Search Results (17,296)

This study presents an Artificial Neural Network (ANN)-based adaptive switching frequency control strategy for series Direct current (DC) motor drives used in battery-powered mining locomotives, aiming to extend the lifetime of critical power-electronic components such as Insulated Gate Bipolar Transistors (IGBTs) and DC bus capacitors. In embedded systems for electric traction, two dominant degradation factors, motor current ripple and IGBT temperature fluctuation, significantly shorten component lifetimes. Conventional fixed switching frequencies impose a trade off: higher frequencies reduce current ripple but increase IGBT losses and temperature, while lower frequencies yield the opposite effect. Consequently, an adaptive variable switching frequency control algorithm is proposed to perform real-time decision making by predicting the optimal switching frequency that minimizes both motor current ripple and IGBT thermal fluctuations. The proposed algorithm was trained with a dataset acquired from current sensors, NTC temperature sensors, and a potentiometer defining the target current (PWM duty). Performance comparisons with a fixed frequency demonstrate that the ANN-driven approach maintains an average current ripple of less than 5% (average) and 10% (maximum), while the lifetime of the IGBT and capacitors improves. A fairness index was defined to quantify the relative lifetime improvement of the IGBT and capacitor, revealing that the proposed variable frequency switching model enhances the overall system performance by up to 13 times compared to fixed-frequency operation. These results confirm that the integration of embedded machine learning and adaptive control algorithms can substantially enhance the durability and efficiency of power-electronic systems in real-time industrial applications. Full article

Open burning of crop residue is a major source of air pollutants in China. While a nationwide straw burning ban implemented in 2016 has proven effective in reducing primary emissions, its impact on ozone (O₃), a key pollutant detrimental to human health, remain poorly quantified. This study aims to assess the impact of straw burning on downwind urban O₃ pollution and to investigate the complex mechanisms governing O₃ changes resulting from transported agricultural fire plumes. Here, using multi-satellite data and ground observations from 2013 to 2020, this study developed a method to identify smoke-affected days and estimate plume transport time over northern China. The results show that the straw burning ban effectively reduced peak concentrations of particulate matter (PM_2.5) during harvest seasons. However, O₃ responses on smoke-affected days were heterogeneous, showing both increases and decreases. The random forest model revealed the meteorological and chemical drivers of O₃ variability. Elevated formaldehyde (HCHO) and temperatures promote O₃ production, while higher NO₂ and relative humidity enhance its titration. Trajectory analysis further decoupled the mechanisms that O₃ and HCHO enhancements were primarily driven by local photochemistry, whereas NO₂ increases were attributable to regional transport and mixing with anthropogenic pollution. This study underscores the necessity for integrated air quality management strategies that account for the complex trade-offs between PM_2.5 and O₃ to fully realize the public health benefits of emission control policies. Full article

Heat stress has emerged as a significant abiotic constraint affecting rice yield and grain quality. In recent years, substantial advancements have been achieved in elucidating molecular regulatory mechanisms and breeding applications pertinent to rice heat tolerance. This review offers a comprehensive examination of the fundamental regulatory pathways involved in rice responses to heat stress, encompassing membrane lipid homeostasis, heat signal transduction, transcriptional regulation, RNA stability and translation, epigenetic modifications, hormone signaling, antioxidant defense, and the protection of reproductive organs. Particular emphasis is placed on the functional mechanisms and breeding potential of pivotal thermotolerance-associated genes and quantitative trait loci (QTLs), such as TT1, TT3, and QT12. Additionally, we summarize recent applications of cutting-edge technologies in the enhancement of heat-tolerant rice varieties, including multi-omics integration, CRISPR/Cas9 genome editing, marker-assisted selection (MAS), and rational design breeding. Finally, we address current challenges, including integrating regulatory mechanisms, developing realistic heat simulation systems, validating the functionality of candidate genes, and managing trait trade-offs. This review provides a theoretical foundation for developing heat-tolerant rice cultivars and offers valuable insights to accelerate the breeding of climate-resilient rice varieties for sustainable production. Full article

Depression affects millions of people worldwide. Traditional management relies heavily on periodic clinical assessments and self-reports, which lack real-time responsiveness and personalization. Despite numerous research prototypes exploring Internet of Things (IoT)-based mental health support, almost none have translated into practical, mainstream solutions. This gap stems from several interrelated challenges, including the absence of robust, flexible, and safe architectural frameworks; the diversity of IoT device ownership; the need for further research across many aspects of technology-based depression support; highly individualized user needs; and ongoing concerns regarding safety and personalization. We aim to develop a reference architecture for IoT-based safe and personalized depression management. We introduce IoTMindCare, integrating current best practices while maintaining the flexibility required to incorporate future research and technology innovations. A structured review of contemporary IoT-based solutions for depression management is used to establish their strengths, limitations, and gaps. Then, following the Attribute-Driven Design (ADD) method, we design IoTMindCare. The Architecture Trade-off Analysis Method (ATAM) is used to evaluate the proposed reference architecture. The proposed reference architecture features a modular, layered logical view design with cross-layer interactions, incorporating expert input to define system components, data flows, and user requirements. Personalization features, including continuous, context-aware feedback and safety-related mechanisms, were designed based on the needs of stakeholders, primarily users and caregivers, throughout the system architecture. ATAM evaluation shows that IoTMindCare supports safety and personalization significantly better than current designs. This work provides a flexible, safe, and extensible architectural foundation for IoT-based depression management systems, enabling the construction of optimal solutions that integrate the most effective current research and technology while remaining adaptable to future advancements. IoTMindCare provides a unifying, aggregation-style reference architecture that consolidates design principles and operational lessons from multiple prior IoT mental-health solutions, enabling these systems to be instantiated, compared, and extended rather than directly competing with any single implementation. Full article

Stock trend prediction is a complex and crucial task due to the dynamic and nonlinear nature of stock price movements. Traditional models struggle to capture the non-stationary and volatile characteristics of financial time series. To address this challenge, we propose the Hybrid Relational Transformer (HRformer), which specifically decomposes time series into multiple components, enabling more accurate modeling of both short-term and long-term dependencies in stock data. The HRformer mainly comprises three key modules: the Multi-Component Decomposition Layer, the Component-wise Temporal Encoder (CTE), and the Inter-Stock Correlation Attention (ISCA). Our approach first employs the Multi-Component Decomposition Layer to decompose the stock sequence into trend, cyclic, and volatility components, each of which is independently modeled by the CTE to capture distinct temporal dynamics. These component representations are then adaptively integrated through the Adaptive Multi-Component Integration (AMCI) mechanism, which dynamically fuses their information. The fused output is subsequently refined by the ISCA module to incorporate inter-stock correlations, leading to more accurate and robust predictions. Extensive experiments on the NASDAQ100 and CSI300 datasets demonstrate that HRformer consistently outperforms state-of-the-art methods, e.g., achieving about 0.83% higher Accuracy and 1.78% higher F1-score than TDformer on NASDAQ100, with Sharpe Ratios of 1.5354 on NASDAQ100 and 0.5398 on CSI300, especially in volatile market conditions. Backtesting results validate its practical utility in real-world trading scenarios, showing its potential to enhance investment decisions and portfolio performance. Full article

This study examines the impact of technological progress on food price dynamics and supply stability across the 27 European Union Member States during 2011–2024. Using a balanced panel dataset, the analysis explores four dependent indicators—consumer food prices, food price inflation, price volatility, and food supply variability—while controlling for trade openness, GDP per capita growth, and population. Technological progress is estimated through panel least squares regression with fixed effects. The results reveal that technological advancement significantly reduces food prices and inflation, suggesting that innovation-driven productivity and efficiency gains stabilize consumer markets. However, its influence on food price volatility and supply variability is statistically insignificant, indicating that innovation alone cannot fully mitigate systemic risks in the European food system. The results provide policy-relevant evidence supporting the integration of technological innovation into food system governance across the European Union. They underline the need for targeted investment and regulatory coordination to translate innovation gains into tangible resilience outcomes, thus offering practical guidance for policymakers and stakeholders involved in implementing the European Green Deal and the Farm to Fork Strategy. Full article

Federated representation learning (FRL) is a promising technique for learning shared data representations that capture general features across decentralized clients without sharing raw data. However, there is a risk of sensitive information leakage from learned representations. The conventional differential privacy (DP) mechanism protects the privacy of the whole data by randomizing (adding noise or random response) at the cost of deteriorating learning performance. Inspired by the fact that some data information may be public or non-private and only sensitive information (e.g., race) should be protected, we investigate the information-theoretic protection on specific sensitive information for FRL. To characterize the trade-off between utility and sensitive information leakage, we adopt mutual information-based metrics to measure utility and sensitive information leakage, and propose a method that maximizes the utility performance, while restricting sensitive information leakage less than any positive value $ϵ$ via the local DP mechanism. Simulation demonstrates that our scheme can achieve the best utility–leakage trade-off among baseline schemes, and more importantly can adjust the trade-off between leakage and utility by controlling the noise level in local DP. Full article

The General AntiParticle Spectrometer is a balloon-borne experiment designed to search for low-energy cosmic-ray antinuclei as a potential indirect signature of dark matter. Over the course of at least three long-duration flights over Antarctica, it will explore the sub-$250 \mathrm{M}\mathrm{e}\mathrm{V}/\mathit{n}$ energy range with sensitivity to antideuterons and antihelium, while also extending antiproton measurements below 100 $\mathrm{M}$$\mathrm{e}\mathrm{V}$. The instrument features a tracker built from more than one thousand lithium-drifted silicon detectors, each read out by a dedicated custom integrated circuit. With the first flight scheduled for the austral summer of 2025, a new prototype chip, ANTARES4, has been developed using a commercial 65 $\mathrm{n}$$\mathrm{m}$ complementary metal-oxide semiconductor process for use in the second flight. It integrates eight independent analog channels, each incorporating a low-noise charge-sensitive amplifier with dynamic signal compression, a CR–RC shaping stage with eight selectable peaking times, and on-chip calibration circuitry. The charge-sensitive amplifier uses metal-oxide semiconductor feedback elements with voltage-dependent capacitance to support the wide input energy range from 10 $\mathrm{k}$$\mathrm{e}\mathrm{V}$ to 100 $\mathrm{M}$$\mathrm{e}\mathrm{V}$. Four alternative feedback implementations are included to compare performance and design trade-offs. Leakage current compensation up to 200 $\mathrm{n}$$\mathrm{A}$ per detector strip is provided by a Krummenacher current–feedback network. This paper presents the design and architecture of ANTARES4, highlighting the motivations, design drivers, and performance requirements that guided its development. Full article

Non-Intrusive Load Monitoring (NILM) enables appliance-level energy analysis from aggregated electrical signals, offering valuable insights for smart energy systems. While most NILM research focuses on high-resource environments, this study evaluates the feasibility of deploying NILM algorithms on constrained edge computing platforms. Two representative models for event detection and for energy disaggregation were trained on a high-end PC and tested on both the PC and two edge devices. A modular software framework using a virtual container and virtual environments ensured reproducibility across platforms. Experiments using datasets under simulated real-time streaming conditions revealed that although all devices achieved consistent detection, classification, and disaggregation performance, edge platforms struggled with real-time inference due to processing latency and memory limitations. This study presents a detailed comparison of execution time, resource usage, and model performance, highlighting the trade-offs associated with NILM deployment on embedded systems and proposing future directions for optimization and integration into smart grids. Full article

This study examined the effects of phosphoric acid (H₃PO₄), polyimide (PI), and lubricants (MoS₂, graphite) on the phase stability, microstructure, and magnetic performance of Fe-5.0 wt.%Si soft magnetic composites (SMCs). Warm compaction (≤550 °C) and annealing at 700 °C were applied to samples prepared under a full factorial design. X-ray diffraction confirmed stable α-Fe(Si) phases without secondary phases. SEM and TEM–EDS revealed interfacial insulating layers mainly composed of Si-O, with localized phosphorus and carbon. Additive composition strongly influenced magnetic and physical properties. Increasing H₃PO₄ and PI reduced the density from 7.50 to 7.27 g/cm³ and lowered the permeability (from 189 at 1 kHz to 156), due to thicker interparticle layers that restricted metallic contact and domain wall motion. In contrast, Q-values rose significantly with frequency: for H₃PO₄ 0.25 wt.% + PI 0.25 wt.% + graphite 0.3 wt.%, Q increased from 0.39 (1 kHz) to 2.91 (10 kHz), reflecting effective eddy current suppression. Lubricant type further influenced performance: graphite consistently outperformed MoS₂, with 0.3 wt.% graphite providing the best balance of high density, permeability, and a frequency-stable Q-value. Overall, Fe-5.0 wt.%Si performance is governed not by bulk phase changes but by the trade-off between densification and insulation at particle interfaces. The optimal combination of low H₃PO₄ and PI with 0.3 wt.% graphite offers practical guidelines for designing high-frequency, high-efficiency motor materials. Full article

Background/Objectives: Spinal fusion with static fixation—surgically joining two or more vertebrae to eliminate motion—is commonly employed to treat degenerative spinal disease. However, the rigidity imposed by static constructs and the increased load on the adjacent segments frequently result in complications such as disc or facet degeneration, spinal stenosis (SS), and segmental instability. This study investigates the effectiveness of pedicle-based dynamic stabilization using the Dynesys system, particularly in a dynamic-transitional optima (DTO) hybrid configuration, in mitigating adjacent segment disease (ASD) and improving clinical outcomes. In this work, we analyzed the mechanical performance and intermediate-term clinical effects of the DTO hybrid lumbar device, focusing on how the load-sharing properties of the Dynesys cord–spacer stabilizers may contribute to junctional complications in individuals with diverse grades of intervertebral disc degeneration. Study Design/Setting: We designed a combined biomechanical finite element (FE) and experimental analysis to predict the clinical outcomes. Patient Sample: Among 115 patients with lumbar SS enrolled for analysis, 31 patients (mean age: 68.5 ± 7.5 years), with or without grade I spondylolisthesis (18/13), underwent a two-level DTO hybrid procedure—L4–L5 static fixation and L3–L4 dynamic stabilization—with minimal decompression to preserve the posterior tension band. Post-surgical follow-ups were conducted for over 48 months (range: 49–82). Outcome Measures: Radiological assessments were performed by two neurosurgeons, one orthopedic surgeon, and one neuroradiologist. The posterior disc height, listhesis distance, and dynamic angular changes were measured pre- and postoperatively to evaluate ASD progression. Methods: Dynamic instrumentation was assigned to the L3–L4 motion segment with lesser disc deterioration, in contrast to the L4–L5 segment, where static fixation was applied due to its greater degree of degeneration. FE analysis was performed under displacement-controlled conditions. Intersegmental motion analysis was conducted under load-controlled conditions in a synthetic model. Results: The DTO hybrid devices reduced stress and motion at the transition segment. However, compensatory biomechanical effects were more pronounced at the adjacent cephalad than the caudal segments. In the biomechanical trade-off zone—where balance between motion preservation and stabilization is critical—the flexible Dynesys cord significantly mitigated stiffness-related issues during flexion. At the L3–L4 transition level, the cord–spacer configuration enhanced dynamic function, increasing motion by 2.7% (rotation) and 12.7% (flexion), reducing disc stress by 4.1% (flexion) and 12.9% (extension), and decreasing the facet contact forces by 4.9% (rotation) and 15.6% (extension). The optimal cord stiffness (50–200 N/mm) aligned with the demands of mild disc degeneration, whereas stiffer cords were more effective for segments with higher degeneration. The pedicle screw motion in dynamic Dynesys systems—primarily caused by axial translation rather than vertical displacement—contributed to screw–vertebra interface stress, influenced by the underlying disc or bone degeneration. Conclusions: Modulating the cord pretension in DTO instrumentation effectively lessened the interface stress occurring at the screw–vertebra junction and adjacent facet joints, contributing to a reduced incidence of pedicle screw loosening, ASD, and revision rates. The modified DTO system, incorporating minimal decompression and preserving the posterior complex at the dynamic level, may be biomechanically favourable and clinically effective for managing transitional degeneration over the mid-term. Full article

Accurate land cover mapping is essential for environmental monitoring and agricultural management. Sentinel-2 imagery, with high spatial resolution and open access, provides valuable opportunities for operational classification. Convolutional neural networks (CNNs) have demonstrated state-of-the-art results, yet their adoption is limited by high computational demands and limited methodological transparency. This study proposes a lightweight CNN for soil–vegetation classification, in Dolj County, Romania. The architecture integrates three convolutional blocks, global average pooling, and dropout, with fewer than 150,000 trainable parameters. A fully documented workflow was implemented, covering preprocessing, patch extraction, training, and evaluation, addressing reproducibility challenges common in deep leaning studies. Experiments on Sentinel-2 imagery achieved 91.2% overall accuracy and a Cohen’s kappa of 0.82. These results are competitive with larger CNNs while reducing computational requirements by over 90%. Comparative analyses showed improvements over an NDVI baseline and a favorable efficiency–accuracy balance relative to heavier CNNs reported in the literature. A complementary ablation analysis confirmed that the adopted three-block architecture provides the optimal trade-off between accuracy and efficiency, empirically validating the robustness of the proposed design. These findings highlight the potential of lightweight, transparent deep learning for scalable and reproducible land cover monitoring, with prospects for extension to multi-class mapping, multi-temporal analysis, and fusion with complementary data such as SAR. This work provides a methodological basis for operational applications in resource-constrained environments. Full article

Auditory stimuli are known to enhance visual target recognition in rapid serial visual presentation (RSVP) tasks, yet the robustness and potential trade-offs of this audiovisual integration (AVI) effect remain debated. Attribute amnesia (AA) refers to the phenomenon in which individuals successfully identify a stimulus for a task, but fail to recall its basic attributes when unexpectedly tested. The present study investigates whether improvements in visual recognition through AVI occur at the expense of auditory information loss, as predicted by the AA framework. Across two RSVP experiments, participants were presented with letter targets embedded among digit distractors. In Experiment 1, an auditory pitch (bass, alto, treble) accompanied the second target (T2); in Experiment 2, an auditory syllable either matched or mismatched the semantic identity of T2. A surprise-test paradigm was used to assess participants’ ability to recall auditory stimuli. The results show that both pitch and semantic attributes were subject to AA, with semantic stimuli recalled more accurately than pitch. Moreover, semantic congruency enhanced T2 identification, highlighting the automatic processing advantage of semantic cues. Post-surprise trials revealed the improved recall of auditory attributes, consistent with the working memory reselection model. Together, these findings suggest that AVI enhances visual recognition by reallocating cognitive resources, but at the cost of a partial loss of irrelevant auditory information. Full article

Decontamination of wastewater, industrial effluents, stormwater, and graywater can be carried out through the use of natural or constructed wetlands. In either case, the natural functions of soil, vegetation, and organisms are widely applied for the treatment of contaminated water. In particular, in the physical design of a constructed wetland, several operational factors must be adjusted with the aim of reducing pollution levels. Although various fully customized design methodologies have been developed and reported in the literature, they often fail to meet the required decontamination objectives. In this context, the application of the NSGA-II evolutionary algorithm is adequate to optimize the physical design of a horizontal subsurface flow wetland for graywater treatment, focusing specifically on the removal of biodegradable organic matter (BOD${}_5$). Four competing objectives are considered: minimizing physical volume and total design cost, while maximizing contaminant removal efficiency and graywater flow rate. Five constraint functions are also incorporated: removal efficiency greater than 95%, physical volume below 1000 m${}^3$, flow rate above 10 m${}^3$/d, a limit on total construction cost of MXN 1,000,000, and maintaining a length-to-width ratio greater than or equal to 2 but less than or equal to 4. The proposed methodology generates a wide set of non-dominated solutions, visualized through Pareto surfaces, which highlight the trade-offs among different objectives. This approach offers the possibility of selecting optimal designs under specific conditions, which underscores the limitations of conventional single-solution models. The results show that the methodology consistently achieved removal efficiencies above 95%, with construction costs within budget and physical volumes below the established limit, offering a more versatile and cost-effective alternative. This work demonstrates that the integration of NSGA-II into wetland design is an effective and adaptable strategy, capable of providing sustainable alternatives for graywater treatment and constituting a valuable decision-making tool. Full article

Municipal solid waste (MSW) recycling is constrained by contamination, heterogeneity, and infrastructure built around material-specific pathways. We introduce effectiveness-normalized greenhouse gas (GHG) emissions as a system-level metric that adjusts reported process burdens by feedstock eligibility (Effectiveness Fraction, EF) and carbon recovery efficiency (CRE) to reflect real-world MSW conditions. Using published LCA data and engineering estimates, we benchmark six pathways, mechanical recycling, PET depolymerization, enzymatic depolymerization, pyrolysis, supercritical water gasification (SCWG), and Regenerative Robust Gasification (RRG), at the scale of mixed MSW. Normalizing for EF and CRE reveals large differences between process-level and system-level performance. Mechanical recycling and PET depolymerization show low process intensities yet high normalized impacts because they can treat only a small share of plastics in MSW. SCWG performs well at broader eligibility. RRG, a plasma-assisted molten-bath approach integrated with methanol synthesis, maintains the lowest normalized impact (~1.6 t CO₂e per ton of recycled polymer) while accepting virtually all organics in MSW and vitrifying inorganics. Modeled methanol yields are ~200–300 gal·t⁻¹ without external hydrogen and up to ~800 gal·t⁻¹ with renewable methane reforming. The metric clarifies trade-offs for policy and investment by rewarding technologies that maximize diversion and carbon retention. We discuss how effectiveness-normalized results can be incorporated into LCA practice and Extended Producer Responsibility (EPR) frameworks and outline research needs in techno-economics, regional scalability, hydrogen sourcing, and uncertainty analysis. Findings support aligning infrastructure and procurement with robust, scalable routes that deliver circular manufacturing from heterogeneous MSW. Full article