Search Results (840)

This paper presents a novel data-driven framework for real-time State of Charge (SOC) estimation in lithium-ion battery systems using a data-driven Prognostics and Health Management (PHM) approach. The method leverages an optimized bidirectional Long Short-Term Memory (Bi-LSTM) network, trained with enhanced datasets filtered via exponentially weighted moving averages (EWMAs) and refined through SHAP-based feature attribution. Compared against a Extended Kalman Filter (EKF) and Unscented Kalman Filter (UKF) across ten diverse drive cycles, the proposed model consistently achieved superior performance, with mean absolute errors (MAEs) as low as 0.40%, outperforming EKF (0.66%) and UKF (1.36%). The Bi-LSTM model also demonstrated higher R² values (up to 0.9999) and narrower 95% confidence intervals, confirming its precision and robustness. Real-time implementation on embedded platforms yielded inference times of 1.3–2.2 s, validating its deployability for edge applications. The framework’s model-free nature makes it adaptable to other nonlinear, time-dependent systems beyond battery SOC estimation. Full article

Structural collapses caused by accidents or disasters could create unexpected radiation shielding, resulting in sharp gradients within the radiation field. Traditional radiation mapping methods often fail to accurately capture these complex variations, making the rapid and precise localization of radiation sources a significant challenge in emergency response scenarios. To address this issue, based on standard Gaussian process regression (GPR) models that primarily utilize a single Gaussian kernel to reflect the inverse-square law in free space, a novel multi-kernel Gaussian process regression (MK-GPR) model is proposed for high-fidelity radiation mapping in environments with physical obstructions. MK-GPR integrates two additional kernel functions with adaptive weighting: one models the attenuation characteristics of intervening materials, and the other captures the energy-dependent penetration behavior of radiation. To validate the model, gamma-ray distributions in complex, shielded environments were simulated using GEometry ANd Tracking 4 (Geant4). Compared with conventional methods, including linear interpolation, nearest-neighbor interpolation, and standard GPR, MK-GPR demonstrated substantial improvements in key evaluation metrics, such as MSE, RMSE, and MAE. Notably, the coefficient of determination (R²) increased to 0.937. For practical deployment, the optimized MK-GPR model was deployed to an RK-3588 edge computing platform and integrated into a mobile robot equipped with a NaI(Tl) detector. Field experiments confirmed the system’s ability to accurately map radiation fields and localize gamma sources. When combined with SLAM, the system achieved localization errors of 10 cm for single sources and 15 cm for dual sources. These results highlight the potential of the proposed approach as an effective and deployable solution for radiation source investigation in post-disaster environments. Full article

Real-time quality monitoring using molten pool images is a critical focus in researching high-quality, intelligent automated welding. To address interference problems in molten pool images under complex welding scenarios (e.g., reflected laser spots from spatter misclassified as porosity defects) and the limited interpretability of deep learning models, this paper proposes a multi-granularity spatiotemporal representation learning algorithm based on the hybrid enhancement of handcrafted and deep learning features. A MobileNetV2 backbone network integrated with a Temporal Shift Module (TSM) is designed to progressively capture the short-term dynamic features of the molten pool and integrate temporal information across both low-level and high-level features. A multi-granularity attention-based feature aggregation module is developed to select key interference-free frames using cross-frame attention, generate multi-granularity features via grouped pooling, and apply the Convolutional Block Attention Module (CBAM) at each granularity level. Finally, these multi-granularity spatiotemporal features are adaptively fused. Meanwhile, an independent branch utilizes the Histogram of Oriented Gradient (HOG) and Scale-Invariant Feature Transform (SIFT) features to extract long-term spatial structural information from historical edge images, enhancing the model’s interpretability. The proposed method achieves an accuracy of 99.187% on a self-constructed dataset. Additionally, it attains a real-time inference speed of 20.983 ms per sample on a hardware platform equipped with an Intel i9-12900H CPU and an RTX 3060 GPU, thus effectively balancing accuracy, speed, and interpretability. Full article

In this study, edge-oxidized graphene oxide (EOGO) was used as an additive in fly ash (FA) geopolymer paste. The effect of EOGO on the properties of the fly ash geopolymer was investigated. EOGO was added to the FA geopolymer at four different percentages (0%, 0.1%, 0.5% and 1%), and the mixture was cured under two different conditions: room curing (~20 °C) and heat curing (~60 °C). To characterize the FA-EOGO geopolymer, multiple laboratory tests were employed, including compressive strength, Free-Free Resonance Column (FFRC), density, water absorption, and setting tests. The FFRC test was used to evaluate the stiffness at small strain (Young’s modulus) via the resonance of the specimen. The mechanical test results showed that the strength and elastic modulus were high during heat curing, and the highest compressive strength and elastic modulus were achieved at 0.1% EOGO. In the physical test, 0.1% EOGO had the highest density and the lowest porosity and water absorption. As a result of the setting time test, as the EOGO content increased, the setting time was shortened. It is concluded that the optimum proportion of EOGO is 0.1% in FA geopolymer paste. Full article

The widespread and persistent occurrence of micropollutants—such as pesticides, pharmaceuticals, heavy metals, personal care products, microplastics, and per- and polyfluoroalkyl substances (PFAS)—has emerged as a critical environmental and public health concern, necessitating the development of highly sensitive, selective, and field-deployable detection technologies. Microfluidic sensors, including biosensors, have gained prominence as versatile and transformative tools for real-time environmental monitoring, enabling precise and rapid detection of trace-level contaminants in complex environmental matrices. Their miniaturized design, low reagent consumption, and compatibility with portable and smartphone-assisted platforms make them particularly suited for on-site applications. Recent breakthroughs in nanomaterials, synthetic recognition elements (e.g., aptamers and molecularly imprinted polymers), and enzyme-free detection strategies have significantly enhanced the performance of these biosensors in terms of sensitivity, specificity, and multiplexing capabilities. Moreover, the integration of artificial intelligence (AI) and machine learning algorithms into microfluidic platforms has opened new frontiers in data analysis, enabling automated signal processing, anomaly detection, and adaptive calibration for improved diagnostic accuracy and reliability. This review presents a comprehensive overview of cutting-edge microfluidic sensor technologies for micropollutant detection, emphasizing fabrication strategies, sensing mechanisms, and their application across diverse pollutant categories. We also address current challenges, such as device robustness, scalability, and potential signal interference, while highlighting emerging solutions including biodegradable substrates, modular integration, and AI-driven interpretive frameworks. Collectively, these innovations underscore the potential of microfluidic sensors to redefine environmental diagnostics and advance sustainable pollution monitoring and management strategies. Full article

With the continuous development of the Internet, people have put forward higher requirements for the stability and availability of the network. Although we constantly strive to take measures to avoid network failures, it is undeniable that network failures are unavoidable. Therefore, in this situation, enhancing the stability and reliability of the network to cope with possible network failures has become particularly crucial. Therefore, researching and developing high fault protection rate intra-domain routing protection schemes has become an important topic and is the subject of this study. This study aims to enhance the resilience and service continuity of networks in the event of failures by proposing innovative routing protection strategies. The existing methods, such as Loop Free Alternative (LFA) and Equal Cost Multiple Paths (ECMP), have some shortcomings in terms of fast fault detection, fault response, and fault recovery processes, such as long fault recovery time, limitations of routing protection strategies, and requirements for network topology. In response to these issues, this article proposes a new routing protection scheme, which is an intra-domain routing protection scheme based on the minimum cross-degree backup path. The core idea of this plan is to find the backup path with the minimum degree of intersection with the optimal path, in order to avoid potential fault areas and minimize the impact of faults on other parts of the network. Through comparative analysis and performance evaluation, this scheme can provide a higher fault protection rate and more reliable routing protection in the network. Especially in complex networks, this scheme has more performance and protection advantages than traditional routing protection methods. The proposed scheme in this paper exhibits a high rate of fault protection across multiple topologies, demonstrating a fault protection rate of 1 in the context of real topology. It performs commendably in terms of path stretch, evidenced by a figure of 1.06 in the case of real topology Ans, suggesting robust path length control capabilities. The mean intersection value is 0 in the majority of the topologies, implying virtually no common edge between the backup and optimal paths. This effectively mitigates the risk of single-point failure. Full article

The automated detection and classification of immature macauba (Acrocomia aculeata) fruits is critical for improving post-harvest processing and quality control. In this study, we present a comparative evaluation of two state-of-the-art YOLO architectures, YOLOv11x and YOLOv12x, trained on the newly constructed VIC01 dataset comprising 1600 annotated images captured under both background-free and natural background conditions. Both models were implemented in PyTorch and trained until the convergence of box regression, classification, and distribution-focal losses. Under an IoU (intersection over union) threshold of 0.50, YOLOv11x and YOLOv12x achieved an identical mean average precision (mAP₅₀) of 0.995 with perfect precision and recall or TPR (true positive rate). Averaged over IoU thresholds from 0.50 to 0.95, YOLOv11x demonstrated superior spatial localization performance (mAP_50–95 = 0.973), while YOLOv12x exhibited robust performance in complex background scenarios, achieving a competitive mAP_50–95. Inference throughput averaged 3.9 ms per image for YOLOv11x and 6.7 ms for YOLOv12x, highlighting a trade-off between speed and architectural complexity. Fused model representations revealed optimized layer fusion and reduced computational overhead (GFLOPs), facilitating efficient deployment. Confusion-matrix analyses confirmed YOLOv11x’s ability to reject background clutter more effectively than YOLOv12x, whereas precision–recall and F1-score curves indicated both models maintain near-perfect detection balance across thresholds. The public release of the VIC01 dataset and trained weights ensures reproducibility and supports future research. Our results underscore the importance of selecting architectures based on application-specific requirements, balancing detection accuracy, background discrimination, and computational constraints. Future work will extend this framework to additional maturation stages, sensor fusion modalities, and lightweight edge-deployment variants. By facilitating precise immature fruit identification, this work contributes to sustainable production and value addition in macauba processing. Full article

This study presents the design and validation of an integrated fire safety system that leverages edge AI, hybrid sensing, and precision suppression to overcome the latency and collateral limitations of conventional smoke detection and sprinkler systems. The proposed platform features a dual-mode sensor array for early fire recognition, motorized ventilation units for rapid smoke extraction, and a 360° directional nozzle for targeted agent discharge using a residue-free clean extinguishing agent. Experimental trials demonstrated an average fire detection time of 5.8 s and complete flame suppression within 13.2 s, with 90% smoke clearance achieved in under 95 s. No false positives were recorded during non-fire simulations, and the system remained fully functional under simulated cloud communication failure, confirming its edge-resilient architecture. A probabilistic risk analysis based on ISO 31000 and NFPA 551 frameworks showed risk reductions of 75.6% in life safety, 58.0% in property damage, and 67.1% in business disruption. The system achieved a composite risk reduction of approximately 73%, shifting the operational risk level into the ALARP region. These findings demonstrate the system’s capacity to provide proactive, energy-efficient, and spatially targeted fire response suitable for high-value infrastructure. The modular design and SmartThings Edge integration further support scalable deployment and real-time system intelligence, establishing a strong foundation for future adaptive fire protection frameworks. Full article

Complex oxides with the general formula Gd₂(Ti_1−xZr_x)₂O₇ are promising candidates for radioactive waste immobilization due to their capacity to withstand radiation by dissipating part of the free energy driving defect creation and phase transitions. In this study, samples with varying zirconium content (xZr = 0.00, 0.15, 0.25, 0.375, 0.56, 0.75, 0.85, 1.00) were synthesized via the sol–gel method and thermally treated at 500 °C to obtain nanosized powders mimicking the defective structure of irradiated materials. Synchrotron-based techniques were employed to investigate their structural properties: High-Resolution X-ray Powder Diffraction (HR-XRPD) was used to assess long-range structure, while Pair Distribution Function (PDF) analysis and Extended X-ray Absorption Fine Structure (EXAFS) spectroscopy provided insights into the local structure. HR-XRPD data revealed that samples with low Zr content (xZr ≤ 0.25) are amorphous. Increasing Zr concentration led to the emergence of a crystalline phase identified as defective fluorite (xZr = 0.375, 0.56). Samples with the highest Zr content (xZr ≥ 0.75) were fully crystalline and exhibited only the fluorite phase. The experimental G(r) functions of the fully crystalline samples in the low r range are suitably fitted by the Weberite structure, mapping the relaxations induced by structural disorder in defective fluorite. These structural insights informed the subsequent EXAFS analysis at the Zr-K and Gd-L₃ edges, confirming the splitting of the cation–cation distances associated with different metal species. Moreover, EXAFS provided a local structural description of the amorphous phases, identifying a consistent Gd-O distance across all compositions. Full article

Handwritten Keyword Spotting (KWS) remains a challenging task, particularly in segmentation-free scenarios where word images must be retrieved and ranked based on their similarity to a query without relying on prior page-level segmentation. Traditional KWS methods primarily focus on visual similarity, often overlooking the underlying semantic relationships between words. In this work, we propose a novel NLP-driven re-ranking approach that refines the initial ranked lists produced by state-of-the-art KWS models. By leveraging semantic embeddings from pre-trained BERT-like Large Language Models (LLMs, e.g., RoBERTa, MPNet, and MiniLM), we introduce a relevance feedback mechanism that improves both verbatim and semantic keyword spotting. Our framework operates in two stages: (1) projecting retrieved word image transcriptions into a semantic space via LLMs and (2) re-ranking the retrieval list using a weighted combination of semantic and exact relevance scores based on pairwise similarities with the query. We evaluate our approach on the widely used George Washington (GW) and IAM collections using two cutting-edge segmentation-free KWS models, which are further integrated into our proposed pipeline. Our results show consistent gains in Mean Average Precision (mAP), with improvements of up to $2.3\%$ (from $94.3\%$ to $96.6\%$) on GW and $3\%$ (from $79.15\%$ to $82.12\%$) on IAM. Even when mAP gains are smaller, qualitative improvements emerge: semantically relevant but inexact matches are retrieved more frequently without compromising exact match recall. We further examine the effect of fine-tuning transformer-based OCR (TrOCR) models on historical GW data to align textual and visual features more effectively. Overall, our findings suggest that semantic feedback can enhance retrieval effectiveness in KWS pipelines, paving the way for lightweight hybrid vision-language approaches in handwritten document analysis. Full article

The electrochemical reduction of carbon monoxide (COER) offers a promising route for generating value-added multi-carbon (C₂₊) products, such as ethanol, but achieving high catalytic performance remains a significant challenge. Herein, we performed comprehensive density functional theory (DFT) computations to evaluate CO-to-ethanol conversion on single metal atoms anchored on graphitic carbon nitride (TM/g–CN). We showed that these metal atoms stably coordinate with edge N sites of g–CN to form active catalytic centers. Screening 20 TM/g–CN candidates, we identified V/g–CN and Zn/g–CN as optimal catalysts: both exhibit low free-energy barriers (<0.50 eV) for the key ^*CO hydrogenation steps and facilitate C–C coupling via an Eley–Rideal mechanism with a negligible kinetic barrier (~0.10 eV) to yield ethanol at low limiting potentials, which explains their superior COER performance. An analysis of d-band centers, charge transfer, and bonding–antibonding orbital distributions revealed the origin of their activity. This work provides theoretical insights and useful guidelines for designing high-performance single-atom COER catalysts. Full article

The entanglement entropy of a random tensor network (RTN) is studied using tools from free probability theory. Random tensor networks are simple toy models that help in understanding the entanglement behavior of a boundary region in the anti-de Sitter/conformal field theory (AdS/CFT) context. These can be regarded as specific probabilistic models for tensors with particular geometry dictated by a graph (or network) structure. First, we introduce a model of RTN obtained by contracting maximally entangled states (corresponding to the edges of the graph) on the tensor product of Gaussian tensors (corresponding to the vertices of the graph). The entanglement spectrum of the resulting random state is analyzed along a given bipartition of the local Hilbert spaces. The limiting eigenvalue distribution of the reduced density operator of the RTN state is provided in the limit of large local dimension. This limiting value is described through a maximum flow optimization problem in a new graph corresponding to the geometry of the RTN and the given bipartition. In the case of series-parallel graphs, an explicit formula for the limiting eigenvalue distribution is provided using classical and free multiplicative convolutions. The physical implications of these results are discussed, allowing the analysis to move beyond the semiclassical regime without any cut assumption, specifically in terms of finite corrections to the average entanglement entropy of the RTN. Full article

Background: Many patients harbor minimal residual disease (MRD)—small clusters of residual tumor cells that survive therapy and evade conventional detection but drive recurrence. Although advances in molecular and computational methods have improved circulating tumor DNA (ctDNA)-based MRD detection, these approaches face challenges: ctDNA shedding fluctuates widely across tumor types, disease stages, and histological features. Additionally, low levels of driver mutations originating from healthy tissues can create background noise, complicating the accurate identification of bona fide tumor-specific signals. These limitations underscore the need for refined technologies to further enhance MRD detection beyond DNA sequences in solid malignancies. Methods: Profiling circulating cell-free mRNA (cfmRNA), which is hyperactive in tumor and non-tumor microenvironments, could address these limitations to inform postoperative surveillance and treatment strategies. This study reported the development of OncoMRD BREAST, a customized, gene signature-informed cfmRNA assay for residual disease monitoring in breast cancer. OncoMRD BREAST introduces several advanced technologies that distinguish it from the existing ctDNA-MRD tests. It builds on the patient-derived gene signature for capturing tumor activities while introducing significant upgrades to its liquid biopsy transcriptomic profiling, digital scoring systems, and tracking capabilities. Results: The OncoMRD BREAST test processes inputs from multiple cutting-edge biomarkers—tumor and non-tumor microenvironment—to provide enhanced awareness of tumor activities in real time. By fusing data from these diverse intra- and inter-cellular networks, OncoMRD BREAST significantly improves the sensitivity and reliability of MRD detection and prognosis analysis, even under challenging and complex conditions. In a proof-of-concept real-world pilot trial, OncoMRD BREAST’s rapid quantification of potential tumor activity helped reduce the risk of incorrect treatment strategies, while advanced predictive analytics contributed to the overall benefits and improved outcomes of patients. Conclusions: By tailoring the assay to individual tumor profiles, we aimed to enhance early identification of residual disease and optimize therapeutic decision-making. OncoMRD BREAST is the world’s first and only gene signature-powered test for monitoring residual disease in solid tumors. Full article

Conventional adhesively bonded joints, such as single-lap, curved-lap, wavy-lap, double-lap, stepped-lap, and scarf joints, are widely used for aerospace, automotive, and medical applications. These adhesively bonded joints exhibit different load transfer mechanisms and stress distributions within adhesive layers, which depend primarily on their geometries and mechanical properties of bonded materials. As such, joint geometry and material properties play a critical role in determining the capability of the joints to withstand high loads, resist fatigue, and absorb energy under impact loading. This paper investigates the effects of geometry and material dissimilarity on the performance of both conventional bonded and interlocking joints under tensile loading based on the information available in the literature. In addition, bonding and load transfer mechanisms were analysed in detail. It was found that stress concentration often occurs at free edges of the adhesive layer due to geometric discontinuities, while most of the load is carried by these regions rather than its centre. Sharp corners further intensify resulting stresses, thereby increasing the risk of joint failure. Adhesives typically resist shear loads better than peel loads, and stiffness mismatches between adherents induce an asymmetric stress distribution. Nonetheless, similar materials promote symmetric load sharing. Among conventional joints, scarf joints provide the most uniform load distribution. In interlocking joints such as dovetail, T-slot, gooseneck, and elliptical types, the outward bending of the female component under tension can lead to mechanical failure. Full article

Obstructive sleep apnea (OSA) is a highly prevalent sleep-related breathing disorder, primarily characterized by recurrent episodes of upper airway obstruction during sleep. Individuals affected by OSA are at increased risk for a variety of adverse health outcomes, particularly neurocognitive impairments and cardiovascular complications, highlighting the clinical significance of this condition. A defining feature of OSA is intermittent hypoxemia, which contributes to the excessive production of reactive oxygen species (ROS) and the subsequent development of oxidative stress. The primary objective of this narrative review was to comprehensively investigate the intricate mechanisms of oxidative stress and elucidate their complex interplay in the development and progression of OSAS. Subsequently, we examined the current literature to identify the most promising biomarkers and pharmacological treatments related to OSA and oxidative stress. We found that biomarkers of oxidative stress have shown potential in assessing disease severity and tracking individual responses to therapy. However, none have yet to be incorporated into standard clinical practice. Continuous positive airway pressure (CPAP) is the gold standard treatment for OSA. Nevertheless, antioxidant therapy has emerged as a potential adjunctive approach that may help address residual dysfunctions not fully resolved by CPAP alone. Both the use of oxidative stress biomarkers and antioxidant-based therapies require further validation through robust clinical studies before they can be routinely implemented in clinical settings. Full article