Search Results (466)

A portable and integrated monitoring device was developed to digitally assess both water levels and surrounding fire-related conditions in forest firefighting water buckets using multi-sensor fusion. The system integrates a hydrostatic liquid-level sensor with temperature–humidity and smoke sensors. Validation was performed through field-oriented experiments conducted under semi-controlled conditions. Water-level measurements were collected over a three-month period under simulated forest conditions and benchmarked against conventional steel-ruler readings. Early-stage fire monitoring experiments were carried out using dry wood and leaf litter under varying wind speeds, wind directions, and representative extreme weather conditions. The device achieved a mean water-level bias of −0.60%, a root-mean-square error of 0.64%, and an overall accuracy of 99.36%. Fire monitoring reached a maximum detection distance of 7.30 m under calm conditions and extended to 16.50 m under strong downwind conditions, with performance decreasing toward crosswind directions. Stable operation was observed during periods of strong winds associated with typhoon events, as well as prolonged high-temperature exposure. The primary novelty of this work lies in the conceptualization of a Collaborative Forest Resource–Hazard Monitoring Architecture. Unlike traditional isolated sensors, our proposed framework utilizes a dual-domain decision-making model that simultaneously assesses water-bucket storage stability and micro-scale fire threats. By implementing a robust ‘sensing–logic–alert’ framework tailored for rugged environments, this study offers a new methodological reference for the intelligent management of forest firefighting resources. Full article

Effective and transparent medical diagnosis relies on accurate and interpretable classification of medical images across multiple modalities. This paper introduces an explainable multi-modal image analysis framework based on a dual-stream architecture that fuses handcrafted descriptors with deep features extracted from a custom MobileNet. Handcrafted descriptors include frequency-domain and texture features, while deep features are summarized using 26 statistical metrics to enhance interpretability. In the fusion stage, complementary features are combined at both the feature and decision levels. Decision-level integration combines calibrated soft voting, weighted voting, and stacking ensembles with optimized classifiers, including decision trees, random forests, gradient boosting, and logistic regression. To further refine performance, a hybrid class-specific feature selection strategy is proposed, combining mutual information, recursive elimination, and random forest importance to select the most discriminative features for each class. This hybrid selection approach eliminates redundancy, improves computational efficiency, and ensures robust classification. Explainability is provided through Local Interpretable Model-Agnostic Explanations, which offer transparent details about the ensemble model’s predictions and link influential handcrafted features to clinically meaningful image characteristics. The framework is validated on three benchmark datasets, i.e., BTTypes (brain MRI), Ultrasound Breast Images, and ACRIMA Retinal Fundus Images, demonstrating generalizability across modalities (MRI, ultrasound, retinal fundus) and disease categories (brain tumor, breast cancer, glaucoma). Full article

Accurate and reliable wind speed measurement is essential for applications such as wind power generation and meteorological monitoring. Data fusion from multiple anemometers mounted on wind measurement towers is a key approach to obtaining high-precision wind speed information. In this study, a hierarchical data fusion strategy is proposed to enhance both the quality and efficiency of multi-sensor fusion on wind measurement towers. At the local fusion stage, multi-sensor wind speed data are denoised and fused using an unscented Kalman filter enhanced with fuzzy logic and a robustness factor (FLR-UKF). At the global decision fusion stage, decision-level fusion is achieved through an extreme learning machine (ELM) neural network optimized by a Q-learning-improved Aquila optimizer (QLIAO-ELM). By incorporating a spiral surrounding attack mechanism and a Q-learning-based adaptive strategy, QLIAO-ELM significantly enhances global search capability and convergence speed, enabling the ELM network to obtain superior parameters within limited computational time. Consequently, the accuracy and efficiency of decision fusion are improved. Experimental results show that, during the local fusion phase, the RMSE of FLR-UKF is reduced by 26.46% to 28.6% compared to the traditional UKF; during the global fusion phase, the RMSE of QLIAO-ELM is reduced by 27.1% and 14.0% compared to ELM and ISSA-ELM, respectively. Full article

Contrast-enhanced spectral mammography (CESM) provides low-energy images acquired in standard craniocaudal (CC) and mediolateral oblique (MLO) views, and clinical interpretation relies on integrating both views. This study proposes a dual-view classification framework that combines deep CNN feature extraction with transformer-based fusion for breast-side classification using low-energy (DM) images from CESM acquisitions (Normal vs. Tumorous; benign and malignant merged). The evaluation was conducted using 5-fold stratified group cross-validation with patient-level grouping to prevent leakage across folds. The final configuration (Model E) integrates dual-backbone feature extraction, transformer fusion, MC-dropout inference for uncertainty estimation, and post hoc logistic calibration. Across the five held-out test folds, Model E achieved a mean accuracy of 96.88% ± 2.39% and a mean F1-score of 97.68% ± 1.66%. The mean ROC-AUC and PR-AUC were 0.9915 ± 0.0098 and 0.9968 ± 0.0029, respectively. Probability quality was supported by a mean Brier score of 0.0236 ± 0.0145 and a mean expected calibration error (ECE) of 0.0334 ± 0.0171. An ablation study (Models A–E) was also reported to quantify the incremental contribution of dual-view input, transformer fusion, and uncertainty calibration. Within the limits of this retrospective single-center setting, these results suggest that dual-view transformer fusion can provide strong discrimination while also producing calibrated probabilities and uncertainty outputs that are relevant for decision support. Full article

Accurate probabilistic forecasting of photovoltaic (PV) power generation is crucial for grid scheduling and renewable energy integration. However, existing approaches often produce prediction intervals with limited calibration accuracy, and the interdependence among meteorological variables is frequently overlooked. This study proposes a probabilistic forecasting framework based on a Multi-scale Temporal–Spatial Attention Quantile Regression Network (MTSA-QRN) and an adaptive calibration mechanism to enhance uncertainty quantification and ensure statistically reliable prediction intervals. The framework employs a dual-pathway architecture: a temporal pathway combining Temporal Convolutional Networks (TCN) and multi-head self-attention to capture hierarchical temporal dependencies, and a spatial pathway based on Graph Attention Networks (GAT) to model nonlinear meteorological correlations. A learnable gated fusion mechanism adaptively integrates temporal–spatial representations, and weather-adaptive modules enhance robustness under diverse atmospheric conditions. Multi-quantile prediction intervals are calibrated using conformalized quantile regression to ensure reliable uncertainty coverage. Experiments on a real-world PV dataset (15 min resolution) demonstrate that the proposed method offers more accurate and sharper uncertainty estimates than competitive benchmarks, supporting risk-aware operational decision-making in power systems. Quantitative evaluation on a real-world 40 MW photovoltaic plant demonstrates that the proposed MTSA-QRN achieves a CRPS of 0.0400 before calibration, representing an improvement of over 55% compared with representative deep learning baselines such as Quantile-GRU, Quantile-LSTM, and Quantile-Transformer. After adaptive calibration, the proposed method attains a reliable empirical coverage close to the nominal level (PICP₉₀ = 0.9053), indicating effective uncertainty calibration. Although the calibrated prediction intervals become wider, the model maintains a competitive CRPS value (0.0453), striking a favorable balance between reliability and probabilistic accuracy. These results demonstrate the effectiveness of the proposed framework for reliable probabilistic photovoltaic power forecasting. Full article

Keeping the levels of soil major nutrients (total nitrogen, TN; available phosphorous, AP; and available potassium, AK) in optimum condition is important to achieve the goals of precision agriculture systems. To address the issues of slow speed and low accuracy in soil nutrient detection, this study developed a prediction model for soil major nutrients content based on an improved Extreme Learning Machine (ELM) algorithm. This model utilizes a soil major nutrients detection system integrating pyrolysis and artificial olfaction. First, the Bootstrap Aggregating (Bagging) ensemble strategy was introduced during the model integration phase to effectively reduce prediction variance through multi-submodel fusion. Second, Generative Adversarial Networks (GAN) were employed for sample augmentation, enhancing the diversity and representativeness of the dataset. Subsequently, a multi-scale convolutional and Efficient Lightweight Attention Network (ELA-Net) was embedded in the feature mapping layer to strengthen the representation capability of soil gas features. Finally, adaptive hyperparameter tuning was achieved using the Adaptive Chaotic Bald Eagle Optimization Algorithm (ACBOA) to enhance the model’s generalization capability. Results demonstrate that this model achieves varying degrees of performance improvement in predicting total nitrogen (R² = 0.894), available phosphorus (R² = 0.728), and available potassium (R² = 0.706). Overall prediction accuracy surpasses traditional models by 8–12%, with significant reductions in both RMSE and MAE. These results demonstrate that the method can rapidly, accurately, and non-destructively estimate key soil nutrients, providing theoretical guidance and practical support for field fertilization, soil fertility assessment, and on-site decision-making in precision agriculture. Full article

Modern port operations face increasing challenges from rapidly changing weather and environmental conditions, requiring accurate short-term forecasting to support safe and efficient maritime activities. This study presents a sensor fusion-based machine learning framework for real-time multi-target nowcasting of wind gust speed, sustained wind speed, and wind direction using heterogeneous data collected at the Port of Livorno from February to November 2025. Using an IoT architecture compliant with the oneM2M standard and deployed at the Port of Livorno, CNIT integrated heterogeneous data from environmental sensors (meteorological stations, anemometers) and vessel-mounted LiDAR systems through feature-level fusion to enhance situational awareness, with gust speed treated as the primary safety-critical variable due to its substantial impact on berthing and crane operations. In addition, a comparative performance analysis of Random Forest, XGBoost, LSTM, Temporal Convolutional Network, Ensemble Neural Network, Transformer models, and a Kalman filter was performed. The results show that XGBoost consistently achieved the highest accuracy across all targets, with near-perfect performance in both single-split testing (R² ≈ 0.999) and five-fold cross-validation (mean R² = 0.9976). Ensemble models exhibited greater robustness than deep learning approaches. The proposed multi-target fusion framework demonstrates strong potential for real-time deployment in Maritime Autonomous Surface Ship (MASS) systems and port decision-support platforms, enabling safer manoeuvring and operational continuity under rapidly varying environmental conditions. Full article

The integration of embedded software in multi-sensor wearable devices has revolutionized real-time monitoring across health, fitness, industrial, and environmental applications. This paper presents a comprehensive approach to designing and implementing embedded software architectures that enable efficient, low-power, and high-accuracy data acquisition and processing from heterogeneous sensor arrays. We explore key challenges such as synchronization of sensor data streams, real-time operating system (RTOS) integration, power management strategies, and wireless communication protocols. The reviewed framework supports modular scalability, allowing for seamless incorporation of additional sensors or features without significant system overhead. Future research directions of the embedded software include Hardware-in-the-Loop and real-world validation, on-device machine learning and edge intelligence, adaptive sensor fusion, energy harvesting and power autonomy, enhanced wireless communications and security, standardization and interoperability, as well as user-centered design and personalization. By adopting this focus, we can highlight the potential of the embedded software to support proactive decision-making and user feedback through edge-level intelligence, paving the way for next-generation wearable monitoring systems. Full article

Gait recognition using wearable sensor data is crucial for healthcare, rehabilitation, and monitoring neurological and musculoskeletal disorders. This study proposes a deep learning framework for gait classification using inertial measurements from four body-mounted IMU sensors (head, lower back, and both feet). The data were collected from a publicly available, clinically annotated dataset comprising 1356 gait trials from 260 individuals with diverse pathologies. The framework, G-MASA-TCN (Gait Multi-Anchor, Space-Aware Temporal Convolutional Network), integrates multi-scale temporal fusion, graph-informed spatial modeling, and residual dilated convolutions to extract discriminative gait signatures. To ensure both high performance and interpretability, Integrated Gradients is incorporated as an explainable AI (XAI) method, providing sensor-level and temporal attributes that reveal the features driving model decisions. The framework is evaluated via repeated cross-validation experiments, reporting detailed metrics with cross-run statistical analysis (mean ± standard deviation) to assess robustness. Results show that G-MASA-TCN achieves 98% classification accuracy for neurological, orthopedic, and healthy cohorts, demonstrating superior stability and resilience compared to baseline architectures, including Gated Recurrent Unit (GRU), Transformer neural networks, and standard TCNs, and 98.4% accuracy in identifying individual subjects based on gait. Furthermore, the model offers clinically meaningful insights into which sensors and gait phases contribute most to its predictions. This work presents an accurate, interpretable, and reliable tool for gait pathology recognition, with potential for translation to real-world clinical settings. Full article

Depression is a prevalent mental disorder that imposes a significant public health burden worldwide. Although multimodal detection methods have shown potential, existing techniques still face two critical bottlenecks: (i) insufficient integration of global patterns and local fluctuations in long-sequence modeling and (ii) static fusion strategies that fail to dynamically adapt to the complementarity and redundancy among modalities. To address these challenges, this paper proposes a dynamic multimodal depression detection framework, DynMultiDep, which combines multi-scale temporal modeling with an adaptive fusion mechanism. The core innovations of DynMultiDep lie in its Multi-scale Temporal Experts Module (MTEM) and Dynamic Multimodal Fusion module (DynMM). On one hand, MTEM employs Mamba experts to extract long-term trend features and utilizes local-window Transformers to capture short-term dynamic fluctuations, achieving adaptive fusion through a long-short routing mechanism. On the other hand, DynMM introduces modality-level and fusion-level dynamic decision-making, selecting critical modality paths and optimizing cross-modal interaction strategies based on input characteristics. The experimental results demonstrate that DynMultiDep outperforms existing state-of-the-art methods in detection performance on two widely used large-scale depression datasets. Full article

The purpose of the present study is to improve the efficiency of phishing web resource detection through multimodal analysis and using methods of explainable artificial intelligence. We propose a late fusion architecture in which independent specialized models process four modalities and are combined using weighted voting. The first branch uses CatBoost for URL features and metadata; the second uses CNN1D for symbolic-level URL representation; the third uses a Transformer based on a pretrained CodeBERT for the homepage HTML code; and the fourth uses EfficientNet-B7 for page screenshot analysis. SHAP, Grad-CAM, and attention matrices are used to interpret decisions; a local LLM generates a consolidated textual explanation. A prototype system based on a microservice architecture, integrated with the SOC, has been developed. This integration enables streaming processing and reproducible validation. Computational experiments using our own updated dataset and the public MTLP dataset show high performance: F1-scores of up to 0.989 on our own dataset and 0.953 on MTLP; multimodal fusion consistently outperforms single-modal baseline models. The practical significance of this approach for zero-day detection and false positive reduction, through feature alignment across modalities and explainability, is demonstrated. All limitations and operational aspects (data drift, adversarial robustness, LLM latency) of the proposed prototype are presented. We also outline areas for further research. Full article

Accurate evaluation of demand response (DR) potential at the individual user level is critical for the effective implementation and optimization of demand response programs. However, existing data-driven methods often suffer from insufficient feature representation, limited characterization of load profile dynamics, and ineffective fusion of heterogeneous features, leading to suboptimal evaluation performance. To address these challenges, this paper proposes a novel demand response potential evaluation method based on multivariate heterogeneous features and a Stacking-based ensemble mechanism. First, multidimensional indicator features are extracted from historical electricity consumption data and external factors (e.g., weather, time-of-use pricing), capturing load shape, variability, and correlation characteristics. Second, to enrich the information space and preserve temporal dynamics, typical daily load profiles are transformed into two-dimensional image features using the Gramian Angular Difference Field (GADF), the Markov Transition Field (MTF), and an Improved Recurrence Plot (IRP), which are then fused into a single RGB image. Third, a differentiated modeling strategy is adopted: scalar indicator features are processed by classical machine learning models (Support Vector Machine, Random Forest, XGBoost), while image features are fed into a deep convolutional neural network (SE-ResNet-20). Finally, a Stacking ensemble learning framework is employed to intelligently integrate the outputs of base learners, with a Decision Tree as the meta-learner, thereby enhancing overall evaluation accuracy and robustness. Experimental results on a real-world dataset demonstrate that the proposed method achieves superior performance compared to individual models and conventional fusion approaches, effectively leveraging both structured indicators and unstructured image representations for high-precision demand response potential evaluation. Full article

Financial statement fraud is a socio-technical risk that arises from coupled organizational, informational, and regulatory processes. To address the Identification Paradox in financial fraud detection, where existing models cannot simultaneously recognize both chronic manipulation and acute outbreaks in financial data, this study proposes the Causal–Temporal Asynchrony (CTA) theory as a process-oriented conceptual framework that guides feature construction and model design in a predictive setting. CTA defines fraud motive as a chronic, multi-period accumulation and fraud action as an acute, single-year event. To operationalize CTA within a predictive setting, we build a deployable Decoupling-Fusion System that encodes CTA as an Acute–Chronic Binary Feature Dimensions schema and performs detection via Decoupling-Fusion FraudNet. Within this system, parallel Long Short-Term Memory networks (LSTM) capture chronic motive signals from longitudinal sequences, while parallel Convolutional Neural Networks (CNN) and a Feed-forward Neural Network (FNN) identify acute action signals from multimodal snapshots; the resulting asynchronous probabilities are integrated via an adaptive decision-level fusion mechanism. Empirical tests on China’s A-share market (2001–2021) show the system (AUC = 0.967) outperforms baseline models. Furthermore, eXplainable AI analysis reveals patterns consistent with the classic fraud triangle (pressure, opportunity and rationalization). This study develops a theory-grounded decision-support system that unifies acute and chronic evidence streams and provides a deployable blueprint for continuous auditing and governance. Full article

In froth flotation, the features of froth images are important information to predict the concentrate grade. However, the froth structure is influenced by multiple factors, such as air flowrate, slurry level, ore properties, reagents, etc., which leads to highly complex and dynamic changes in the image features. Additionally, issues such as the immeasurability of ore properties and measurement errors pose significant uncertainties including aleatoric uncertainty (intrinsic variability from ore fluctuations and sensor noise) and epistemic uncertainty (incomplete feature representation and local data heterogeneity) and generalization challenges for prediction models. This paper proposes an uncertainty quantification regression framework based on cross-modal interaction fusion, which integrates the complementary advantages of Selective Kernel Networks (SKNet) and Vision Transformers (ViT). By designing a cross-modal interaction module, the method achieves deep fusion of local and global features, reducing epistemic uncertainty caused by incomplete feature expression in single-models. Meanwhile, by combining adaptive calibrated quantile regression—using exponential moving average (EMA) to track real-time coverage and adjust parameters dynamically—the prediction interval coverage is optimized, addressing the inability of static quantile regression to adapt to aleatoric uncertainty. And through the localized conformal prediction module, sensitivity to local data distributions is enhanced, avoiding the limitation of global conformal methods in ignoring local heterogeneity. Experimental results demonstrate that this method significantly improves the robustness of uncertainty estimation while maintaining high prediction accuracy, providing strong support for intelligent optimization and decision-making in industrial flotation processes. Full article

A 3D multi-modal detection method using a monocular camera and LiDAR has drawn much attention due to its low cost and strong applicability, making it highly valuable for autonomous driving and unmanned aerial vehicles (UAVs). However, conventional fusion approaches relying on static arithmetic operations often fail to adapt to dynamic, complex scenarios. Furthermore, existing ROI alignment techniques, such as local projection and cross-attention, are inadequate for mitigating the feature misalignment triggered by depth estimation noise in pseudo-point clouds. To address these issues, this paper proposes a pseudo-point-based 3D object detection method that achieves biased fusion of multi-modal data. First, a meta-weight fusion module dynamically generates fusion weights based on global context, adaptively balancing the contributions of point clouds and images. Second, a module combining bidirectional cross-attention and a gating filter mechanism is introduced to eliminate the ROI feature misalignment caused by depth completion noise. Finally, a class-agnostic box fusion strategy is introduced to aggregate highly overlapping detection boxes at the decision level, improving localization accuracy. Experiments on the KITTI dataset show that the proposed method achieves APs of 92.22%, 85.03%, and 82.25% on Easy, Moderate, and Hard difficulty levels, respectively, demonstrating leading performance. Ablation studies further validate the effectiveness and computational efficiency of each module. Full article