Search Results (617)

Human fertility is declining across industrialised populations, while the incidence of hormone-dependent reproductive cancers rises. Endocrine-disrupting chemicals (EDCs) and structurally related emerging pollutants are implicated in both. These two outcomes are generally reviewed as separate studies. This review argues that they are two latencies of a single molecular toxicology. The compounds concerned are structurally diverse: phthalates, bisphenols, per- and polyfluoroalkyl substances (PFASs), pesticides, polychlorinated biphenyls (PCBs) and dioxins, brominated and organophosphate flame retardants, pharmaceuticals and personal-care products (PPCPs), and micro- and nanoplastics. They nonetheless converge on a limited repertoire of molecular lesions. These include the disruption of hypothalamic–pituitary–gonadal (HPG) signalling through kisspeptin/GnRH and gonadotropin gene expression and interference at nuclear and membrane hormone receptors (ERα/β, AR, GPER, thyroid receptors, AhR, PPARγ). They also include the inhibition of steroidogenesis at StAR and the CYP11A1–CYP17A1–CYP19A1/3β-HSD/17β-HSD cascade and reactive-oxygen-species generation with mitochondrial dysfunction and Keap1–Nrf2 disruption. Epigenetic reprogramming through DNA methylation, histone modification and non-coding RNAs, together with crosstalk with metabolic and immune signalling, completes the set. These lesions produce measurable cytotoxic and genotoxic damage to gametes and the early embryo: sperm DNA fragmentation and 8-oxo-dG accumulation, blood–testis-barrier breakdown, oocyte meiotic-spindle defects, and granulosa-cell apoptosis and pyroptosis. The same receptor, oxidative and genotoxic hubs drive hormone-dependent reproductive carcinogenesis over longer latencies. The review makes three contributions. First, it traces these shared hubs continuously from fertility impairment to malignancy rather than treating them as separate fields. Second, it grades the certainty of the human evidence class by class, so that robust associations can be distinguished from provisional ones. Third, it integrates pseudo-persistent pollutants alongside the classical persistent compounds. These are micro- and nanoplastics, which act as both toxicants and vectors for adsorbed co-contaminants, and pharmaceutical and personal-care residues sustained by continuous wastewater input. Their inclusion demonstrates that chronic low-dose exposure does not require chemical persistence. We conclude with mitigation strategies and an explicit account of what the current evidence base cannot yet support. Full article

While variable-latency AI dialogue processing proceeds asynchronously, the VR state continues to be updated frame by frame. In this study, we implemented “guard-mediated LLM-VR scene control” for breastfeeding support training, in which LLM-derived ActionClass and PhaseCandidate are not used to update the phase directly, but are instead provided as inputs to phase-specific deterministic guards together with the VR state and the ActionClass-mediated mother-avatar state on the Unity side. We conducted human-in-the-loop sessions with 20 nursing students, post-study guard/Turn-ID tests, a 2-expert semantic evaluation of 639 cases, and state-binding and local guard sensitivity analyses. There were no failures preventing session continuation, and the median of the participant-specific median latencies was 2.61 s. Phase guards matched a separately implemented specification-based oracle in 1600 out of 1600 cases, and the Turn-ID gate eliminated cross-turn stale results. LLM-expert strict agreement was 74.5%/77.8% for ActionClass and 16.9%/35.2% for PhaseCandidate. The local outcome divergence in cases where semantic labels differed between the two experts was 85.2% in Phase 1, 1.9% in Phase 2, and 56.8% in Phase 3; the impact of semantic variability differed across phases. In particular, the current topology, which uses PhaseCandidate as a standalone transition-enabling disjunct, requires reevaluation. This study is a technical pilot and mechanism-specific characterization; it does not evaluate educational effectiveness, clinical efficacy or safety, or superiority over alternative architectures. Full article

The rapid electrification of transportation has accelerated the demand for high-power electric vehicle (EV)-charging systems with high efficiency, compact size, galvanic isolation, and flexible scalability. Conventional isolated EV chargers typically adopt cascaded AC–DC and DC–DC conversion stages, which require additional semiconductor devices, passive components, and bulky dc-link capacitors, thereby increasing system complexity and limiting power density. This paper proposes a scalable three-phase modular parallel quasi-single-stage isolated single-ended primary-inductor converter (SEPIC) for high-power EV fast-charging applications. The proposed converter integrates power factor correction, voltage regulation, and high-frequency isolation within a unified SEPIC-based conversion cell, eliminating the intermediate dc-link capacitor while reducing the number of magnetic components and power conversion stages. By employing a Δ-connected three-phase input and input/output-parallel modular configuration, the proposed architecture provides a flexible power expansion approach based on a 9 kW basic module, with the potential to extend to higher power levels, such as 54 kW, through paralleling multiple identical modules. The operating principle, steady-state characteristics, continuous conduction mode (CCM)/discontinuous conduction mode (DCM) transition mechanism, current-sharing behavior, and control strategy are systematically investigated. An 18 kW prototype consisting of two parallel modules is experimentally validated under 380 V three-phase AC input and 400 V DC output conditions. The experimental results demonstrate a peak efficiency of 97.5%, a rated efficiency of 97.3%, a power factor (PF) of 0.999, and an input current total harmonic distortion (THD) of 2.55%, confirming the effectiveness and scalability of the proposed converter for high-power EV fast-charging applications. Full article

In Global Navigation Satellite System (GNSS)-denied indoor environments, ultra-wideband (UWB) localization of unmanned ground vehicles (UGVs) is challenged by position-dependent anchor visibility and mixed line-of-sight (LOS)/non-line-of-sight (NLOS) ranging. This study proposes a soft continuous confidence weighting method within an adaptive Kalman filter (AKF)-based UWB/inertial measurement unit (IMU) localization framework. The vehicle model uses motor pulse increments and IMU yaw-rate measurements as inputs and outputs vehicle position and heading estimates. Virtual forward–backward iteration converts inconsistencies between the current UWB ranges and tag–anchor geometry into terminal virtual-anchor displacements. A half-Gaussian function then maps each displacement to a continuous confidence coefficient. The resulting coefficients are incorporated into weighted least-squares (WLS) and AKF localization, while the UWB measurement-noise covariance is adaptively updated using the range innovations. The proposed method was evaluated through static calibration and dynamic localization experiments. These experiments compared soft and hard weighting schemes and assessed the contribution of AKF fusion. These results indicate that the method proposed in this study improves localization accuracy, robustness, and temporal continuity under position-dependent anchor visibility and mixed LOS/NLOS conditions. Full article

The conductor temperature of cross-linked polyethylene (XLPE) high-voltage direct-current (HVDC) cables governs ampacity assessment and insulation life management, yet it cannot be measured in service, and finite-element simulation is too expensive for real-time use. This paper presents a physics-informed neural network (PINN) that embeds the transient heat-conduction equation, a temperature-dependent Joule source, and the boundary and initial conditions into the training loss of a neural surrogate. Three ingredients adapt the framework to power cables: a layer-aware material mapping over the eight heterogeneous cable layers; an electro-thermal coupling through the temperature dependence of the conductor conductivity, handled during training by a convergent Picard-type evaluation of the Joule source; and a condition-embedding input treating the load current and ambient temperature as continuous parameters so that a single network covers the admissible current–ambient envelope of the studied cable configuration. Validated against finite-element references under fifteen operating conditions, the model attains a root-mean-square error of $0.0024$ K (mean over five training seeds) on a held-out condition relative to a finite-element reference whose mesh-discretization error a refinement study bounds at about $0.04$ K while reducing the governing-equation residual by approximately 28-fold relative to an identically sized data-driven network at statistically indistinguishable pointwise accuracy. The physics prior also renders degradation under training-data reduction more graceful and improves extrapolation to unseen ambient temperatures, whereas current extrapolation remains the most challenging transfer. The differentiable surrogate identifies the load current and the unmeasurable conductor hotspot from ten surface sensors within seconds, at below 9 ms per ${10}^5$ queries. A loss-weight sensitivity study and a three-dimensional cable-end-effect case on a second material configuration are also reported. All reference data are numerical; experimental cable-loop validation remains for future work. Full article

Object detection in UAV aerial images remains fundamentally constrained by extremely small object scales, strong background interference, and progressive structural information degradation along the feature extraction pipeline. Current small-object detection methods suffer from two fundamental deficiencies rooted in their convolutional feature extraction pipelines: the smoothing effect of strided convolutions in early layers, which attenuates fine-grained details before backbone processing, and the feature overwriting phenomenon, where sequential transformations progressively erase structural information from earlier layers. We propose PFE-Det (Progressive Feature Evolution Detector), built upon the DEIM framework and constructing a continuous optimization pathway across three stages. A Feature Adaptive Enhancement Network (FAENet) is adopted as a front-end preprocessor to decouple high- and low-frequency components via a Laplacian pyramid at the input stage, mitigating early-layer smoothing at the input. A Multi-Receptive-Field Adaptive Fusion Module (MFAM) is designed to reorganize single-path features into structure-retaining and progressive enhancement paths and is further coupled with hierarchical receptive-field modeling, suppressing feature overwriting through multi-scale context modeling. A Multi-Path Gated State Space Modeling Block (MG-SSM Block) couples HSM-SSD-based long-range dependency extraction with Convolutional Gated Linear Units (CGLU) for adaptive feature selection in the encoder. Experiments on VisDrone2019 demonstrate an AP of 0.225 and an ${\mathrm{AP}}_s$ of 0.134, yielding a 13.5% relative improvement in small-object precision over the baseline. Cross-dataset evaluations on DIOR and UAVVaste, each under independent training and testing, further support the effectiveness of progressive feature evolution for UAV small-object detection. Full article

Urban drainage overflows can release a large fraction of event-scale pollutants during the early stage, yet current control remains largely driven by hydraulic signals rather than pollutant-release dynamics. This study created a dissolved organic matter (DOM)-based fluorescence fingerprint method to precisely identify the shift from pollutant flushing to dilution or ongoing input, helping determine the timing of first-flush transitions and potential interception. Fourteen wet-weather overflow events from seven drainage systems in Shanghai and Changzhou were investigated using excitation–emission matrix fluorescence spectroscopy, combined with non-negative matrix factorization, random forest feature screening, principal component analysis, mass–volume (M(V)) curve analysis, and Pettitt change-point detection. Five macro-scale fluorescence fingerprints were resolved, representing protein-like, fulvic-like, and humic-like components. Protein-like fingerprints dominated rapid event-scale variations, while fulvic-like and humic-like fingerprints reflected continuous surface-derived input and stable background contribution, respectively. Peak-shift trajectories revealed three fluorescence-evolution modes: directional red-shift migration, peak-position stability, and weak, non-directional variability, reflecting different source-release dynamics and DOM compositional adjustments during overflow. Random forest screening identified 20 high-importance fine-scale fluorescence fingerprints, with 90% concentrated in protein-like regions linked to sewage-derived and labile DOM. Compared with macro-scale fingerprints and conventional water quality indicators, fine-scale fluorescence fingerprints showed clearer stage separation, stronger consistency with M(V)-based cumulative response patterns, and more distinct first-flush interception timing. This timing marked the transition from early concentrated pollutant release to dilution or sustained input, whereas macro-scale fingerprints indicated broader transition intervals and conventional indicators showed delayed responses. These findings highlight the potential of fine-scale fluorescence fingerprints to support future fluorescence-assisted overflow control by improving transition identification and targeted interception decisions. Full article

Accurate voltage degradation prediction is essential for health management and lifetime extension of proton exchange membrane fuel cell (PEMFC) systems. During long-term constant-current operation, stack voltage evolves nonlinearly and is influenced by coupled variations in temperature, pressure, flow rate, and humidity, while many deep learning-based models lack physical interpretability. This study proposes an interpretable Transformer-based framework for PEMFC voltage degradation prediction under constant-current operation. The framework integrates outlier correction, interpolation, Savitzky–Golay filtering, Z-score normalization, sliding-window reconstruction, Transformer-based prediction, and feature-ablation interpretation. Using multivariate sensor measurements and historical voltage as inputs, the Transformer was compared with RNN, LSTM, and GRU baselines under identical preprocessing and evaluation conditions. The models were evaluated chronologically by continuously applying the sliding-window model over the held-out final 20% of the aging sequence. The Transformer achieved the best performance, with MAE of 8.2 × 10⁻⁴, RMSE of 1.18 × 10⁻³, MAPE of 0.0256%, and R² of 0.9961. Compared with the second-best RNN model, it reduced MAE, RMSE, and MAPE by 9.89%, 7.81%, and 9.86%, respectively. Feature ablation showed that flow- and pressure-related variables contributed 50.08% and 26.78% of the total importance, respectively. Full article

Less attention has been paid to soils with potentially toxic elements (PTEs) below agricultural land risk screening values, even though they continue to accumulate these elements. In this study, four typical areas in Ningxia were selected to determine the concentrations of cadmium (Cd) and arsenic (As) in 176 samples collected from 130 sampling sites across seven matrices. A refined mass balance model was developed by partitioning irrigation input into suspended-solid and supernatant phases and crop removal into grain and straw components. The model was used to analyze the contributions of various input and output factors and to predict future soil Cd/As concentrations and their related health risks via soil exposure. The input fluxes of Cd and As in the four regions ranged from 1.31 to 3.25 g·ha⁻¹·yr⁻¹ and 71.43 to 146.99 g·ha⁻¹·yr⁻¹, respectively, mainly contributed by irrigation water (40~64%), especially suspended solids in irrigation water, and atmospheric deposition (23~40%). The output fluxes of Cd and As were 0.89~1.43 g·ha⁻¹·yr⁻¹ and 13.81~33.86 g·ha⁻¹·yr⁻¹, respectively, dominated by crop harvesting (36~82%). The differences in input and output fluxes were mainly caused by the regional industrial structure and agricultural planting structure. The predicted results showed that soil Cd and As concentrations in all regions would not exceed regulatory limits after 100 years in the current scenario. A health risk assessment based on soil ingestion, dermal contact, and inhalation showed that the hazard indices for Cd and As were negligible, but their total carcinogenic risk reached notable levels. Over time, Cd-specific carcinogenic risk for children increased in several scenarios and transitioned from negligible to notable risk, with soil ingestion being the dominant exposure pathway. According to the results, targeted mitigation strategies, including the regulation of atmospheric deposition, optimization of irrigation water quality, and adoption of straw off-field practices, show potential to effectively limit the accumulation of potentially toxic elements in agricultural soils. Full article

Real-time grid monitoring of photovoltaic (PV) systems with battery storage requires continuous access to voltage, current, and power states. Traditional simulation models obtain these responses by calculating switching events, which limits the scale of real-time simulation. This paper proposes a Multi-Feature Dynamic Reconstruction (MFDR) method that combines environmental inputs, averaged converter states, and frequency-domain electrical variables. On the DC side, PV output is calculated from irradiance and temperature, while the bidirectional battery converter is represented by a low-frequency reconstruction model. On the AC side, dynamic phasors are used to convert the grid-connected inverter into a frequency-domain Norton equivalent for reconstructing its port voltage and current responses. Controller hardware-in-the-loop tests are conducted under different operating conditions. The reported peak and normalized tracking errors of the evaluated transient quantities remain below 3%. In the four-core IEEE 118-bus case, the average per-core CPU utilization decreases from 70.43% to 41.08%, while the maximum step execution time decreases from 38 μs to 27 μs. The model with 1069 state variables also operates within the fixed 50 μs simulation step. The results show that the MFDR method reduces the computational demand of real-time grid monitoring while retaining the voltage, current, and power responses. Full article

To address the challenge of reconstructing statistical wind pressure fields of membrane structures under sparse measurement conditions, this study proposes an FF-Res-GP-PINN framework based on wind tunnel test data from a ridge–valley membrane structure. The proposed framework integrates Gaussian Fourier feature mapping, a residual network, spatial gradient penalty, and Laplacian smoothing regularization. The model takes the spatial coordinates and incoming wind direction as inputs and outputs the mean and fluctuating wind pressure coefficients. A total of 675 samples under three wind directions, namely 0°, 45°, and 90°, were used for training and validation. The results show that compared with the conventional PINN, the proposed FF-Res-GP-PINN reduces the validation MSE from 0.226 to 0.155, corresponding to a reduction of 31.4%. The ablation study indicates that Gaussian Fourier feature mapping contributes most significantly to the improvement of pointwise prediction accuracy, whereas the spatial gradient penalty and Laplacian smoothing regularization mainly enhance the spatial continuity of the predicted wind pressure field. Compared with the Fourier-MLP model, the FF-Res-GP-PINN reduces the average gradient norm, Laplacian smoothing residual, and spatial oscillation index by approximately 98.5%, 96.3% and 62.2%, respectively. The sensitivity analysis of regularization weights further demonstrates that increasing λ_grad and λ_lap can enhance spatial smoothness, but may also increase the pointwise prediction error. Therefore, λ_grad = 2 × 10⁻³ and λ_lap = 1 × 10⁻⁴ were adopted in this study as a compromise between prediction accuracy and spatial regularization for the current dataset. The findings indicate that the proposed framework provides a feasible approach for reconstructing statistical wind pressure fields of membrane structures with a specific geometry and tested wind directions. Nevertheless, its predictive capability for unseen wind directions, different structural geometries, and transient wind pressure fields requires further validation. Full article

Accurate estimation of the state of health (SOH) of lithium-ion batteries is crucial for battery management systems (BMS), as it directly affects system safety, operational reliability, and remaining useful life prediction. Traditional SOH estimation methods generally rely on complete charge–discharge cycle data, which limits their practicality in real-time and online applications. To address this limitation, this paper proposes a novel voltage-charge increment curve-based dual-branch bidirectional long short-term memory network (VCIC-DB-BiLSTM) for battery SOH estimation under incomplete data. First, non-uniformly sampled battery current-voltage data are processed into standardized sequences with equal voltage intervals via voltage-charge increment curves based on ampere-hour integration and cubic spline interpolation. Subsequently, sliding window segmentation is applied to extract fixed-length curve segments from continuous voltage intervals as input features, while the corresponding complete voltage interval curves are used as labels. Finally, the VCIC-DB-BiLSTM network is designed, which uses a dual-branch structure to integrate feature extraction from both patch-processed and raw data, combined with bidirectional sequential modeling. Experimental validation on four benchmark datasets, CALCE, Oxford, XJTU, and TJU, demonstrates that the proposed method achieves competitive performance in SOH estimation under incomplete discharge data conditions, confirming its effectiveness and practical applicability. Full article

This study presents the design, manufacturing, and prototype-level evaluation of a high-efficiency interior permanent magnet synchronous motor (IPMSM) developed for compact electric vehicle traction applications. The proposed motor employs a 12-slot/10-pole spoke-type rotor topology and was evaluated in terms of electromagnetic performance, mechanical integrity, and thermal behavior. The slot–pole and winding configuration was assessed as part of the design evaluation, and the manufactured prototype was experimentally tested under different operating conditions. The experimental results were compared with numerical simulations using line-to-line back-EMF, efficiency maps, phase current–torque characteristics, and output power variation. At the nominal operating point of 7000 rpm and 3.5 Nm, the prototype delivered 2.5 kW output power with an experimental efficiency of 90.7%. The deviations between experimental and simulation results were 1.17% for phase current, 0.48% for line-to-line back-EMF, 1.18% for input power, and 1.20% for efficiency. Mechanical static structural finite element analysis indicated a rotor safety factor of 3.61 under the maximum centrifugal loading condition, while the resulting structural deformation remained sufficiently low to avoid adverse effects on air-gap alignment. In addition, the rotor incorporated an adhesive-free, mechanically disassemblable magnet-retention structure, which was mechanically evaluated under centrifugal loading and showed no magnet displacement, structural damage, or bolt-preload loss after testing. Thermal analysis and continuous-load experimental testing showed that the winding temperature remained around 80 °C under passive cooling conditions. Overall, the results demonstrate that the manufactured IPMSM prototype provides consistent electromagnetic performance, adequate mechanical reliability, and thermally safe operation for compact electric vehicle applications. Full article

Mastication is a fundamental rhythmic motor behavior controlled by a brainstem central pattern generator (CPG) located within the pontine and medullary reticular formations. Coordinated activation of jaw-opening and jaw-closing muscles is generated by this network and continuously refined through sensory feedback from periodontal mechanoreceptors and muscle spindles, together with descending inputs from the cortical masticatory area (CMA), basal ganglia, and cerebellum. Thus, mastication is regulated by distributed neural circuits rather than a single central locus. At the cellular level, the rhythmic activity of the masticatory CPG depends on the coordinated action of voltage-gated and ligand-gated ion channels. Recent electrophysiological and computational studies have identified candidate conductances that are proposed to underlie rhythm generation. Persistent sodium currents are proposed to facilitate burst initiation, whereas T-type calcium channels are thought to promote burst activation through post-inhibitory rebound. HCN channels may contribute to rhythmic timing, while calcium-activated potassium channels are thought to participate in burst termination. This review summarizes the hierarchical neural control of mastication and the biophysical mechanisms by which ion channels shape CPG rhythmogenesis. It also discusses the impact of channelopathies and neurodegenerative disorders on masticatory function, highlighting potential ion-channel-targeted therapeutic approaches for temporomandibular disorders, bruxism, and impaired mastication. Full article

This paper proposes a new hybrid switched inductor and switched capacitor (HSISC) buck–boost DC–DC converter. For the duty ratio above 28%, the proposed converter operates as a boost converter; otherwise, it acts as a buck converter. Compared with conventional buck–boost converters, incorporating the hybrid switched inductor and switched capacitor (HSISC), the network yields a substantial voltage gain at lower duty ratios. Being a non-isolated topology, high-frequency transformers and the associated issues are absent. Additionally, the proposed topology has the merits of continuous input current, making it suitable for renewable energy integration and vehicle-to-grid (V2G) applications, a wide range of duty ratio for boost operation, and ease of control as there are only two modes of operation with switches operating in a complementary manner. Operational analysis for the two modes, necessary mathematical derivations for component design, and a steady-state analysis of the converter are reported. The experimental findings for the converter, which were conducted at a duty ratio of 0.05 to 0.5 at a switching frequency of 10 kHz, are reported. The presented results provide proof-of-concept validation based on analytical and simulation studies, demonstrating the feasibility and operational characteristics of the proposed converter. Full article