Search Results (1,585)

This study addresses the challenges of complex noise and short speech in civil aviation air-ground communication scenarios and proposes a novel speaker identification model, Chrono-ECAPA-TDNN (CET). The aim of the study is to enhance the accuracy and robustness of speaker identification in these environments. The CET model incorporates three key components: the Chrono Block module, the speaker embedding extraction module, and the optimized loss function module. The Chrono Block module utilizes parallel branching architecture, Bi-LSTM, and multi-head attention mechanisms to effectively extract both global and local features, addressing the challenge of short speech. The speaker embedding extraction module aggregates features from the Chrono Block and employs self-attention statistical pooling to generate robust speaker embeddings. The loss function module introduces the Sub-center AAM-Softmax loss, which improves feature compactness and class separation. To further improve robustness, data augmentation techniques such as speed perturbation, spectral masking, and random noise suppression are applied. Pretraining on the VoxCeleb2 dataset and testing on the air-ground communication dataset, the CET model achieves 9.81% EER and 88.62% accuracy, outperforming the baseline ECAPA-TDNN model by 1.53% in EER and 2.19% in accuracy. The model also demonstrates strong performance on four cross-domain datasets, highlighting its broad potential for real-time applications. Full article

With the rapid development and increasing demands of civil aviation, the accurate detection of key aircraft components and staff on airport aprons is of great significance for ensuring the safety of flights and improving the operational efficiency of airports. However, the existing detection models for airport aprons are relatively scarce, and their accuracy is insufficient. Based on YOLOv5, we propose an improved object detection algorithm, called DT-YOLO, to address these issues. We first built a dataset called AAD-dataset for airport apron scenes by randomly sampling and capturing surveillance videos taken from the real world to support our research. We then introduced a novel module named D-CTR in the backbone, which integrates the global feature extraction capability of Transformers with the limited receptive field of convolutional neural networks (CNNs) to enhance the feature representation ability and overall performance. A dropout layer was introduced to reduce redundant and noisy features, prevent overfitting, and improve the model’s generalization ability. In addition, we utilized deformable convolutions in CNNs to extract features from multi-scale and deformed objects, further enhancing the model’s adaptability and detection accuracy. In terms of loss function design, we modified GIoULoss to address its discontinuities and instability in certain scenes, which effectively mitigated gradient explosion and improved the stability of the model. Finally, experiments were conducted on the self-built AAD-dataset. The results demonstrated that DT-YOLO significantly improved the mean average precision (mAP). Specifically, the mAP increased by 2.6 on the AAD-dataset; moreover, other metrics also showed a certain degree of improvement, including detection speed, AP50, AP75, and so on, which comprehensively proves that DT-YOLO can be applied for real-time object detection in airport aprons, ensuring the safe operation of aircraft and efficient management of airports. Full article

This study develops an MTSP model for multi-UAV delivery optimization from a central hub, proposing a hybrid algorithm that integrates genetic simulated annealing-enhanced clustering with an improved Hopfield neural network to minimize the total flight distance. The proposed methodology initially employs an enhanced fuzzy C-means clustering technique integrated with genetic simulated annealing (GSA) to effectively partition the MTSP formulation into multiple discrete traveling salesman problem (TSP) instances. The subsequent phase implements an enhanced Hopfield neural network (HNN) architecture incorporating three key modifications: data normalization procedures, adaptive step-size control mechanisms, and simulated annealing integration, collectively improving the TSP solution quality and computational efficiency. The proposed algorithm’s effectiveness is validated through comprehensive case studies, demonstrating significant performance improvements in the computational efficiency and solution quality compared to conventional methods. The results show that during clustering, the improved clustering algorithm is more stable in its clustering effect. With regard to path optimization, the improved neural network algorithm has a higher computational efficiency and makes it easier to obtain the global optimal solution. Compared with the genetic algorithm and ant colony algorithm, its iteration times, path length, and delivery time are reduced to varying degrees. To sum up, the hybrid optimization algorithm has obvious advantages for solving a multi-UAV collaborative distribution path optimization problem. Full article

In the context of climate change, the increasing urgency to mitigate environmental impacts has driven firms to adopt green supply chain strategies. Existing research primarily focuses on either carbon tax or emission trading schemes, leaving a gap in understanding the combined impact of hybrid carbon policies. This study addresses this gap by developing a dual-tier supply chain model with a manufacturer and retailer, exploring the effects of a carbon tax, emission trading, and a hybrid policy on emission reduction strategies and pricing decisions. Using a reverse inductive method within a Stackelberg game framework, we identify optimal strategies for emission reduction and profit maximization under each policy scenario. Results indicate that the hybrid policy achieves the lowest unit carbon emissions when the manufacturer’s initial pollution level is below a critical threshold. This research contributes to the literature by providing actionable insights into the strategic advantages of hybrid carbon policies for firms seeking both profitability and sustainability in green supply chains. Full article

Monitoring the curing process is crucial for guiding and optimizing the curing procedures of composite material repair patches. Traditional embedded online monitoring methods are limited in their ability to track the curing process of these patches. This paper presents a composite material curing monitoring platform designed using dielectric methods. It integrates temperature control, pressure control, dielectric signal acquisition, control and display modules, and is specifically tailored for bag molding curing of repair patches. The platform measures the ionic viscosity of T300 2019B composites, analyzes the curing index, and correlates it with DSC-cured degree tests. The results indicate that the multiple ionic viscosity curves obtained from monitoring exhibit consistent trends, with correlation coefficients between curves exceeding 0.96. The changes in curing index align with the changes in curing degree, demonstrating that the platform can reliably and accurately monitor the ionic viscosity of repair patches. This platform enables effective monitoring of the ionic viscosity during the curing process of composite material repair patches. Full article

Motor current analysis is useful for ensuring the safety and reliability of electromechanical systems. However, for gearboxes, the commonly used methods of detecting faulty frequency sidebands are easily disturbed by installation errors, inherent harmonics, and fundamental frequency with high amplitude. Aiming at this problem, this study presents instantaneous frequency polarview (IFpolarview), which diagnoses faults based on motor angle and motor current frequency modulation (FM) features. Firstly, to address the problem of the limited analysis order of higher-order synchrosqueezing transform (HSST), the higher-order multisynchrosqueezing transform (HMSST) is introduced to improve the instantaneous frequency (IF) estimation accuracy and reveal the transient fault features from the motor current without further increasing the order and algorithm difficulty. Then, based on the motor angle and accurate motor current IF extracted from HMSST, the IFpolarview is proposed to visualize gear faults through detecting the FM of motor current synchronized with the faulty gear mesh. In the simulation, the IF estimation error of HMSST is 2.51%, which is smaller than other methods. The experimental results show that the HMSST has the smallest Rényi entropy value of 9.13, implying that the most aggregated time–frequency representation (TFR) of the energy is obtained. HMSST can enhance the resolution of fault characteristics, and IFpolarview concentrates the abnormal IF fluctuations with periodicity into a small angular interval, which highlights the fault features and demonstrates greater intuitiveness and reliability in comparison to the frequency sideband detection method. Full article

Taxi time plays a critical role in airport capacity, aircraft fuel consumption, and emissions. It is defined as the time from touchdown to the gate and from the gate to liftoff. The International Civil Aviation Organization (ICAO) established a standard taxi/ground idle time-in-mode (TIM) of 26 min in the landing and take-off (LTO) cycle for modeling turbine engine aircraft emissions. However, actual taxi times vary significantly across airports. While a simplified standard streamlines emissions modeling, the 26 min assumption may not accurately reflect real-world conditions. While using airport-specific taxi times may not always be practical, hub classifications of U.S. commercial airports may affect taxi time and serve as a compromise between airport-specific taxi times and a simplified standard. Therefore, this study statistically analyzed Federal Aviation Administration (FAA) data from 71 U.S. commercial hub airports to compare reported taxi times with the ICAO’s standard and assess the influence of airport hub classifications. The exploratory findings indicate that the 26 min ICAO taxi/idle TIM does not represent reported taxi times at 70 of the 71 sampled airports. Moreover, total taxi time varied by hub classification: small-hub airports had a mean taxi time of 19.82 min (median: 18 min), medium-hub airports had a mean taxi time of 19.72 min (median: 18.25 min), and large hubs had a mean taxi time of 26.98 min (median: 25.08 min). When hub classifications were ignored, the overall mean taxi time was 23.78 min (median: 22 min), indicating a statistically significant difference between the ICAO’s standard 26 min assumption and the observed taxi times at most airports. Full article

The traffic behavior characteristics within university campuses have received limited scholarly attention, despite their distinct differences from external road networks. These differences include the predominance of non-motorized vehicles and pedestrians in traffic flow composition, as well as traffic peaks primarily coinciding with class transition periods. To investigate the riding behavior of cyclists on university campuses, this study examines cyclist attention, proposes a novel method for constructing a rider attention recognition framework, utilizes a hierarchical ordered logistic model to analyze the factors influencing attention, and evaluates the model’s performance. The findings reveal that traffic density and riding style significantly influence cyclists’ eye-tracking characteristics, which serve as indicators of their attention levels. The covariates of lane gaze time and the coefficient of variation in pupil diameter exhibited significant effects, indicating that a hierarchical ordered logistic model incorporating these covariates can more effectively capture the impact of influencing factors on cyclist attention. Moreover, the hierarchical ordered logistic model achieved a 7.22% improvement in predictive performance compared to the standard ordered logistic model. Additionally, cyclists exhibiting a “conservative” riding style were found to be more attentive than those adopting a “aggressive” riding style. Similarly, cyclists navigating “sparse” traffic conditions were more likely to maintain attention compared to those in “dense” traffic scenarios. These findings provide valuable insights into the riding behavior of university campus cyclists and have significant implications for improving traffic safety within such environments. Full article

With the rapid development of e-commerce, logistics UAVs (unmanned aerial vehicles) have shown great potential in the field of urban logistics. However, the large-scale operation of logistics UAVs has brought challenges to air traffic management, and the competitiveness of UAV logistics is still weak compared with traditional ground logistics. Therefore, this paper constructs a double-layer route network structure that separates logistics transshipment from terminal delivery. In the delivery layer, a door-to-door distribution mode is adopted, and the transshipment node service location model is constructed, so as to obtain the location of the transshipment node and the service relationship. In the transshipment layer, the index of the route betweenness standard deviation (BSD) is introduced to construct the route network planning model. In order to solve the above model, a double-layer algorithm was designed. In the upper layer, the multi-objective simulated annealing algorithm (MOSA) is used to solve the transshipment node service location issue, and in the lower layer, the multi-objective non-dominated sorting genetic algorithm II (NSGA-II) is adopted to solve the network planning model. Based on real obstacle data and the demand situation, the double-layer network was constructed through simulation experiments. To verify the network rationality, actual flights were carried out on some routes, and it was found that the gap between the UAV’s autonomous flight route time and the theoretical calculations was relatively small. The simulation results show that compared with the single-layer network, the total distance with the double-layer network was reduced by 62.5% and the structural intersection was reduced by 96.9%. Compared with the minimum spanning tree (MST) algorithm, the total task flight distance obtained with the NSGA-II was reduced by 42.4%. The BSD factors can mitigate potential congestion risks. The route network proposed in this paper can effectively reduce the number of intersections and make the UAV passing volume more balanced. Full article

Hearing impairments affect approximately 1.5 billion individuals worldwide, highlighting the critical need for effective communication tools between deaf and hearing populations. Traditional sign language translation (SLT) models predominantly rely on gloss-based methods, which convert visual sign language inputs into intermediate gloss sequences before generating textual translations. However, these methods are constrained by their reliance on extensive annotated data, susceptibility to error propagation, and inadequate handling of low-frequency or unseen sign language vocabulary, thus limiting their scalability and practical application. Drawing upon multimodal translation theory, this study proposes the open-vocabulary sign language translation (OVSLT) method, designed to overcome these challenges by integrating open-vocabulary principles. OVSLT introduces two pivotal modules: Enhanced Caption Generation and Description (CGD), and Grid Feature Grouping with Advanced Alignment Techniques. The Enhanced CGD module employs a GPT model enhanced with a Negative Retriever and Semantic Retrieval-Augmented Features (SRAF) to produce semantically rich textual descriptions of sign gestures. In parallel, the Grid Feature Grouping module applies Grid Feature Grouping, contrastive learning, feature-discriminative contrastive loss, and balanced region loss scaling to refine visual feature representations, ensuring robust alignment with textual descriptions. We evaluated OVSLT on the PHOENIX-14T and CSLDaily datasets. The results demonstrated a ROUGE score of 29.6% on the PHOENIX-14T dataset and 30.72% on the CSLDaily dataset, significantly outperforming existing models. These findings underscore the versatility and effectiveness of OVSLT, showcasing the potential of open-vocabulary approaches to surpass the limitations of traditional SLT systems and contribute to the evolving field of multimodal translation. Full article

Bearing-vibration signals, characterized by strong non-stationarity, typically consist of multiple components. The periodic pulses related to bearing faults are frequently obscured by surrounding noise, and early bearing-fault vibrations are feeble, which complicates the extraction of inherent fault characteristics. The aim of this research is to develop an effective method for extracting early-fault characteristic frequencies in rolling bearings. VMD, short for variational mode decomposition, is an innovative technique rooted in the classical Wiener filter for analyzing signals that include multiple components. However, applying VMD to process real non-stationary signals still poses several challenges. A key challenge is that the internal parameters of VMD require manual setting prior to use. Aiming to mitigate this limitation, this paper introduces an enhanced variational mode decomposition approach utilizing the Chaotic Harris Hawk Optimization (CHHO) method. Average energy entropy is used as the optimization criterion in the CHHO–VMD algorithm to ascertain both the ideal mode count and its corresponding penalty factor. The original signal is further broken down into intrinsic mode functions (IMFs), with each IMF corresponding to a different frequency interval. In addition, IMF components are selected based on kurtosis and cross-correlation criteria to reconstruct fault signals. Finally, envelope demodulation is performed to reveal the fault characteristic frequencies. Experimental findings demonstrate that, as opposed to alternative techniques, this approach achieves superior performance in extracting early-fault frequencies in rolling bearings, offering a novel solution for early-fault feature extraction. Full article

Faults in the DC drive circuit of UAV electro-hydrostatic actuators directly affect the flight safety of a UAV. An integrated learning and Bayesian network-based fault diagnosis strategy is proposed to address the problems of early fault diagnosis, poor unbalanced data processing performance, and lack of interpretability in intelligent fault diagnosis in engineering practice. In the data preprocessing stage, Pearson coefficients are used for feature correlation analysis, and XGBoost performs feature screening to extract key features from the collected DC drive circuit data. This process effectively saves computational resources while significantly reducing the risk of overfitting. The optimal weak learner selection for the high-performance boosting integrated learner is identified through comparative validation. The performance of the proposed diagnostic strategy is fully verified by setting up different comparison algorithms in two experimental circuits. The experimental results show that the strategy outperforms the comparison algorithms in various scenarios such as data balancing, data imbalance, early-stage faults, and high noise; in particular, it shows a significant advantage in diagnosing data imbalance and early-stage faults. The interpretable fault diagnosis of UAV DC drive circuits is realized by the interpretation strategy of Bayesian networks, which provides the necessary theoretical and methodological support for practical engineering operations. Full article

Enhancing the effectiveness of aviation engine borescope inspection is critical for flight safety. Statistics indicate that engine defects contribute to 20% of mechanical-related flight accidents, while existing defect detection and segmentation models for borescope images suffer from a low operational efficiency and suboptimal accuracy. To address these challenges, this study proposes a Visual State Space with Multi-directional Feature Fusion Mamba (VMmamba) model and constructs a real-world borescope defect dataset. First, a feature compensation module with symmetrical diagonal feature optimization fusion is developed to enhance the feature representation capabilities, expand the receptive fields, and improve the feature extraction of the model. Second, a content-aware upsampling module is introduced to restructure contextual information for complex scene understanding. Finally, the learning process is optimized by integrating Smooth L1 Loss with Focal Loss to strengthen defect recognition. The experimental results demonstrate that VMmamba achieves a 43.4% detection mAP and 36.4% segmentation mAP on our dataset, outperforming state-of-the-art models by 2.3% and 1.4%, respectively, while maintaining a 29.2 FPS inference speed. This framework provides an efficient and accurate solution for borescope defect analysis, offering significant practical value for aviation maintenance and safety-critical decision making. Full article

To achieve intelligent manufacturing and improve the machining quality of machine tools, this paper proposes an interpretable machining size prediction model combining eXtreme Gradient Boosting (XGBoost), autoencoder (AE), and Shapley additive explanation (SHAP) analysis. In this study, XGBoost was used to establish an evaluation system for the actual machining size of computer numerical control (CNC) machine tools. The XGBoost model was combined with SHAP approximation to effectively capture local and global features in the data using autoencoders and transform the preprocessed data into more representative feature vectors. Grey correlation analysis (GRA) and principal component analysis (PCA) were used to reduce the dimensions of the original data features, and the synthetic minority overstimulation technique of the Gaussian noise regression (SMOGN) method was used to deal with the problem of data imbalance. Taking the actual size of the machine tool as the response parameter, based on the size parameters in the milling process of the CNC machine tool, the effectiveness of the model is verified. The experimental results show that the proposed AE-XGBoost model is superior to the traditional XGBoost method, and the prediction accuracy of the model is 7.11% higher than that of the traditional method. The subsequent SHAP analysis reveals the importance and interrelationship of features and provides a reliable decision support system for machine tool processing personnel, helping to improve processing quality and achieve intelligent manufacturing. Full article

The formation of ice due to global climate change poses challenges across multiple industries. Traditional anti-icing technologies often suffer from low efficiency, high energy consumption, and environmental pollution. Photothermal and hydrophobic surfaces with nano-micro structures (PHS-NMSs) offer innovative solutions to these challenges due to their exceptional optical absorption, heat conversion capabilities, and unique surface water hydrophobic characteristics. This paper reviews the research progress of PHS-NMSs in their anti-icing applications. It introduces the mechanisms of ice prevention, fabrication methods, and pathways for performance optimization of PHS-NMSs. The anti-icing performance of PHS-NMSs in different application scenarios is also discussed. Additionally, the paper provides insights into the challenges and future development directions in this field. Full article