Search Results (785)

Ship trajectory prediction plays an important role in numerous maritime applications and services. With the development of deep learning technology, the deep learning prediction method based on Automatic Identification System (AIS) data has become one of the hot topics in current maritime traffic research. However, as current models always concatenate dynamic information with distinct meanings (such as position, ship speed, and heading) into a single integrated input when processing trajectory point information as input, it becomes difficult for the models to grasp the correlations between different types of dynamic information of trajectory points and the specific information contained in each type of dynamic information itself. Aiming at the problem of insufficient modeling of the relationships among dynamic information in ship trajectory prediction, we propose the Multi-dimensional Attribute Relationship Transformer (MART) model. This model introduces a simulated trajectory training strategy to obtain the Association Loss (AssLoss) for learning the associations among different types of dynamic information; and it uses the Distance Loss (DisLoss) to integrate the relative distance information of the attribute embedding encoding to assist the model in understanding the relationships among different values in the dynamic information. We test the model on two AIS datasets, and the experiments show this model outperforms existing models. In the 15 h long-term prediction task, compared with other models, the MART model improves the prediction accuracy by 9.5% on the Danish Waters Dataset and by 15.4% on the Northern European Dataset. This study reveals the importance of the relationship between attributes and the relative distance of attribute values in spatiotemporal sequence modeling. Full article

The rapid development of wireless communication technology is leading to increasingly scarce spectrum resources, making efficient utilization a critical challenge. This paper proposes a Convolutional Neural Network–Long Short-Term Memory-Integrated Gradient-Weighted Class Activation Mapping (GC-CNN-LSTM) model, aimed at enhancing the accuracy of long-term spectrum prediction across multiple frequency bands and improving model interpretability. First, we achieve multi-frequency long-term spectrum prediction using a CNN-LSTM and compare its performance against models including LSTM, GRU, CNN, Transformer, and CNN-LSTM-Attention. Next, we use an improved Grad-CAM method to explain the model and obtain global heatmaps in the time–frequency domain. Finally, based on these interpretable results, we optimize the input data by selecting high-importance frequency points and removing low-importance time segments, thereby enhancing prediction accuracy. The simulation results show that the Grad-CAM-based approach achieves good interpretability, reducing RMSE and MAPE by 6.22% and 4.25%, respectively, compared to CNN-LSTM, while a similar optimization using SHapley Additive exPlanations (SHAP) achieves reductions of 0.86% and 3.55%. Full article

This study presents an optimized methodology for the placement of water film thickness sensors, integrating information theory with experimental validation. Initially, the two-dimensional shallow-water equations are employed to simulate the spatiotemporal evolution of water film thickness across the entire runway, providing a comprehensive foundational dataset. By applying information entropy theory, the total information content at each runway grid point is quantified. Analysis indicates that grid points with higher total information content generally correspond to regions of greater water film thickness. The optimal placement for a single sensor is determined by identifying the location that maximizes total information content, and its effectiveness is validated through controlled rain–fog experiments. The results demonstrate that positioning a single sensor at a site with higher water film thickness reduces the overall mean estimation error by 57%, thereby enhancing prediction accuracy. By extending the single-sensor placement framework, the total information content across all runway points is recalculated, and additional rain–fog experiments are conducted to verify the optimal locations. By incorporating a correlation coefficient–distance (C–D) model to define each sensor’s influence radius, a collaborative multi-sensor placement strategy is developed and implemented at Seletar Airport, Singapore. The findings show that sensor locations with higher water film thickness correspond to increased total information content, and that expanding the number of deployed sensors further improves estimation accuracy. Compared with conventional placement approaches, which rely on subjective judgment and long-term operational experience, the proposed method enhances estimation accuracy by over 23% when deploying two sensors. These results provide a robust basis for the strategic placement of runway water film thickness sensors and contribute to more precise assessments of pavement surface conditions. Full article

Habitat fragmentation represents a significant threat to biodiversity, particularly the survival of wild species. Constructing and optimizing ecological networks are critical for promoting sustainable biodiversity, especially in the conservation of unmanaged wildlife. To address this, this study focused on designing and optimizing an ecological network tailored to the preservation of the Xinjiang desert lacertid lizard (Eremias multiocellata). This study integrated a dual-model approach, applying the InVEST model for habitat quality assessment and the MaxEnt model for suitable habitat prediction. An overlay analysis identified 15 core ecological source areas spanning 126,044 km², primarily located in the desert–grassland transition zones of the central and western study areas. A total of 34 ecological corridors were established utilizing the minimum cumulative resistance model, totaling 3764 km in length. These include 11 long corridors, 17 short corridors, and 6 potential corridors. Additionally, 100 strategic points were identified: 41 pinch points, 38 barrier points, and 21 stepping stones. This study identifies priority areas and obstacles affecting the ecological connectivity of the species’ habitats and highlights the importance of small habitat patches for long-term species dispersal and habitat expansion, providing more comprehensive guidance for sustainable development and species conservation. Furthermore, the methodology provides valuable insights into biodiversity conservation and the optimization of the natural habitat spatial layout in desert ecosystems, along with novel methods for managing and conserving other unmonitored animal species in various ecosystems. Full article

Miller’s rule originated as an empirical relation between the nonlinear and linear optical coefficients of materials. It is now accepted as a useful tool for guiding experiments and computational materials discovery, but its theoretical foundation had long been limited to a derivation for the classical Lorentz model with a weak anharmonic perturbation. Recently, we developed a mathematical framework which enabled us to prove that Miller’s rule is equally valid for quantum anharmonic oscillators, despite different dynamics due to zero-point fluctuations and further quantum-mechanical effects. However, our previous derivation applied only to one-dimensional oscillators and to the special case of second- and third-harmonic generation in a monochromatic electric field. Here we extend the proof to three-dimensional quantum anharmonic oscillators and also treat all orders of the nonlinear response to an arbitrary multi-frequency field. This makes the results applicable to a much larger range of physical systems and nonlinear optical processes. The obtained generalized Miller formulae rigorously express all tensor elements of the frequency-dependent nonlinear susceptibilities in terms of the linear susceptibility and thus allow a computationally inexpensive quantitative prediction of arbitrary parametric frequency-mixing processes from a small initial dataset. Full article

As an important traffic and transportation roadway, tunnel engineering is widely used in important fields such as highways, railways, water conservancy, subways and mining. It is limited by complex geological conditions, harsh construction environments and poor robustness of the monitoring system. If the construction process and monitoring method are not properly designed, it will often directly induce disasters such as tunnel deformation, collapse, leakage and rockburst. This seriously threatens the safety of tunnel construction and operation and the protection of the regional ecological environment. Therefore, based on distributed fiber optic sensing technology, the full–cycle spatiotemporally continuous sensing information of the tunnel structure is obtained in real time. Accordingly, the health status of the tunnel is dynamically grasped, which is of great significance to ensure the intrinsic safety of the whole life cycle for the tunnel project. Firstly, this manuscript systematically sorts out the development and evolution process of the theory and technology of structural health monitoring in tunnel engineering. The scope of application, advantages and disadvantages of mainstream tunnel engineering monitoring equipment and main optical fiber technology are compared and analyzed from the two dimensions of equipment and technology. This provides a new path for clarifying the key points and difficulties of tunnel engineering monitoring. Secondly, the mechanism of action of four typical optical fiber sensing technologies and their application in tunnel engineering are introduced in detail. On this basis, a spatiotemporal continuous perception method for tunnel engineering based on DFOS is proposed. It provides new ideas for safety monitoring and early warning of tunnel engineering structures throughout the life cycle. Finally, a high–speed rail tunnel in northern China is used as the research object to carry out tunnel structure health monitoring. The dynamic changes in the average strain of the tunnel section measurement points during the pouring and curing period and the backfilling period are compared. The force deformation characteristics of different positions of tunnels in different periods have been mastered. Accordingly, scientific guidance is provided for the dynamic adjustment of tunnel engineering construction plans and disaster emergency prevention and control. At the same time, in view of the development and upgrading of new sensors, large models and support processes, an innovative tunnel engineering monitoring method integrating “acoustic, optical and electromagnetic” model is proposed, combining with various machine learning algorithms to train the long–term monitoring data of tunnel engineering. Based on this, a risk assessment model for potential hazards in tunnel engineering is developed. Thus, the potential and disaster effects of future disasters in tunnel engineering are predicted, and the level of disaster prevention, mitigation and relief of tunnel engineering is continuously improved. Full article

Droughts are becoming more frequent, severe, and impactful across the globe. Agroecosystems, which are human-made ecosystems with high water demand that provide essential ecosystem services, are vulnerable to extreme droughts. Although water use efficiency in agriculture has increased in rec ent decades, drought management should be based on long-term, proactive strategies rather than crisis management. The AgrHyS network of sites in French Brittany collects high-resolution soil moisture data from agronomic stations and catchments to improve understanding of temporal soil moisture dynamics and enhance water use efficiency. Frequent mapping of soil moisture and plant water stress is crucial for assessing water stress risk in the context of global warming. Although satellite remote sensing provides reliable, periodic global data on surface soil moisture, it does so at a very coarse spatial resolution. The intrinsic spatial heterogeneity of surface soil moisture requires a higher spatial resolution in order to address upcoming challenges on a local scale. Drones are an excellent tool for upscaling point measurements to catchment level using different onboard cameras. In this study, we evaluated the potential of multispectral images, thermal images and LiDAR data captured in several concurrent drone flights for high-resolution mapping of soil moisture spatial variability, using in situ point measurements of soil water content and plant water stress in both agricultural areas and natural ecosystems. Statistical models were fitted to map soil water content in two areas: a natural marshland and a grassland-covered agricultural field. Our results demonstrate the statistical significance of topography, land surface temperature and red band reflectance in the natural area for retrieving soil water content. In contrast, the grasslands were best predicted by the transformed normalised difference vegetation index (TNDVI). Full article

Accurate wave forecasting under typhoon conditions is essential for coastal safety in the Pearl River Estuary. This study explores the use of Random Forest (RF) and Long Short-Term Memory (LSTM) models to predict significant wave heights, using SWAN-simulated data from 87 historical typhoon events. Ten representative typhoons were reserved for independent testing. Results show that the LSTM model outperforms RF in 3 h forecasts, achieving a lower mean RMSE and higher R², particularly in capturing wave peaks under highly dynamic conditions. For 6 h forecasts, both models exhibit decreased accuracy, with RF performing slightly better in stable scenarios, while LSTM remains more responsive in complex wave evolution. Generalization tests at three nearby stations demonstrate that both models, especially LSTM, retain strong predictive skill beyond the training location. These findings highlight the potential of combining numerical wave models with machine learning for short-term, data-driven wave forecasting in typhoon-prone and observation-sparse regions. The study also points to future improvements through integration of wind field predictors, model updating strategies, and ensemble meteorological data. Full article

Developing structural materials capable of maintaining integrity under extreme irradiation conditions is a cornerstone challenge for advancing sustainable nuclear energy technologies. The complexity and severity of radiation-induced microstructural changes—spanning multiple length and timescales—pose significant hurdles for purely experimental approaches. This review critically evaluates recent advancements in atomistic modeling, emphasizing its transformative potential to decipher fundamental mechanisms driving microstructural evolution in irradiated alloys. Atomistic simulations, such as molecular dynamics (MD), have successfully unveiled initial defect formation processes at picosecond scales. However, the inherent temporal limitations of conventional MD necessitate advanced methodologies capable of exploring slower, thermally activated defect kinetics. We specifically traced the development of powerful potential energy landscape (PEL) exploration algorithms, which enable the simulation of high-barrier, rare events of defect evolution processes that govern long-term material degradation. The review systematically examines point defect behaviors in various crystal structures—BCC, FCC, and HCP metals—and elucidates their characteristic defect dynamics, respectively. Additionally, it highlights the pronounced effects of chemical complexity in concentrated solid-solution alloys and high-entropy alloys, notably their sluggish diffusion and enhanced defect recombination, underpinning their superior radiation tolerance. Further, the interaction of extended defects with mechanical stresses and their mechanistic implications for material properties are discussed, highlighting the critical interplay between thermal activation and strain rate in defect evolution. Special attention is dedicated to the diverse mechanisms of dislocation–obstacle interactions, as well as the behaviors of metastable grain boundaries under far-from-equilibrium environments. The integration of data-driven methods and machine learning with atomistic modeling is also explored, showcasing their roles in developing quantum-accurate potentials, automating defect analysis, and enabling efficient surrogate models for predictive design. This comprehensive review also outlines future research directions and fundamental questions, paving the way toward autonomous materials’ discovery in extreme environments. Full article

Background: Information on childhood body composition is critical to understanding children’s growth, development, and long-term health outcomes. BMI metrics, however, have several limitations for assessing and understanding changes in BMI. Therefore, understanding the influence of various body composition factors (covariates) that are linked to, and influence, BMI over time in growing children is important. This study aims to determine sex differences in longitudinal changes in covariates of BMI from 6 to 13 years. Methods: Participants (N = 332, 160 boys 172 girls) from North West Province in South Africa were assessed longitudinally at the following three time-points during their primary years of schooling: Grade 1 (6–7 years); Grade 4 (9–10 years); and Grade 7 (12–13 years). Covariates included: stature (cm); body weight (kg); sub-scapular-, calf-, and triceps skinfolds (mm); body fat percentage (%), relaxed forearm, waist and mid-upper arm circumferences; percentage fat weight; and percentage muscle weight. Correlational analysis and multiple stepwise regression analysis in SPSS analyzed the significance of the contributions of the different covariates to changes in BMI from 6 to 12 years. Results: Different covariates influence BMI in boys and girls at different ages and the covariates also change over time in boys and girls. Weight had the strongest influence on the BMI of boys and girls, although the prediction value decreased over time. Weight and stature were consistently the strongest BMI predictors across all ages in boys. In girls, a broader range of variables influences BMI from a younger age, where slightly higher BMI correlations with fat-related variables emerged, and the percentage of fat weight distribution was a strong influential factor. These findings indicate a more in-depth analysis of BMI to determine sound intervention strategies. Full article

Accurately predicting stock prices is crucial for investment and risk management, but the non-stationarity of the financial market and the complex correlations among stocks pose challenges to traditional models (ARIMA, LSTM, XGBoost), resulting in difficulties in effectively capturing dynamic patterns and limited prediction accuracy. To this end, this paper proposes the Financial Spatio-Temporal Graph Attention Network (FSTGAT), with the following core innovations: temporal modelling through gated causal convolution to avoid future information leakage and capture long- and short-term fluctuations; enhanced spatial correlation learning by adopting the Dynamic Graph Attention Mechanism (GATv2) that incorporates industry information; designing the Multiple-Input-Multiple-Output (MIMO) architecture of industry grouping for the simultaneous learning of intra-group synergistic and inter-group influence; symmetrically fusing spatio-temporal modules to construct a hierarchical feature extraction framework. Experiments in the commercial banking and metals sectors of the New York Stock Exchange (NYSE) show that FSTGAT significantly outperforms the benchmark model, especially in high-volatility scenarios, where the prediction error is reduced by 45–69%, and can accurately capture price turning points. This study confirms the potential of graph neural networks to model the structure of financial interconnections, providing an effective tool for stock forecasting in non-stationary markets, and its forecasting accuracy and industry correlation capturing ability can support portfolio optimization, risk management improvement and supply chain decision guidance. Full article

Background/Objectives: Vitamin D deficiency is prevalent in higher-latitude regions and among older adults, and has been linked to depressive symptoms and cognitive decline, although the neurobiological link remains unclear. Brain-derived neurotrophic factor (BDNF) may be a key modulator and mediator of vitamin D-related neuroprotection. Methods: Selected databases (2009–2025) were searched for specific studies reporting vitamin D exposure, BDNF, and mood or cognitive outcomes. Risk of bias was appraised with RoB 2, Newcastle–Ottawa Scale or SYRCLE. Results: Thirteen studies were included. High-dose vitamin D improves mood primarily when levels are low. Supplementation of at least 2000 IU/day for 12 weeks reduced BDI scores by 1.7–7.6 points and increased BDNF levels by ~7%. Each 1 ng/mL increase in 25(OH)D levels decreased the likelihood of depressive symptoms, especially when BDNF levels were high. In animal studies vitamin D increases hippocampal BDNF and reverses stress-induced depressive behavioral deficits. Adequate vitamin D intake is associated with improved cognitive performance and a dose-dependent increase in BDNF. Each 10 ng/mL increase in 25(OH)D was associated with a 0.6-point increase in MMSE scores and a 15% increase in serum BDNF. Low vitamin D status in children may predict cognitive decline. Animal studies have shown that supplementation with 500–10,000 IU/kg for at least 3 weeks increased hippocampal BDNF and improved biochemical markers of aging. Conclusions: Vitamin D supplementation may support mood and cognition via BDNF modulation, especially in people with insufficient vitamin D levels (<30 ng/mL), but long-term, adequately powered studies with objective tools are required. Full article

Type 2 diabetes (T2D) is a chronic metabolic disorder requiring effective management to avoid complications. Metformin is a first-line drug agent and is routinely prescribed for the control of glycemia, but its underlying dynamics are complicated and not fully quantified. This paper formulates a control-oriented and interpretable mathematical model that integrates metformin dynamics into a classic beta-cell–insulin–glucose (BIG) regulation system. The paper’s applicability to theoretical and clinical settings is enhanced by rigorous mathematical analysis, which guarantees the model is globally bounded, well-posed, and biologically meaningful. One of the key features of the study is its global stability analysis using Lyapunov functions, which demonstrates the asymptotic stability of critical equilibrium points under realistic physiological constraints. These findings support the predictive reliability of the model in explaining long-term glycemic regulation. Bifurcation analysis also clarifies the dynamic interplay between glucose production and utilization by identifying parameter thresholds that signify transitions between homeostasis and pathological states. Residual analysis, which detects Gaussian-distributed errors, underlines the robustness of the fitting process and suggests possible refinements by including temporal effects. Sensitivity analysis highlights the predominant effect of the initial dose of metformin on long-term glucose regulation and provides practical guidance for optimizing individual treatment. Furthermore, changing the two considered metformin parameters from their optimal values—altering the dose by ±50% and the decay rate by ±20%—demonstrates the flexibility of the model in simulating glycemic responses, confirming its adaptability and its potential for optimizing personalized treatment strategies. Full article

Background: Enkephalins are endogenous opioid peptides that modulate cardiovascular and renal function and are overexpressed in patients with acute heart failure (AHF). Although biologically active enkephalins lack a favorable biomarker profile, their stable surrogate proenkephalin 119–159 (PENK) appears to display prognostic value in AHF settings. The aim of the present study was to evaluate the role of point-of-care (POC) PENK in predicting mortality and worsening renal function (WRF) in patients presenting to the emergency department (ED) with AHF. Methods: In this single-center observational study, 107 patients presenting to the ED with AHF were prospectively enrolled. We measured PENK levels upon ED presentation with a commercially available POC immunoassay and investigated their association with WRF within 48 h and all-cause mortality during a 1-year follow-up. Results: The patients had a mean age of 72 ± 13 years, and 58% were men. Moreover, 62% had acutely decompensated chronic heart failure (HF), 24% had pulmonary edema, 9% had cardiogenic shock, and 5% had right HF. The median PENK levels were 111 [60–193] pmol/L. PENK was independently associated with WRF (adjusted OR, 95% CI: 15.4 [2.0–120.2]; p = 0.009), with levels of ≥90.5 pmol/L identified as the optimal cut-off for predicting WRF (AUC: 0.690; p < 0.001). PENK was also an independent predictor of short- and long-term all-cause mortality, with an optimal cut-off of ≥95.8 pmol/L (AUC for 30-day, 90-day, and 1-year mortality: 0.717, 0.723, and 0.724, respectively; all p < 0.001). Conclusions: In patients presenting to the ED with AHF, POC PENK may serve as an early prognostic marker of WRF and short- and long-term mortality. Full article

In the modern era of HIV treatment, people co-infected with HIV and HBV still face poor liver outcomes, including liver fibrosis, liver cirrhosis, and hepatocellular carcinoma. We investigated baseline characteristics and long-term liver function outcomes in 435 people living with HIV and HBV co-infection, focusing on HCC-associated point mutations (PMs) and PreS region deletion mutations. PMs were present in 72.9% of participants and were associated with male predominance, lower HBV genotype C prevalence, reduced HBV DNA and HBeAg levels, and higher HBsAg and HBeAb positivity. However, PMs did not significantly impact liver function or fibrosis progression over six years of ART follow-up. In contrast, PreS deletions were found in 21.8% of cases and stratified into PreS1, PreS2, and PreS1+2 deletions. PreS2 and PreS1+2 deletions were linked to older age, higher HBsAg and AFP levels, elevated liver enzymes, and lower platelet counts. These groups also exhibited significantly worse liver fibrosis markers (APRI and FIB-4), with PreS2 deletions consistently showing the highest values throughout the follow-up. Despite the initial improvement with ART, patients with PreS2 and PreS1+2 deletions maintained higher fibrosis and cirrhosis risks over six years. In summary, while PMs were not predictive of liver disease progression, PreS deletion mutations (especially in the PreS2 region) were associated with poorer liver outcomes, indicating their potential as biomarkers for fibrosis risk in co-infected individuals with long-term ART. Full article