Applied Sciences

15 pages, 9492 KB

Open AccessArticle

Evaluation of Quality Weld Deposit on Different Types of Rails

by Michal Bucko, Lucie Krejci, Ivo Hlavaty, Jindrich Kozak, Petr Mohyla, Ondrej Sopr, Petr Samek and Martina Gree

Appl. Sci. 2026, 16(2), 690; https://doi.org/10.3390/app16020690 - 9 Jan 2026

Abstract

Welding of high-carbon rail steels is widely applied in railway maintenance to restore worn rail surfaces and extend service life. However, the weldability of these steels is limited by their high carbon content and susceptibility to brittle microstructures in the heat-affected zone. This [...] Read more.

Welding of high-carbon rail steels is widely applied in railway maintenance to restore worn rail surfaces and extend service life. However, the weldability of these steels is limited by their high carbon content and susceptibility to brittle microstructures in the heat-affected zone. This paper evaluates the quality of weld deposits applied to different grades of railway rails (UIC 1100, UIC 900A, and UIC HSH) using submerged arc welding (SAW) and flux-cored arc welding (FCAW) technologies with various filler materials. Weld quality was assessed through macrostructural examination, HV30 hardness measurements, and microstructural analysis. The results show that inappropriate combinations of filler materials and welding parameters lead to excessive hardness and martensitic structures, which are undesirable for in-service performance. In contrast, selected multi-layer welding procedures produced bainitic or tempered microstructures with favourable hardness distributions. Based on the experimental results, optimal welding procedures and filler material combinations for rail renovation are proposed. Full article

(This article belongs to the Special Issue Design and Optimization of Computer Numerical Control (CNC) Machining Technology)

► Show Figures

Figure 1

20 pages, 4244 KB

Open AccessArticle

UG-Net: An Unsupervised-Guided Framework for Railway Foreign Object Detection

by Zhuowen Tian and Jinbai Zou

Appl. Sci. 2026, 16(2), 689; https://doi.org/10.3390/app16020689 - 9 Jan 2026

Abstract

Foreign object intrusion severely threatens railway safety. Existing methods struggle with open-set categories, high annotation costs, and poor label-efficient generalization. To address these issues, we propose UG-Net, an unsupervised-guided label-efficient detection framework. The core idea is a two-stage strategy: first, a masked autoencoder [...] Read more.

Foreign object intrusion severely threatens railway safety. Existing methods struggle with open-set categories, high annotation costs, and poor label-efficient generalization. To address these issues, we propose UG-Net, an unsupervised-guided label-efficient detection framework. The core idea is a two-stage strategy: first, a masked autoencoder (MAE) learns “normality” priors from unlabeled data and generates a spatial attention mask via a deep feature difference strategy; then, this mask is fused as a fourth channel into a lightweight YOLOv8n detector. This approach effectively alleviates reliance on manual annotations. On a self-constructed railway dataset, UG-Net achieved 94.56% mAP@0.5 using only 200 labeled samples, significantly outperforming the YOLOv8n baseline (86.91%). The framework provides a label-efficient solution for industrial anomaly detection. Full article

(This article belongs to the Section Computing and Artificial Intelligence)

► Show Figures

Figure 1

14 pages, 3529 KB

Open AccessArticle

Study on Non-Contact Defect Detection Using the Laser Ultrasonic Method for Friction Stir-Welded Cu–Al Dissimilar Material Joints

by Kazufumi Nomura, Shogo Ishifuro and Satoru Asai

Appl. Sci. 2026, 16(2), 688; https://doi.org/10.3390/app16020688 - 9 Jan 2026

Abstract

Ensuring friction stir welding (FSW) joint quality typically relies on ultrasonic testing (UT) and radiographic testing (RT), but achieving complete coverage is challenging, and echo-based defect discrimination becomes difficult in dissimilar joints. Laser ultrasonics is a promising non-contact technique that remotely assesses weld [...] Read more.

Ensuring friction stir welding (FSW) joint quality typically relies on ultrasonic testing (UT) and radiographic testing (RT), but achieving complete coverage is challenging, and echo-based defect discrimination becomes difficult in dissimilar joints. Laser ultrasonics is a promising non-contact technique that remotely assesses weld quality and provides high spatial resolution at the generation and detection points. This study establishes a laser-ultrasonic method for defect detection in dissimilar Cu–Al FSW joints. Slit-like artificial defects (0.1–2.5 mm deep in 5 mm thick plates) were introduced at the Al-side interface of specimens fabricated with an Al-offset tool. Experiments and numerical simulations were used to evaluate wave modes and irradiation configurations, focusing on intensity-attenuation ratios of specific wave types, including longitudinal and Rayleigh waves. On the non-slit surface, attenuation of reflected longitudinal waves enabled detection of defects ≥0.5 mm deep. On the slit surface, Rayleigh-wave attenuation allowed identification of defects as shallow as 0.1 mm, although slit-side irradiation may be less practical during joining. These results demonstrate that defect identification in dissimilar materials can be achieved by evaluating wave-intensity attenuation rather than relying solely on the presence of reflected echoes, suggesting potential for implementing laser ultrasonics in in-process monitoring of FSW joints. Full article

(This article belongs to the Special Issue Industrial Applications of Laser Ultrasonics)

► Show Figures

Figure 1

1 pages, 145 KB

Open AccessCorrection

Correction: Aurin et al. Remote Sensing of CDOM, CDOM Spectral Slope, and Dissolved Organic Carbon in the Global Ocean. Appl. Sci. 2018, 8, 2687

by Dirk Aurin, Antonio Mannino and David J. Lary

Appl. Sci. 2026, 16(2), 687; https://doi.org/10.3390/app16020687 - 9 Jan 2026

Abstract

In the original publication [...] Full article

24 pages, 2366 KB

Open AccessArticle

Hybrid Modeling of Wave Propagation in a 1D Bar: Integrating Peridynamics and Finite Element Methods for Enhanced Dynamic Analysis

by Laxman Khanal, Mijia Yang and Evan J. Pineda

Appl. Sci. 2026, 16(2), 686; https://doi.org/10.3390/app16020686 - 8 Jan 2026

Abstract

This study analyzes a hybrid computational framework that combines peridynamics (PD) and the finite element (FE) method to model wave propagation in a one-dimensional bar, focusing on their integration for enhanced accuracy and efficiency. The analysis investigates PD’s ability to capture non-local interactions [...] Read more.

This study analyzes a hybrid computational framework that combines peridynamics (PD) and the finite element (FE) method to model wave propagation in a one-dimensional bar, focusing on their integration for enhanced accuracy and efficiency. The analysis investigates PD’s ability to capture non-local interactions in regions near loading points, with computationally efficient coarse discretization in other areas through finite element methods. The dynamic response to symmetric and asymmetric axial loading, including loading and unloading phases, is analyzed through time-dependent external forces, solving displacement, velocity, and acceleration fields at each time step. The effects of PD-specific parameters, such as the horizon size, and the FE–PD node spacing size ratios on the performance of the hybrid model in wave propagation are investigated. Additionally, the study examines the von Neumann stability for PD to ensure stability and reliability, offering a robust framework for integrating PD and FE in dynamic analyses. Full article

(This article belongs to the Special Issue Advances in AI and Multiphysics Modelling)

15 pages, 5995 KB

Open AccessArticle

A Multi-Scale Soft-Thresholding Attention Network for Diabetic Retinopathy Recognition

by Xin Ma, Linfeng Sui, Ruixuan Chen, Taiyo Maeda and Jianting Cao

Appl. Sci. 2026, 16(2), 685; https://doi.org/10.3390/app16020685 - 8 Jan 2026

Abstract

Diabetic retinopathy (DR) is a major cause of preventable vision loss, and its early detection is essential for timely clinical intervention. However, existing deep learning-based DR recognition methods still face two fundamental challenges: substantial lesion-scale variability and significant background noise in retinal fundus [...] Read more.

Diabetic retinopathy (DR) is a major cause of preventable vision loss, and its early detection is essential for timely clinical intervention. However, existing deep learning-based DR recognition methods still face two fundamental challenges: substantial lesion-scale variability and significant background noise in retinal fundus images. To address these issues, we propose a lightweight framework named Multi-Scale Soft-Thresholding Attention Network (MSA-Net). The model integrates three components: (1) parallel multi-scale convolutional branches to capture lesions of different spatial sizes; (2) a soft-thresholding attention module to suppress noise-dominated responses; and (3) hierarchical feature fusion to enhance cross-layer representation consistency. A squeeze-and-excitation module is further incorporated for channel recalibration. On the APTOS 2019 dataset, MSA-Net achieves 97.54% accuracy and 0.991 AUC-ROC for binary DR recognition. We further evaluate five-class DR grading on APTOS2019 with 5-fold stratified cross-validation, achieving 82.71 ± 1.25% accuracy and 0.8937 ± 0.0142 QWK, indicating stable performance for ordinal severity classification. With only 4.54 M parameters, MSA-Net remains lightweight and suitable for deployment in resource-constrained DR screening environments. Full article

► Show Figures

Figure 1

21 pages, 3734 KB

Open AccessArticle

DHAG-Net: A Small Object Semantic Segmentation Network Integrating Edge Supervision and Dense Hybrid Dilated Convolution

by Qin Qin, Huyuan Shen, Qing Wang, Qun Yang and Xin Wang

Appl. Sci. 2026, 16(2), 684; https://doi.org/10.3390/app16020684 - 8 Jan 2026

Abstract

Small-object semantic segmentation remains challenging in urban driving scenes due to limited pixel occupancy, blurred boundaries, and the constraints imposed by lightweight deployment. To address these issues, this paper presents a lightweight semantic segmentation framework that enhances boundary awareness and contextual representation while [...] Read more.

Small-object semantic segmentation remains challenging in urban driving scenes due to limited pixel occupancy, blurred boundaries, and the constraints imposed by lightweight deployment. To address these issues, this paper presents a lightweight semantic segmentation framework that enhances boundary awareness and contextual representation while maintaining computational efficiency. The proposed method integrates an edge-supervised boundary gating module to emphasize object boundaries, an efficient multi-scale context aggregation strategy to mitigate scale variation, and a lightweight feature enhancement mechanism for effective feature fusion. Edge supervision is introduced as an auxiliary regularization signal and does not require additional manual annotations. Extensive experiments conducted on multiple benchmark datasets, including Cityscapes, CamVid, PASCAL VOC 2012, and IDDA, demonstrate that the proposed framework consistently improves segmentation performance, particularly for small-object categories, while preserving a favorable balance between accuracy and efficiency. Full article

(This article belongs to the Section Computing and Artificial Intelligence)

► Show Figures

Figure 1

18 pages, 19599 KB

Open AccessArticle

A Semi-Supervised Approach to Microseismic Source Localization with Masked Pre-Training and Residual Convolutional Autoencoder

by Zhe Wang, Xiangbo Gong, Qiao Cheng, Zhuo Xu, Zhiyu Cao and Xiaolong Li

Appl. Sci. 2026, 16(2), 683; https://doi.org/10.3390/app16020683 - 8 Jan 2026

Abstract

Microseismic monitoring is extensively applied in hydraulic fracturing and mineral extraction, with accurate event localization being a critical component. Recently, deep learning approaches have shown promise for microseismic event localization; however, most of these supervised methods depend on large, labeled datasets, which are [...] Read more.

Microseismic monitoring is extensively applied in hydraulic fracturing and mineral extraction, with accurate event localization being a critical component. Recently, deep learning approaches have shown promise for microseismic event localization; however, most of these supervised methods depend on large, labeled datasets, which are costly and challenging to acquire. To mitigate this issue, we propose a semi-supervised approach based on a residual convolutional autoencoder (RCAE) for automated microseismic localization, designed to leverage limited labeled data effectively and improve source localization accuracy even with small sample sizes. Our method employs pre-training by masking and reconstructing unlabeled seismic records, while integrating residual connections within the encoder to enhance feature extraction from seismic signals. This enables high localization accuracy with minimal labeled data, resulting in significant cost savings. Experimental results indicate that our method surpasses purely supervised approaches on both a 2D salt dome model and a 3D homogeneous half-space model, validating its effectiveness in microseismic localization. Further comparisons with baseline models highlight the method’s advantages, providing an innovative solution for improving cost-efficiency in practical applications. Full article

(This article belongs to the Special Issue Machine Learning Applications in Seismology: 2nd Edition)

17 pages, 1356 KB

Open AccessArticle

Syngas Production and Heavy Metals Distribution During the Gasification of Biomass from Phytoremediation Poplar Prunings: A Case Study

by Enrico Paris, Debora Mignogna, Cristina Di Fiore, Pasquale Avino, Domenico Borello, Luigi Iannitti, Monica Carnevale and Francesco Gallucci

Appl. Sci. 2026, 16(2), 682; https://doi.org/10.3390/app16020682 - 8 Jan 2026

Abstract

The present study investigates the potential of poplar (Populus spp.) biomass from phytoremediation plantations as a feedstock for downdraft fixed bed gasification. The biomass was characterized in terms of moisture, ash content, elemental composition (C, H, N, O), and calorific values (HHV [...] Read more.

The present study investigates the potential of poplar (Populus spp.) biomass from phytoremediation plantations as a feedstock for downdraft fixed bed gasification. The biomass was characterized in terms of moisture, ash content, elemental composition (C, H, N, O), and calorific values (HHV and LHV), confirming its suitability for thermochemical conversion. Gasification tests yielded a volumetric syngas production of 1.79 Nm³ kg⁻¹ biomass with an average composition of H₂ 14.58 vol%, CO 16.68 vol%, and CH₄ 4.74 vol%, demonstrating energy content appropriate for both thermal and chemical applications. Alkali and alkaline earth metals (AAEM), particularly Ca (273 mg kg⁻¹) and Mg (731 mg kg⁻¹), naturally present enhanced tar reforming and promoted reactive gas formation, whereas heavy metals such as Cd (0.27 mg kg⁻¹), Pb (0.02 mg kg⁻¹), and Bi (0.01 mg kg⁻¹) were detected only in trace amounts, posing minimal environmental risk. The results indicate that poplar pruning residues from phytoremediation sites can be a renewable and sustainable energy resource, transforming a waste stream into a process input. In this perspective, the integration of soil remediation with syngas production constitutes a tangible model of circular economy, based on the efficient use of resources through the synergy between environmental remediation and the valorization and sustainable management of marginal biomass—i.e., pruning residues—generating environmental, energetic, and economic benefits along the entire value chain. Full article

► Show Figures

Figure 1

18 pages, 8939 KB

Open AccessArticle

Research on the Temporal and Spatial Evolution Patterns of Vegetation Cover in Zhaogu Mining Area Based on kNDVI

by Congying Liu, Hebing Zhang, Zhichao Chen, He Qin, Xueqing Liu and Yiheng Jiao

Appl. Sci. 2026, 16(2), 681; https://doi.org/10.3390/app16020681 - 8 Jan 2026

Abstract

Extensive coal mining activities can exert substantial negative impacts on surface ecosystems. Vegetation indices are widely recognized as effective indicators of land ecological conditions and provide valuable insights into long-term ecological changes in mining areas. In this study, the Zhaogu mining area of [...] Read more.

Extensive coal mining activities can exert substantial negative impacts on surface ecosystems. Vegetation indices are widely recognized as effective indicators of land ecological conditions and provide valuable insights into long-term ecological changes in mining areas. In this study, the Zhaogu mining area of the Jiaozuo Coalfield was selected as the study site. Using the Google Earth Engine (GEE) platform, the Kernel Normalized Difference Vegetation Index (kNDVI) was constructed to generate a vegetation dataset covering the period from 2010 to 2024. The temporal dynamics and future trends of vegetation coverage were analyzed using Theil–Sen median trend analysis, the Mann–Kendall test, the Hurst index, and residual analysis. Furthermore, the relative contributions of climatic factors and human activities to vegetation changes were quantitatively assessed. The results indicate that: (1) vegetation coverage in the Zhaogu mining area exhibits an overall improving trend, affecting approximately 77.1% of the study area, while slight degradation is mainly concentrated in the southeastern region, accounting for about 15.2%; (2) vegetation dynamics are predominantly characterized by low and relatively low fluctuations, covering approximately 78.5% of the region, whereas areas with high fluctuations are limited and mainly distributed in zones with intensive mining activities; although the current vegetation trend is generally increasing, future projections suggest a potential decline in approximately 55.8% of the area; and (3) vegetation changes in the Zhaogu mining area are jointly influenced by climatic factors and human activities, with climatic factors promoting vegetation growth in approximately 70.6% of the study area, while human activities exert inhibitory effects in about 24.2%, particularly in regions affected by mining operations and urban expansion. Full article

► Show Figures

Figure 1

22 pages, 338 KB

Open AccessReview

Can ChatGPT Replace the Teacher in Assessment? A Review of Research on the Use of Large Language Models in Grading and Providing Feedback

by Marcin Jukiewicz and Michał Wyrwa

Appl. Sci. 2026, 16(2), 680; https://doi.org/10.3390/app16020680 - 8 Jan 2026

Abstract

This article presents a systematic review of empirical research on the use of large language models (LLMs) for automated grading of student work and providing feedback. The study aimed to determine the extent to which generative artificial intelligence models, such as ChatGPT, can [...] Read more.

This article presents a systematic review of empirical research on the use of large language models (LLMs) for automated grading of student work and providing feedback. The study aimed to determine the extent to which generative artificial intelligence models, such as ChatGPT, can replace teachers in the assessment process. The review was conducted in accordance with PRISMA guidelines and predefined inclusion criteria; ultimately, 42 empirical studies were included in the analysis. The results of the review indicate that the effectiveness of LLMs in grading is varied. These models perform well on closed-ended tasks and short-answer questions, often achieving accuracy comparable to human evaluators. However, they struggle with assessing complex, open-ended, or subjective assignments that require in-depth analysis or creativity. The quality of the prompts provided to the model and the use of detailed scoring rubrics significantly influence the accuracy and consistency of the grades generated by LLMs. The findings suggest that LLMs can support teachers by accelerating the grading process and delivering rapid feedback at scale, but they cannot fully replace human judgment. The highest effectiveness is achieved in hybrid assessment systems that combine AI-driven automatic grading with teacher oversight and verification. Full article

(This article belongs to the Special Issue Artificial Intelligence in Education: Latest Advances and Prospects)

► Show Figures

Figure 1

24 pages, 1272 KB

Open AccessSystematic Review

How Extended Reality Is Shaping Smart Cities: A Systematic Literature Review

by Marina Ricci, Nicola Mosca, Moh Rafik and Maria Di Summa

Appl. Sci. 2026, 16(2), 679; https://doi.org/10.3390/app16020679 - 8 Jan 2026

Abstract

XR technologies enhance the sustainable development of urban areas by merging digital and physical worlds. In smart city contexts, XR has been applied in miscellaneous ways, from supporting urban planning and design through immersive visualization, to improving traffic and navigation services via real-time [...] Read more.

XR technologies enhance the sustainable development of urban areas by merging digital and physical worlds. In smart city contexts, XR has been applied in miscellaneous ways, from supporting urban planning and design through immersive visualization, to improving traffic and navigation services via real-time overlays, and to enhancing public safety and emergency response through simulation and situational support. However, the literature does not clearly categorize XR application domains in smart cities, interaction methods, and types of sensory feedback. This study presents an SLR reported in accordance with the PRISMA 2020 guidelines. We included 92 studies published between 2009 and 2024, proposing a classification of application domains, interaction modalities, and sensory feedback. We searched Scopus, Web of Science, and IEEE Xplore using predefined search terms and eligibility criteria. This review offers a comprehensive overview of nearly 20 years of XR research in smart cities, highlighting established practices and guiding future application development and research directions. Full article

(This article belongs to the Special Issue Extended Reality (XR): Recent Advances and Emerging Trends)

► Show Figures

Figure 1

20 pages, 3259 KB

Open AccessArticle

Green Transportation Planning for Smart Cities: Digital Twins and Real-Time Traffic Optimization in Urban Mobility Networks

by Marek Lis and Maksymilian Mądziel

Appl. Sci. 2026, 16(2), 678; https://doi.org/10.3390/app16020678 - 8 Jan 2026

Abstract

This paper proposes a comprehensive framework for integrating Digital Twins (DT) with real-time traffic optimization systems to enhance urban mobility management in Smart Cities. Using the Pobitno Roundabout in Rzeszów as a case study, we established a calibrated microsimulation model (validated via the [...] Read more.

This paper proposes a comprehensive framework for integrating Digital Twins (DT) with real-time traffic optimization systems to enhance urban mobility management in Smart Cities. Using the Pobitno Roundabout in Rzeszów as a case study, we established a calibrated microsimulation model (validated via the GEH statistic) that serves as the core of the proposed Digital Twin. The study goes beyond static scenario analysis by introducing an Adaptive Inflow Metering (AIM) logic designed to interact with IoT sensor data. While traditional geometrical upgrades (e.g., turbo-roundabouts) were analyzed, simulation results revealed that geometrical changes alone—without dynamic control—may fail under peak load conditions (resulting in LOS F). Consequently, the research demonstrates how the DT framework allows for the testing of “Software-in-the-Loop” (SiL) solutions where Python-based algorithms dynamically adjust inflow parameters to prevent gridlock. The findings confirm that combining physical infrastructure changes with digital, real-time optimization algorithms is essential for achieving sustainable “green transport” goals and reducing emissions in congested urban nodes. Full article

(This article belongs to the Special Issue Green Transportation and Pollution Control)

► Show Figures

Figure 1

21 pages, 4583 KB

Open AccessArticle

Magnitude Scaling and Real-Time Performance Assessment for an ElarmS-Based Early Warning System: The Case of the 2025 Silivri (Istanbul) Earthquake (Mw = 6.2)

by Emrah Budakoğlu, Süleyman Tunç, Berna Tunç and Deniz Çaka

Appl. Sci. 2026, 16(2), 677; https://doi.org/10.3390/app16020677 - 8 Jan 2026

Abstract

This study develops and evaluates a regionally calibrated magnitude scaling and early warning framework based on the ElarmS–EPIC algorithm using the 23 April 2025 Silivri (Istanbul) Earthquake (Mw = 6.2) scenario. A comprehensive dataset comprising the mainshock and its aftershocks was used to [...] Read more.

This study develops and evaluates a regionally calibrated magnitude scaling and early warning framework based on the ElarmS–EPIC algorithm using the 23 April 2025 Silivri (Istanbul) Earthquake (Mw = 6.2) scenario. A comprehensive dataset comprising the mainshock and its aftershocks was used to derive local regression relationships between earthquake magnitude (Mw) and the peak displacement amplitude (

P d

) and predominant period (

T_{p}^{m a x}

) parameters. Replay simulations were conducted to assess real-time performance, and the results of the regional models were compared with those of the default EPIC configuration. The results indicate that the

P d

-based magnitude estimation model produces faster and more stable results than the

T_{p}^{m a x}

-based approach, significantly improving accuracy and operational reliability. The region-specific

P d

–Mw scaling provided higher consistency with catalog magnitudes compared to the default EPIC relationships. The calculated distribution of warning times shows that the system can provide actionable warning times of 5–9 s in districts near the epicenter (e.g., Silivri, Avcılar, Beylikdüzü) and 20–50 s in more distant districts and city centers (e.g., Kadıköy, Pendik, Bursa, Sakarya). These values demonstrate that a regionally optimized early warning system can provide critical decision-making time for automatic safety systems and emergency responses in the densely populated Marmara Region. Overall, this study emphasizes the importance of regional calibration in improving earthquake early warning (EEW) performance in Türkiye. The findings show that the success of EEW systems depends on station density, network latency, data transmission speed, processing capacity, and algorithmic optimization. The proposed

P d

-based regional framework provides a scientifically robust and operationally applicable foundation for future EEW implementations in Istanbul and the Marmara Region. Full article

(This article belongs to the Section Earth Sciences)

► Show Figures

Figure 1

21 pages, 2160 KB

Open AccessArticle

Integrating Ski Material Properties with Skier Dynamics for a Personalization Algorithm

by Paulina Maślanka, Ryszard Korycki and Tomasz Józefiak

Appl. Sci. 2026, 16(2), 676; https://doi.org/10.3390/app16020676 - 8 Jan 2026

Abstract

The aim of this study is to present an integrated approach to ski personalization by combining a detailed material and geometric characterization of ski structure with real-time on-slope dynamic performance data. Longitudinal and torsional stiffness were selected as the basic material parameters for [...] Read more.

The aim of this study is to present an integrated approach to ski personalization by combining a detailed material and geometric characterization of ski structure with real-time on-slope dynamic performance data. Longitudinal and torsional stiffness were selected as the basic material parameters for a wide range of skis with different application profiles. Acceleration graphs for turns were synchronized with video footage recorded by a drone, which provided real-time monitoring of the skiers’ actual course and facilitated the precise correlation between kinematic data and on-slope performance. To estimate style similarity between different skiers, the Pearson correlation coefficient was utilized due to its robust mathematical properties. Comprehensive analyses helped to finally formulate a multistage ski personalization algorithm. Full article

19 pages, 2055 KB

Open AccessArticle

Detection of Mobile Phone Use While Driving Supported by Artificial Intelligence

by Gustavo Caiza, Adriana Guanuche and Carlos Villafuerte

Appl. Sci. 2026, 16(2), 675; https://doi.org/10.3390/app16020675 - 8 Jan 2026

Abstract

Driver distraction, particularly mobile phone use while driving, remains one of the leading causes of road traffic incidents worldwide. In response to this issue and leveraging recent technological advances and increased access to intelligent systems, this research presents the development of an application [...] Read more.

Driver distraction, particularly mobile phone use while driving, remains one of the leading causes of road traffic incidents worldwide. In response to this issue and leveraging recent technological advances and increased access to intelligent systems, this research presents the development of an application running on an intelligent embedded architecture for the automatic detection of mobile phone use by drivers, integrating computer vision, inertial sensing, and edge computing. The system, based on the YOLOv8n model deployed on a Jetson Xavier NX 16Gb—Nvidia, combines real-time inference with an inertial activation mechanism and cloud storage via Firebase Firestore, enabling event capture and traceability through a lightweight web-based HMI interface. The proposed solution achieved an overall accuracy of 81%, an inference rate of 12.8 FPS, and an average power consumption of 8.4 W, demonstrating a balanced trade-off between performance, energy efficiency, and thermal stability. Experimental tests under diverse driving scenarios validated the effectiveness of the system, with its best performance recorded during daytime driving—83.3% correct detections—attributed to stable illumination and enhanced edge discriminability. These results confirm the feasibility of embedded artificial intelligence systems as effective tools for preventing driver distraction and advancing intelligent road safety. Full article

(This article belongs to the Section Electrical, Electronics and Communications Engineering)

20 pages, 16754 KB

Open AccessArticle

GSA-cGAN: A Geospatial-Aware Conditional Wasserstein Generative Adversarial Network for Mineral Resources Interpolation

by Hosang Han and Jangwon Suh

Appl. Sci. 2026, 16(2), 674; https://doi.org/10.3390/app16020674 - 8 Jan 2026

Abstract

In the context of mineral resource exploration, spatial prediction must cope with heterogeneous, non-normal data distributions and limited sampling. While conventional geostatistics and standard machine learning provide baseline estimates, they often suffer from excessive smoothing or fail to capture continuous spatial dependencies. This [...] Read more.

In the context of mineral resource exploration, spatial prediction must cope with heterogeneous, non-normal data distributions and limited sampling. While conventional geostatistics and standard machine learning provide baseline estimates, they often suffer from excessive smoothing or fail to capture continuous spatial dependencies. This study proposes a geospatially aware Wasserstein conditional Generative Adversarial Network (GSA-cGAN) to complement existing workflows for multivariate mineral interpolation. The framework augments a baseline cGAN with WGAN-GP for stable adversarial training, CoordConv to encode absolute spatial coordinates and Self-Attention to capture long-range spatial dependencies. Eight model configurations were trained on 272 samples from a mineralized zone in the Taebaek Mountains, Korea, and strictly benchmarked against Ordinary/Universal Kriging and multivariate machine learning baselines (Random Forest, XGBoost). Under the adopted experimental design, the full GSA-cGAN achieved the lowest test root mean squared error and highest coefficient of determination, demonstrating a significant performance improvement over the baselines. Furthermore, distribution analysis confirmed that the model effectively overcomes the smoothing limitations of regression-based methods, generating high-resolution 10 m × 10 m maps that preserve statistical variance, hotspot anomalies, and complex spatial patterns. The results indicate that deep generative models can serve as practical decision-support tools for identifying drilling targets and prioritizing follow-up exploration in geologically complex settings. Full article

(This article belongs to the Special Issue Geographic Visualization: Evaluation and Monitoring of Geohazards—3rd Edition)

► Show Figures

Figure 1

22 pages, 1253 KB

Open AccessArticle

Ammonia/Ethane Blends Combustion and Oxidation: Experimental and Numerical Study

by Ksenia N. Osipova, Vladislav V. Matyushkov, Alexander V. Drakon, Stanislav A. Trubachev and Andrey G. Shmakov

Appl. Sci. 2026, 16(2), 673; https://doi.org/10.3390/app16020673 - 8 Jan 2026

Abstract

Ammonia is considered as a promising hydrogen carrier and a carbon-free fuel. Methods for improving ammonia combustion characteristics often involve its co-firing with more reactive fuels (natural gas, biofuels, etc.). Among the natural gas components, ethane is second most abundant. Therefore, the development [...] Read more.

Ammonia is considered as a promising hydrogen carrier and a carbon-free fuel. Methods for improving ammonia combustion characteristics often involve its co-firing with more reactive fuels (natural gas, biofuels, etc.). Among the natural gas components, ethane is second most abundant. Therefore, the development of detailed chemical–kinetic mechanisms that accurately consider the interactions between ammonia and each component of natural gas is very important. Such mechanisms must be based on experimental data obtained under a wide range of conditions. In this work, NH₃/C₂H₆/O₂/Ar blends were studied in JSR (φ = 0.5–2.0, p = 1 atm, τ = 1 s, T = 800–1300 K) and in a shock tube (p = 7.3–8.6 atm, T = 1260–1590 K). Additionally, the structure of premixed flames was investigated (φ = 0.8–1.2, p = 1–5 atm). Eleven recently published detailed chemical–kinetic mechanisms were tested. The model Shrestha-2025 was updated to achieve better agreement with the entire set of experimental data. The effect of p and φ on intermediate species concentration was analyzed. Ammonia and ethane consumption pathways were also examined. Full article

(This article belongs to the Special Issue Ammonia and Hydrogen as Energy Carriers: Challenges and Applications)

18 pages, 1346 KB

Open AccessArticle

ALEX: Adaptive Log-Embedded Extent Layer for Low-Amplification SQLite Writes on Flash Storage

by Youngmi Baek and Jung Kyu Park

Appl. Sci. 2026, 16(2), 672; https://doi.org/10.3390/app16020672 - 8 Jan 2026

Abstract

Efficient metadata and page management are essential for sustaining database performance on modern flash-based storage. However, conventional SQLite configurations—rollback journal and WAL—often trigger excessive small writes and frequent synchronization events, leading to high write amplification and degraded tail latency, particularly on UFS and [...] Read more.

Efficient metadata and page management are essential for sustaining database performance on modern flash-based storage. However, conventional SQLite configurations—rollback journal and WAL—often trigger excessive small writes and frequent synchronization events, leading to high write amplification and degraded tail latency, particularly on UFS and NVMe devices. This study introduces ALEX (Adaptive Log-Embedded Extent Layer), a lightweight VFS-level extension that coalesces scattered 4 KB page updates into sequential, page-aligned extents while embedding compact log records for recovery. The proposed design reduces redundant writes through in-memory page deduplication, minimizes fdatasync()frequency by flushing multi-page extents, and preserves full SQLite compatibility. We evaluate ALEX on both Linux NVMe SSDs and Android UFS storage under controlled workloads. Results show that ALEX significantly lowers write amplification, reduces sync counts, and improves p95–p99 write latency compared with baseline SQLite modes. The approach consistently achieves near-sequential write patterns without modifying SQLite internals. These findings demonstrate that lightweight extent-based coalescing can provide substantial efficiency gains for embedded and mobile database systems, offering a practical direction for enhancing SQLite performance on flash devices. Full article

► Show Figures

Figure 1

Journal Description

Applied Sciences

Latest Articles

Journal Menu

Journal Browser

Highly Accessed Articles

Latest Books

E-Mail Alert

News

Topics

Conferences

Special Issues

Topical Collections

Further Information

Guidelines

MDPI Initiatives

Follow MDPI