Search Results (28,647)

Acquiring information about tree species has special relevance for the management and preservation of forests. Remote sensing has been widely used for this task, especially at landscape scales. However, less research has been done to acquire this information at the single-tree level. In recent decades, mobile laser scanner (MLS) platforms have progressively attracted interest in the forestry sector in regard to obtaining single-tree information. In this study, we examine the feasibility of differentiating between seven different tree species (Castanea sativa, Eucalyptus globulus, Eucalyptus nitens, Pinus pinaster, Pinus radiata, Quercus robur and Chamaecyparis lawsoniana) using MLS point clouds. The classifications were performed using a machine learning algorithm, random forest (RF), and a deep learning algorithm, namely, a 2D Convolutional Neural Network (CNN). The RF algorithm was used to identify tree species at the pixel level using raster layers of statistics that reflect the vertical distribution of the points within single-tree point clouds. The 2D CNN algorithm constructed multi-view 2D profiles of single-tree point clouds by rotating the point clouds around a single axis. Different 2D image sizes were tested. The 2D CNN algorithm outperformed the RF algorithm and yielded an Overall Accuracy of 93%. We also found that image size affected both the accuracy metrics obtained and the amount of training time needed. Of the tree species studied, Pinus radiata and Quercus robur had the lowest classification accuracies (with F-Scores of 83% and 87%, respectively). According to these results, MLS point clouds can be efficiently used in the forest sector to perform individual-tree species recognition through fully automated procedures. Full article

Saturated diving plays an important role in fields such as navigation operations, ocean development, military oceanography, and maritime rescue and is indispensable for the marine economy. Heliumspeech communication is an essential component of deep-sea saturated diving operations and serves as the sole means of communication in such environments. This paper presents a heliumspeech unscrambling method for saturated diving based on phonetic multi-objective optimization using deep learning. The method consists of a heliumspeech correction network and a heliumspeech unscrambling network. First, a phonetic multi-objective optimization algorithm is used to design the correction network, which reduces the demand for large heliumspeech training datasets. Then, a saturated diving working language heliumspeech corpus is used to train the heliumspeech unscrambling network. Finally, the unscrambling network processes heliumspeech using a cognitive transfer learning algorithm. During unscrambling, the network continuously converts unscrambled heliumspeech into sample data and adds them to a supervised database, forming a closed-loop control system that dynamically adjusts the network parameters to adapt to variations in heliumspeech signals. This approach not only reduces the deep learning neural network’s reliance on large training datasets but also enhances unscrambling performance, particularly under dynamically changing diving depths. Simulation experiments demonstrate that the method effectively unscrambles heliumspeech with a low word error rate and fast convergence. Full article

This paper introduces Hybrid Physics-Augmented Neural Network (HyPA-Net), a hybrid modeling framework that integrates physics-based linear time-invariant models with artificial neural networks (ANNs) to address dynamic system modeling when only partial physical knowledge is available. The approach leverages the interpretability and robustness of established physical models while using ANNs—such as long short-term memory architectures—to capture unknown or nonlinear system behaviors. The methodology normalizes state and input variables for compatibility with ANN training and expands traditional recursive state-space equations for efficient backpropagation over sequences. Vehicle dynamics, specifically using a rear-wheel steering test case, validate the proposed framework. Various HyPA-Net configurations are benchmarked against pure physics-based and pure data-driven models, demonstrating improved prediction accuracy and model flexibility. The experimental results in this application confirm that hybrid models yield superior performance over strict physical approaches and can implicitly approximate submodel dynamics within a unified, yet modular, architecture, opening avenues for applications in domains where partial physics-based knowledge is available but insufficient on its own. Full article

This study presents a comprehensive analysis and predictive modeling framework for the axial compressive strength (f_cc) and ultimate axial strain (ε_cu) of concrete columns confined within fiber-reinforced polymer (FRP) systems. Large databases comprising 3312 samples for f_cc and 3319 for ε_cu were compiled from the literature, encompassing a wide range of key variables, including unconfined concrete strength from 7 MPa to 204 MPa and diverse FRP confinement configurations. The datasets were subjected to extensive statistical and multivariate analyses to identify the primary factors influencing axial behavior and guide feature selection for predictive modeling. Three groups of machine learning (ML) algorithms were subsequently considered: (i) artificial neural networks (including multilayer perceptrons with one and two hidden layers), (ii) kernel-based models (Gaussian process regression and support vector regression), and (iii) tree-based ensemble models (gradient boosting machine, eXtreme gradient boosting, and light gradient boosting machine). Hyperparameters were optimized using grid search cross-validation, while feature importance analyses were performed to quantify the contribution of each input variable. Among all ML models, eXtreme gradient boosting demonstrated superior predictive performance, effectively capturing the nonlinear and multivariate interactions governing confinement effectiveness. Comparative analysis with the top performing regression-based formulations further highlighted the accuracy, robustness, and generalization capability of the eXtreme gradient boosting model. The findings provide a data-driven and interpretable framework for the design and prediction of FRP-confined concrete columns. Full article

In wheat threshing, reducing total grain loss and energy consumption is crucial for both economic and sustainable food security. This study investigates the effects of threshing drum type (straight and helical row), drum peripheral speed (36.73–48.98 m s⁻¹), and drum-concave clearance (35–50 mm) on total grain loss and specific fuel consumption in a stationary threshing machine using a full factorial design. The optimum machine settings (drum type, peripheral speed and drum–concave clearance) that simultaneously minimize these two outputs were then determined. We systematically compared three surrogate modelling approaches—Response Surface Model (RSM), Gaussian Process Regression (GPR), and Artificial Neural Network (ANN)—to identify the most effective method for small-dataset optimization in threshing machine design. The best model was selected through cross-validation, and optimization was performed using the NSGA-II multi-objective genetic algorithm. GPR yielded the highest prediction accuracy for both outputs (R² in prediction data: 0.99 for total grain loss and 0.91 for specific fuel consumption). Multi-objective optimization revealed a conflict between the two objectives; the best balance was achieved for the helical drum at a peripheral speed of approximately 41.5 m s⁻¹ and a drum–concave clearance of 50 mm (predicted total grain loss approximately 3.7%, specific fuel consumption approximately 2.98 mL kg⁻¹). Compared to the straight-row drum, the helical drum provided lower losses and fuel consumption, as well as approximately 3.5 times wider safe operating range. It should be noted that this optimum was predicted by the surrogate model and agreed closely with the best measured treatment; it was not confirmed by an independent validation experiment. The results demonstrated that combining a surrogate model with a genetic algorithm is an effective tool for optimizing threshing machine parameters. Full article

Accurately predicting time-difference signals is a key prerequisite for extracting effective temporal characteristics from the output of RTD fluxgate sensors and enhancing the measurement reliability of the sensors. However, the nonlinear characteristics and temporal dependencies of residence time difference (RTD)-fluxgate time-difference signals make accurate prediction difficult. In this study, an improved Transformer neural network model is proposed for RTD-fluxgate time-difference signals. By introducing positional encoding, the proposed method incorporates temporal position information into the Transformer model, enabling effective extraction of temporal features from sensor output signals. To alleviate overfitting during model training, dropout and weight decay strategies are introduced to improve the generalization capability of the prediction model. The predicted time-difference signals are analyzed and the key temporal features are extracted for sensor signal processing applications. The proposed method is compared with feedforward neural networks and long short-term memory networks. Experimental results obtained under the same constant magnetic field demonstrate that the proposed method improves cosine similarity (CS) by 37% and 4%, and reduces mean square error (MSE) by 29% and 2%, respectively. The results verify the effectiveness of the proposed approach for RTD-fluxgate time-difference signal prediction and provide a data-driven method for sensor signal analysis and performance evaluation in magnetic measurement applications. This method provides a novel approach for predicting the time-difference signals of RTD fluxgate sensors, offering a reliable reference for subsequent signal processing and improving the accuracy and consistency of magnetic field data. Full article

Drone-based chemical spraying in agriculture faces challenges related to operator health hazards, spray deposition, application efficiency, and environmental safety. Electrostatic charging system is a novel technology that minimizes off-target spraying losses and increases droplet deposition on the plant canopy. In electrostatic spraying, chemical consumption and application rates are reduced due to the uniform distribution and enhanced deposition of charged droplets on plant surfaces, thereby improving spraying efficacy. In this study, a drone-based air-assisted electrostatic sprayer was developed to investigate the effect of operational parameters that include forward speed, discharge rate, applied voltage (charged condition) and propeller speed on the droplet density (drops/cm²) in a cotton crop under laboratory conditions. The charge-to-mass ratio (CMR) of the developed air-assisted electrostatic nozzle was found in the range between 1.8–2.5 mC/kg. An Artificial Neural Network–Ant Colony Optimization (ANN–ACO) method was used to optimize the operational parameters for obtaining the highest charged droplet density on the plant canopy surfaces. Results showed that the charged droplet density was significantly affected by discharge rate (DR) and applied voltage (AV) followed by forward speed (FS) and was slightly influenced by the propeller speed (PS). Optimal performance was achieved at FS = 2.58 km/h, PS = 1204 rpm, DR = 558.45 mL/min and AV = 6.18 kV under the charged conditions. At these optimized parameters, an average charged droplet density of 185.83 ± 4.25 drops/cm² (mean ± SE) was achieved. For the uncharged conditions, the optimal performance was achieved at FS = 2.80 km/h, PS = 1065 rpm and DR = 556.75 mL/min, corresponding to an average droplet density of 108.49 ± 2.15 drops/cm². The integration of the ANN–ACO optimization algorithm with the drone-based air-assisted electrostatic spraying system can enhance precision chemical application on the cotton, improving efficiency and sustainability. Full article

Reliable in situ monitoring is essential for the improvement of process supervision, quality assurance, and machine-state recognition in additive manufacturing of smart composite systems. This study presents a non-invasive acoustic–vibration side-channel monitoring framework for identifying FDM printing cases using experimental recordings from two printers, Bambu Lab A1 mini and Bambu Lab P1P. Four representative printing cases were investigated: simple key, hard key, two keys, and retraction test. Raw acoustic and vibration signals were converted into interval-level statistical features, including six acoustic descriptors and 18 vibration descriptors extracted from the X, Y, and Z axes. A baseline deep neural network (DNN) and an enhanced residual attention deep neural network (RA-DNN) were implemented under acoustic-only, vibration-only, and fused acoustic–vibration input conditions using stratified five-fold cross-validation. The fused acoustic–vibration features achieved the best performance, with the RA-DNN reaching 96.75% accuracy, 96.91% precision, 96.75% recall, and 96.70% F1-score for the A1 mini, and 98.19% accuracy, 98.25% precision, 98.19% recall, and 98.18% F1-score for the P1P. These results indicate that acoustic–vibration side-channel signals can provide effective process signatures for intelligent and quality-aware FDM monitoring. Full article

Russia’s invasion of Ukraine has posed serious risks to global food security, by causing substantial crop yield losses since 2022. Accurate yield estimation helps policymakers to plan compensation, yet modelling yields in conflict regions remains challenging due to significant non-meteorological disruptions. This study proposes a novel framework to quantify war-related crop yield losses by comparing estimations derived from meteorological data, representing weather-driven yield variability, with those based on Earth observation (EO) data, reflecting actual crop conditions influenced by both weather and conflict. Thus, meteorologically based yield estimates are expected to exceed those derived from EO data, with the difference indicating war-related losses. Both, meteorological- and EO-based models, are developed using transfer learning (TL) to estimate yields of maize, winter wheat, and spring barley. Models are initially trained on EU country data and subsequently finetuned with Ukrainian data. Their performance is compared to two non-TL approaches: Extreme Gradient Boosting (XGB) and Artificial Neural Network (ANN) to test their reliability. Results show crop yield losses for maize; however, since we do not detect losses in the other crops, we conclude that simply comparing meteorological- and EO-based models proves insufficient to fully isolate conflict effects due to strong interactions of EO and meteorological data. Nevertheless, TL substantially enhances prediction accuracy (R² around 0.7), exceeding alternative models by 0.05–0.2 across crops. These findings demonstrate the value of TL for yield modelling in data-scarce environments and underscore the need for improved methodologies to quantify conflict-induced agricultural losses. Full article

In biomedical research and clinical practices, Graph Neural Networks (GNNs) are playing an increasingly important role and have been applied to the problems of disease pathway detection, gene–disease relation prediction, etc. They show great potential for biomedical predictions; however, there are interpretability issues when used on complex datasets like gene expression data. Current explainability methods such as GNNExplainer are designed to explain individual instances, not the whole network. The absence of transparency hinders trust and limits the clinical/biomedical implementation of GNNs. Additionally, more interpretable models like GNN-SubNet and XGDAG do not fulfill the expectation of a clear picture for the entire network. This research addresses the limitation of the network-wide explainability of GNNs by introducing a GNN-based BioGraphEX model that incorporates interpretable methods at two levels, instance-level and network-wide level, such as gradient-based methods and SHAP (Shapley Additive Explanations). Using the GSE25097 biomedical dataset, the model achieves an accuracy of 85% and an F1 Score of 0.82, surpassing baseline methods in both predictive performance and interpretability. These results address the limitations of existing models like GNN-SubNet and XGDAG by providing both instance-level and network-wide insights. Metrics like Explanation Fidelity (83%) further validated the robustness of the explanations. Full article

Wireless power transfer (WPT) has emerged as a promising technology for electric vehicle (EV) charging owing to its capability to provide convenient, safe, and contactless energy transfer. This paper presents a comprehensive review of recent advances in wireless EV charging systems. Different charging architectures, including static, quasi-dynamic, dynamic, and bidirectional configurations, are discussed. Basic and hybrid compensation topologies are critically examined with emphasis on their operating characteristics and suitability for EV applications. Various magnetic coupler structures, ranging from conventional circular coils to double-D, quadrature, bipolar, and multi-coil configurations, are reviewed in terms of coupling performance and misalignment tolerance. In addition, conventional, advanced, and intelligent control strategies, including frequency, phase-shift, duty-cycle, model predictive, adaptive, sliding mode, fuzzy logic, artificial neural network, and reinforcement learning approaches, are comparatively analyzed. Finally, the major technical challenges, electromagnetic compatibility and safety issues, economic barriers, emerging technologies, and future research directions are highlighted. This review provides a comprehensive reference for researchers and engineers and offers insights into the development of highly efficient, intelligent, and sustainable wireless EV charging infrastructures. Full article

Hyperspectral imaging is a well-established remote sensing technique for material detection and classification, relying on hundreds of reflectance bands to exploit the spectral signatures of surface materials. Although hyperspectral imagery has demonstrated excellent capabilities for material identification, its operational implementation may be constrained in some applications due to data volume and processing requirements. Consequently, there is growing interest in evaluating the capability of lower-dimensional multispectral imagery for material detection tasks. In this sense, this article proposes as its contribution the comparative evaluation of machine learning models and neural networks for asbestos–cement detection on VNIR imagery. For the development of this research, the CRISP-DM methodology was adapted into four phases: P1. Business and data understanding; P2. Data preparation; P3. Modelling and evaluation; P4. Model deployment. At the results level, three datasets with different numbers of bands were constructed, which were structured by adding to the original dataset an additional layer with the NDVI and two additional layers with the PCA components of the original image. Across the three datasets, four machine learning models and one neural network model were tuned and evaluated, yielding as a result that in all three datasets the KNN and neural network models achieved the best performance. Likewise, it was found that the detection capability of the models improved with the inclusion of the additional bands. The proposed approach serves as a reference to be extrapolated by research centres and universities for the detection of asbestos and other materials in VNIR images, with a view toward integration into resource-constrained systems and specifically into environmental monitoring systems. Full article

The digital transformation of manufacturing associated with Industry 4.0 and the human-centric paradigm of Industry 5.0 requires advanced predictive tools that can improve the performance, reliability, and sustainability of tribological systems. In this context, accurate prediction of friction behavior in rolling contacts is crucial for intelligent monitoring and optimization of bearing components. This study presents a machine learning-based methodology for predicting the coefficient of friction in rolling contact of 100Cr6 steel bearing discs as a function of surface roughness and rolling distance parameters. Experimental studies were conducted using discs with different surface topography under controlled rolling contact conditions. Surface roughness characteristics and rolling distance data were correlated with experimentally measured friction coefficients to create a comprehensive dataset for artificial intelligence (AI) modeling. Several dozen machine learning (ML) algorithms, including random forest, support vector regression, and artificial neural networks, were developed and comparatively evaluated to capture nonlinear relationships between operational and surface parameters. The predictive ability of the models was assessed using statistical metrics such as the coefficient of determination (R²), mean absolute error (MAE), and root mean square error (RMSE). The obtained results demonstrate that ML methods provide high prediction accuracy and effectively identify the combined effects of surface roughness and rolling distance on rolling contact friction. Feature importance analysis revealed that roughness parameters dominate friction behavior during the run-in phase, while rolling distance becomes increasingly important under stabilized operating conditions. The proposed approach supports the development of intelligent tribological systems, predictive maintenance strategies, and data-driven decision-making frameworks aligned with Industry 4.0 and Industry 5.0 concepts. The presented methodology can contribute to the implementation of intelligent manufacturing solutions, the sustainable operation of bearing systems, and AI-assisted monitoring of machine components in modern industrial environments. Full article

Precise step length estimation (SLE) is a key necessity for not only navigation systems design but also gait health monitoring in neurological conditions. Among existing solutions, non-invasive inertial sensor-based approaches operating without dedicated infrastructure are more cost-effective. Many such methods, however, rely on bodily affixed inertial sensors rather than freely held smartphone sensors. Traditional signal processing approaches, on the other hand, offer varying accuracy across diverse gait patterns due to user parameter calibration. This study, thus, proposes a regression-based SLE framework employing eight regression algorithms: linear regression (LR), k-nearest neighbors, support vector machine, decision tree, elastic network, random forest, histogram-based gradient boosting (HGB) regressor, and artificial neural network (ANN). Their extensive and rigorous evaluations across varied window sizes, using a dataset collected in normal and fast walking modes with two device positions (hand-held and trouser-pocket) during three evaluation scenarios, demonstrate the HGB regressor’s outstanding performance, achieving the lowest mean absolute error (MAE) below 1 cm across four different contexts under leave-one-out cross-validation-based evaluation and three in the seen test evaluations. Moreover, the findings report the ANN’s exceptional generalization capacity over other models and the previous method IRT-SD-SLE in unseen test evaluations, with an MAE not exceeding $6.3$ cm. The extensive evaluations of training and testing times reveal the highest computational efficiency for LR, moderate efficiency for the HGB regressor, and the highest training cost for the ANN, indicating a clear trade-off between MAE and computational expense. Additionally, this study includes an insightful discussion on the performance results, including the trade-offs between accuracy and efficiency. Full article

The rapid proliferation of drones has raised growing concerns regarding low-altitude airspace safety. In practical UAV communication scenarios, relying solely on a single time-domain IQ signal is often insufficient to fully characterize UAV RF fingerprints. Moreover, increasingly subtle hardware differences among UAV transmitters make RF fingerprint identification more challenging. Other representations, such as STFT-based features, are commonly converted into image-like inputs for neural networks, increasing deployment complexity on edge devices. To address these challenges, this paper proposes a UAV recognition framework based on multi-domain signal representations. The proposed framework employs a multi-domain input strategy and structural reparameterization to reduce the number of parameters, computational cost, and deployment latency. Experiments under AWGN conditions demonstrate that the proposed model achieves superior recognition performance in both UAV classification and individual identification tasks. The proposed model is further deployed on the Ascend 910B platform to verify its deployment feasibility. Full article