Search Results (27,921)

Road traffic accident severity prediction is crucial for implementing effective safety measures and proactive traffic management strategies. Existing methods often treat this as a nominal classification problem and use traditional feature selection techniques. However, ordinal classification methods that account for the ordered nature of accident severity (e.g., slight < serious < fatal injuries) in feature selection still need to be investigated thoroughly. In this study, we propose a novel approach, the Ordinal Random Tree with Rank-Oriented Feature Selection (ORT-ROFS), which utilizes the inherent ordering of class labels both in the feature selection and prediction stages for accident severity classification. The proposed approach enhances the model performance by separately determining feature importance based on severity levels. The experiments demonstrated the effectiveness of ORT-ROFS with an accuracy of 87.19%. According to the results, the proposed method improved prediction accuracy by 10.81% over state-of-the-art studies on average on different train–test split ratios. In addition, it achieved an average improvement of 4.58% in accuracy over traditional methods. These findings suggest that ORT-ROFS is a promising approach for accurate accident severity prediction, supporting road safety planning and intervention strategies. Full article

This study refers to the application of an advanced tool in the form of numerical modelling in order to develop a low-waste hot die forging technology to produce a connecting rod forging. The technology aims at ensuring a limited amount of the charge material is necessary to produce one forging, as well as minimizing forging forces, and thus the electric energy consumption. The study includes a verification of the current production technology, which constituted the basis for the construction and development of a numerical model. A new construction of the forging tools was developed, with an additional pre-roughing pass (0X). The new process consists of die forging in the pre-roughing pass (0X), the roughing pass (1X) and the finishing impression (2X). Numerical modelling was subsequently conducted with the use of the Forge 3.0 NxT software. A detailed analysis was conducted on the accuracy of the tool impression filling (including the pre-roughing pass) by the deformed material, the distribution of temperatures for the forgings and the plastic deformations, as well as the courses of forging forces and energy. The results were verified under industrial conditions and compared with the forgings obtained in the previous technology (a roughing pass and a finishing impression). As a result of introducing the pre-roughing pass 0X, the forces were distributed between three impressions, including the especially developed pre-roughing pass. It was confirmed that the abovementioned changes in terms of forging tool construction had a positive effect on relieving the roughing pass and the finishing impression as well as limiting the charge material, and they also lowered the process energy consumption by 10%. This study also validated the relevance of using FE modelling to verify processes under virtual conditions before being implemented under industrial conditions. Therefore, the proposed approach based on multi-variant numerical simulations can be successfully used to improve other manufacturing processes in terms of reducing energy and material consumption and increasing tool service life. Full article

Using the low-voltage and low-current platform (220 VAC-10 A), this paper selected the Mayr arc theoretical model and the improved control theory model as a theoretical basis and built a single-phase low-voltage AC series arc model based on Simulink. The simulation results showed that arc dissipation power directly determined arc voltage amplitude, arc time constant influenced arc voltage waveform, and arc current was mainly determined by load resistance. Because the arc length parameter can be set by the improved control arc theory model, the arc can be drawn only at the micro-distance of two electrodes, which is more suitable for describing the arc characteristics of low voltage and low current. A scheme of large ratio reducer for permanent magnet brushless DC motor was developed, which was combined with the stepless governor controlled by PWM and the positive and negative switch to realize the adjustment of the two-electrode micro-distance. The collection and analysis of arc voltage and arc current under pure resistance, resistive load, and multi-branch load were completed. The experimental results also verified that the Mayr arc and improved control theory arc have good accuracy in describing low voltage and low current characteristics, which improves data support for later fault identification and removal. Full article

To enhance the positioning accuracy of autonomous underwater vehicles (AUVs), a new adaptive filtering algorithm (RHAUKF) is proposed. The most widely used filtering algorithm is the traditional Unscented Kalman Filter or the Adaptive Robust UKF (ARUKF). Excessive noise interference may cause a decrease in filtering accuracy and is highly likely to result in divergence by means of the traditional Unscented Kalman Filter, resulting in an increase in uncertainty factors during submersible mission execution. An estimation model for system noise, the adaptive Unscented Kalman Filter (UKF) algorithm was derived in light of the maximum likelihood criterion and optimized by applying the rolling-horizon estimation method, using the Newton–Raphson algorithm for the maximum likelihood estimation of noise statistics, and it was verified by simulation experiments using the Lie group inertial navigation error model. The results indicate that, compared with the UKF algorithm and the ARUKF, the improved algorithm reduces attitude angle errors by 45%, speed errors by 44%, and three-dimensional position errors by 47%. It can better cope with complex underwater environments, effectively address the problems of low filtering accuracy and even divergence, and improve the stability of submersibles. Full article

This paper proposes a transformer temperature early warning method based on an adaptive sliding window and stacking ensemble learning algorithm, aiming to improve the accuracy and robustness of temperature prediction. The transformer temperature early warning system is crucial for ensuring the safe operation of the power system, and temperature prediction, as the foundation of early warning, directly affects the early warning effectiveness. This paper analyzes the characteristics of transformer temperature using support vector regression, random forest, and gradient boosting regression as base learners and ridge regression as the meta-learner to construct a stacking model. At the same time, Bayesian optimization is used to automatically adjust the sliding window size, achieving adaptive sliding window processing. The experimental results indicate that the temperature prediction method based on adaptive sliding window and stacking significantly reduces prediction errors, enhances the model’s adaptability and generalization ability, and provides more reliable technical support for transformer fault warning. Full article

As a result of global climate change and human production activities, algal blooms are occurring in aquatic environments. The problem of eutrophication in water bodies is becoming increasingly severe, affecting the safety of drinking water sources. In this study, an algal bloom risk index model combining the Improved Fuzzy Analytic Hierarchy Process (IFAHP), Entropy Weight Method (EWM), and Game Theory (GT) was proposed for the Shanxi Reservoir based on the TOPSIS method. After the seasonal and spatial variability in algal bloom risk from 2022 to 2023 was analyzed, an adaptive simplification of the algal bloom risk index calculation was proposed to optimize the model. To enhance its practical applicability, this study proposed an adaptive simplification of the algal bloom risk index calculation based on an improved TOPSIS approach. The error indexes R² for the four seasons and the annual analysis were 0.9884, 0.9968, 0.9906, 0.9946, and 0.9972, respectively. Additionally, the RMSE, MAE, and MRE values were all below 0.035, indicating the method’s high accuracy. Using the adaptively simplified risk index, a risk grading and a spatial delineation of risk areas in Shanxi Reservoir were conducted. A comparison with traditional risk classification methods showed that the error in the risk levels did not exceed one grade, demonstrating the effectiveness of the proposed calculation model and risk grading approach. This study provides valuable guidance for the prevention and control of algal blooms in reservoir-type drinking water sources, contributing to the protection of drinking water sources and public health. Full article

Efficient and reliable corn (Zea mays L.) yield prediction is important for varietal selection by plant breeders and management decision-making by growers. Unlike prior studies that focus mainly on county-level or controlled laboratory-scale areas, this study targets a production-scale area, better representing real-world agricultural conditions and offering more practical relevance for farmers. Therefore, the objective of our study was to determine the best combination of vegetation indices and abiotic factors for predicting corn yield in a rain-fed, production-scale area, identify the most suitable corn growth stage for yield estimation using machine learning, and identify the most effective machine learning model for corn yield estimation. Our study used high-resolution (6 cm) aerial multispectral imagery. Sixty-two different predictors, including soil properties (sand, silt, and clay percentages), slope, spectral bands (red, green, blue, red-edge, NIR), vegetation indices (GNDRE, NDRE, TGI), color-space indices, and wavelengths were derived from the multispectral data collected at the seven (V4, V5, V6, V7, V9, V12, and V14/VT) growth stages of corn. Four regression and machine learning algorithms were evaluated for yield prediction: linear regression, random forest, extreme gradient boosting, and gradient boosting regressor. A total of 6865 yield values were used for model training and 1716 for validation. Results show that, using random forest method, the V14/VT stage had the best yield predictions (RMSE of 0.52 Mg/ha for a mean yield of 10.19 Mg/ha), and yield estimation at V6 stage was still feasible. We concluded that integrating abiotic factors, such as slope and soil properties, significantly improved model accuracy. Among vegetation indices, TGI, HUE, and GNDRE performed better. Results from this study can help farmers or crop consultants plan ahead for future logistics through enhanced early-season yield predictions and support farm profitability and sustainability. Full article

This study proposes a comprehensive framework for the identification of nonlinear dynamics in Unmanned Aerial Vehicles (UAVs), integrating data-driven methodologies with theoretical modeling approaches. Two principal techniques are employed: Proportional-Derivative (PD)-based control input approximation and Sparse Identification of Nonlinear Dynamics (SINDy). Addressing the inherent platform constraints—where control inputs are restricted to specific attitude angles and z-axis velocities—thrust and torque are approximated via a PD controller, which serves as a practical intermediary for facilitating nonlinear system identification. Both methodologies leverage data-driven strategies to construct compact and interpretable models from experimental data, capturing significant nonlinearities with high fidelity. The resulting models are rigorously evaluated within a Model Predictive Control (MPC) framework, demonstrating their efficacy in precise trajectory tracking. Furthermore, the integration of data-driven insights enhances the accuracy of the identified models and improves control performance. This framework offers a robust and adaptable solution for analyzing UAV dynamics under realistic operational conditions, emphasizing the comparative strengths and applicability of each modeling approach. Full article

A gesture recognition method is proposed in this paper, which leverages millimeter-wave radar point clouds, primarily for identifying six basic human gestures. First, the raw radar signals collected by the MIMO millimeter-wave radar are converted into 3D point cloud sequences using a microcontroller integrated into the radar’s baseband processor. Next, based on the SequentialPointNet network, a multiscale feature extraction module is proposed in this paper, which enhances the network’s ability to extract local and global features through convolutional layers at different scales. This compensates for the lack of feature understanding capability caused by single-scale convolution kernels. Moreover, the CBAM in the network is replaced with GAM, which effectively enhances the extraction of global features by more precisely modeling global contextual information, thereby increasing the network’s focus on global features. A separable MLP structure is introduced into the network. The separable MLP operation is used to separately extract local point cloud features and neighborhood features, and then fuse these features, significantly improving the model’s performance. The effectiveness of the proposed method is confirmed through experiments, achieving a 99.5% accuracy in recognizing six fundamental human gestures, effectively distinguishing between gesture categories, and confirming the potential of millimeter-wave radar 3D point clouds in recognizing gestures. Full article

In this study, a novel hybrid heat source model was developed to simulate the welding temperature field in the heat-affected zone (HAZ) of X80 pipeline steel. This model replicates welding conditions with high accuracy and allows flexible three-dimensional adjustments to suit various scenarios. Its development involved the innovative integration of microstructural crystallography information with a multi-scale calibration and validation methodology. The methodology focused on three critical aspects: the weld interface morphology, the location of the Ac₁ temperature, and the size of prior austenite grains (PAG). The morphology of the weld interface was calibrated to align closely with experimental observations. The model’s prediction of the Ac₁ location in actual welded joints exhibited a deviation of less than ±0.3 mm. Furthermore, comparisons of reconstructed PAG sizes between thermal simulation samples and actual HAZ samples revealed minimal discrepancies (5 μm). Validation results confirmed that the calibrated model accurately describes the welding temperature field, with reconstructed PAG size differences between simulation and experimental results being less than 9 μm. These findings validate the accuracy of the calibrated model in predicting welding temperature fields. This research introduces a novel framework for the development of heat source models, offering a robust foundation for improving welding performance and controlling microstructure in different regions during the welding process of high-strength low-alloy (HSLA) steel. Full article

The INterpolated FLOod Surface (INFLOS) tool was developed to meet the operational needs of the Copernicus Emergency Management Service (CEMS) Rapid Mapping (RM) component, which delivers critical crisis information within hours during and after disasters. With increasing demand for accurate and real-time flood depth estimates, INFLOS provides a rapid, adaptable solution for estimating floodwater depth across diverse flood scenarios, using remotely sensed data and high-resolution Digital Terrain Models (DTMs). INFLOS calculates flood depth by interpolating water surface elevation from sample points along flooded area boundaries, derived from satellite imagery. This tool is capable of delivering flood depth estimates in a rapid mapping context, leveraging a multistep interpolation and filtering process for improved accuracy. Tested across fourteen regions in Europe and South America, INFLOS has been successfully integrated into CEMS RM operations. The tool’s computational optimisations further enhance efficiency, improving computation times by up to 15-fold, compared to similar techniques. Indeed, it is able to process areas of up to 6000 ha in a median time of 5.2 min, and up to 30 min at most. In conclusion, INFLOS is currently operational and consistently generates flood depth products quickly, supporting real-time emergency management and reinforcing the CEMS RM portfolio. Full article

Underwater object detection using side-scan sonar (SSS) remains a significant challenge in marine exploration, especially for small objects. Conventional methods for small object detection face various obstacles, such as difficulties in feature extraction and the considerable impact of noise on detection accuracy. To address these issues, this study proposes an improved YOLOv11 network named YOLOv11-SDC. Specifically,a new Sparse Feature (SF) module is proposed, replacing the Spatial Pyramid Pooling Fast (SPPF) module from the original YOLOv11 architecture to enhance object feature selection. Furthermore, the proposed YOLOv11-SDC integrates a Dilated Reparam Block (DRB) with a C3k2 module to broaden the model’s receptive field. A Content-Guided Attention Fusion (CGAF) module is also incorporated prior to the detection module to assign appropriate weights to various feature maps, thereby emphasizing the relevant object information. Experimental results clearly demonstrate the superiority of YOLOv11-SDC over several iterations of YOLO versions in detection performance. The proposed method was validated through extensive real-world experiments, yielding a precision of 0.934, recall of 0.698, [email protected] of 0.825, and [email protected]:0.95 of 0.598. In conclusion, the improved YOLOv11-SDC offers a promising solution for detecting small objects in SSS images, showing substantial potential for marine applications. Full article

Accurate vegetation mapping is essential for monitoring biodiversity and managing habitats, particularly in the context of increasing environmental pressures and conservation needs. Ground truthing plays a crucial role in ensuring the accuracy of supervised remote sensing maps, as it provides the high-quality reference data needed for model training and validation. However, traditional ground truthing methods are labor-intensive, time-consuming and restricted in spatial coverage, posing challenges for large-scale or complex landscapes. The advent of drone technology offers an efficient and cost-effective solution to these limitations, enabling the rapid collection of high-resolution imagery even in remote or inaccessible areas. This study proposes an approach to enhance the efficiency of supervised vegetation mapping in complex landscapes, integrating Multivariate Functional Principal Component Analysis (MFPCA) applied to the Sentinel-2 time series with drone-based ground truthing. Unlike traditional ground truthing activities, drone truthing enabled the generation of large, spatially balanced reference datasets, which are critical for machine learning classification systems. These datasets improved classification accuracy by ensuring a comprehensive representation of vegetation spectral variability, enabling the classifier to identify the key phenological patterns that best characterize and distinguish different vegetation types across the landscape. The proposed methodology achieves a classification accuracy of 92.59%, significantly exceeding the commonly reported thresholds for habitat mapping. This approach, characterized by its efficiency, repeatability and adaptability, aligns seamlessly with key environmental monitoring and conservation policies, such as the Habitats Directive. By integrating advanced remote sensing with drone-based technologies, it offers a scalable and cost-effective solution to the challenges of biodiversity monitoring, enabling timely updates and supporting effective habitat management in diverse and complex environments. Full article

Tea bud localization detection not only ensures tea quality, improves picking efficiency, and advances intelligent harvesting, but also fosters tea industry upgrades and enhances economic benefits. To solve the problem of the high computational complexity of deep learning detection models, we developed the Tea Bud DSCF-YOLOv8n (TBF-YOLOv8n)lightweight detection model. Improvement of the Cross Stage Partial Bottleneck Module with Two Convolutions(C2f) module via efficient Distributed Shift Convolution (DSConv) yields the C2f module with DSConv(DSCf)module, which reduces the model’s size. Additionally, the coordinate attention (CA) mechanism is incorporated to mitigate interference from irrelevant factors, thereby improving mean accuracy. Furthermore, the SIOU_Loss (SCYLLA-IOU_Loss) function and the Dynamic Sample(DySample)up-sampling operator are implemented to accelerate convergence and enhance both average precision and detection accuracy. The experimental results show that compared to the YOLOv8n model, the TBF-YOLOv8n model has a 3.7% increase in accuracy, a 1.1% increase in average accuracy, a 44.4% reduction in gigabit floating point operations (GFLOPs), and a 13.4% reduction in the total number of parameters included in the model. In comparison experiments with a variety of lightweight detection models, the TBF-YOLOv8n still performs well in terms of detection accuracy while remaining more lightweight. In conclusion, the TBF-YOLOv8n model achieves a commendable balance between efficiency and precision, offering valuable insights for advancing intelligent tea bud harvesting technologies. Full article

This paper proposes an innovative algorithm for forecasting the motion response of floating offshore wind turbines by employing force-to-motion transfer functions and state-space models. Traditional numerical integration techniques, such as the Newmark-β method, frequently struggle with inefficiencies due to the heavy computational demands of convolution integrals in the Cummins equation. Our new method tackles these challenges by converting the problem into a system output calculation, thereby eliminating convolutions and potentially enhancing computational efficiency. The procedure begins with the estimation of force-to-motion transfer functions derived from the hydrostatic and hydrodynamic characteristics of the wind turbine. These transfer functions are then utilized to construct state-space models, which compactly represent the system dynamics. Motion responses resulting from initial conditions and wave forces are calculated using these state-space models, leveraging their poles and residues. We validated the proposed method by comparing its calculated responses to those obtained via the Newmark-β method. Initial tests on a single-degree-of-freedom (SDOF) system demonstrated that our algorithm accurately predicts motion responses. Further validation involved a numerical model of a spar-type floating offshore wind turbine, showing high accuracy in predicting responses to both regular and irregular wave conditions, closely aligning with results from conventional methods. Additionally, we assessed the efficiency of our algorithm over various simulation durations, confirming its superior performance compared to traditional time-domain methods. This efficiency is particularly advantageous for long-duration simulations. The proposed approach provides a robust and efficient alternative for predicting motion responses in floating offshore wind turbines, combining high accuracy with improved computational performance. It represents a promising tool for enhancing the development and evaluation of offshore wind energy systems. Full article