Search Results (133)

Precise geometric registration between high-resolution satellite imagery and aerial orthophotos is essential for generating high-definition (HD) maps that support autonomous vehicle navigation. This study presents a comprehensive evaluation of multi-spectral band performance for image registration between KOMPSAT-3A satellite imagery (0.55 m resolution) and VWorld aerial orthophotos (0.25 m resolution) across seven patch size configurations. Five feature-matching approaches were systematically compared: LightGlue with CLAHE preprocessing, edge-based FFT methods (with and without CLAHE), and SIFT-based methods (with and without CLAHE). Two additional detector-free deep matchers, LoFTR and RoMa, were further integrated into the same pipeline for comparison. The experimental results reveal significant variations in registration accuracy across spectral bands, with the panchromatic-derived products (SPECPAN and EMPPAN) and luminance composite BT601 image demonstrating superior stability compared to individual visible and NIR bands. LightGlue achieved consistently high inlier counts (averaging 1100+ matched points) across all spectral bands and patch configurations, while SIFT with CLAHE preprocessing yielded the lowest matching RMSE (averaging 1.55 pixels). Among all matchers, the detector-free methods produced the densest and most stable correspondences, with LoFTR giving the best transformation stability, whereas edge-based methods were markedly less stable. However, an independent assessment against network GNSS check points showed a registration accuracy of approximately 2.8 m that was statistically similar across all matchers, indicating that matcher selection mainly affects correspondence density and transformation stability rather than independent geodetic accuracy. The achieved meter-level accuracy is suitable for HD map preprocessing and candidate GCP generation rather than final lane-level mapping, and the reported guidance is specific to the tested KOMPSAT-3A/VWorld setting. Full article

Reliable visual detection of small floating objects on the water surface is a prerequisite for environmental monitoring and clean-up tasks performed by unmanned surface vehicles (USVs) on inland waterways. Such scenes are routinely degraded by low illumination at dawn and dusk, strong specular reflections, ripple-induced clutter, and large object-scale variations, which together cause missed detections, false alarms, and unstable localization. Aiming at these practical challenges, this study conducts a scenario-oriented optimization and experimental validation based on the lightweight YOLOv11n detector. We integrate multiple mature attention mechanisms, regression loss functions and data augmentation strategies to develop an improved scheme, YOLOv11-LREP, for floating object detection. The detailed optimizations are as follows: (i) a Coordinate Attention (CoordAtt) module is inserted at the top of the backbone to enhance positional encoding and highlight obstacle-related semantic regions; (ii) three Efficient Channel Attention (ECA) modules are embedded at the multi-scale fusion nodes of the Neck so that reflection- and ripple-induced spurious channel responses can be suppressed at almost no extra cost; (iii) the Powerful-IoU (PIoU) loss replaces the original regression loss to enforce four-side boundary alignment and stabilize convergence on small, blurred-edge targets; and (iv) a joint low-light and reflection augmentation strategy, together with CutMix region-level mixing, broadens the training distribution along the illumination and occlusion axes. Experiments on the public FloW-Img dataset, split into 1200 training and 800 validation images (2024 instances) and run under a fixed random seed (seed = 0, deterministic = true), show that YOLOv11-LREP attains AP₅₀ = 80.1%, AP_50:95 = 38.5%, and AP_S = 24.3% with only 2.84 M parameters and 9.3 GFLOPs. On an NVIDIA RTX 4060 Laptop GPU, the model runs at 3.3 ms total per 640 × 640 image (≈303 FPS), satisfying real-time perception requirements while retaining lightweight deployability. The ablation results indicate that different components contribute differently to localization accuracy, small-object sensitivity, and robustness, and that the final configuration provides a balanced trade-off rather than the best value for every individual metric. A systematic threshold sensitivity analysis (F₁ fluctuation < 0.2%) demonstrates the stability of the final model. Full article

The increasing penetration of renewable energy sources and converter-interfaced loads has intensified the need for fast and reliable grid-support services. Although electric vehicle (EV) battery chargers have emerged as promising resources for Vehicle-to-Grid (V2G) applications, existing solutions typically focus on individual services such as virtual inertia or frequency regulation, while limited attention has been given to the coordinated provision of multiple ancillary services within a unified framework. Furthermore, the use of batteries alone for fast frequency support may accelerate battery degradation due to frequent high-power transients. To address these challenges, this paper proposes a hybrid energy storage-based EV battery charger architecture and a coordinated multi-timescale control strategy capable of simultaneously providing virtual inertia support, long-term frequency regulation, reactive power compensation, and harmonic mitigation. The proposed approach utilizes a DC-link capacitor to deliver fast inertial response while the battery supplies sustained frequency support, thereby reducing battery stress and improving energy management efficiency. An enhanced frequency estimation method based on a phase-locked loop combined with a low-pass filter is also introduced to improve dynamic performance. Simulation results demonstrate the effectiveness of the proposed strategy under various grid disturbances. The system achieves an equivalent virtual inertia constant of approximately 1.85 s and delivers up to 786 W of transient inertial support within 80 ms during frequency events. The enhanced frequency estimation method significantly reduces transient overshoot, while harmonic compensation limits the grid current and voltage total harmonic distortion to 1.50% and 3.23%, respectively. In addition, the controller provides up to 400 VAR of reactive power support during voltage disturbances while maintaining stable battery operation. These results demonstrate that the proposed EV battery charger can function as a multifunctional grid-support resource, enhancing frequency stability, voltage regulation, power quality, and overall V2G capability in future smart grids. Full article

Under the trend of intelligent transportation and connected vehicles, real-time control plays a vital role in improving bus system efficiency. Existing bus control strategies typically treat buses as homogeneous points and achieve system equilibrium by maintaining consistent headways. However, this simplification overlooks differences in dynamic responses and the evolution of powertrain lifespan arising from vehicle heterogeneity. It converts the sparse constraint problem, which is intended to ensure timely arrival, into a hard constraint on the vehicle trajectory over the entire time horizon, thereby excessively restricting individual optimal evolutionary paths and causing the optimization process to become trapped in a local optimum. To this end, this paper proposes SMATD3, a multi-agent cooperative control algorithm that accounts for vehicle heterogeneity. By adopting a centralized training and decentralized execution paradigm and avoiding the specification of a fixed inter-vehicle spacing target, the algorithm enables each vehicle to adaptively adjust its speed control strategy according to its own dynamic characteristics, thereby achieving the coordinated optimization of system equilibrium and individual objectives. The simulation results indicate that the proposed method can effectively suppress bus tailgating and achieve the coordinated multi-objective optimization of operational stability, passenger travel efficiency, energy consumption, and battery health. From a sustainability perspective, improved headway regularity and service reliability can enhance public transit attractiveness and support mode shift, while smoother energy use and reduced battery degradation lower lifecycle impacts. Full article

Accurate prediction of loan defaults is essential for financial institutions seeking to minimize credit losses and maintain portfolio stability. In the vehicle financing segment of emerging markets, real-world datasets frequently exhibit extreme class imbalance ratios that far exceed those encountered in standard benchmark corpora, posing severe challenges for conventional machine learning pipelines. This study introduces a gradient boosting framework integrating Weight of Evidence (WoE) transformation, Bayesian hyperparameter optimization, and three complementary classifiers—Extreme Gradient Boosting (XGBoost), Light Gradient Boosting Machine (LightGBM), and Categorical Boosting (CatBoost)—to predict vehicle loan default risk. The methodology is evaluated on a large-scale, fully anonymized Turkish vehicle loan dataset ($N=207,572$) with an extreme imbalance ratio of 1:1133 (183 defaults versus 207,389 non-defaults). A strict three-way data partition (60% training, 20% validation, 20% test) is adopted to ensure leakage-free model selection and unbiased performance estimation. A multi-stage experimental pipeline is developed encompassing: (i) statistical feature selection via Mann–Whitney U and chi-square tests with adaptive thresholding, (ii) a comparative analysis of seven resampling strategies including Synthetic Minority Oversampling Technique (SMOTE) variants, Adaptive Synthetic Sampling (ADASYN), and focal loss weighting, (iii) a greedy forward selection ensemble procedure for heterogeneous model fusion, and (iv) a systematic training-set size sensitivity analysis across eight majority undersampling ratios. Under the leakage-free evaluation protocol, the highest-AUC individual model (LightGBM with SMOTE-ENN) achieves an Area Under the Curve (AUC) Receiver Operating Characteristic (ROC) of 0.710 (95% bootstrap CI: 0.614–0.798), while CatBoost with cost-sensitive weighting exhibits superior operational metrics (KS $=0.389$, PR-AUC $=0.011$). The greedy ensemble procedure exhibits high selection instability with only 37 validation-set positives, providing a methodological finding on the minimum sample requirements for reliable ensemble construction under extreme scarcity. Ablation results confirm that WoE encoding contributes 3.1 percentage points to the overall AUC gain. Tree SHAP-based interpretability analysis identifies the financing-to-age ratio, WoE-encoded occupation group, and log financing amount as the primary predictive drivers, with cross-model stability confirmed via Spearman rank correlation. A decision support analysis provides precision–recall curves, a Brier score of 0.0082, reliability diagrams, and threshold-dependent performance at operationally plausible review rates. Fairness evaluation across gender and marital status subgroups demonstrates that threshold-dependent metrics such as Disparate Impact Ratio and Equalized Odds Gap are inherently compromised under extreme minority scarcity, whereas rank-based subgroup AUC analysis with bootstrap 95% confidence intervals preserves meaningful discriminative assessment. These findings provide an empirically validated framework for credit default prediction in highly imbalanced and data-scarce financial environments. Full article

On-site hydrogen refueling stations (HRS) face significant operational challenges due to the stochastic nature of hydrogen demand, creating a severe supply–demand mismatch. Under traditional pressure-based hysteresis control, this volatility forces Proton Exchange Membrane (PEM) electrolyzers into frequent start–stop cycles, accelerating degradation and reducing efficiency. In response, this study introduces an automated control framework integrating macroscopic gas-state modeling with deep-learning-based demand prediction. First, a real-gas thermodynamic model was established. Monte Carlo simulations of 100 random filling scenarios identified a robust design benchmark of 4.5 kg per vehicle. A low filling stability coefficient (5.02%) confirmed that individual thermodynamic fluctuations are negligible, validating a traffic-flow-driven demand approach. Next, a deep Long Short-Term Memory (LSTM) network was developed to forecast short-term demand. Trained on an 8784 h dataset exhibiting “double-peak” traffic patterns, the model achieved high precision on the unseen test set, yielding a Root Mean Square Error (RMSE) of 6.75 kg and a normalized RMSE (nRMSE) of 0.0987, explaining 82% of the demand variance. Finally, an LSTM-informed demand-following control strategy was formulated to enable proactive, thermally bounded operation alongside a novel “Hot Standby” mechanism. Maintaining a minimal 3.0 kg/h holding current during idle periods sustains stack temperatures above 60 °C, effectively mitigating thermal stress. Comparative simulations over 1464 h demonstrated that the proposed framework reduces detrimental cold start–stop cycles by 98.4% (from 61 to 1) and suppresses power output fluctuations by 40.7% compared to the traditional baseline. These results confirm that data-driven control significantly enhances operational stability, facilitates grid integration, and extends core equipment service life. Full article

To address the power quality deterioration caused by the large-scale integration of grid-following (GFL) electric vehicle charging stations, this paper proposes a comprehensive assessment method based on relative entropy distance fusion weighting and a dynamic ideal solution VIKOR algorithm. First, a multi-dimensional power quality evaluation system is constructed, focusing on key indicators such as voltage deviation, frequency deviation, three-phase imbalance, and harmonic distortion, to accommodate the operational characteristics of vehicle-to-grid (V2G) under grid-following and grid-forming (GFM) interaction scenarios. Building on this, the three-scale analytic hierarchy process (AHP) is employed to determine subjective weights, while the divergence-maximized entropy weight method is used to derive objective weights. The relative entropy distance model is then applied to achieve adaptive fusion of subjective and objective weights, resulting in an optimal combined weighting. Subsequently, a dynamic ideal solution mechanism is introduced into the VIKOR algorithm, where the range of the ideal solution is adjusted based on the indicator weights to enhance the discrimination of key indicators. By comprehensively calculating the group utility value, individual regret value, and compromise evaluation index, accurate ranking and performance assessment of different mitigation schemes are achieved. Using measured data from a vehicle-grid interaction demonstration base for analysis, the results demonstrate that the proposed method can effectively quantify the actual effects of various mitigation schemes, providing decision-making support for power grid safety and stability under high penetration of renewable energy and converter-interfaced generation. Full article

With the large-scale integration of electric vehicles (EVs) into the power grid, uncoordinated charging behavior has aggravated load fluctuations in the power system. Deep reinforcement learning can optimize EV charging and discharging strategies through dynamic decision-making, thereby alleviating the operational pressure imposed on the grid by load variations. However, under large-scale EV integration scenarios, challenges still remain, including the excessively high dimensionality of the state space and the resulting decline in training efficiency. In addition, the coupling between existing clustering methods and dynamic scheduling mechanisms is still insufficiently tight. To address these issues, this study proposes a cluster-based deep reinforcement learning method for EV charging and discharging scheduling, referred to as CDRL. First, a probabilistic behavioral model is constructed based on EV charging transaction data to characterize the stochasticity of user charging behavior. A Density–Centroid Hybrid Clustering (DCHC) method is then adopted to cluster the charging behavior characteristics of EVs. Subsequently, at the cluster level, a day-ahead base load forecasting model is introduced, and the forecasting results are fed into a mixed-integer linear programming (MILP) model to generate the charging and discharging power allocation tasks for each cluster. At the individual level, the EV charging and discharging process is formulated as a Markov decision process (MDP), and a deep Q-network (DQN) is employed for policy learning, thereby achieving the decomposition of cluster-level tasks into individual scheduling decisions. The simulation results demonstrate that the proposed method can effectively reduce charging costs and smooth system load fluctuations while improving training convergence speed and policy stability. Full article

The rising market penetration of new energy vehicles (NEVs) is transforming urban traffic into a heterogeneous mix of battery electric (BEVs), hybrid electric (HEVs), and conventional fuel vehicles (FVs). For analytical brevity, traditional internal combustion engine vehicles (ICEVs) are hereafter referred to as ‘fuel vehicles (FVs)’ in the discussion of New Energy Vehicle (NEV) networks. This research investigates the efficacy of centralized coordination for NEVs within a localized region, as opposed to individualized speed control, in enhancing the mitigation of traffic congestion. Evaluating traffic efficiency and decarbonization strategies in such settings often requires extensive random sampling and Monte Carlo simulations over a large set of parameter combinations. However, conventional microscopic traffic simulators, which rely on fine-grained modeling of vehicle dynamics and signal control, incur prohibitive computational time when scaled to large networks and numerous experimental scenarios. In this study, battery electric vehicles and hybrid electric vehicles are designed as density-aware vehicles, whose movement speed is adaptively adjusted according to the regional traffic density in their vicinity and the control parameter β. In contrast, fuel vehicles adopt a stochastic movement speed and, together with other vehicle types, exhibit either movement or stoppage in the lattice environment. This density-driven speed-adaptive control and lattice arbitration mechanism is intended to reproduce, in a simplified yet extensible manner, changes in mobility and traffic-flow stability under high-density traffic conditions. The simulation results indicate that, under the same Manhattan road network and vehicle-density conditions, tuning the β parameter of new energy vehicles to reduce their movement speed in high-density areas and to mitigate abrupt position changes can suppress traffic-flow oscillations, delay the onset of the congestion phase transition, and promote spatial equilibrium of traffic flow. Meanwhile, this study develops simplified energy-consumption and carbon emission models for battery electric vehicles, hybrid electric vehicles, and fuel vehicles, demonstrating that incorporating a speed-adaptive density strategy into mixed traffic flow not only helps alleviate abnormal congestion but also reduces potential energy use and carbon emissions caused by congestion and stop-and-go behavior. From a sensing and practical perspective, the proposed framework assumes that future connected and autonomous vehicles (CAVs) can estimate vehicle states and local traffic density through GNSS–IMU multi-sensor fusion and V2X communications, indicating methodological consistency between the proposed model and real-world CAV sensing capabilities and making it a suitable and effective experimental platform for investigating the relationships among new energy vehicle penetration, density-control strategies, and carbon footprint. Full article

The Yellow River Basin contains abundant coal resources; however, its ecological environment is inherently fragile, and vegetation degradation has been further intensified by extensive mining activities. Accurate classification of individual tree species in mining-affected areas is therefore essential for assessing ecological conditions and establishing a scientific foundation for targeted restoration and sustainable management. To address this need, an evaluated machine learning framework was developed and evaluated for individual tree species classification in a coal mining area of the Yellow River Basin using integrated unmanned aerial vehicle (UAV) data. A comprehensive feature set was constructed by extracting 278 attributes per tree. These attributes included 224 spectral bands and 29 hyperspectral indices derived from hyperspectral imagery, 24 textural metrics obtained from RGB orthophotos, and one canopy height feature generated from a LiDAR-derived model. Based on ground-truth data from 1095 individual trees, seven machine learning algorithms were trained and systematically compared: Random Forest (RF), Support Vector Machine (SVM), K-Nearest Neighbors (KNN), Decision Tree (DT), Gradient Boosting (GB), Logistic Regression (LR), and XGBoost. Statistical significance testing using 5 × 5 repeated cross-validation, together with the Friedman test and post hoc Nemenyi test, and additional model stability analysis consistently identified XGBoost as the optimal classifier. On an independent test set, XGBoost achieved high accuracy (Overall Accuracy = 0.897, Kappa = 0.811) with an efficient training time of 2.36 s. Further analysis demonstrated the critical and complementary roles of hyperspectral and structural features in species discrimination. The optimized model was subsequently applied to generate a detailed wall-to-wall tree species map across the entire mining area. Overall, this study presents a statistically informed comparison of classifiers for multi-source feature-based species discrimination and delivers an evaluated and practical pipeline for effective vegetation monitoring. The proposed framework provides a scientific tool for assessing and managing ecological recovery in complex mining environments, particularly within ecologically sensitive regions such as the Yellow River Basin. Full article

This paper presents a distributed hierarchical model predictive control (MPC) framework designed to ensure dynamic consensus and stability in nonlinear vehicle platoons, addressing challenges posed by mixed communication topologies and hard constraints. By directed graph modeling of the mixed communication topologies, the dynamic consensus goal for the platoon is defined by the inter-vehicle distances between the host and its neighbors, whereas the stability criterion for an individual vehicle is expressed as a positive definite function of its position and velocity deviations. Then, a contractive constraint is elegantly designed to correlate these two objectives in a hierarchical model predictive control framework, where the lower layer optimizes the stability objective and the upper layer optimizes the dynamic consensus objective. The conditions ensuring stability and string stability for the vehicle platoon are shown to be only dependent on the deviations of the host vehicle, which achieves dynamic consensus and string stability simultaneously for nonlinear vehicle platoons. Several representative scenarios are used to validated the performance of the proposed strategy. Full article

With the widespread integration of smart buildings equipped with distributed photovoltaics (PV) and electric vehicles (EVs), distribution networks face significant challenges arising from source-load fluctuations. Conventional centralized dispatch approaches are constrained by communication bottlenecks and data privacy requirements. These limitations make it difficult to achieve global coordination while preserving the autonomy of individual entities. This paper proposes a hierarchical coordination framework for the coordinated operation of distribution networks and smart buildings. The distribution management system (DMS) and building energy management systems (BEMSs) perform independent optimization within their respective domains. Only aggregated boundary power information is exchanged to protect data privacy, enabling cross-entity coordination under information boundary constraints. Building-side models incorporating thermal dynamics, EV charging and discharging, and PV generation are developed, along with a distribution network power flow model. To solve the coordinated optimization problem, an Anderson-accelerated alternating direction method of multipliers (AA-ADMM) is introduced. A safeguarding mechanism based on combined residuals is incorporated to enhance convergence efficiency and stability. Case studies on the IEEE 33-bus test system demonstrate that compared with the uncoordinated baseline, the proposed method reduces network loss by 12.1% and lowers PV curtailment from 9.20% to 0.52%, while improving voltage profiles without significantly compromising occupant comfort or EV travel requirements. In addition, AA-ADMM achieves convergence with up to 66% fewer iterations than standard ADMM. Full article

Background/Objectives: Plasmids are key vehicles for the dissemination of antimicrobial resistance (AMR), yet their contribution to the global resistome architecture of Escherichia coli remains poorly resolved. This study aimed to quantify how plasmid backbones shape the distribution, mobility, and stabilization of resistance genes across diverse phylogenetic backgrounds. Methods: We analyze 9700 high-quality genomes spanning major phylogroups and sequence types. Plasmidome reconstruction was integrated with lineage-resolved antimicrobial resistance gene (ARG) mapping to characterize plasmid–ARG associations and evolutionary patterns. Results: Although most antimicrobial resistance genes (ARGs) are chromosomal, plasmids disproportionately encode clinically important determinants including bla_NDM-5, mcr-1.1, and multiple bla_CTX-M alleles that show strong, recurrent associations with a restricted set of backbone families, most notably IncX3, IncX4, IncI, and IncF. These conserved plasmid–gene modules recur across phylogenetic backgrounds and continental scales. We identify a marked divergence in evolutionary strategies: generalist phylogroups (A, B1, D) maintain plasmid-rich and highly diverse resistomes, whereas globally dominant Extraintestinal Pathogenic E. coli (ExPEC) clones such as ST131 and ST410 exhibit reduced plasmid dependency and frequent chromosomal integration of extended-spectrum β-lactamase (ESBL) genes, particularly bla_CTX-M-15, consistent with a shift toward vertically stabilized resistomes. By integrating plasmidome reconstruction with lineage-resolved ARG mapping, this study delivers the most extensive plasmid-focused resistome analysis to date, revealing highly modular plasmid–ARG networks structured around a small number of high-risk backbone types. These backbones account for the majority of globally relevant ARGs, including 64.6% of bla_NDM-5 and 76.4% of mcr-1.1 detections. Conclusions: Together, our findings establish plasmid lineages rather than individual genes or clones as central units of AMR dissemination and critical targets for future genomic surveillance and intervention strategies. Full article

The adoption of electric vehicles (EVs) is a promising strategy for reducing emissions and promoting sustainable mobility. The increasing adoption of EVs has created a demand for efficient and sustainable charging infrastructure. The integration of rooftop solar-powered EV charging stations into distribution networks is a promising solution for reducing carbon emissions and improving grid efficiency. This integration also introduces challenges, such as power quality issues, grid instability, and the impact of environmental factors on solar generation. This study proposes a novel system that integrates a smart control algorithm for a central battery management system (CBMS) with advanced bidirectional DC-DC converters for optimised power distribution. Unlike existing systems that focus on individual components, this study combines real-time environmental monitoring with adaptive power management algorithms to handle variations in generation owing to solar irradiance, temperature, and shading, and ensure maximum power harvesting. This study also presents the role of the DC-to-DC converter integrated with a smart charging control and CBMS in smart grid-enabled EV charging station. The proposed system was validated using MATLAB 2025b Simulink simulations. This study demonstrates an improvement in overall grid stability and highlights the potential of DC-DC converter technologies for smart grid applications and decarbonisation efforts. Full article

Forest pests and diseases are some of the major disturbances affecting the stability of forest ecosystems. Accurate identification of insect-infested trees is therefore crucial for assessing forest health and implementing precision forestry management. This study focuses on stand-level detection of cypress trees (Cupressus funebris Endl.) that were affected by the cypress bark beetle (Phloeosinus aubei Perris), and the framework enables individual tree segmentation, insect-infested tree detection, and stand infestation assessment. Firstly, individual trees were extracted from Light Detection and Ranging (LiDAR) point cloud data using the layer-stacking seed point algorithm. Based on the segmented tree crowns, four vegetation indices (Visible Atmospherically Resistant Index (VARI), Visible-band Difference Vegetation Index (VDVI), Red-Green Index (RGI), and Color Index of Vegetation Extraction (CIVE)) were calculated from Unmanned Aerial Vehicle (UAV) RGB imagery. Insect-infested cypress trees were extracted through threshold segmentation. Through visual interpretation, the optimal vegetation index was determined and the infestation rate at the stand level was calculated. Based on the above framework, a total of 1368 trees were identified in the cypress stand, with a segmentation Precision of 82.51%, a Recall of 80.00%, and an F1-score of 81.24%. RGI achieved the best performance (Precision = 100.00%, Recall = 86.96%, F1-score = 93.02%) and identified 20 infested trees, accounting for 1.46% of the cypress stand. Supplementary experiments further confirm the superiority of the RGI index and the μ ± 2σ thresholding method. These results demonstrate that the proposed method enables rapid detection of the infested cypress trees, effective monitoring of stand health and infestation severity, thereby supporting informed decision-making in pest control and forest management. Full article