Search Results (3,188)

Driving style, reflected in recurrent patterns of acceleration, braking, speed selection, following distance, gear use, and lane-changing behavior, plays an important role in road safety and is also associated with fuel consumption, emissions, passenger comfort, and vehicle wear. This paper presents a structured review of recent research on driving style analysis, with particular emphasis on its relationship with road-safety outcomes and risk indicators. Following a PRISMA-oriented methodology, studies published between 2015 and 2025 were identified, screened, and synthesized to examine how driving styles are defined, detected, classified, and evaluated. The review shows a clear shift toward data-driven approaches, including feature-based machine learning and representation-learning methods using support vector machines, ensemble models, convolutional neural networks, recurrent neural networks, and hybrid deep learning architectures. Common data sources include smartphone inertial and GNSS signals, CAN/OBD vehicle data, telematics platforms, naturalistic driving datasets, and camera-based perception systems. Safety impact is most often assessed through crashes, near-miss events, traffic conflicts, time-to-collision measures, harsh maneuvers, and composite risk scores. Across the reviewed literature, aggressive and unstable driving patterns are generally associated with reduced safety margins and increased risk, although comparability remains limited by inconsistent label definitions, heterogeneous datasets, indirect safety proxies, and varied validation protocols. The paper also proposes a reproducible database architecture linking drivers, trips, driving events, and safety events to support transparent analysis, benchmark development, and future implementation in fleet monitoring, driver feedback, and connected vehicle applications. Full article

The integration of zeolitic imidazolate framework-8 (ZIF-8) into polymer matrices has created a versatile class of nanocomposites with potential applications in gas separation, water purification, food packaging, sensing, catalysis, energy systems, and environmental remediation. However, their performance is governed by complex and strongly coupled variables, including ZIF-8 particle size, morphology, defect density, surface chemistry, filler loading, polymer compatibility, interfacial adhesion, dispersion state, and processing conditions. To organize this complexity, the review introduces a hierarchical design framework that distinguishes controllable synthesis and processing inputs, experimentally measurable intermediate material states, and condition-dependent performance outputs, thereby providing a structured basis for machine-learning-ready data representation. Conventional trial-and-error approaches are therefore often inefficient and provide limited capacity to identify transferable structure–processing–property relationships. This review examines the emerging role of machine learning (ML) in the rational design and optimization of ZIF-8/polymer nanocomposites for sustainable applications. Particular attention is given to the construction of material descriptors, selection of predictive algorithms, interpretation of feature importance, optimization of synthesis and processing parameters, and prediction of mechanical, thermal, barrier, transport, adsorption, catalytic, and antimicrobial properties. The review further discusses how supervised learning, explainable artificial intelligence, active learning, Bayesian optimization, transfer learning, and physics-informed models can support material screening and multi-objective optimization across performance, cost, energy consumption, environmental impact, and end-of-life considerations. Current limitations, including small and heterogeneous datasets, inconsistent reporting, insufficient negative results, limited model interpretability, and weak experimental validation, are critically evaluated. A future framework is proposed that integrates standardized databases, high-throughput experimentation, multiscale characterization, life-cycle indicators, uncertainty quantification, and closed-loop machine learning. Such an approach could accelerate the transition from empirical formulation toward data-driven, interpretable, and sustainability-oriented design of ZIF-8/polymer nanocomposites. Full article

As smart sensing applications grow rapidly, the IoT edge architectures need to support low latency, make decisions with little memory, power, and communication resources while preserving the privacy of the data. But traditional cloud-based IoT solutions come with transmission delay, increased energy consumption, and privacy issues because of the constant transfer of raw data. In this paper, we propose a privacy-preserving TinyML-driven IoT edge architecture that enables real-time smart sensing and autonomous AI-based control. The proposed framework includes on-device sensor pre-processing, lightweight TinyML inference, adaptive model selection, encrypted feature-level communication, trust-aware decision validation, and local control execution. Raw data streams from the sensors are also processed locally, and only compact encrypted features or a summary of the decisions are sent if necessary to minimize privacy exposure. The experimental evaluation reveals that the proposed architecture has an accuracy of 97.4%, an F1 score of 96.9%, and a secure-event detection rate of 98.1% and reduces the inference latency by 14.8%, the energy consumption by 4.1 mJ per inference and the amount of data transmitted by 74.5% compared to conventional edge-cloud processing. Results show that the proposed architecture is a scalable, privacy-aware, and energy-efficient solution for real-time autonomous IoT control of smart environments with limited resources. Full article

Post-disaster reconstruction of educational facilities is frequently driven by heuristic decision-making that prioritises speed over long-term sustainability, resilience or climate compatibility. To address this gap, this study proposes a BIM-LCA decision-support framework for the evaluation and prioritisation of school retrofit strategies in post-disaster contexts. The framework integrates a BIM-derived building energy model with life cycle assessment based on EN 15978-compliant material take-offs and explicitly accounts for future climate projections. A two-storey school building in Damascus, Syria, classified under the Köppen-Geiger hot semi-arid climate zone, serves as the case study. Three retrofit scenarios are systematically evaluated against the status quo, namely shallow retrofit (external painting and shading), advanced retrofit (compliant with Passivhaus EnerPHit hot-climate standards) and deep retrofit (EnerPHit with photovoltaic integration). Simulations conducted in DesignBuilder v7 (DesignBuilder Software Ltd., Stroud, UK) assess three performance dimensions including operational and embodied energy, operational and embodied carbon footprint, and financial metrics including Net Present Value (NPV) and Marginal Abatement Cost (MAC). Future climate conditions for horizons 2030, 2050, and 2080 are generated using Meteonorm v8 (Meteotest AG, Bern, Switzerland) software under Representative Concentration Pathways RCP 2.6, RCP 4.5, and RCP 8.5. Results under the deep retrofit, on-site photovoltaic generation delivers net-positive energy performance, with an annual surplus of 27.15 MWh and net-negative operational carbon of −14,428 kgCO₂e. The advanced retrofit realises a 32% decline in operational energy consumption at a MAC of £0.89/kgCO₂e, rendering it the most favourable financial strategy under stable inflation-adjusted energy prices. Sensitivity analysis shows this ranking inverts towards the deep retrofit under sustained energy-price growth, and towards the shallow retrofit under a high cost of capital. Under RCP8.5 by 2080, cooling demand rises by up to 82% in the advanced and deep retrofits relative to their respective present-day values. The shallow retrofit records the lowest cooling demand among the retrofit options but remains approximately 9% above the contemporaneous status quo. These findings underscore the necessity of climate-adaptive, scenario-aware decision frameworks for post-disaster reconstruction, moving beyond static energy optimisation toward long-term resilience planning. Full article

Background: Anticoagulant treatment is central to the prevention and management of thromboembolic cardiovascular disease, and contemporary practice has increasingly shifted toward direct oral anticoagulants. Objectives: To quantify ambulatory anticoagulant utilization trends in Greece during 2018 to 2022 and to forecast utilization through 2030 using standardized drug-utilization metrics. Methods: A retrospective pharmacoepidemiological analysis of community-pharmacy anticoagulant sales data in Greece was performed using IQVIA Greece data as a proxy for population-level drug consumption. Utilization was standardized as defined daily doses per 1000 inhabitants per day. Drug-specific temporal trends were evaluated, and seasonal autoregressive integrated moving average models were developed to generate forecasts through 2030. Results: Total sales-based anticoagulant consumption increased by 31.1%, from 24.84 defined daily doses per 1000 inhabitants per day in 2018 to 32.57 in 2022, corresponding to an average annual increase of 1.57. This increase was primarily driven by direct oral anticoagulants, particularly apixaban, which increased from 5.03 ± 0.36 to 9.91 ± 0.46, and rivaroxaban, which increased from 6.84 ± 0.32 to 9.38 ± 0.29. In contrast, acenocoumarol declined by 37.6%, from 4.02 ± 0.17 to 2.51 ± 0.08. Among parenteral anticoagulants, enoxaparin increased from 1.87 ± 0.11 to 2.65 ± 0.15, whereas nadroparin decreased from 0.06 ± 0.01 to 0.02 ± 0.01. By 2030, apixaban and rivaroxaban are forecast to reach 19.61 (95% CI 18.45 to 20.77) and 9.29 (95% CI 5.09 to 13.49), respectively, whereas acenocoumarol is projected to decline to 0.97 (95% CI 0.72 to 1.23). Conclusions: Sales-based anticoagulant consumption shifted toward direct oral anticoagulants, with concurrent declines in vitamin K antagonists and selected parenteral agents, trends that may inform cardiovascular care planning, reimbursement policy, and implementation of evidence-based anticoagulation strategies. Full article

Federated learning enables model training in edge and distributed service environments without directly sharing raw data. In long-running edge federated services, fixed collaboration structures may become inefficient under workload drift, whereas frequent regrouping can incur migration overhead, group churn, and additional privacy-budget consumption. This paper studies budget-prioritized dynamic regrouping under workload drift, privacy-budget constraints, and migration or reconfiguration cost. We formulate a dynamic regrouping problem that jointly captures workload pressure, remaining privacy budget, service utility, and regrouping cost. We propose Budget-Prioritized Dynamic Regrouping (BP-DR), a triggered local method that evaluates single-node candidate operations and commits at most one regrouping operation per time slot. A candidate is accepted only when it is privacy-budget feasible and its utility improvement exceeds a threshold combining an anti-oscillation margin, migration or reconfiguration cost, and privacy-budget opportunity cost. We derive this trigger from a one-step local comparison and establish its monotonicity with respect to migration or reconfiguration cost and remaining privacy budget. Trace-driven experiments based on Alibaba Cluster Trace 2018 show that BP-DR maintains competitive migration-adjusted utility while controlling regrouping activity across dynamic and stress-test settings. FLamby Fed-Heart-Disease validation further shows similar learning performance across the compared methods while demonstrating the integration of BP-DR with group-aware federated training. Full article

Electrochemical CO₂ reduction reaction (ECO₂RR) is a promising technology for converting rapidly rising atmospheric CO₂—driven by fossil fuel consumption and industrial processes—into a circular carbon economy. In particular, ECO₂RR is expected to enable renewable-based long-duration energy storage (LDES) systems through the highly efficient conversion of CO₂ into high-value multi-carbon (C2⁺) compounds. However, scaling ECO₂RR to the industrial level remains challenging because, under high-current operation, the CO₂ consumption rate exceeds its supply rate, causing a sharp decline in local CO₂ concentration. The resulting increase in local pH promotes both carbonate formation and electrode flooding within the gas diffusion electrode (GDE), creating a wetting-induced mass transfer bottleneck. To address this challenge, this review categorizes and analyzes recent strategies for CO₂ microenvironment engineering that overcome mass transfer limitations at high-current densities, focusing on two complementary approaches: (1) enhancing gas-phase CO₂ supply while suppressing flooding through nano/microscale hydrophobic polymers and structural gradient designs, and (2) enhancing active CO supply in the liquid phase through electrolyte composition optimization. We further show that these strategies are not mutually independent but create complementary structural and chemical synergies, and we propose future directions for simultaneously improving high-current operability and C2⁺ product selectivity. Full article

The supply chain ecosystem (SCE) operating within an urban environment is characterised by continuous interactions among manufacturing, logistics, service, information, and energy flows across multiple smart city manufacturing-service clusters (SCMSCs). Within the SCE, intelligent production buildings (IPBs) emerge as multifunctional multistorey production-service infrastructures developed under conditions of limited urban land availability and increasing demand for localised manufacturing and service integration. These buildings operate under heterogeneous and dynamically changing energy-demand conditions, substantially complicating energy consumption management. This study develops a knowledge-driven coordination approach for the energy consumption management of IPBs operating within SCMSCs from the perspective of the urban SCE. IPBs are conceptualised as distributed environments with finite building-level power supply system capacity, where manufacturing, logistics, service, and digital processes dynamically compete for shared energy resources. A hierarchical representation of the SCMSC energy environment is proposed, capturing distributed interactions and heterogeneous electricity-demand profiles across interconnected clusters. An information-analytical system integrating monitoring, data acquisition, analysis, ML-based demand prediction, planning, and decision-support functions is developed to support predictive electricity-demand coordination. The proposed framework combines IoT-enabled monitoring with digital-twin-supported synchronisation of energy states for distributed coordination among IPBs. The proposed approach is evaluated through scenario-based analysis of an IPB operating within an urban manufacturing-service environment. The results indicate the potential of knowledge-driven coordination to improve energy-capacity utilisation, mitigate peak-load formation, and enhance operational stability within SCMSCs. Full article

Ports are progressively evolving from traditional logistics nodes into integrated energy ecosystems, characterised by increasing electrification of maritime and land-based operations, the deployment of renewable energy sources, and the emergence of new and highly variable energy demand profiles. In this context, the main challenge is not only the availability of renewable energy but also the capacity of port energy systems to provide sufficient electrical power, flexibility, and resilience under increasing operational constraints. These issues are particularly relevant in Mediterranean ports, where limited grid capacity, infrastructure constraints, load variability, and interactions with surrounding urban areas strongly influence energy planning strategies. This paper proposes an integrated framework for the development of sustainable port energy hubs based on renewable generation, energy storage, green hydrogen systems, port microgrids, and intelligent energy management strategies (EMS). The main novelty lies in the integration of these energy vectors within a unified framework that explicitly accounts for the specific operational and infrastructure constraints of Mediterranean ports. The proposed approach aims to optimise the interaction between energy production, distribution, storage, and consumption, with particular attention to the role of hydrogen as a long-duration energy storage vector and as an energy carrier for selected port logistics applications. Through a data-driven Port Energy Baseline Assessment (PEBA), port operational characteristics are translated into quantified energy demand and power requirements, providing the basis for power adequacy assessment and the evaluation of alternative transition pathways. An illustrative application to a representative Mediterranean port, characterized by a peak electricity demand of 42 MW, illustrates how the framework quantifies power requirements, assesses power adequacy under infrastructure constraints, and compares alternative transition pathways based on renewable generation, battery storage, and hydrogen. Full article

Plant-based beverages (PBBs) can be fortified to contribute to vitamin D intake. The nutritional composition of vitamin D-fortified PBBs (FPBBs) and consumer use of PBBs are examined through a three-phase technique that included a 2010–2025 national registry retrospective study (vitamin D FPBBs = 167), an on-shelf market survey (vitamin D FPBBs = 111), and a consumer survey on 225 subjects. Coconut-based PBBs had the highest vitamin D mean (0.84 μg/100 mL) in the registry database, although in the market survey oat-based PBBs had the highest vitamin D average (0.91 μg/100 mL). The compositional heterogeneity of FPBB matrices was confirmed by non-parametric Kruskal–Wallis tests across five broader botanical categories (soy, nuts, coconut, grains, blended), for carbohydrate content (χ² = 52.43, df = 4, p < 0.001), protein content (χ² = 55.85, df = 4, p < 0.001), lipid content (χ² = 26.94, df = 4, p < 0.001), and calorie content (χ² = 34.31, df = 4, p < 0.001). In contrast, vitamin D content did not differ significantly across these categories (χ² = 8.21, df = 4, p = 0.084). Dietary preferences were the main drivers of daily use (48.3% of consumers), while the subgroup driven by general health benefits or vitamin D fortification tended to consume these products mainly weekly (33.3%) or occasionally (42.8%). The mean serving size reported by respondents was 157.14 mL. Since vitamin D status was not biochemically assessed, this study does not evaluate whether current consumption addresses population-level vitamin D status. These findings highlight the need for front-of-pack labelling on nutritional density and fortification. Full article

Water treatment plants are essential urban infrastructure with direct implications for public health and everyday life. Data-driven management has received growing attention in drinking water treatment, particularly for optimizing chemical dosing to improve operational efficiency, reduce costs, and ease operator workload. AI-based prediction of coagulant dosage has therefore become an active research topic. Existing studies, however, have focused mainly on model architecture, with less attention to data validity and cost control. In practice, many plants face data-quality problems, including inconsistent dosing records under similar water-quality conditions. Conventional data cleaning may also remove large portions of the dataset, which can weaken model reliability. This study proposes an artificial intelligence (AI) modeling framework for coagulation dosing that handles anomalous data, emphasizes data quality assurance, and combines cost-oriented feedforward prediction with feedback control. A genetic algorithm-optimized backpropagation (GA-BP) neural network was first evaluated on controlled laboratory data and full-scale plant data using the same core model architecture, allowing the effects of model configuration to be separated from those of data quality. Historical plant records were subsequently cleaned through expert-guided validation, approximate time-delay alignment, and turbidity-based classification of operating conditions. Settled-water turbidity was then used as a feedback signal to dynamically adjust subsequent coagulant dosage and assess the resulting chemical savings. Changes in the input structure produced only modest improvements in full-scale prediction performance (R² = 0.53–0.72). In contrast, data cleaning and process-based data organization markedly improved predictive performance, with R² values increasing to 0.927–0.969. Standalone AI models achieved only moderate dosage reductions, while their integration with real-time turbidity feedback provided the best cost-control performance. The model-based control strategy reduced average coagulant consumption by 10.37%, with a maximum reduction of 21.33% at a settled-water turbidity target of 1.9 nephelometric turbidity units (NTU). Across the evaluated feedback-control scenarios, manual dosing was up to 32.83% higher than the corresponding feedback-controlled dosage. Overall, AI models can fit coagulation-dosing data and predict coagulant dosage with sufficient accuracy, but data quality assurance remains the main factor determining model performance. Effective cost control also requires real-time turbidity-based feedback regulation rather than model outputs alone. Full article

This paper analyzes the systemic threat of election-related Foreign Information Manipulation and Interference (FIMI) and Generative Artificial Intelligence (GAI) across recent Latin American electoral cycles in Guatemala (2023), Ecuador (2025), and Honduras (2025). Integrating the FIMI behavioral taxonomy with the DISARM framework and systems dynamics, we compare the transition from traditional digital propaganda to industrialized “algorithmic campaigns” driven by deepfakes, voice cloning, and media impersonation. Methodologically, we audit a database of documented media impersonations alongside international electoral observation reports. Our comparative analysis suggests that even if synthetic content cannot be shown to directly alter individual votes, it appears to reshape the emotional and symbolic climate of elections, which appears to exacerbate affective polarization and institutional distrust. We contextualize these dynamics within the slop economy—a structural digital divide in information quality between digital elites and digital commoners. Arguing that top-down regulations are insufficient against these decentralized threats, we propose a decentralized democratic defense model based on Jean Cloutier’s émirec concept, introducing cognitive friction into digital consumption to foster bottom-up epistemic resilience. Full article

This study proposes an integrated Industry 5.0 framework that compares and combines generative artificial intelligence (GenAI) with traditional machine learning (ML) to enhance the sustainability of manufacturing systems. This framework supports energy-efficient process optimisation, waste minimisation, material recycling and environmentally friendly production planning through data-driven decision-making. GenAI is more commonly used to generate sustainable process configurations and alternative eco-design solutions, whilst traditional ML models predict energy consumption, emissions and material losses in real time. It is proposed that a life cycle assessment (LCA) be carried out to estimate the environmental impact at all stages of production. Preliminary analyses indicate significant potential for reducing resource consumption, improving the circular economy, and supporting more sustainable and resilient manufacturing ecosystems and supply chains. Full article

Unmanned aerial vehicles (UAVs) are progressively evolving from remotely operated platforms into intelligent autonomous systems. This research addresses the role of Artificial Intelligence (AI) in advancing UAV capabilities and expanding their applications across diverse sectors. Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 methodology, studies retrieved from Scopus and relevant academic books and book chapters were screened, resulting in 217 publications retained for final analysis. The analysis introduces a three-layer framework linking AI techniques, UAV functional capabilities, and application domains. The first layer covers the list of adopted AI and ML approaches in the UAV applications. The second layer maps these approaches to key UAV capabilities, including perception, autonomous navigation, control and stability, swarm coordination, communication, and energy optimization. The third layer examines applications in agriculture, logistics, disaster response, environmental monitoring, surveillance, defense, and wireless network systems. The findings show that deep learning enhances aerial perception, reinforcement learning supports adaptive navigation and control, federated learning improves distributed intelligence, and swarm intelligence enables cooperative multi-UAV missions. Despite these advances, AI-enabled UAVs still face challenges related to energy consumption, onboard computation, data availability, communication reliability, safety, ethics, privacy, and regulation. Future progress is expected to be driven by edge AI, Tiny Machine Learning (TinyML), quantum-inspired optimization, explainable artificial intelligence (XAI), human-AI collaboration, and robust swarm coordination. Overall, this review provides a structured synthesis of AI-enabled UAV research and identifies key directions for future innovation. Full article

Co-phase traction power supply system (CTPSS) integrated with photovoltaic (PV) and hybrid energy storage system (HESS) can enhance power supply capacity while facilitating the on-site consumption of renewable energy. This integration provides a pathway toward green development and the achievement of the “carbon peaking and carbon neutrality” goal. However, the stochastic nature of PV poses challenges to safe and economical operation of the system. Existing energy management methods remain limited in simultaneously balancing operational economy and extreme risks caused by PV uncertainty, making it difficult to achieve an effective trade-off between operating cost and constraint violation risk. To address this issue, a risk-sensitive optimal scheduling framework integrating probabilistic PV forecasting, correlated scenario generation, and risk-aware optimization is developed. First, a PV probabilistic prediction model based on parallel TCN-BiLSTM-Attention is proposed, which extracts multiscale local features and long-range temporal features from PV for accurate uncertainty quantification. Second, a t-Copula PV scenario generation method driven by weather classification and temporal correlation is proposed. On this basis, a day-ahead optimal scheduling strategy integrating chance constraints programming (CCP) and conditional value at risk (CVaR) is established to simultaneously control constraint violation risk and extreme economic risk. The proposed prediction model can accurately quantify uncertainty, achieving an R² value of 0.9979. When the allowable power supply loss probability is 0.5%, the proposed scheme’s operating cost is reduced by 1.36% compared with the baseline scheme. The results demonstrate that the proposed framework can effectively coordinate operating economy and extreme-risk control, thereby improving the risk-sensitive operational performance of CTPSS with PV and HESS. The proposed strategy balances economy and extreme risk, providing a reference for safe, low-carbon and economical operation of CTPSS with PV and HESS. Full article