, reducing seed-to-seed variability roughly fourfold. The gain persists under sensor noise up to 20% of the per-channel standard deviation, and a parameter-matched 1D-CNN backbone confirms it is not tied to the recurrent architecture. The setting is semi-supervised, with respect to which target domain labels are available for the source and the intermediate near-target range, while the target sea states contribute unlabeled windows only. Because it requires no labels at the target sea states and only the standard six-degree-of-freedom motion sensors, the method can be extended to the severe, unseen simulated sea states that an asset meets in service; labeled near-target simulations from a calibrated platform model are still required. The present study is a simulation-based feasibility study: all evidence derives from OpenFAST and MoorDyn simulations of a single platform, and validation against tank-test or field measurements remains to be carried out. Full article

Accurate sea state estimation is of great significance to marine engineering safety and disaster prevention and mitigation. This paper proposes a new sea state estimation architecture integrating multi-scale 1D-CNN, transformer encoder, and the MC-DropKey mechanism. First, the front-end multi-scale 1D-CNN is used in sequence modeling to extract high-quality local features. Then, the back-end transformer encoder is used to capture global long-range temporal dependencies. In addition, the MC-DropKey mechanism is introduced. It achieves dynamic uncertainty quantification while maintaining high-precision estimation. Comprehensive comparative and ablation experiments show that the proposed model significantly outperforms the baseline models in wave height and wave direction estimation tasks, achieving the lowest MAE of 0.117 m and CAE of 5.212°. Moreover, the uncertainty intervals output by the model effectively quantify estimation confidence, reaching a 96.67% prediction interval coverage probability. The obtained results provide strong technical support for scientific scheduling and decision-making in complex marine environments. Full article

Forest disturbances significantly affect the terrestrial carbon cycle, yet high-resolution detection, driver attribution, and carbon loss quantification remain challenging in cloudy and complex terrains. Here, we investigated the Northeast China and Southwest Hengduan Mountains forest regions from 2021 to 2024. We developed a Bayesian Model Averaging (BMA) framework integrating multi-source remote sensing (Sentinel-1/2, Landsat 8/9) and multi-algorithm ensembles (LandTrendr, CCDC, 1D-CNN) to extract 10 m disturbance features. Automated driver attribution and carbon loss quantification were achieved utilizing the Fire Information for Resource Management System (FIRMS), Dynamic World, and GEDI L4B LiDAR data. Validation yielded overall spatial accuracies of 91.15% in the Northeast and 89.62% in the Hengduan Mountains, with corresponding ensemble F1-Scores of 0.92 in both regions. Results indicated the disturbed area in the Northeast (1084.58 ha) significantly exceeded the Hengduan region (133.48 ha). Natural degradation dominated both regions (Northeast: 72.25%; Hengduan: 88.43%), though the Northeast experienced more wildfires and anthropogenic activities. Topographically, Northeast disturbances clustered on low-lying, gentle landscapes, whereas Hengduan events occurred on steep, high-altitude terrains. Due to denser per-pixel carbon storage, the Hengduan area exhibited higher carbon emission costs per unit area. Ultimately, this framework provides a quantitative technical foundation supporting high-resolution forest conservation and spatial evaluations for carbon neutrality commitments. Full article

Qinghai Lake, the largest endorheic saline lake in China, has undergone a pronounced hydrological regime shift from a multi-decadal decline to a rapid post-2004 recovery, reflecting strong hydroclimatic non-stationarity in the northeastern Tibetan Plateau (TP). This paper supplements the current water level and lake area status of Qinghai Lake to provide basic background for future prediction. Reliable forecasting of such climate sensitive lake systems remains difficult because conventional statistical models often fail to capture non-linear fluctuations, whereas standalone deep learning models may overlook long-term deterministic evolution. To address this challenge, we developed a serial decomposition GeoAI framework that integrates autoregressive integrated moving average (ARIMA), one-dimensional convolutional neural networks (1D-CNNs), and long short-term memory (LSTM) networks for non-stationary water level forecasting. Using annual water level observations from 1960 to 2025, the ARIMA component was first used to extract the low-frequency deterministic trend, after which the CNN-LSTM module reconstructed the nonlinear residual variability. The model was trained on the 1960–2012 period and validated over 2013–2025, which represents the most dynamic expansion stage of Qinghai Lake. The hybrid framework outperformed the benchmark models, achieving a Root Mean Square Error (RMSE) of 0.2033 m, Mean Absolute Error (MAE) of 0.1727 m, and Mean Squared Error (MSE) of 0.0413 m² during validation. The decomposition strategy effectively reduced phase lag and amplitude attenuation, improving both predictive accuracy and process interpretability. Multi-step forecasting for 2026–2056 suggests that Qinghai Lake will continue to rise, reaching approximately 3204.08 m by 2056, although the growth rate is projected to slow as negative hydrological feedback strengthen. By explicitly separating deterministic climate scale signals from nonlinear short-term variability, the proposed framework provides a robust and transferable geoinformation based tool for forecasting water level dynamics and supporting adaptive management in climate sensitive, data scarce lake basins. Full article

Alpha/beta (α/β) pulse shape discrimination (PSD) in liquid scintillation counting (LSC) is fundamentally limited by the charge comparison method (CCM) at low energies, where the entire tritium (³H) beta spectrum resides (0–18.6 keVee). The CCM figure-of-merit drops below 0.6 in this region, rendering it inadequate for simultaneous tritium and natural uranium alpha monitoring in nuclear power plant (NPP) liquid effluents. We present a one-dimensional convolutional neural network (1D-CNN) trained on an 80,000-waveform physics-based simulation dataset using established scintillation parameters for Ultima Gold AB. The proposed network achieves 97.4% overall classification accuracy and an area under the receiver operating characteristic curve (AUC) of 0.9981 on the held-out test set, representing improvements of 13.8 percentage points and 0.046 AUC over CCM. In the critical 0–18.6 keVee region, CNN accuracy exceeds 95% compared to below 60% for CCM—a greater than 35 percentage point improvement. Pulse amplitude discrimination (PAD), evaluated as a preliminary screening method, exhibits a 6.3% alpha spillover rate into the beta window, exceeding the regulatory limit of 3%. Gradient-weighted class activation maps (Grad-CAM) confirm that the network exploits physically meaningful pulse features rather than simulation artefacts. A comprehensive background suppression strategy combining dual-SiPM coincidence (24× reduction), anti-coincidence guard detector (5.8× reduction), composite passive shielding (10× reduction), and CNN-assisted discrimination reduces the system equivalent background to 1.83 ± 0.12 cpm, yielding a tritium minimum detectable activity (MDA) of 0.21 Bq/mL (10 mL sample, 30 min count), which satisfies the GB 14587 reference limit of 0.5 Bq/mL. After 8-bit post-training quantisation, the model achieves sub-microsecond inference latency on an embedded Xilinx Artix-7 Field-programmable gate array(FPGA), enabling real-time deployment in portable online monitoring systems. Full article

Salinity stress remains a major bottleneck for agriculture in arid regions. While Opuntia ficus-indica is known for its resilience, its young cladodes maintain a misleadingly healthy visual appearance and stable biomass even under heavy saline pressure, making traditional vegetation indices and standard statistics unreliable for early diagnosis. The objective of this study was to develop a non-destructive phenotyping framework for the early detection of latent salinity stress in young Opuntia cladodes. Controlled experiments were conducted using hyperspectral data cubes (400–1000 nm) acquired from plants exposed to six distinct salinity levels ranging from 2 to 21 dS m⁻¹. Our methodology integrates these high-dimensional spatial–spectral data with a tailor-made 3D Convolutional Neural Network (3D-CNN). Seven physiological vegetation indices—NDVI, PRI, WI, PSRI, MCARI, SIPI, and NDRE were extracted to track sub-clinical shifts and processed as a volumetric depth dimension within the network to preserve spatial–spectral integrity. The optimized 3D-CNN framework achieved a validation accuracy of 99.7% and a weighted F1-score of 99.1%, delivering 100% precision at critical stress thresholds (13 and 21 dS m⁻¹). Spatial confidence maps (Softmax > 0.95) further confirmed the high reliability of the diagnostic output. Requiring a training duration of approximately 8 s, this framework provides a robust basis for precision early-warning irrigation systems to sustain Opuntia cultivation in challenging environments. Full article

Fatty acid abundance is a key parameter for evaluating the quality of edible oils. This study developed a rapid and non-destructive method for predicting palmitic acid content in edible oils by combining visible-near-infrared (Vis-NIR) reflectance spectroscopy with deep learning models. A total of 1740 reflectance spectra in the range of 350–2500 nm were collected from 87 brands of edible oils, including peanut, soybean, corn, sunflower, rapeseed, sesame, and olive oils. Reference values of palmitic acid content were determined via gas chromatography–mass spectrometry (GC-MS). Two conventional machine learning models (SVR and KNN) and four deep learning models (1D-CNN, 1D-ResNet, 1D-Inception, and 1D-Inception-ResNet) were developed and compared using both full-spectrum data and CARS selected characteristic wavelengths. Among the full-spectrum models, the designed 1D-ResNet model achieved the best performance, with the determination coefficient of prediction ($R_p^2$) of 0.9027 and the root mean square error of prediction ($RMS{E}_p$) of 1.13 in the prediction dataset. The proposed 1D-Inception-ResNet model yielded the best prediction results based on the 91 selected informative wavelengths via competitive adaptive reweighted sampling (CARS), achieving an $R_p^2$ of 0.9825 and an $RMS{E}_p$ of 0.4804 in the prediction dataset. The experimental results indicated that Vis-NIR reflectance spectroscopy combined with informative wavelength selection and deep learning models provided an effective strategy for rapid prediction of palmitic acid content in edible oils. Full article

Body area networks (BANs) require secure intra-body communication, yet sensor nodes are too resource-constrained for conventional public-key cryptography, and pre-shared key schemes conflict with plug-and-play clinical workflows. This paper introduces PhysioKey, a TinyML-based key agreement framework that derives symmetric session keys from physiological signals without pre-shared secrets or trusted third parties. A lightweight 1D-CNN (6320 parameters, INT8-quantized, 31.2 KB flash) extracts embeddings from ECG and PPG windows on ARM Cortex-M4 class devices, which are reconciled through fuzzy commitment with BCH error-correcting codes. Patient-level 5-fold cross-validation on PTB-XL (500 patients, dual-ECG) achieves EER of $7.8\%\pm 0.8\%$ with ROC AUC $0.978\pm 0.004$; on BIDMC (53 patients, ECG + PPG), a dual-encoder architecture reduces cross-modal EER to $30.6\%\pm 1.2\%$. Since standalone PhysioKey yields only 7–24 effective key bits, the recommended deployment mode is a hybrid PhysioKey + ECDH protocol providing 128-bit security while PhysioKey adds physical on-body authentication; standalone operation suits energy-constrained scenarios with its $27\times$ advantage over ECDH. HKDF-SHA-256 post-processing yields session keys passing all six NIST SP 800-22 tests (≥96% at the 1024-bit level). Full article

Knee joint moment is an important biomechanical parameter for sports assessment, rehabilitation monitoring, and human–machine interaction. However, direct measurement is often restricted to laboratory-based settings. Surface electromyography (sEMG) offers a non-invasive alternative for indirect joint moment estimation, but many existing deep learning models remain too computationally demanding for potential wearable edge deployment. To address this gap, this study proposes Topo2DCNN-LSTM, a lightweight two-dimensional (2D) convolutional neural network model, designed for sagittal-plane knee flexion moment estimation. The model used a feature-based sequential representation, transforming raw sEMG signals into compact Root Mean Square (RMS) feature frames. The input was processed by a lightweight 2D convolutional neural network (CNN) encoder and paired with long short-term memory (LSTM) units. The model was trained on a public walking dataset of healthy subjects with synchronized sEMG and joint kinetics at two treadmill speeds. When compared with selected deep learning baselines, the quantized model achieved a mean RMS Error of 0.088 ± 0.020 Nm/kg at 1.2 m/s and 0.114 ± 0.034 Nm/kg at 1.8 m/s. On a SparkFun Thing Plus–SAMD51, it achieved an average inference latency of 28 ms using 71,316 bytes of random-access memory (RAM) and 257,172 bytes of flash. These results support its use as a proof of concept for personalized unilateral knee moment estimation with isolated on-device inference feasibility under resource-constrained and limited walking conditions. Full article

A method for simultaneous classification of fabric material and color based on hyperspectral imaging and visual detection is proposed. Fabric material classification is performed using hyperspectral imaging (HSI) combined with a one-dimensional convolutional neural network (1D-CNN), while fabric color recognition is achieved using an red-green-blue (RGB) camera and a color classification model. Material and color features from the same fabric sample are matched to realize synchronous classification. Experiments were conducted on three fabric materials (cotton, polyester, and cotton–polyester blend) and eight colors. At a conveyor speed of 1 m/s, the sorting success rates reach 95.0% for cotton, 97.5% for polyester, and 85.0% for cotton–polyester blended fabrics. The proposed method demonstrates reliable performance for single-material fabrics and good industrial applicability for automated fabric sorting. Full article

Leaf mass per area (LMA) plays an important role in vegetation productivity, carbon cycling, and remote sensing-based ecosystem monitoring. However, remotely predicting LMA from hyperspectral reflectance remains challenging due to the weak and strongly overlapping spectral response of LMA and spectral variability across species. To address these limitations, this study proposed an integrated framework that combines a fractional-order spectral derivative (FOD) with a one-dimensional convolutional neural network (1D-CNN) to enhance LMA prediction accuracy and cross-species generalization. Leaf hyperspectral reflectance was processed using FOD with 0–2 orders, and the relationship between FOD-enhanced spectra and LMA was analyzed. Model performance was assessed using (i) overall prediction accuracy by an 8:2 random split between training and test sets, and (ii) cross-species generalization through leave-one-species-out validation. The results demonstrated that the 1D-CNN using a 1.5-order derivative achieved the best performance (R² = 0.85; RMSE = 11.57 g/m²), outperforming common machine-learning models including partial least squares regression (PLSR), random forest (RF), and support vector regression (SVR). The proposed method also demonstrated great generalization in cross-species prediction. These results indicate that integrating FOD with 1D-CNN effectively enhances LMA-related spectral information and improves LMA prediction across various species. It provides a promising pathway for applying airborne and satellite hyperspectral images in vegetation biochemical parameter mapping, crop monitoring, and ecological assessment. Full article

Accurate upper-limb motor assessment is critical for post-stroke rehabilitation but relies on subjective clinical scales. This study proposes the Action-Aware Multimodal Wavelet Fusion Network (AMWFNet), integrating surface electromyography (sEMG) and robotic kinematics for automated Fugl-Meyer Assessment (FMA-UE)-aligned quantification. Continuous Wavelet Transform (CWT) converts heterogeneous signals into unified time-frequency scalograms. A learnable modality gating mechanism dynamically weights physiological and kinematic features, while action embeddings encode task contexts across 18 standardized reaching tasks. Validated on 40 participants (20 post-stroke, 20 healthy), AMWFNet achieved 94.68% accuracy in six-class classification, outperforming baselines by 9.17% (Random Forest: 85.51%, SVM: 85.30%, 1D-CNN: 91.21%). The lightweight architecture (1.27 M parameters, 922 ms inference) enables real-time assessment-training integration in rehabilitation robots, providing an objective, efficient solution. Full article

Background: Ground contact (GC) detection is essential for sprint performance analysis. Inertial measurement units (IMUs) enable field-based assessment, but their reliability during sprint acceleration remains limited when using heuristic and recently used machine learning algorithms. This study introduces a deep learning one-dimensional convolutional neural network (1D-CNN) to improve GC event and GC times detection in sprint acceleration. Methods: Twelve sprint-trained athletes performed 60 m sprints while bilateral shank-mounted IMUs (1125 Hz) and synchronized high-speed video (250 Hz) captured the first 15 m. Video-derived GC events served as reference labels for model training, validation, and testing, using resultant acceleration and angular velocity as model inputs. Results: The optimized model (18 inception blocks, window = 100, stride = 15) achieved mean Hausdorff distances ≤ 6 ms and 100% precision and recall for both validation and test datasets (Rand Index ≥ 0.977). Agreement with video references was excellent (bias < 1 ms, limits of agreement ± 15 ms, r > 0.90, p < 0.001). Conclusions: The 1D-CNN surpassed heuristic and prior machine learning approaches in the sprint acceleration phase, offering robust, near-perfect GC detection. These findings highlight the promise of deep learning-based time-series models for reliable, real-world biomechanical monitoring in sprint acceleration tasks. Full article