Search Results (1,163)

Gravity and gravity gradient data serve as fundamental inputs for geophysical resource exploration and geological structure analysis. However, traditional denoising methods—including wavelet transforms, moving averages, and low-pass filtering—exhibit signal loss and limited adaptability under complex, non-stationary noise conditions. To address these challenges, this study proposes an improved U-Net deep learning framework that integrates multi-scale feature extraction and attention mechanisms. Furthermore, a Laplace consistency constraint is introduced into the loss function to enhance denoising performance and physical interpretability. Notably, the datasets used in this study are generated by the authors, involving simulations of subsurface prism distributions with realistic density perturbations (±20% of typical rock densities) and the addition of controlled Gaussian noise (5%, 10%, 15%, and 30%) to simulate field-like conditions, ensuring the diversity and physical relevance of training samples. Experimental validation on these synthetic datasets and real field datasets demonstrates the superiority of the proposed method over conventional techniques. For noise levels of 5%, 10%, 15%, and 30% in test sets, the improved U-Net achieves Peak Signal-to-Noise Ratios (PSNR) of 59.13 dB, 52.03 dB, 48.62 dB, and 48.81 dB, respectively, outperforming wavelet transforms, moving averages, and low-pass filtering by 10–30 dB. In multi-component gravity gradient denoising, our method excels in detail preservation and noise suppression, improving Structural Similarity Index (SSIM) by 15–25%. Field data tests further confirm enhanced identification of key geological anomalies and overall data quality improvement. In summary, the improved U-Net not only delivers quantitative advancements in gravity data denoising but also provides a novel approach for high-precision geophysical data preprocessing. Full article

This study investigates the fabrication and residual stress behavior of a 0.5 wt.% graphene-reinforced AA2195 aluminum matrix composite, developed for advanced aerospace structural applications. The composite was synthesized via squeeze casting, followed by a multi-pass hot rolling process and subsequent T8 heat treatment. The evolution of residual stress was systematically examined after each rolling pass and during thermal treatments. The successful incorporation of graphene into the matrix was confirmed through Energy-Dispersive Spectroscopy (EDS) analysis. Residual stress measurements after each pass revealed a progressive increase in compressive stress, reaching a maximum of −68 MPa after the fourth hot rolling pass. Prior to the fifth pass, a solution treatment at 530 °C was performed to dissolve coarse precipitates and relieve internal stresses. Cold rolling during the fifth pass reduced the compressive residual stress to −40 MPa, and subsequent artificial aging at 180 °C for 48 h further decreased it to −23 MPa due to recovery and stress relaxation mechanisms. Compared to the unreinforced AA2195 alloy in the T8 condition, which exhibited a tensile residual stress of +29 MPa, the graphene-reinforced composite in the same condition retained a compressive residual stress of −23 MPa. This represents a net improvement of 52 MPa, highlighting the composite’s superior capability to retain compressive residual stress. The presence of graphene significantly influenced the stress distribution by introducing thermal expansion mismatch and acting as a barrier to dislocation motion. Overall, the composite demonstrated enhanced residual stress characteristics, making it a promising candidate for lightweight, fatigue-resistant aerospace components. Full article

This systematic review examines state-of-the-art artificial intelligence applications in energy systems, assessing their performance, real-world deployments and transformative potential. Guided by PRISMA 2020, we searched Web of Science, IEEE Xplore, ScienceDirect, SpringerLink, and Google Scholar for English-language studies published between January 2015 and January 2025 that reported novel AI uses in energy, empirical results, or significant theoretical advances and passed peer review. After title–abstract screening and full-text assessment, it was determined that 129 of 3000 records met the inclusion criteria. The methodological quality, reproducibility and real-world validation were appraised, and the findings were synthesised narratively around four critical themes: reinforcement learning (35 studies), multi-agent systems (28), planning under uncertainty (25), and AI for resilience (22), with a further 19 studies covering other areas. Notable outcomes include DeepMind-based reinforcement learning cutting data centre cooling energy by 40%, multi-agent control boosting virtual power plant revenue by 28%, AI-enhanced planning slashing the computation time by 87% without sacrificing solution quality, battery management AI raising efficiency by 30%, and machine learning accelerating hydrogen catalyst discovery 200,000-fold. Across domains, AI consistently outperformed traditional techniques. The review is limited by its English-only scope, potential under-representation of proprietary industrial work, and the inevitable lag between rapid AI advances and peer-reviewed publication. Overall, the evidence positions AI as a pivotal enabler of cleaner, more reliable, and efficient energy systems, though progress will depend on data quality, computational resources, legacy system integration, equity considerations, and interdisciplinary collaboration. No formal review protocol was registered because this study is a comprehensive state-of-the-art assessment rather than a clinical intervention analysis. Full article

Structural damage in bridges is typically a local phenomenon. When a vehicle passes over the damage location, it induces a local response, which is highly sensitive to the damage. The interaction between the bridge and moving vehicle is a non-stationary time-varying process. The local damage can be accurately identified by analyzing the time-varying characteristics of the bridge response subjected to a moving vehicle. Synchrosqueezing transform, a reassignment method used to sharpen time–frequency representations, offers an effective tool to decompose the non-stationary signal into distinct components. This paper proposes a novel method based on multi-synchrosqueenzing transform to extract the time-varying characteristics of the vehicle–bridge interaction systems for bridge structural health monitoring. A vehicle–bridge interaction model is built to simulate the bridge under moving vehicles. Different damage scenarios of concrete bridges have been simulated. The effect of bridge damage parameters, the vehicle speed, the road surface roughness on the time-varying characteristics of the vehicle–bridge interaction system is studied. Numerical and experimental results demonstrate that the proposed method efficiently and accurately extracts the time-varying features of the vehicle–bridge interaction system, which could serve as potential indicators of structural damage in bridges. Full article

Treatment of Mycobacterium tuberculosis requires multi-drug regimens, and resistance to any individual antibiotic can compromise outcomes. For slow-growing organisms like M. tuberculosis, rapid detection of resistance-conferring mutations enables timely initiation of effective therapy. Conversely, confirming wild-type status in resistance-associated genes supports confidence in standard regimens. We developed an amplicon-based next generation sequencing (amplicon tNGS) assay on the Illumina platform targeting eight genes linked to resistance to isoniazid, rifampin, ethambutol, pyrazinamide, and fluoroquinolones. Sequencing results were analyzed using a custom bioinformatics pipeline. Forty-seven samples were used for assay development, and 37 additional samples underwent post-implementation clinical validation. Compared to whole genome sequencing (WGS), amplicon tNGS demonstrated 97.7% sensitivity, 98.9% specificity, and 98.7% overall accuracy for variant detection in targeted regions. Resistance prediction showed 79.3% concordance with WGS; discrepancies were primarily due to mutations outside of target regions. Among post-implementation samples, 27/37 passed quality metrics for all targets, with 95.7% concordance between amplicon tNGS results and final susceptibility results. This assay is now in use in our laboratory and offers significantly faster turnaround than both WGS and phenotypic methods on cultured isolates, enabling more rapid, informed treatment decisions for tuberculosis patients. Full article

The rapid progress of Large Language Models (LLMs) has greatly improved natural language tasks like code generation, boosting developer productivity. However, challenges persist. Generated code often appears “pseudo-correct”—passing functional tests but plagued by inefficiency or redundant structures. Many models rely on outdated methods like greedy selection, which trap them in local optima, limiting their ability to explore better solutions. We propose AnnCoder, a multi-agent framework that mimics the human “try-fix-adapt” cycle through closed-loop optimization. By combining the exploratory power of simulated annealing with the targeted evolution of genetic algorithms, AnnCoder balances wide-ranging searches and local refinements, dramatically increasing the likelihood of finding globally optimal solutions. We speculate that traditional approaches may struggle due to narrow optimization focuses. AnnCoder addresses this by introducing dynamic multi-criteria scoring, weighing functional correctness, efficiency (e.g., runtime/memory), and readability. Its adaptive temperature control dynamically modulates the cooling schedule, slowing cooling when solutions are diverse to encourage exploration, then accelerating convergence as they stabilize. This design elegantly avoids the pitfalls of earlier models by synergistically combining global exploration with local optimization capabilities. After conducting thorough experiments with multiple LLMs analyses across four problem-solving and program synthesis benchmarks—AnnCoder showcased remarkable code generation capabilities—HumanEval 90.85%, MBPP 90.68%, HumanEval-ET 85.37%, and EvalPlus 84.8%. AnnCoder has outstanding advantages in solving general programming problems. Moreover, our method consistently delivers superior performance across various programming languages. Full article

Multi-agent large language models promise flexible, modular architectures for delivering personalized educational content. Drawing on a pilot randomized controlled trial with middle school students (n = 23), we introduce a two-agent GPT-4 framework in which a Profiler agent infers learner-specific preferences and a Rewrite agent dynamically adapts science passages via an explicit message-passing protocol. We implement structured system and user prompts as inter-agent communication schemas to enable real-time content adaptation. The results of an ordinal logistic regression analysis hinted that students may be more likely to prefer texts aligned with their profile, demonstrating the feasibility of multi-agent system-driven personalization and highlighting the need for additional work to build upon this pilot study. Beyond empirical validation, we present a modular multi-agent architecture detailing agent roles, communication interfaces, and scalability considerations. We discuss design best practices, ethical safeguards, and pathways for extending this framework to collaborative agent networks—such as feedback-analysis agents—in K-12 settings. These results advance both our theoretical and applied understanding of multi-agent LLM systems for personalized learning. Full article

In the field of modern optical communication, radar signal processing and optical sensors, true time delay technology, as a key means of signal processing, can achieve the accurate control of the time delay of optical signals. This study presents a novel design that integrates a 2 × 2 Multi-Mode Interference (MMI) structure with a Mach–Zehnder modulator on a silicon nitride–lithium niobate (SiN-LiNbO₃) heterogeneous integrated optical platform. This configuration enables the selective interruption of optical wave paths. The upper path passes through an ultralow-loss Archimedes’ spiral waveguide delay line made of silicon nitride, where the five spiral structures provide delays of 10 ps, 20 ps, 40 ps, 80 ps, and 160 ps, respectively. In contrast, the lower path is straight through, without introducing an additional delay. By applying an electrical voltage, the state of the SiN-LiNbO₃ switch can be altered, facilitating the switching and reconfiguration of optical paths and ultimately enabling the combination of various delay values. Simulation results demonstrate that the proposed optical true delay line achieves a discrete, adjustable delay ranging from 10 ps to 310 ps with a step size of 10 ps. The delay loss is less than 0.013 dB/ps, the response speed reaches the order of ns, and the 3 dB-EO bandwidth is broader than 67 GHz. In comparison to other optical switches optical true delay lines in terms of the parameters of delay range, minimum adjustable delay, and delay loss, the proposed optical waveguide digital adjustable true delay line, which is based on an optical switch and an Archimedes’ spiral structure, has outstanding advantages in response speed and delay loss. Full article

LiDAR point clouds, as direct carriers of 3D spatial information, comprehensively record the geometric features and spatial topological relationships of object surfaces, providing intelligent systems with rich 3D scene representation capability. However, current point cloud semantic segmentation methods primarily extract features through operations such as convolution and pooling, yet fail to adequately consider sparse features that significantly influence the final results of point cloud-based scene perception, resulting in insufficient feature representation capability. To address these problems, a sparse feature dynamic graph convolutional neural network, abbreviated as SFDGNet, is constructed in this paper for LiDAR point clouds of complex scenes. In the context of this paper, sparse features refer to feature representations in which only a small number of activation units or channels exhibit significant responses during the forward pass of the model. First, a sparse feature regularization method was used to motivate the network model to learn the sparsified feature weight matrix. Next, a split edge convolution module, abbreviated as SEConv, was designed to extract the local features of the point cloud from multiple neighborhoods by dividing the input feature channels, and to effectively learn sparse features to avoid feature redundancy. Finally, a multi-neighborhood feature fusion strategy was developed that combines the attention mechanism to fuse the local features of different neighborhoods and obtain global features with fine-grained information. Taking S3DIS and ScanNet v2 datasets, we evaluated the feasibility and effectiveness of SFDGNet by comparing it with six typical semantic segmentation models. Compared with the benchmark model DGCNN, SFDGNet improved overall accuracy $\left(OA\right)$, mean accuracy $\left(mAcc\right)$, mean intersection over union $\left(mIoU\right)$, and $sparsity$ by $1.8\%$, $3.7\%, 3.5\%$, and $85.5\%$ on the S3DIS dataset, respectively. The $mIoU$ on the ScanNet v2 validation set, $mIoU$ on the test set, and $sparsity$ were improved by $3.2\%, 7.0\%$, and $54.5\%$, respectively. Full article

The output energy of Laser Active Optical Systems (LAOSs), in which image brightness is amplified within the laser-active medium, is always higher than the input energy. This contrasts with conventional optical systems (OSs). As a result, a LAOS enables the creation of laser beams with tailored energy distribution across the aperture, making them ideal for material processing applications. This concept was first successfully implemented using metal vapor lasers as the gain medium. In these systems, material processing was achieved by using a laser beam that either carried the required energy profile or the image of the object itself. Later, other laser media were utilized for LAOSs, including barium vapor, strontium vapor, excimer XeCl lasers, and solid-state media. Additionally, during the development of these systems, several modifications were introduced. For example, Space-Time Light Modulators (STLMs) and CCD cameras were incorporated, along with the use of multipass amplifiers, disk-shaped or thin-disk (TD) solid-state laser amplifiers, and other advancements. These techniques have significantly expanded the range of power, energy, pulse durations, and operating wavelengths. Currently, TD laser amplifiers and STLMs based on Digital Light Processor (DLP) technology or Digital Micromirror Devices (DMDs) enhance the potential to develop LAOS devices for Subtractive and Additive Technologies (ST, AT), applicable in both macromachining (cutting, welding, drilling) and micro-nano processing. This review presents comparable characteristics and requirements for these various LAOS applications. Full article

This study examines the microstructural evolution, mechanical properties, and wear behavior of medium-carbon dual-phase steel (AISI 1040) processed via Multi-Axis Compression (MAC). The DP steel was produced through inter-critical annealing at 745 °C, followed by MAC at 500 °C, resulting in a refined grain microstructure. Optical micrographs confirmed the presence of ferrite and martensite phases after annealing, with significant grain refinement observed following MAC. The average grain size decreased from 66 ± 4 μm to 18 ± 1 μm after nine MAC passes. Mechanical testing revealed substantial improvements in hardness (from 145 ± 9 HV to 298 ± 18 HV) and ultimate tensile strength (from 557 ± 33 MPa to 738 ± 44 MPa), attributed to strain hardening and the Hall–Petch effect. Fractographic analysis revealed a ductile failure mode in the annealed sample, while DP0 and DP9 exhibited a mixed fracture mode. Both DP0 and DP9 samples demonstrated superior wear resistance compared to the annealed sample. However, the DP9 sample exhibited slightly lower wear resistance than DP0, likely due to the fragmentation of martensite induced by high accumulated strain, which could act as crack initiation sites during sliding wear. Furthermore, wear resistance was significantly enhanced due to the combined effects of the DP structure and Severe Plastic Deformation (SPD). These findings highlight the potential of MAC processing for developing high-performance steels suitable for lightweight automotive applications. Full article

With water availability being one of the world’s major challenges, this study aims to propose a Zero Liquid Discharge (ZLD) system for treating saline effluents from an urban wastewater treatment plant (UWWTP), thereby supplementing into the existing water cycle. The system, which employs membrane (nanofiltration and reverse osmosis) and thermal technologies (multi-effect distillation evaporator and vacuum crystallizer), has been installed and operated in Cyprus at Larnaca’s WWTP, for the desalination of the tertiary treated water, producing high-quality reclaimed water. The nanofiltration (NF) unit at the plant operated with an inflow concentration ranging from 2500 to 3000 ppm. The performance of the installed NF90-4040 membranes was evaluated based on permeability and flux. Among two NF operation series, the second—operating at 75–85% recovery and 2500 mg/L TDS—showed improved membrane performance, with stable permeability (7.32 × 10⁻¹⁰ to 7.77 × 10⁻¹⁰ m·s⁻¹·Pa⁻¹) and flux (6.34 × 10⁻⁴ to 6.67 × 10⁻⁴ m/s). The optimal NF operating rate was 75% recovery, which achieved high divalent ion rejection (more than 99.5%). The reverse osmosis (RO) unit operated in a two-pass configuration, achieving water recoveries of 90–94% in the first pass and 76–84% in the second. This setup resulted in high rejection rates of approximately 99.99% for all major ions (Cl⁻, Na⁺, Ca²⁺, and Mg²⁺), reducing the permeate total dissolved solids (TDS) to below 35 mg/L. The installed multi-effect distillation (MED) unit operated under vacuum and under various inflow and steady-state conditions, achieving over 60% water recovery and producing high-quality distillate water (TDS < 12 mg/L). The vacuum crystallizer (VC) further concentrated the MED concentrate stream (MEDC) and the NF concentrate stream (NFC) flows, resulting in distilled water and recovered salts. The MEDC process produced salts with a purity of up to 81% NaCl., while the NFC stream produced mixed salts containing approximately 46% calcium salts (mainly as sulfates and chlorides), 13% magnesium salts (mainly as sulfates and chlorides), and 38% sodium salts. Overall, the ZLD system consumed 12 kWh/m³, with thermal units accounting for around 86% of this usage. The RO unit proved to be the most energy-efficient component, contributing 71% of the total water recovery. Full article

Soil organic carbon (SOC) monitoring is central to carbon farming Monitoring, Reporting, and Verification (MRV), yet high laboratory costs and sparse sampling limit its scalability. We present the first independent field validation of the Stenon FarmLab multi-sensor probe across 100 temperate European arable-soil samples, benchmarking its default outputs and a simple pH-corrected model against three laboratory reference methods: acid-treated TOC, temperature-differentiated TOC (SoliTOC), and total carbon dry combustion. Uncorrected FarmLab algorithms systematically overestimated SOC by +0.20% to +0.27% (SD = 0.25–0.28%), while pH adjustment reduced bias to +0.11% and tightened precision to SD = 0.23%. Volumetric moisture had no significant effect on measurement error (r = −0.14, p = 0.16). Bland–Altman and Deming regression demonstrated improved agreement after pH correction, but formal equivalence testing (accuracy, precision, concordance) showed that no in-field model fully matched laboratory standards—the pH-corrected variant passed accuracy and concordance evaluation yet failed the precision criterion (p = 0.0087). At ~EUR 3–4 per measurement versus ~EUR 44 for lab analysis, FarmLab facilitates dense spatial sampling. We recommend a hybrid monitoring strategy combining routine, pH-corrected in-field mapping with laboratory-based recalibrations alongside expanded calibration libraries, integrated bulk density measurement, and adaptive machine learning to achieve both high-resolution and certification-grade rigor. Full article

Recognizing hand actions and poses from first-person RGB videos is crucial for applications like human–computer interaction. However, the recognition accuracy is often affected by factors such as occlusion and blurring. In this study, we propose a unified framework for action recognition and hand pose estimation in first-person RGB videos. The framework consists of two main modules: the Hand Pose Estimation Module and the Action Recognition Module. In the Hand Pose Estimation Module, each video frame is fed into a multi-layer transformer encoder after passing through a feature extractor. The hand pose results and object categories for each frame are obtained through multi-layer perceptron prediction using a dual residual network structure. The above prediction results are concatenated with the feature information corresponding to each frame for subsequent action recognition tasks. In the Action Recognition Module, the feature vectors from each frame are aggregated by a multi-layer transformer encoder to capture the temporal information of the hand between video frames and obtain the motion trajectory. The final output is the category of hand movements in consecutive video frames. We conducted experiments on two publicly available datasets, FPHA and H2O, and the results show that our method achieves significant improvements on both datasets, with action recognition accuracies of 94.82% and 87.92%, respectively. Full article

As urban traffic systems become increasingly complex, traditional traffic data collection methods based on fixed detectors face challenges such as poor data quality and acquisition difficulties. Traditional methods also lack the ability to capture complete vehicle path information essential for signal optimization. While vehicle trajectory data can provide rich spatiotemporal information, its sampling characteristics present new technical challenges for traffic parameter extraction. This study addresses the key issue of extracting traffic parameters suitable for signal timing optimization from sampled trajectory data by proposing a comprehensive method for traffic parameter extraction and analysis based on vehicle trajectory data. The method comprises five modules: data preprocessing, basic feature processing, exploratory data analysis, key feature extraction, and data visualization. An innovative algorithm is proposed to identify which intersections vehicles pass through, effectively solving the challenge of mapping GPS points to road network nodes. A dual calculation method based on instantaneous speed and time difference is adopted, improving parameter estimation accuracy through multi-source data fusion. A highly automated processing toolchain based on Python and MATLAB is developed. The method advances the state of the art through a novel polygon-based trajectory mapping algorithm and a systematic multi-source parameter extraction framework specifically designed for signal control optimization. Validation using actual trajectory data containing 2.48 million records successfully eliminated 30.80% redundant data and accurately identified complete paths for 7252 vehicles. The extracted multi-dimensional parameters, including link flow, average speed, travel time, and OD matrices, accurately reflect network operational status, identifying congestion hotspots, tidal traffic characteristics, and unstable road segments. The research outcomes provide a feasible technical solution for areas lacking traditional detection equipment. The extracted parameters can directly support signal optimization applications such as traffic signal coordination, timing optimization, and congestion management, providing crucial support for implementing data-driven intelligent traffic control. This research presents a theoretical framework validated with real-world data, providing a foundation for future implementation in operational signal control systems. Full article