alexandria engineering journal最新文献

{"title":"MDMB-YOLO: A liquid crystal display defect detection method using Multi-Differential Fusion and multi-branch feature pyramid","authors":"Shi Luo ,&nbsp;Xiaoyue Chen ,&nbsp;Sheng Zheng ,&nbsp;Yuxin Zhao","doi":"10.1016/j.aej.2026.01.035","DOIUrl":"10.1016/j.aej.2026.01.035","url":null,"abstract":"<div><div>To address the challenges of coexisting defects at multiple scales and the tendency for small defects to be missed in LCD defect detection, this paper proposes a novel detection algorithm. The method first designs a Multi-Differential Fusion Module (MDFM), which enhances sensitivity to small defects (especially dot defects) by integrating multiple differential sensing strategies. Second, a multi-branch fusion efficient feature pyramid network (MFEFPN) is constructed. Leveraging a multi-branch structure and efficient fusion mechanisms, this network effectively mitigates information loss and feature interference issues inherent in traditional feature pyramid networks. To further balance accuracy and computational efficiency, we designed an Adaptive Shared Lightweight Detection Head (ASLD), which maintains excellent detection accuracy while significantly reducing the number of parameters and computational complexity (GFLOPs) through a parameter-sharing mechanism. Additionally, geometric constraint terms are incorporated into the loss function to further enhance the localization capability of defect boundaries. Experimental results show that the proposed MDMB-YOLO achieves an accuracy of 85.2%, with a 4.4% improvement in accuracy, a 3.3% improvement in recall rate, a 2.8% improvement in mAP50, and a 0.9% improvement in mAP50-95 compared to the baseline model. The number of parameters and GFLOPs were reduced by 23.3% and 8%, respectively, compared to the baseline model, indicating that this approach offers both accuracy and efficiency advantages in LCD defect detection tasks. The dataset used in this study has been publicly released, and we encourage its use for related research in accordance with the platform’s terms at:<span><span>https://aistudio.baidu.com/dataset/detail/358247/settings</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":7484,"journal":{"name":"alexandria engineering journal","volume":"137 ","pages":"Pages 521-536"},"PeriodicalIF":6.8,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146074949","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"ST-Former: A transformer-based temporal-scene fusion-driven auditory experience analysis model","authors":"Yanxi Shen ,&nbsp;Anran Li","doi":"10.1016/j.aej.2025.12.068","DOIUrl":"10.1016/j.aej.2025.12.068","url":null,"abstract":"<div><div>In the digital age, audio experience analysis and personalized recommendations have become core requirements for intelligent interaction. However, traditional methods struggle to simultaneously address audio temporal dynamic capture and scene semantic context fusion. This study proposes the ST-Former model, constructing an integrated architecture of “temporal dynamic capture - multimodal scene fusion - cross-task collaborative modeling.” Through the collaborative design of TCN, Transformer, and a multimodal scene perception module (integrating CLIP and DINOv2), it efficiently solves audio sentiment classification and multi-label scene recognition tasks. Experimental results show that ST-Former achieves 85.2% accuracy and 82.4% Macro-F1 in the IEMOCAP sentiment classification task, and 81.5% mAP@1, 83.7% mAP@5, and 84.6% mAP@10 in the FSD50K multi-label scene recognition task, significantly outperforming existing state-of-the-art models in all metrics. Ablation experiments validate the collaborative value of the core modules: removing TCN, the scene perception module, or Transformer resulted in a 5.6%, 6.4%, and 14.8% decrease in overall performance, respectively. This model, through an innovative combination of temporal modeling and multimodal semantic fusion, provides an efficient new paradigm for multi-task audio analysis and lays the technical foundation for the practical deployment of personalized recommendation systems.</div></div>","PeriodicalId":7484,"journal":{"name":"alexandria engineering journal","volume":"137 ","pages":"Pages 1-13"},"PeriodicalIF":6.8,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145993578","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Indoor mobile robot localization system based on ORB-SLAM3 and multi-sensor fusion","authors":"Siyong Fu,&nbsp;Qinghua Zhao,&nbsp;Qiuxiang Tao,&nbsp;Hesheng Liu,&nbsp;Qing Wang,&nbsp;Danjuan Liu","doi":"10.1016/j.aej.2026.01.029","DOIUrl":"10.1016/j.aej.2026.01.029","url":null,"abstract":"<div><div>Indoor localization is a fundamental capability for autonomous mobile robots operating in complex indoor environments, where visual degradation, sensor noise, and rotational motion often lead to accumulated drift. This paper presents MFL-SLAM, a practical multi-sensor fusion localization system that extends the ORB-SLAM3 framework by explicitly integrating wheel odometry with visual–inertial SLAM. Unlike conventional visual–inertial approaches, MFL-SLAM employs an Extended Kalman Filter (EKF) to tightly fuse Inertial Measurement Unit (IMU) and wheel odometry, effectively compensating for vibration-induced inertial drift and wheel slippage during rotational motion. The EKF fusion output is then incorporated as a prior in a nonlinear optimization back-end together with RGB-D visual constraints, enabling accurate and globally consistent pose estimation. Extensive experiments demonstrate that MFL-SLAM achieves a 47.3 % reduction in relative pose error compared to ORB-SLAM3 and reduces the average localization error to 0.29 m, outperforming ORB-SLAM2 and LIO-SAM across small- and large-scale indoor environments. These results indicate that the proposed fusion strategy provides a robust and deployable solution for reliable indoor mobile robot localization.</div></div>","PeriodicalId":7484,"journal":{"name":"alexandria engineering journal","volume":"137 ","pages":"Pages 194-205"},"PeriodicalIF":6.8,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146036659","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Corrigendum to “Guardian-AI: A novel deep learning based deepfake detection model in images” [Alex. Eng. J. 126 (2025) 507–514]","authors":"Hadeel Alsolai ,&nbsp;Khalid Mahmood ,&nbsp;Asma Alshuhail ,&nbsp;Achraf Ben Miled ,&nbsp;Mohammed Alqahtani ,&nbsp;Abdulrhman Alshareef ,&nbsp;Fouad Shoie Alallah ,&nbsp;Bandar M. Alghamdi","doi":"10.1016/j.aej.2026.02.007","DOIUrl":"10.1016/j.aej.2026.02.007","url":null,"abstract":"","PeriodicalId":7484,"journal":{"name":"alexandria engineering journal","volume":"138 ","pages":"Page 36"},"PeriodicalIF":6.8,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146186315","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Fracture evolution and anisotropic mechanical properties of layered rock based on discrete element modeling and experimental study","authors":"Minglang Zou ,&nbsp;Yan Zhang ,&nbsp;Shaojun Li ,&nbsp;Tianbin Li ,&nbsp;Guoqiang Zhu ,&nbsp;Yining Zhang","doi":"10.1016/j.aej.2026.01.042","DOIUrl":"10.1016/j.aej.2026.01.042","url":null,"abstract":"<div><div>Since numerous geotechnical activities crossing layered rock masses, a comprehensive understanding of their mechanical behavior is crucial for engineering stability assessment. This paper establishes transversely isotropic numerical models with multiple bedding angles using PFC3D. After calibrating the meso-parameters to verify model validity, multi-confining pressure triaxial compression simulations are conducted. The results show that both peak strength and elastic modulus exhibit a “U” shaped with bedding angle increase, divided at 60°; under the same confining pressure, the maximum differences reaching 42.2 % and 26.6 %, respectively; failure characteristics vary significantly with bedding angle, manifesting as axial splitting (0°), shear sliding along bedding planes (30°–60°), and mixed tension-shear failure (90°). The sudden increase in acoustic emission (AE) event count and energy can serve as reliable precursors of rock failure, while the proposed energy competition factor <em>β</em> (ratio of slip energy to bond energy) can effectively characterize rock failure evolution. CT observations indicate that confining pressure suppresses micro-crack initiation and propagation. Tunnel excavation simulations based on discrete element method further demonstrate that bedding angle plays a significant controlling role in surrounding rock deformation patterns and support requirements. The research findings provide important insights for stability assessment and support design in layered rock masses.</div></div>","PeriodicalId":7484,"journal":{"name":"alexandria engineering journal","volume":"137 ","pages":"Pages 443-467"},"PeriodicalIF":6.8,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146074946","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Data sharing in privacy preserving intelligent transportation system based on blockchain and low-rank gradient compression","authors":"Xiaowei Li ,&nbsp;Xin Li ,&nbsp;Xiaoxue Feng","doi":"10.1016/j.aej.2026.02.004","DOIUrl":"10.1016/j.aej.2026.02.004","url":null,"abstract":"<div><div>Intelligent Transportation Systems (ITS) utilize advanced technologies such as the Internet of Things (IoT), Cloud Computing (CC), and Artificial Intelligence (AI) to enhance transportation services. A critical aspect of ITS is secure and efficient data sharing, where ensuring data privacy remains a primary research focus. This study proposes a novel data sharing scheme that integrates federated learning with blockchain technology. The scheme employs low-rank gradient compression and utilizes blockchain for secure training data storage, enabling participants to update a global model collaboratively without a central server while managing data access via smart contracts. Furthermore, it incorporates a lightweight blockchain consensus mechanism for efficient global model aggregation from roadside nodes, meeting the stringent efficiency and real-time demands of connected vehicle environments. Experimental results demonstrate the scheme’s feasibility and superior performance, showing a 10% reduction in communication overhead compared to existing methods, while maintaining robustness against malicious vehicle activities.</div></div>","PeriodicalId":7484,"journal":{"name":"alexandria engineering journal","volume":"138 ","pages":"Pages 152-160"},"PeriodicalIF":6.8,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146186310","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Mathematical framework for robust autonomous driving via stochastic graph optimization and CVaR policies","authors":"Xinghua Liu ,&nbsp;Quanrong Fang","doi":"10.1016/j.aej.2026.01.037","DOIUrl":"10.1016/j.aej.2026.01.037","url":null,"abstract":"<div><div>Autonomous driving under stochastic sensor failures poses a mathematically challenging problem of learning with partially observed, dynamically structured data. Existing approaches often lack rigorous mechanisms to guarantee robustness, particularly in the presence of missing features and uncertain communication reliability. To overcome these limitations, we propose a mathematically grounded unified optimization framework that integrates V2X-based cooperative sensing, graph-regularized feature imputation, and risk-aware decision generation. The pipeline is optimized via a hybrid strategy combining stochastic gradient descent, variance reduction, and the Alternating Direction Method of Multipliers (ADMM)-based consensus, ensuring stable convergence for both convex and non-convex components. Empirical validation shows that our framework reduces perception error by 18.7 %, improves decision consistency by 15.3 %, and accelerates convergence by 1.8 × compared with state-of-the-art baselines under stochastic sensor failures across diverse driving environments. Beyond these empirical gains, the explicit incorporation of reliability-weighted graph embeddings and Conditional Value-at-Risk (CVaR)-based optimization provides theoretical robustness guarantees against rare but high-risk events. Unlike risk-neutral formulations, CVaR minimizes the expected tail loss, making it particularly suitable for safety-critical autonomous driving systems where uncertainty and risk interact. Overall, the proposed method delivers both mathematical rigor and practical deployability, offering a concise and extensible blueprint for robust and reliable optimization in distributed autonomous driving systems.</div></div>","PeriodicalId":7484,"journal":{"name":"alexandria engineering journal","volume":"138 ","pages":"Pages 128-140"},"PeriodicalIF":6.8,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146186276","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Innovative structure for energy harvesting: Layered PVDF/TPU nanofibers with synergistic piezoelectric, dielectric, and ferroelectric enhancement","authors":"Germein Magdy ,&nbsp;Safaa Elkhoby ,&nbsp;Asmaa Hegazy ,&nbsp;Mahmoud Salama ,&nbsp;Aya Hamed ,&nbsp;Nader Shehata ,&nbsp;Ishac Kandas ,&nbsp;Ahmed H. Hassanin","doi":"10.1016/j.aej.2026.02.009","DOIUrl":"10.1016/j.aej.2026.02.009","url":null,"abstract":"<div><div>The fabrication process is critical for optimizing the piezoelectricity of PVDF/TPU nanofibers. This work illustrates the merits of layer-by-layer (LBL) electrospinning in optimizing the piezoelectric behavior of PVDF/TPU nanofibers for energy harvesting. In comparison with solution blending (SB), LBL electrospinning promotes molecular orientation and β-phase crystallization, which contributes to improved piezoelectric activity. The 8 LBL sample showed the optimal mechanical and electrical characteristics, with the highest output voltage (116.27 ± 5.4 V/mm) at applied force 0.5 N, along with piezoelectric coefficient (d₃₃ = 38.52 ± 0.92 pC/N), the output current (I_p–p = 1.6 μA), and surface charge density (SCD = 0.57 μC/m²). Mechanical characterization revealed that 8 LBL withstood high strains due to slippage of fibers, and their failure was observed at higher stackings (10 and 12 LBL), reducing their efficiency. The dielectric constant increased with stacking from 14.7 ± 0.85 at 6 LBL to 28.9 ± 1.93 at 12 LBL due to the interfacial polarization. While ferroelectric characterization pointed towards the 8 LBL as the optimum architecture with a coercive field of 562 kV/mm and energy storage efficiency of 67.5 %. LBL stacking of nanofibers leads to synergistic triboelectric–piezoelectric coupling along with an optimum mechanical design. The 8 LBL sample exhibited the best properties among all architectures and points towards the potential of LBL electrospun PVDF/TPU nanofibers in high-efficiency flexible and self-powered electronics. Compared with recently reported PVDF/TPU and hybrid PVDF systems, which primarily rely on solution blending or filler incorporation to enhance performance, this work demonstrates an architecture-driven strategy in which controlled layer-by-layer electrospinning enables simultaneous optimization of piezoelectric output, mechanical durability, and interfacial charge dynamics. The results establish LBL electrospun PVDF/TPU nanofibers as a scalable and tunable platform for next-generation self-powered and wearable energy harvesting devices.</div></div>","PeriodicalId":7484,"journal":{"name":"alexandria engineering journal","volume":"138 ","pages":"Pages 81-95"},"PeriodicalIF":6.8,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146186278","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Numerical treatments of (IntVarFrac) order partial differential equations for cancer tumor disease based on non-singular kernel","authors":"N.H. Sweilam ,&nbsp;S.M. Al–Mekhlafi ,&nbsp;A. Ahmed ,&nbsp;E.M. Abo-Eldahab ,&nbsp;Nehaya R. Alsenaideh","doi":"10.1016/j.aej.2026.01.015","DOIUrl":"10.1016/j.aej.2026.01.015","url":null,"abstract":"<div><div>Cancer remains one of the most challenging medical conditions, requiring sophisticated mathematical models to accurately describe its dynamics and treatment responses. Traditional integer-order differential equations often fail to capture the complexities of tumor growth, immune system interactions, and therapeutic effects. In this study, we propose an advanced cancer tumor model based on integer-variable-fractional order partial differential equations with nonsingular kernels. This model incorporates memory dependent properties to more effectively represent the biological behavior of tumor progression under chemotherapy. Using the Atangana–Baleanu Caputo fractional operator, we extend the classical system of coupled partial differential equations to an IntVarFrac framework. The theoretical analysis of the proposed model is presented, and a numerical scheme based on the Atangana–Baleanu fractional Newton polynomial method is employed to obtain approximate solutions. Numerical simulations illustrate the influence of fractional-order parameters on tumor dynamics and the effectiveness of chemotherapy treatment. The results demonstrate that the inclusion of memory effects through fractional derivatives provides a more accurate. Although the model has not yet been validated with clinical tumor data, it captures key tumor–immune dynamics reported in the literature, supporting its biological relevance.</div></div>","PeriodicalId":7484,"journal":{"name":"alexandria engineering journal","volume":"137 ","pages":"Pages 299-311"},"PeriodicalIF":6.8,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146074934","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Efficient human pose estimation in complex coal mining scenes via Keypoint Partitioning Adaptive Convolution","authors":"Jin Wu ,&nbsp;Huaping Zhou ,&nbsp;Xiangrui Meng ,&nbsp;Tao Wu","doi":"10.1016/j.aej.2026.01.041","DOIUrl":"10.1016/j.aej.2026.01.041","url":null,"abstract":"<div><div>Human pose estimation (HPE) is crucial for underground mining safety, but it suffers from uneven brightness, occlusions from dense equipment, complex backgrounds, and limited computational resources. To address these challenges, we propose a novel Keypoint-Adaptive Convolutional Network (KAnet) for accurate miner pose estimation. KAnet integrates our newly proposed content-adaptive convolution method called Keypoint Partitioning Adaptive Convolution (KAconv), which adaptively partitions feature maps based on semantic similarity and generates region-specific dynamic filters. This design enables the model to handle complex and variable spatial information distribution effectively. Additionally, we introduce an Attention-Based Cross-Layer Feature Fusion (ACFF) module to enhance multi-scale feature fusion and improve robustness against occlusion and illumination variations. To further optimize model efficiency, we present the Pruning-guided Adaptive Filtering Knowledge Distillation (PAF-KD), which leverages channel importance ranking for efficient model compression while preserving essential feature representations. We validate the effectiveness of KAnet using the newly developed Miner-Pose dataset, a large-scale dataset of miner poses in coal mines. Experimental results demonstrate that KAnet outperforms current state-of-the-art methods in both accuracy and robustness in complex mining scenarios.</div></div>","PeriodicalId":7484,"journal":{"name":"alexandria engineering journal","volume":"137 ","pages":"Pages 312-328"},"PeriodicalIF":6.8,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146074936","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}