Search Results (20,814)

The Centralized Teacher with Decentralized Student (CTDS) framework is a multi-agent reinforcement learning (MARL) approach that utilizes knowledge distillation within the Centralized Training with Decentralized Execution (CTDE) paradigm. In this framework, a teacher module learns optimal Q-values using global observations and distills this knowledge to a student module that operates with only local information. However, CTDS has limitations including inefficient knowledge distillation processes and performance gaps between teacher and student modules. This paper proposes the evolutionary sampling method that employs genetic algorithms to optimize selective knowledge distillation in CTDS frameworks. Our approach utilizes a selective sampling strategy that focuses on samples with large Q-value differences between teacher and student models. The genetic algorithm optimizes adaptive sampling ratios through evolutionary processes, where the chromosome represent sampling ratio sequences. This evolutionary optimization discovers optimal adaptive sampling sequences that minimize teacher–student performance gaps. Experimental validation in the $StarCraft$ Multi-Agent Challenge (SMAC) environment confirms that our method achieved superior performance compared to the existing CTDS methods. This approach addresses the inefficiency in knowledge distillation and performance gap issues while improving overall performance through the genetic algorithm. Full article

In order to determine thermally safe driving parameters of ignition coils for hydrogen internal combustion engines (ICE), a reliable estimation of internal power losses is essential. These losses include resistive winding losses, magnetic core losses due to hysteresis and eddy currents, dielectric losses in the insulation, and electronic switching losses. Direct experimental assessment is difficult because the components are inaccessible, while conventional computer-aided engineering (CAE) approaches face challenges such as the need for accurate input data, the need for detailed 3D models, long computation times, and uncertainties in loss prediction for complex structures. To address these limitations, we propose an artificial intelligence (AI)-based framework for estimating internal losses from external temperature measurements. The method relies on an artificial neural network (ANN), trained to capture the relationship between external coil temperatures and internal power losses. The trained model is then employed within an optimization process to identify losses corresponding to experimental temperature values. Validation is performed by introducing the identified power losses into a CAE thermal model to compare predicted and experimental temperatures. The results show excellent agreement, with errors below 3% across the −30°C to 125°C range. This demonstrates that the proposed hybrid ANN–CAE approach achieves high accuracy while reducing experimental effort and computational demand. Furthermore, the methodology allows for a straightforward determination of the coil safe operating area (SOA). Starting from estimates derived from fitted linear trends, the SOA limits can be efficiently refined through iterative verification with the CAE model. Overall, the ANN–CAE framework provides a robust and practical tool to accelerate thermal analysis and support coil development for hydrogen ICE applications. Full article

Driven by the urgent need to decarbonize the logistics sector—where conventional vehicles exhibit high energy consumption and emissions, posing significant environmental sustainability challenges—electrification represents a pivotal strategy for reducing emissions and achieving sustainable urban freight transport. Despite rising global electric vehicle sales, the penetration rate of electric logistics vehicles (ELVs) remains comparatively low, impeding progress toward sustainable logistics objectives. Battery-swapping mode (BSM) has emerged as a potential solution to enhance operational efficiency and economic viability, thereby accelerating sustainable adoption. This model improves ELV operational efficiency through rapid battery swaps at centralized stations. This study constructs a tripartite evolutionary game model involving government, consumers, and BSM-ELV manufacturers to analyze market dynamics under diverse strategies. Key considerations include market scale, government environmental benefits, battery leasing/purchasing costs, lifecycle cost analysis (via discount rates), and resource efficiency (reserve battery ratio $\lambda$). MATLAB-2021b-based simulations predict participant strategy evolution paths. Findings reveal that market size and manufacturer expectations significantly influence governmental and manufacturing strategies. Crucially, incorporating discount rates demonstrates that battery leasing reduces consumer enterprises’ initial investment, enhancing economic sustainability and cash flow while offering superior total cost of ownership. Furthermore, gradual reduction of government subsidies effectively stimulates market self-regulation, incentivizes leasing adoption, and bolsters long-term economic/operational sustainability. Market feedback can guide policy adjustments toward fiscally sustainable support mechanisms. This study proposes the following management implications for advancing sustainable logistics: 1. Governments should phase out subsidies systematically to foster market resilience; 2. Manufacturers must invest in BSM R&D to improve efficiency and resource circularity; 3. Consumer enterprises can achieve economic benefits and emission reductions by adopting BSM-ELVs. Full article

Understory-cultivated Panax ginseng possesses high pharmacological and economic value; however, its visual quality grading predominantly relies on subjective manual assessment, constraining industrial scalability. To address challenges including fine-grained morphological variations, boundary ambiguity, and complex natural backgrounds, this study proposes Ginseng-YOLO, a lightweight and deployment-friendly object detection model for automated ginseng grade classification. The model is built on the YOLOv11n (You Only Look Once11n) framework and integrates three complementary components: (1) C2-LWA, a cross-stage local window attention module that enhances discrimination of key visual features, such as primary root contours and fibrous textures; (2) ADown, a non-parametric downsampling mechanism that substitutes convolution operations with parallel pooling, markedly reducing computational complexity; and (3) Slide Loss, a piecewise IoU-weighted loss function designed to emphasize learning from samples with ambiguous or irregular boundaries. Experimental results on a curated multi-grade ginseng dataset indicate that Ginseng-YOLO achieves a Precision of 84.9%, a Recall of 83.9%, and an mAP@50 of 88.7%, outperforming YOLOv11n and other state-of-the-art variants. The model maintains a compact footprint, with 2.0 M parameters, 5.3 GFLOPs, and 4.6 MB model size, supporting real-time deployment on edge devices. Ablation studies further confirm the synergistic contributions of the proposed modules in enhancing feature representation, architectural efficiency, and training robustness. Successful deployment on the NVIDIA Jetson Nano demonstrates practical real-time inference capability under limited computational resources. This work provides a scalable approach for intelligent grading of forest-grown ginseng and offers methodological insights for the design of lightweight models in medicinal plants and agricultural applications. Full article

This paper introduces the ProVANT Simulator, a comprehensive environment for developing and validating control algorithms for Unmanned Aerial Vehicles (UAVs). Built on the Gazebo physics engine and integrated with the Robot Operating System (ROS), it enables reliable Software-in-the-Loop (SIL) and Hardware-in-the-Loop (HIL) testing. Addressing key challenges such as modeling complex multi-body dynamics, simulating disturbances, and supporting real-time implementation, the framework features a modular architecture, an intuitive graphical interface, and versatile capabilities for modeling, control, and hardware validation. Case studies demonstrate its effectiveness across various UAV configurations, including quadrotors, tilt-rotors, and unmanned aerial manipulators, highlighting its applications in aggressive maneuvers, load transportation, and trajectory tracking under disturbances. Serving both academic research and industrial development, the ProVANT Simulator reduces prototyping costs, development time, and associated risks. Full article

Non-cooperative spatial target detection plays a vital role in enabling autonomous on-orbit servicing and maintaining space situational awareness (SSA). However, due to the limited computational resources of onboard embedded systems and the complexity of spaceborne imaging environments, where spacecraft images often contain small targets that are easily obscured by background noise and characterized by low local information entropy, many existing object detection frameworks struggle to achieve high accuracy with low computational cost. To address this challenge, we propose YOLO-GRBI, an enhanced detection network designed to balance accuracy and efficiency. A reparameterized ELAN backbone is adopted to improve feature reuse and facilitate gradient propagation. The BiFormer and C2f-iAFF modules are introduced to enhance attention to salient targets, reducing false positives and false negatives. GSConv and VoV-GSCSP modules are integrated into the neck to reduce convolution operations and computational redundancy while preserving information entropy. YOLO-GRBI employs the focal loss for classification and confidence prediction to address class imbalance. Experiments on a self-constructed spacecraft dataset show that YOLO-GRBI outperforms the baseline YOLOv8n, achieving a 4.9% increase in mAP@0.5 and a 6.0% boost in mAP@0.5:0.95, while further reducing model complexity and inference latency. Full article

As a consequence of the gradual exhaustion of apatite ore reserves, intensive comminution has been implemented in mineral processing operations to enhance phosphorus liberation. Consequently, improving the flotation efficiency of fine particles has remained a persistent challenge within the phosphate industry. The performance of flotation columns is strongly affected by the interaction between gas (bubble) and particle. The present research was designed to evaluate how certain process variables and chemical dosages influence gas holdup and its correlation with the column flotation performance of fine particles derived from a low-grade apatite ore. Column flotation experiments were conducted employing a factorial experimental approach to evaluate the effects of air flow rate, surfactant concentration, collector dosage, and depressant dosage on gas holdup, P₂O₅ grade, and recovery. The results made it possible to identify the levels of gas holdup that lead to appropriate values of P₂O₅ grade and recovery simultaneously, and their relation with the operating variables and reagent dosage. Gas holdup values higher than 23.5% led to the desired values of P₂O₅ grade (>30%) and recovery (>60%) simultaneously. Statistical models were developed with high correlation coefficients (R² > 0.98) to predict P₂O₅ grade and recovery as functions of the operating variables. This research provides a comprehensive framework of the gas holdup effect on column flotation systems, offering significant potential for improving the economic viability of low-grade phosphate ore processing. Full article

Accurate photovoltaic (PV) power output forecasting is critical for ensuring stable operation of modern power systems, yet it is constrained by high-dimensional redundancy in input weather data and the inherent heterogeneity of output scenarios. To address these challenges, this paper proposes a novel cascaded data-driven forecasting approach that enhances forecasting accuracy through systematically improving and optimizing the feature extraction, scenario clustering, and temporal modeling. Firstly, guided by weather data–PV power output correlations, the Deep Autoencoder (DAE) is enhanced by integrating Pearson Correlation Coefficient loss, reconstruction loss, and Kullback–Leibler divergence sparsity penalty into a multi-objective loss function to extract key weather factors. Secondly, the Fuzzy C-Means (FCM) algorithm is comprehensively refined through Mahalanobis distance-based sample similarity measurement, max–min dissimilarity principle for initial center selection, and Partition Entropy Index-driven optimal cluster determination to effectively cluster complex PV power output scenarios. Thirdly, a Long Short-Term Memory–Temporal Pattern Attention (LSTM–TPA) model is constructed. It utilizes the gating mechanism and TPA to capture time-dependent relationships between key weather factors and PV power output within each scenario, thereby heightening the sensitivity to key weather dynamics. Validation using actual data from distributed PV power plants demonstrates that: (1) The enhanced DAE eliminates redundant data while strengthening feature representation, thereby enabling extraction of key weather factors. (2) The enhanced FCM achieves marked improvements in both the Silhouette Coefficient and Calinski–Harabasz Index, consequently generating distinct typical output scenarios. (3) The constructed LSTM–TPA model adaptively adjusts the forecasting weights and obtains superior capability in capturing fine-grained temporal features. The proposed approach significantly outperforms conventional approaches (CNN–LSTM, ARIMA–LSTM), exhibiting the highest forecasting accuracy (97.986%), optimal evaluation metrics (such as Mean Absolute Error, etc.), and exceptional generalization capability. This novel cascaded data-driven model has achieved a comprehensive improvement in the accuracy and robustness of PV power output forecasting through step-by-step collaborative optimization. Full article

Accurate State of Health (SOH) estimation is critical for ensuring the safe operation of lithium-ion batteries. However, current data-driven approaches face significant challenges: insufficient feature extraction and ambiguous physical meaning compromise prediction accuracy, while initialization sensitivity to noise undermines stability; the inherent nonlinearity and temporal complexity of battery degradation data further lead to slow convergence or susceptibility to local optima. To address these limitations, this study proposes an enhanced CNN-BIGRU model. The model replaces conventional random initialization with a Hiking Optimization Algorithm (HOA) to identify superior initial weights, significantly improving early training stability. Furthermore, it integrates an Attention mechanism to dynamically weight features, strengthening the capture of key degradation characteristics. Rigorous experimental validation, utilizing multi-dimensional features extracted from the NASA dataset, demonstrates the model’s superior convergence speed and prediction accuracy compared to the CNN-BIGRU-Attention benchmark. Compared with other methods, the HOA-CNN-BIRGU-Attention model proposed in this study has a higher prediction accuracy and better robustness under different conditions, and the RMSEs on the NASA dataset are all controlled within 0.01, with R2 kept above 0.91. The RMSEs on the University of Maryland dataset are all below 0.006, with R2 kept above 0.98. Compared with the CNN-BIGRU-ATTENTION baseline model without HOA optimization, the RMSE is reduced by at least 0.15% across different battery groups in the NASA dataset. Full article

With the advancement of China’s industrialization, demand for express freight transportation has been rising. However, high-speed rail freight faces challenges, such as relatively low transport efficiency and lower revenues, compared with air and road modes. To address these issues, this paper focuses on freight train operations. First, it analyzes key influencing factors, including operating costs and benefits. Next, it conducts a comprehensive assessment of train consist capacity, freight node capacity, transport demand, and the number of freight services, and formulates an operational planning model that maximizes rail revenue, minimizes intermediate stops, and satisfies freight demand. Finally, an Improved Golden Jackal Optimization–based Genetic Algorithm (IGJOGA) is proposed to solve the model. Simulation results indicate that IGJOGA achieves higher solution efficiency than a traditional genetic algorithm for the freight train operation planning problem, and the results can provide a practical reference for freight train set operation schemes. Full article

Weapon systems have transitioned from hardware-centered designs to software-driven platforms, introducing new cybersecurity risks, including software manipulation and cyberattacks. To address these challenges, this study proposes an improved manufacturing readiness level assessment (MRLA) method that integrates cybersecurity capabilities into the evaluation process to address the gaps in hardware-focused practices in South Korea. Based on the MITRE adversarial tactics, techniques, and common knowledge, and the defensive cybersecurity framework, this study identified security requirements, assessed vulnerabilities, and constructed exploratory testing scenarios using defense trees. These methods evaluate system resilience, the effectiveness of security controls, and response capabilities under diverse attack scenarios. The proposed MRLA approach incorporates cyberattacks and defense scenarios that may occur in operational environments. This approach was validated through a case study involving unmanned vehicle systems, where the modified MRLA successfully identified and mitigated critical cybersecurity threats. Consequently, the target operational mode summary/mission profile of a weapon system can be revised based on practical considerations, enhancing the cybersecurity assessments and thereby improving the operational readiness of weapon systems through scenario-based, realistic evaluation frameworks. The findings of this study demonstrate the practical utility of incorporating cybersecurity evaluations into MRLA, contributing to more resilient defense systems. Full article

Cutting operational costs is a critical component for transportation agencies. To reduce these costs, agencies must optimize their scheduling. Typically, the total operating costs of transport include vehicle expenses and driver wages. Solving such tasks is complex, and optimal planning is usually broken down into multiple stages. These stages can include vehicle scheduling, driver shift planning, and driver assignment. This paper focuses specifically on developing a near-optimal driver schedule for a specified set of vehicle schedules. It shows how to efficiently assign drivers to predetermined optimal vehicle routes while ensuring compliance with regulatory constraints on driving hours. We address this challenge using a mathematical model based on the set covering problem, building on a framework established perviously. The set covering problem is typically formulated as an integer programming problem, solvable through column generation techniques. Our algorithm combines this method with heuristics, taking into account the practical aspects of the problem. The article also presents a computational analysis of the method using benchmark and real data. Full article

A smart irrigation system based on soil moisture sensors supported by photovoltaic energy is an innovation to address water use efficiency in the agricultural sector, especially in remote areas. This technology utilizes photovoltaic panels as a renewable energy source to operate water pumps, while soil moisture sensors provide real-time data that is used to automatically manage irrigation according to plant needs. This technology not only increases the efficiency of water and energy use but also supports environmental conservation by reducing dependence on fossil fuels. This research was conducted using a Systematic Literature Review (SLR) approach guided by the PRISMA framework to analyze trends, benefits, and challenges in implementing this technology. The analysis results show that this system offers various advantages, including energy efficiency, reduced carbon emissions, and ease of management through the integration of Internet of Things (IoT) technology. Several challenges remain, such as high initial investment costs, limited network access, and obstacles. Technical matters related to installation and maintenance. Various solutions have been proposed, including providing subsidies for small farmers, implementing radiofrequency modules, and using modular designs to simplify implementation. This study contributes to the development of a conceptual framework that can be adapted to various geographic and socio-economic conditions. Potential further developments include the integration of artificial intelligence and additional sensors to increase efficiency and support the sustainability of the agricultural sector globally. Full article

With the accelerating global energy transition, wind power has become a core pillar of renewable energy systems. However, its inherent intermittency and volatility pose significant challenges to the safe, stable, and economical operation of power grids—making ultra-short-term wind power prediction a critical technical link in optimizing grid scheduling and promoting large-scale wind power integration. Current forecasting techniques are plagued by problems like the inadequate representation of features, the poor separation of features, and the challenging clarity of deep learning models. This study introduces a method for the prediction of wind energy using spatiotemporal contrastive learning, employing seasonal trend decomposition to encapsulate the diverse characteristics of time series. A contrastive learning framework and a feature disentanglement loss function are employed to effectively decouple spatiotemporal features. Data on geographical positions are integrated to simulate spatial correlations, and a convolutional network of spatiotemporal graphs, integrated with a multi-head attention system, is crafted to improve the clarity. The proposed method is validated using operational data from two actual wind farms in Northwestern China. The research indicates that, compared with typical baselines (e.g., STGCN), this method reduces the RMSE by up to 38.47% and the MAE by up to 44.71% for ultra-short-term wind power prediction, markedly enhancing the prediction precision and offering a more efficient way to forecast wind power. Full article

The assessment of pig welfare and health at abattoirs is crucial for ensuring both animal well-being and food safety. Traditional assessment methods often rely on human observation, which is time-consuming, subjective, and difficult to scale in high-throughput facilities. This systematic review addresses a crucial gap by identifying and evaluating non-invasive human-free diagnostic methods applicable in commercial settings. Following PRISMA guidelines, a total of 102 articles met the inclusion criteria. Thirteen distinct methods were identified and classified into three categories: biological sample analysis (5 methods; n = 80 articles), imaging and computer vision systems (4 methods; n = 19), and physiological and other sensors (4 methods; n = 24). Some articles assessed more than one method and are therefore counted in multiple categories. While no method achieved both high implementation and practicality, blood analysis for glucose and lactate, convolutional neural networks for lesion detection, and automated camera-based systems emerged as the most promising for practical integration into the slaughter flowline. Most techniques still face challenges related to automation, operator independence, and standardisation. Overall, this review highlights the growing potential of non-invasive methods in pig welfare evaluation and underscores the need for continued development and validation to facilitate their adoption into routine abattoir practices. Full article