Search

This study introduces a novel algorithm for the detection of three-cell coupling faults, March ML3C, along with a simulator, TCFS. The March ML3C algorithm targets the detection of single-port, static, and unlinked three-cell coupling faults within bit-oriented random access memory (RAM). March ML3C improves the coverage of three-cell coupling faults to 100%, with a complexity of only 58n (n is the number of memory cells), which is lower than that of most similar algorithms. Furthermore, to solve the problem of none of the existing tools supporting the simulation of three-cell coupling faults, this study designs and implements TCFS, which can quickly simulate all March-type algorithms and calculate the coverages for three-cell coupling faults, thus achieving the purpose of software simulation of the algorithms. The simulation results of March ML3C and other similar algorithms using TCFS are also presented in this paper.

The task of detecting surface defects on substation equipment faces several challenges, including a variety of target categories, the scarcity of original defect image data, complex environmental conditions, low accuracy in existing algorithms, as well as notable issues with false alarms and missed detections. Overcoming these obstacles is crucial for the successful implementation of intelligent inspection systems for substations. To address the problem of limited original data, we first employ the method of ADD-GAN to augment the image training set. Furthermore, this paper proposes a target detection model called YOLO-SD to detect various equipment defects in complex real-world scenarios. In order to enhance the network’s feature extraction capabilities in the presence of complex backgrounds and to improve detection accuracy, a novel deep perceptual feature extraction module named C3+ was introduced in this research. Furthermore, we incorporates SimAM into the neck network of YOLO-SD. This integration not only bolsters the network’s learning capacity but also equips it with the capability to autonomously learn and dynamically fine-tune attention weights to suit different input scenarios. To tackle the challenges posed by variations in size and shapes of different substation equipment defects in images, a novel fusion loss function NWD-CIoU is designed. The improvements enhance the accuracy and robustness of YOLO-SD in defect target detection across different scales. The experiment demonstrated that the YOLO-SD model achieved an mAP@0.5 of 90.3% and mAP@0.5:0.95 of 63.9% in detecting defects in substation equipment. The F1 score reached 81.1%, IoU value was 90.5%. This model realized accurate detection of multi-scale substation defect targets, reaching the state-of-the-art level in substation defects detection.

Pansharpening is a fundamental and crucial image processing task for many remote sensing applications, which generates a high-resolution multispectral image by fusing a low-resolution multispectral image and a high-resolution panchromatic image. Recently, vision transformers have been introduced into the pansharpening task for utilizing global contextual information. However, long-range and local dependencies modeling and multiscale feature learning are all essential to the pansharpening task. Learning and exploiting these various information raises a big challenge and limits the performance and efficiency of existing pansharpening methods. To solve this issue, we propose a pansharpening network based on multiscale embedding and dual attention transformers (MDPNet). Specifically, a multiscale embedding block is proposed to embed multiscale information of the images into vectors. Thus, transformers only need to process a multispectral embedding sequence and a panchromatic embedding sequence to efficiently use multiscale information. Furthermore, an additive hybrid attention transformer is proposed to fuse the embedding sequences in an additive injection manner. Finally, a channel self-attention transformer is proposed to utilize channel correlations for high-quality detail generation. Experiments over QuickBird and WorldView-3 datasets demonstrate the proposed MDPNet outperforms state-of-the-art methods visually and quantitatively with low running time. Ablation studies further verify the effectiveness of the proposed multiscale embedding and transformers in pansharpening.

Reducing urban transportation carbon emissions is a fashionable task around the globe. Electric cars, fuel cell cars, hydrogen cars, restriction policies, etc. are all aimed at this goal. But they are all expensive to make them into reality. Transportation management can also provide some contribution to this, like optimizing operation rules such as Left-turn Waiting Area Twice Startup. From the stop line to the Left-turn Waiting Area (LWA), one vehicle will start twice with more emissions due to the engine’s inefficient operating range. We proposed an Accurate Left-turn One-time Startup Model (ALOSM) based on the LWA rule, which includes the intersection geometry parameters, traffic light timing schemes, vehicle start speeds, vehicle operation speeds, and fuel vehicle ratios. Make the Left-turn Vehicle (LV) waiting behind the stop line during the through vehicle light green. Calculate the left-turn green light adjustment duration and make the LV and the last through vehicle across at one moment. With real collected data in a symbolic intersection in Xi’an, China, the simulation result shows vehicle emission would be reduced by 13% and traffic efficiency would increase by 5% in this single intersection. The whole city of 5 million vehicles would reduce 2,087.88 kg CO emission in one day due to estimation.

When AC system operates asymmetrically, poor structure and control parameters of the converter station controller will have a negative impact on the stable operation of the flexible DC power grids. In this paper, a small-signal model of flexible DC power grids under asymmetric conditions is proposed, considering the internal characteristics of converter station, positive-sequence current vector controller based on voltage-droop, negative-sequence current suppressor based on quasi-proportional integral resonant (quasi-PIR), and DC power grids. Then, the influence of control parameters in voltage-droop controller and quasi-PIR suppressor on the system stability is analyzed by using Lyapunov stability theory and the root locus method, and their reasonable ranges of values are determined. Finally, the Zhangbei four-terminal flexible DC grid project is simulated in PSCAD/EMTDC to verify the accuracy of the presented small-signal model and the conclusions about the impact of parameters on the system stability, which provides a theoretical basis for the stability analysis and the tuning of control parameters in flexible DC power grids under asymmetric conditions.

Although reinforcement learning (RL) methodologies exhibit potential in addressing decision-making and planning problems in autonomous driving, ensuring the safety of the vehicle under all circumstances remains a formidable challenge in practical applications. Current RL methods are predominantly driven by singular reward mechanisms, frequently encountering difficulties in balancing multiple sub-rewards such as safety, comfort, and efficiency. To address these limitations, this paper introduces a constraint-driven safety RL method, applied to decision-making and planning policy in highway scenarios. This method ensures decisions maximize performance rewards within the bounds of safety constraints, exhibiting exceptional robustness. Initially, the framework reformulates the autonomous driving decision-making problem as a Constrained Markov Decision Process (CMDP) within the safety RL framework. It then introduces a Multi-Level Safety-Constrained Policy Optimization (MLSCPO) method, incorporating a cost function to address safety constraints. Ultimately, simulated tests conducted within the CARLA environment demonstrate that the proposed method MLSCPO outperforms the current advanced safe reinforcement learning policy, Proximal Policy Optimization with Lagrangian (PPO-Lag) and the traditional stable longitudinal and lateral autonomous driving model, Intelligent Driver Model with Minimization of Overall Braking Induced by Lane Changes (IDM+MOBIL). Compared to the classic IDM+MOBIL method, the proposed approach not only achieves efficient driving but also offers a better driving experience. In comparison with the reinforcement learning method PPO-Lag, it significantly enhances safety while ensuring driving efficiency, achieving a zero-collision rate. In the future, we will integrate the aforementioned potential expansion plans to enhance the usability and generalization capabilities of the method in real-world applications.

PointNet, the first deep learning framework for processing 3D point clouds, suffers from significant information loss during point cloud classification due to its Max pooling method, leading to decreased performance and accuracy when handling large-scale point clouds. In this paper, we propose an improved method for PointNet classification accuracy—a Functional Multi-head Pooling encoding model. Based on the permutation-invariant functional concept, this model constructs richer multi-head invariant feature vectors, significantly enhancing recognition classification scores, and introducing a new model—Functional Multi-head Pooling (FMP)—capable of reducing overfitting. Due to the non-monotonicity, continuous differentiability, and controllable amplitude of functionals, we conducted experiments using cosine and sine trigonometric functions to replace the baseline Max method. The new approach also involves aggregating the results of the cosine and sine transformations, which allows for retaining more features during training compared to PointNet, thereby enhancing the training model’s capabilities. Our three rounds of experiments on two datasets including ModelNet40 have proven the effectiveness of the functional approach. The new method can increase accuracy by up to 3.25% on the same datasets and significantly improving generalization capabilities.

To solve the problem of tuning the weight coefficient matrices of Linear Quadratic Regulator (LQR) for vehicle active suspension system, a multi-strategy Advanced Snake Optimizer-Linear Quadratic Regulator (ASO-LQR) is proposed based on the Snake Optimizer (SO). Firstly, the dynamic model of four-freedom-degree half vehicle active suspension is established, and the LQR is set up. Secondly, the multi-strategy Advanced Snake Optimizer (ASO) is designed to address the problems of slow convergence speed and low optimization accuracy of the SO. Wherein the quality of initial snake swarm is improved through the good-point set and oppositional learning; the algorithm’s capabilities of global exploration and local exploitation are enhanced by utilizing the adaptive oscillating weight; the crossover-mutation operator of Genetic Algorithm (GA) for parallel search of solution space as well as the Levy Flight perturbation of optimal solutions is introduced to avoid local optima. Then, the ASO is combined with the LQR by a fitness function including suspension performance indexes to reduce system vibration. Finally, compared with other algorithms, such as SO, GA, Adaptive Particle Swarm Optimization (APSO), et al., the proposed ASO has higher optimization accuracy and faster convergence speed. On the above basis, through the simulation tests of C-class road, bump road and speed generalization on the C-class road, the ASO-LQR-controlled active suspension (ASO-LQR-ASS) can improve vibration damping performance more effectively.

The rapid growth of Internet of Things (IoT) devices and communication standards has led to an increasing demand for data security, particularly with limited hardware resources. In addition to conventional software-level data encryption, physical-layer security techniques, such as device-specific radio frequency fingerprints (RFFs), are emerging as promising solutions. This article first summarizes prior arts on timestamped RFFs generation and reconfigurable power amplifier (PA) designs. Following that, an innovative 2-stage PA incorporating a reconfigurable class A stage with a Doherty amplifier, designed in 65-nm CMOS to generate 4096 timestamped RFFs without introducing in-band power variation, is presented. Multiple 3-bit resistive digital-to-analog converters (RDACs) are applied to control body biasing units within the two-stage PA, facilitating the generation of massive and distinguishable RFFs. Subsequently, time-varying unequally spaced multitone (USMT) techniques are proposed to further elevate the count of available timestamped RFFs from 4096 to 16 384. The validation results of RFFs utilizing 64-QAM WiFi-6E advertising packets, employing time-varying USMT transmitted within the 5.39–5.41-GHz band, confirm the successful generation of 16 384 distinct RFF patterns. Moreover, the measurement results demonstrate that more than 11 504 RFFs among the generated patterns can be classified with an accuracy exceeding 99%.

To realize the effective reuse of the existing model measurement process planning and to improve the efficiency of the preparation of the measurement program, this paper proposes an on-machine measurement model partial retrieval method based on geometric features. First, the B-Rep information and measurement procedures of the model are extracted, and combined with the analysis of the geometric features of the on-machine measurement, the model is transformed into an attribute adjacency labeling graph; next, the index tree is constructed by using hash coding, and the result set is obtained through the filtering-verification framework to realize the synthesis similarity evaluation of the geometrical level of the on-machine measurement model; lastly, different reuse methods of the measurement process planning are selected based on the model’s similarity value. Fast programming based on partial retrieval of on-machine measurement models is realized by directly selecting programs, calling macros, and assigning values to variables. The experimental results show that the method can realize the effective reuse of the existing model measurement process planning, which can improve the efficiency of the preparation of the measurement program and meet the actual needs of on-machine measurement.

Diffractive optical imaging spectroscopy as a promising miniaturized and high throughput portable spectral imaging technique suffers from the problem of low precision and slow speed, which limits its wide use in various applications. To reconstruct the diffractive spectral image more accurately and fast, a three-dimensional spectrum recovery algorithm is proposed in this paper. The algorithm takes advantage of a neural network for image reconstruction which consists of a U-Net architecture with 3D convolutional layers to improve the processing precision and speed. Numerical experiments are conducted to prove its effectiveness. It is shown that the mean peak signal-to-noise ratio (MPSNR) of the recovered image relative to the original image is improved by 1.8 dB in comparison to other traditional methods. In addition, the obtained mean structural similarity (MSSIM) of 0.91 meets the standard of discrimination to human eyes. Moreover, the algorithm runs in just 0.36 s, which is faster than other traditional methods. 3D convolutional networks play a critical role in performance improvement. Improvements in processing speed and accuracy have greatly benefited the realization and application of diffractive optical imaging spectroscopy. The new algorithm with high accuracy and fast speed has a great potential application in diffraction lens spectroscopy and paves a new way for emerging more portable spectral imaging technique.

Optical synthetic aperture imaging (OSAI) systems consist of multiple subapertures arranged in a specific spatial pattern to achieve high resolution imaging comparable to that of a large aperture while reducing costs. However, due to the sparsity of the apertures and co-phase errors, these systems inevitably suffer from image degradation and blurring. Traditional nonblind deconvolution methods are commonly used for image deblurring in OSAI systems, but they require accurate prior knowledge, which, if inaccurate, can affect image restoration quality. Recent deep learning-based methods have achieved remarkable results, but they do not consider the impact of co-phase errors in OSAI systems and require large amounts of real datasets for training. This study proposes a semiblind unsupervised learning method named RPIR for image restoration in OSAI systems with co-phase errors. RPIR is based on the traditional maximum a posteriori (MAP) model and utilizes a multiscale neural network that does not require training to capture the input blur kernel errors, which are then used as the residual prior term of the MAP model. The data term and the prior term are solved using an alternating minimization algorithm. Consequently, RPIR can effectively address the issue of inaccurate blur kernels caused by variations in co-phase errors in OSAI systems. Experimental results demonstrate that RPIR significantly improves image resolution and detail clarity in OSAI systems with complex co-phase errors, outperforming traditional deconvolution methods and other unsupervised deep learning methods.

As the frontier of urbanization, urban–rural fringes (URFs) transitionally connect urban construction regions to the rural hinterland, and its identification is significant for the study of urbanization-related socioeconomic changes and human dynamics. Previous research on URF identification has predominantly relied on remote sensing data, which often provides a uniform overhead perspective with limited spatial resolution. As an additional data source, street view images (SVIs) offer a valuable human-related perspective, efficiently capturing intricate transitions from urban to rural areas. However, the abundant visual information offered by SVIs has often been overlooked and multimodal techniques have seldom been explored to integrate multisource data for delineating URFs. To address this gap, this study proposes a transformed-based multimodal methodology for identifying URFs, which includes a street view panorama classifier and a remote sensing classification model. In the study area of Beijing, the experimental results indicate that an URF with a total area of 731.24 $\text{km}^{2}$ surrounds urban cores, primarily located between the fourth and sixth ring roads. The effectiveness of the proposed method is demonstrated through comparative experiments with traditional URF identification methods. In addition, a series of ablation studies demonstrate the efficacy of incorporating multisource data. Based on the delineated URFs in Beijing, this research introduced points of interest data and commuting data to analyze the socioeconomic characteristics of URFs. The findings indicate that URFs are characterized by longer commuting distances and less diverse restaurant consumption patterns compared to more urbanized regions. This study enables the accurate identification of URFs through the transform-based multimodal approach integrating SVIs. Furthermore, it provides a human-centric comprehension of URFs, which is essential for informing strategies of urban planning and development.

As the connectivity between Electronic Control Units (ECUs) and the external environment intensifies, the security and safeguarding of In-Vehicle Networks (IVNs) have become pressing issues. The Controller Area Network (CAN) bus, the most commonly employed vehicular network protocol, is particularly vulnerable due to its inherent lack of security measures, making it susceptible to various attacks. In this paper, we introduce a novel intrusion detection model for CAN networks, named IVN-ViT. The proposed IVN-ViT model utilizes an enhanced Vision Transformer architecture (Edge-ViT), incorporating self-supervised learning with Cutout data augmentation, Particle Swarm Optimization (PSO) for feature selection, and a fine-tuning strategy that merges Parameter Instance Discrimination (PID) with supervised loss. This not only enhances the model’s convergence speed and predictive accuracy but also meets the lightweight requirements of CAN networks. Experimental results on the “HCRL-car hacking” dataset and the “X-CANIDS” dataset demonstrate that our model achieves over 99% accuracy across all types of attacks. In a simulated vehicular environment, the detection latency for each message averages approximately 1.04ms, fully meeting the first-time requirements of CAN networks. The experiments validate that our proposed method significantly improves model accuracy while ensuring efficient operation within the limited computational resources available in vehicle systems.

The evaluation of auto insurance risks is a fundamental task for financial institutions, crucial for setting equitable premiums and managing risks effectively. Traditional machine learning algorithms often struggle to capture the intricate relationships necessary for accurate risk assessment. In contrast, deep learning methods, while capable of processing complex data structures, lack the ability to model feature interactions and provide interpretability, which are essential for transparent decision-making in the insurance industry. To address these challenges, we introduce the Actuarial Transformer (AT)–a pioneering model that leverages the self-attention mechanism of the Transformer architecture to meticulously map feature interactions. The AT integrates advanced residual models with tree-based methods, enhancing its predictive accuracy. Additionally, it incorporates the SHAP (SHapley Additive exPlanations) model, which uses Shapley values from cooperative game theory to ensure interpretability and transparency in its risk assessments. Our empirical analysis, conducted on a representative dataset of auto insurance risks, demonstrates the AT’s superior performance in risk prediction. The SHAP analysis further validates the model’s ability to prioritize features logically, providing clear insights into the decision-making process. The AT not only improves the precision of auto insurance risk evaluations but also enhances the interpretability of these evaluations, making it a valuable tool for both industry practitioners and clients.

In the recent advancements of recommendation systems, the integration of deep learning models has significantly enhanced prediction accuracy and user experience. This paper introduces a novel model called GRATB, which enhances the existing BERT4Rec framework by incorporating a self-attention mechanism based on time-mixing attention mechanism and dynamic head fusion. Additionally, it introduces a new ’comment tower’ to integrate user review information. By employing a hybrid approach, our model is able to incorporate both the interaction sequence between users and items, as well as sentiment-rich review data. This enables us to obtain a more comprehensive understanding of user preferences. Empirical evidence derived from extensive experimentation demonstrates that GRATB exhibits substantial performance improvements over existing state-of-the-art models across multiple evaluation metrics. GRATB achieves significant enhancements in two critical performance indicators: Hit Ratio (HR) and Normalized Discounted Cumulative Gain (NDCG), with average increases of 2.41% and 3.25%, respectively.

Dim and small target detection plays an important role in infrared target recognition systems. In this paper, we present a dim and small target detection algorithm based on improved bilateral filtering and Gaussian motion probability estimation, aiming to improve the detection efficiency of the detection system. First, a bilateral filtering algorithm based on image patch analysis is proposed to complete the background modeling, compare with single pixel, image patch contains more neighborhood information. Then, we use the Gaussian process combining the target position of consecutive $n$ frames to predict the target position of the $(n+1)\text{th}$ frame, and the target energy is accumulated along the trajectory direction at the same time. Finally, we construct the grayscale probability model to realize the multi-frame correlation detection, which combining the grayscale features and the motion characteristics of the target. Six scenes and eleven comparison algorithms are selected for experiments, experimental results show the effectiveness and robustness of the proposed algorithm.

Object detection in Unmanned Aerial Vehicle (UAVs) faces a significant challenge in computer vision. Traditional methods are difficult to model object appearance feature with large scale variations and viewpoint differences, when drones fly at different altitudes and capture images from diverse shooting angles. To address this issue, we propose a Learnable Cross-scale Sparse Attention (LCSA) guided feature fusion method to improve the performance of UAV object detection. Specifically, the LCSA feature fusion module enables each point in a feature map to aggregate discriminative information from a set of points with learnable offsets in neighbor feature maps. It enhances local discriminative features of the object by facilitating semantic information interaction across multiple feature maps. The LCSA can function as a novel neck method that complements the existing neck methods and is also transplantable to different object detection frameworks. Moreover, we also employ a scale-aware loss function to integrate the normalized Wasserstein distance with CIoU in order to improve the incompatibility of IoU for objects with large scale variance. Experimental results on the SeaDroneSeev2 and VisDrone2019-DET datasets show that the proposed method achieves superior performance. At a resolution of 640*640, our method achieves 81.9% AP50 and 47.4% AP on SeaDroneSeev2, surpassing baseline 4.9% and 4.8%, achieves state-of-the-art performance. Furthermore, our method outperforms baseline by 5% AP on VisDrone2019-DET. Code will be available at https://github.com/qch777/LSACF.

The terahertz range, as a type of electromagnetic wave with wavelengths between microwaves and the infrared band, has the characteristics of penetration, low energy and a stable absorption spectrum of specific substances, and is widely used in non-destructive testing, human security inspections, biological tissue diagnoses and military detection. In particular, terahertz wave 3D imaging technology can detect the internal information of the target of detection, and it has become the focus of current research. This study carried out research on 3D reconstruction and object detection algorithms based on terahertz images. In view of the problem that the MVS (Multi-ViewStereo) series of 3D reconstruction algorithms ignore the context information between the cost layers and have unsatisfactory reconstruction effects when used on complex regions, an improved MVSNet 3D reconstruction algorithm FCTMVSNet(Feature and Cost Transformer Depth Inference for Unstructured Multi-view Stereo) based on Transformer is proposed here. A structured object recognition algorithm was designed to provide theoretical support for subsequent terahertz image-based object detection algorithms.

The accurate modeling of multi-temporal correlation of photovoltaic stations output is important to achieve the precise power system reliability. However, the existing studies is mainly focused on the hourly temporal correlation between adjacent interval outputs of photovoltaic station. In this paper, a novel multi-temporal correlation modeling method for photovoltaic station output is proposed by using an improved nonparametric disaggregation (INPD) technique, and the multi-temporal characteristic representing the interactive correlation between the total daily output and hourly temporal outputs are deeply explored. Moreover, a three-stage sampling method for the INPD based power system reliability evaluation is further presented considering the multi-temporal correlation of photovoltaic stations output, which includes 1) the random sampling of daily photovoltaic station output; 2) the random disaggregation of daily output to the hourly temporal outputs and 3) the random sampling of power device. Finally, several numerical tests are conducted for a modified IEEE-RTS79 to validate the proposed method.

A refined bidirectional A $^{\ast }$ algorithm proposes to address the issue of imprecise path planning for unmanned mining vehicles navigating through complex open-pit mining terrains. To ensure the smooth traversal of these vehicles, a gradient factor incorporates into the cost function to circumvent obstacles along the pathway. Additionally, a weighted coefficient introduces into the heuristic function to fine-tune the combination of Euclidean and Manhattan distances, enhancing the accuracy of path distance measurement and ultimately leading to an optimal path. The enhanced algorithm adeptly simultaneously explores from both the starting and end points, significantly reducing search time and improving path planning efficiency. Utilizing the established map of the mining environment, an experiment for unmanned mining vehicle path planning devises, and a simulation test of the global path planning algorithm conducts using the MATLAB platform. The experimental results demonstrate that the refined bidirectional A $^{\ast }$ algorithm exhibits accelerated search speed and superior path planning effectiveness.

Synthetic Aperture Radar (SAR) imaging technology is crucial for maritime vessel monitoring. However, the inherent characteristics of SAR images, such as limited feature resolution and speckle noise, pose a series of challenges for ship detection. Although existing research has made great progress in improving the model detection accuracy, it often comes at the cost of reduced efficiency and increased computational complexity. To address this issue, we propose a lightweight SAR ship detection model based on YOLOv5s, named GS-YOLO. Firstly, we adopt C3GhostV2 modules as the lightweight backbone to enhance the computational efficiency which is based on GhostNetV2 and GConv. In the neck, Squeeze and Excitation (SE) attention is employed to strengthen the feature extraction ability for small targets and improve the SAR ship detection precision. Furthermore, we develop a novel XIoU loss function to reinforce the accuracy and robustness of our model. Comparative experiment results on the HRSID dataset show that the proposed GS-YOLO achieves a significant increase in detection precision from 88.2% to 92.7% and the mean Average Precision (mAP) from 90.5% to 94.3%. In addition, the parameters are effectively reduced, achieving a good balance between detection speed and accuracy.

Sea surface temperature (SST) is a critical factor in the interaction between the ocean and the atmosphere, directly influencing global climate patterns and the dynamic changes in marine ecosystems. Accurate prediction of SST is of great significance for assessing and managing global climate change and maintaining marine ecological balance. However, existing SST prediction methods face challenges such as low accuracy, short prediction periods, and significant errors. This paper proposes an innovative deep learning prediction method, Ocean Temperature Cycle Fusion and Analysis Model (OTCFM), constructed based on datasets from the South China Sea and the East China Sea. This approach aims to accurately capture and predict the cyclical variations and variability in ocean temperature data to provide more precise forecasts of ocean temperatures. Firstly, based on observations of SST’s seasonal and periodic variations, we present a periodic partitioning strategy to decompose complex temperature changes into intra-period and inter-period variations. Secondly, we propose the Ocean Unit to capture both long-term and short-term small-scale changes, moving beyond the inherent attributes of the dataset’s frequency and time domain characteristics to extract intra-period and inter-period feature changes simultaneously. Finally, by stacking the Ocean Units using residual connections, we alleviate the gradient vanishing problem and achieve more accurate long-term and short-term SST predictions. In this study, data from the East China Sea and the South China Sea with different spatial distribution patterns are selected for predictive analysis of the National Oceanic and Atmospheric Administration (NOAA) data from September 1, 1981, to June 7, 2023, with a total of 15,408 data. The experimental results show that OTCFM can accurately capture the evolution patterns of SST data in the spatial and temporal processes under different experimental conditions. The MAE values on the East China Sea and South China Sea SST datasets are improved by 19.08% and 19.52%, respectively, compared with the convolutional long short-term memory neural network (ConvLSTM), which improves the accuracy of the long and short-term prediction of SST time series and has a far-reaching impact on the subsequent promotion of sustainable marine resource management and environmental protection.

Visual localization is a significant problem in computer vision and robotics, involving estimating the six degrees of freedom pose of a camera relative to a known environment based on captured images. Most DL-based visual localization methods exhibit poor generalization capabilities. While some scene-independent visual localization methods demonstrate satisfactory generalization, they often suffer from low localization accuracy. To address the issue of low accuracy in scene-independent methods and the indiscriminate fusion of channel and spatial information when using neural networks for feature extraction, we propose a visual localization method ST-PixLoc, which effectively leverages image edge gradient information using the PixLoc framework. Firstly, we optimize the gradients of input images to enhance the weighting of gradient values within the image. Secondly, we employ ResNet50 as the feature extraction network’s downsampling layers to enhance feature extraction capability, while introducing channel attention mechanisms in the upsampling layers of the feature extraction network. Notably, this mechanism focuses on relevant information and resolves the aforementioned indiscriminate fusion problem. Lastly, based on the feature maps of the considered images, we compute feature residuals and optimize the initial pose using optimization algorithms. Additionally, we optimize the loss function to improve the model’s accuracy in complex scenes. Experimental results demonstrate that the proposed method achieves high-precision localization. The average rotation and translation errors on the indoor 7-Scenes dataset increased by 6.9% and 9.7%, respectively, while those on the outdoor Cambridge Landmarks dataset increased by 16.7% and 28.2%, validating the effectiveness of the proposed approach.

The changing marine environment profoundly affects synthetic aperture radar (SAR) imaging quality. And the current deep neural network model cannot consider the variation of environmental factors. As sea conditions intensify, distinct variations in sea surface scattering characteristics emerge, leading to diverse feature distributions in images across different sea conditions, which variability poses constraints on the extraction capabilities of deep learning models. To address the above-mentioned issues, the random sea condition adaptive perception modulation network (RSC-APMN) is proposed to establish a coupled relationship with sea condition levels for adaptive enhancement of SAR imagery and semantic segmentation. Leveraging geographic coordinates and acquisition time of SAR images, RSC-APMN employs the sea condition level assessment method to estimate actual wind speeds and determine sea condition levels. The sea condition adaptive nonlinear decay module adjusts the decay ratio of different regions in SAR images under varying sea conditions based on the grayscale intensity and density characteristics of aquaculture targets, which maximizes the retention of target information while suppressing interference, such as sea clutter. To effectively utilize information from both original and enhanced images and address challenges posed by varying background noise and image blurring due to diffuse or Bragg scattering under different sea conditions, we designed the complex environments omnidirectional perception segmentation module, which ensures robust semantic segmentation in random sea conditions. Experiments demonstrate the effectiveness of the proposed approach based on marine aquaculture under sea conditions levels ranging from 0 to 5, with experimental data composed of GaoFen-3 images from Ningde, Fujian.

The Shuttle-Based Storage and Retrieval System (SBS/RS) faces challenges of low efficiency due to the constraints of single outbound or inbound operations. To overcome this limitation, a scheme enabling simultaneous outbound and inbound operations has been proposed. This involved developing a physical model that incorporates costs related to outbound cargo urgency, shelf stability, time, and warehouse busyness. The model was solved using a Genetic Algorithm with priority selection, adaptive operators, and a decay factor (GA-DF). Experimental results, validated across various environments, demonstrate that the proposed GA-DF algorithm achieves 50% higher efficiency compared to the IOSA algorithm when the shelf occupancy rate is 50% and multiple cost environments are considered. Additionally, GA-DF outperforms the Simulated Annealing algorithm, traditional Genetic Algorithm, and IOSA algorithm in optimizing storage and retrieval locations, significantly enhancing system optimization. This provides a crucial reference for optimizing such systems, particularly in dynamic and complex warehousing environments. The GA-DF algorithm’s applicability and advantages have been widely recognized through further validation, highlighting its potential to drive improvements in warehousing system efficiency and optimization strategies.

The secure and stable operation of power transmission lines is essential for electrical systems. Given that abnormal targets such as bird’s nests and defective insulators may lead to transmission failures, timely detection of these targets is imperative. This paper introduces the YOLO-PowerLite model, an advanced lightweight object detection model based on YOLOv8n, designed for efficient, real-time detection on resource-constrained unmanned aerial vehicles (UAVs) equipped with edge computing platforms. In the feature fusion module, YOLO-PowerLite incorporates the innovative C2f_AK module, significantly reducing the number of parameters and enhancing the adaptability and fusion capability of features at different scales. Meanwhile, the adoption of the Bidirectional Feature Pyramid Network (BiFPN) further optimizes the efficiency and effectiveness of feature processing. In addition, the newly designed lightweight detection head significantly reduces the number of parameters and computational requirements. The integration of the Coordinate Attention mechanism in the backbone network enhances the model’s ability to focus on and recognize abnormal targets in complex backgrounds. Experimental results show that YOLO-PowerLite achieves a mAP@0.5 of 94.2%, maintaining the accuracy of the original YOLOv8n while significantly reducing parameters, FLOPs, and model size by 42.3%, 30.9%, and 40.4%, respectively. Comparative analysis shows that YOLO-PowerLite surpasses other mainstream lightweight models in detection accuracy and computational efficiency. Deployment on the NVIDIA Jetson Xavier NX platform demonstrates an average processing time of 31.2 milliseconds per frame, highlighting its potential for real-time applications in monitoring transmission lines.

With the popularity of electronic devices, the number of malware has increased dramatically, posing a serious threat to the digital world. Accurately identifying malware has become a research focus. However, there are many difficulties in the research, such as insufficient algorithm generalization ability, unbalanced datasets, and long processing and identification times. To address these problems, this study proposes a malware detection framework (VBDN) based on a convolutional neural network (CNN). The framework incorporates data visualization, balanced adoption, data augmentation, and convolutional neural network techniques to achieve over 90% accuracy in classifying malware on all four open-source datasets. The experimental work has two other contributions: first, it not only focuses on the overall recognition effect of the algorithm during the research process, but also on the recognition effect of each category with the help of a confusion matrix, which provides useful information for cybersecurity personnel, researchers, and others to carry out subsequent targeted research. Secondly, the balanced approach adopted in this paper has the following advantages: no need to construct a new dataset, consumes fewer hardware resources, automatically evaluates the sampling weights, etc. Additionally, to enhance the generalization ability of the algorithm and alleviate the overfitting problem, this paper employs data augmentation techniques to improve the adopted method. By comparing with several state-of-the-art algorithms, it can be observed that the VBDN framework proposed in this paper achieves the desired results in time with acceptable accuracy.

Quantum computing is commonly considered one highly efficient computing method with the potential to revolutionize computation technology and solve problems that are currently unsolvable. However, due to the limitation of hardware equipment and an immature experimental base, quantum technology is still in its early stages and is far from achieving the expected performance, especially in solving large-scale complex problems. To break through these barriers, we propose a parallelized quantum annealing algorithm based on Lagrangian relaxation. The proposed algorithm divides the large-scale problem into several small problems and then employs multiple quantum computers to solve them. Our proposed approach overcomes the limited number of qubits and allows quantum computing to solve largescale optimization problems. Additionally, we incorporate a local search method to ensure this Lagrangian relaxation based quantum algorithm achieves an optimal solution. We use the proposed parallelized quantum annealing algorithm to solve optimal scheduling problems in network function virtualization networks. The problem is expressed in a linear optimization model that is NP-hard. Our proposed algorithm presents excellent time performance in solving this virtualized network functions scheduling problem, compared with the Lagrangian relaxation based classical algorithm.

The structure of plant communities and their response to temperature variations are an essential basis for evaluating the ecological structure and function of tea plantations. However, field surveys and quantitative evaluation of plant communities and ecotea plantations remain challenging. In this study, a novel strategy was proposed for rapid surveillance of plant community structure and its response to changes in weather conditions in tea plantations. This strategy aims to construct the normalized tea plantation ecological index (NTEI) by synergizing environmental parameters with multisource remote sensing data; establish the fitting and inversion model of NTEI by cascading the Fourier function with the convolutional neural networks gate recurrent unit (CNN-GRU) network; and evaluate the variability of the plant community in tea plantations by analyzing the variation characteristics of the NTEI and the measured temperature. The study revealed the following: First, the NTEI can objectively characterize the plant communities of tea plantations, and its variation characteristics were consistent with the changes in vegetation phenology and temperature; second, the Fourier function has the potential to quantify NTEI, and it is fitting R2 for the NTEI of nine plant communities ranged from 0.840 to 0.921; third, the CNN-GRU has the most advantage in establishing the prediction model of NTEI, and its prediction accuracy was Rp2 = 0.955 and RMSEP = 0.314; and fourth, the plant communities with high species richness increased regional ecological stability, had a strong buffering capacity against temperature changes, and had less variability in NTEI. The results provide significant guidance for building plant community structures and improving the ecological benefits of tea plantations.

In low signal-to-noise ratio (SNR) conditions, to address the problem of poor speech enhancement effect of traditional neural networks, this paper combines Convolution Recurrent Neural Network (CRN) with Gated Linear Units (GLU) to extract speech characteristics. Meanwhile, considering the advantages of the Empirical Mode Decomposition (EMD) algorithm, this paper put forward a model for speech enhancement that is based on Adaptive Mean Median Empirical Modal Decomposition and Multilayer Gated Convolutional Recurrent Neural Networks (ME-MGCRN). The model first uses the improved EMD algorithm to split the original noisy speech into low-frequency and high-frequency components. After that, the noisy speech is denoised using noise correlation and completes the feature extraction process. Finally, the extracted feature information is fed into the MGCRN network to realize speech enhancement further. This paper analyzes the performance of the Adaptive Mean Median Empirical Modal Decomposition algorithm using the LibriSpeech ASR dataset. This paper compares and investigates the enhanced performance of the ME-MGCRN model with the baseline model and the traditional model in terms of evaluation metrics such as the perceptual evaluation of speech quality (PESQ) and short-time objective intelligence (STOI). The research indicates that the ME-MGCRN model proposed in this paper has improved in the evaluation indexes, such as PESQ and STOI, respectively, by 0.22% and 5.6% than the baseline model and the traditional model, and the speech enhancement effect is better when Huber is selected as the loss function.

Exploiting machine learning techniques to automatically classify multispectral remote sensing imagery plays a significant role in deriving changes on the Earth’s surface. However, the computation power required to manage large Earth observation data and apply sophisticated machine learning models for this analysis purpose has become an intractable bottleneck. Leveraging quantum computing provides a possibility to tackle this challenge in the future. This article focuses on land cover classification by analyzing Sentinel-2 images with quantum computing. Two hybrid quantum-classical deep learning frameworks are proposed. Both models exploit quantum computing to extract features efficiently from multispectral images and classical computing for final classification. As proof of concept, numerical simulation results on the LCZ42 dataset through the TensorFlow Quantum platform verify our models' validity. The experiments indicate that our models can extract features more effectively compared with their classical counterparts, specifically, the convolutional neural network (CNN) model. Our models demonstrated improvements, with an average test accuracy increase of 4.5% and 3.3%, respectively, in comparison to the CNN model. In addition, our proposed models exhibit better transferability and robustness than CNN models.

The refracted, reflected, and scattered signals from global navigation satellite systems (GNSSs) have been successfully used for remote sensing (RS) of the Earth's surface and atmosphere, demonstrating their potential for sensing the atmosphere, oceans, land surface and cryosphere. As an emerging RS technology, spaceborne GNSS reflectometry (GNSS-R) has been widely used in the field of land RS, with unique advantages such as high spatial and temporal resolution, low observation cost, wide coverage, and all-weather operation. In recent years, the successful launch of several spaceborne GNSS-R satellites has opened new opportunities in this field. The aim of this article is to provide a comprehensive overview of the breakthrough applications of spaceborne GNSS-R technology in the key areas of soil moisture, soil salinity, vegetation monitoring, soil freezing and thawing, flood detection, inland water bodies, wetland dynamics, forest fire disturbances, and land topography measurements, as well as possible future research directions. In addition, this article also emphasizes the contribution and role of China's recently launched Tianmu-1 constellation (as of 5 January 2024, 23 satellites have been in stable operation in orbit) in the development of spaceborne GNSS-R technology and explores the future development direction and prospects of this technology. By comprehensively analyzing the latest progress and innovative achievements of relevant research, the significant impact of this technology on land RS applications in environmental monitoring, resource management, and disaster warning has been highlighted. Through a comprehensive summary of these fields, the enormous potential of spaceborne GNSS-R technology in promoting sustainable development and ecological protection has been revealed.

Pneumatic artificial muscle (PAM) has been widely used in rehabilitation and other fields as a flexible and safe actuator. In this paper, a PAM-actuated wearable exoskeleton robot is developed for upper limb rehabilitation. However, accurate modeling and control of the PAM are difficult due to complex hysteresis. To solve this problem, this paper proposes an active neural network method for hysteresis compensation, where a neural network (NN) is utilized as the hysteresis compensator and unscented Kalman filtering is used to estimate the weights and approximation error of the NN in real time. Compared with other inversion-based methods, the NN is directly used as the hysteresis compensator without needing inversion. Additionally, the proposed method does not require pre-training of the NN since the weights can be dynamically updated. To verify the effectiveness and robustness of the proposed method, a series of experiments have been conducted on the self-built exoskeleton robot. Compared with other popular control methods, the proposed method can track the desired trajectory faster, and tracking accuracy is gradually improved through iterative learning and updating.

With the increasing demand for seamless urban transportation, there is a growing need to establish a high-quality transportation service system that improves the efficiency of comprehensive passenger transportation hubs and provides superior passenger transport services. The taxi travel system at Terminal T2 of Shanghai Hongqiao Airport, as a crucial component of transportation in a world-class hub city, has a significant impact on travelers’ satisfaction with transportation services. This study utilizes fieldwork and interviews to investigate the composition and current operational modes of the taxi travel system at Terminal T2. Using the simulation software FlexSim, we model and simulate the existing system, revealing the critical need for enhanced system performance, particularly in terms of stay time in queue and passenger output. Subsequently, we propose a performance optimization strategy, based on process analysis, involving the introduction of push-pull control to enhance supply and demand equilibrium, and the refinement of the boarding process to improve bottleneck capability. Our simulation experiments show that process enhancements, including a 5-second reduction in boarding activity and the implementation of push-pull control, can elevate the number of passengers served by 6.4% and reduce the average waiting time for passengers and taxis by 60 seconds and 110 seconds, respectively, thereby significantly heightening travel satisfactions.

Recently, the number of birds along the 330kV transmission lines in the northwest region has gradually increased. The frequency of birds nesting and excreting on the towers has been steadily rising, with bird droppings increasingly becoming the primary factor causing bird-related tripping faults. A three-dimensional model of the tower cross arm, continuous bird droppings, and insulators was constructed to simulate and analyze the impact characteristics of bird droplet diameter, length, and relative position to insulators on the spatial electric field near the insulators. Experimental corrections were made, and based on simulation and experimental results, a bird-proof baffle was designed. Research results indicate that under conditions of single-ended grounding and suspended droplets of bird droppings, the average electric field strength at the lower end of the bird droppings and the grading ring was extracted to define the protective area against irregular-shaped bird droppings. After experimental validation, a correction coefficient of 1.4 was obtained based on the experimental results. A square protective baffle with a diversion function was designed based on simulation and experimental results. Bird droppings were placed at the baffle’s diversion groove, and the average electric field strength between the lower end of the bird droppings and the grading ring was consistently lower than the breakdown field strength of 2.98 kV/cm. The research results of this paper can serve as a basis for the subsequent designing of physical models of bird-proof baffles and prevention of bird-related flashover faults on V-type insulators.

This study addresses significant challenges within the field of remote sensing monitoring, including operational inefficiency, data confidentiality concerns, high hardware costs, and issues with data management and distribution. To tackle these problems, we introduce a synergistic remote sensing monitoring framework that leverages a bimodal cloud infrastructure, facilitated by cloud services to provide on-demand resource allocation and efficient management. Our research focuses on designing and developing an integrated operating system that optimizes remote sensing monitoring processes and enhances operational efficiency through the use of a mutual scheduling mechanism and rapid data indexing capabilities. The system is underpinned by a dual-state cloud service mechanism, combining the Memory Cloud (Flash Cloud) known for its high-speed data processing and the Storage Cloud (Persistent Cloud) for long-term data retention. This dual-state approach establishes a multi-level caching system to ensure quick access to frequently requested spatial data. Additionally, a two-tier security system is implemented to safeguard data integrity and confidentiality. Our “YunYao” geographic information service rendering engine, operating on this dual-state cloud platform, demonstrated remarkable performance advantages over mainstream platforms in identical testing environments. Specifically, it outperformed ArcGIS Desktop by over two times, exceeded GeoServer by more than four times, and was over seven times faster than ArcGIS Server in rendering speeds. Experimental and practical applications have shown that our system streamlines routine workflows and enhances work efficiency, making it a critical reference for remote sensing monitoring. Furthermore, a comparative analysis was conducted to quantitatively demonstrate the superior performance of our method in handling large volumes of remote sensing data(Including satellite imagery and UAV imagery). Despite these advancements, the integration of cloud service technology in the field of satellite remote sensing requires further development, particularly regarding the establishment of private clouds and the internal collaborative computing mechanisms within the remote sensing domain. Our research paves the way for future advancements and the eventual full integration of cloud service models into remote sensing monitoring.

Underwater vehicles are crucial for advancing underwater geological exploration, military training, and the marine industry. However, the accuracy of lithium-ion battery SOC estimation algorithms for these vehicles deteriorates with depth changes due to variations in seawater temperature, leading to limited generalization. This paper proposes a SOC estimation framework that combines the UKF with the MLP-Mixer neural network for lithium-ion batteries in underwater vehicles. The proposed framework employs the MLP-Mixer neural network to predict battery terminal voltage, incorporating hierarchical temperature compensation based on temperature data from the BMS. This straightforward and efficient compensation method enhances prediction accuracy. Subsequently, the framework combines the MLP-Mixer neural network with the UKF to form a new estimator, MM-UKF. The MM-UKF estimator’s prediction accuracy, temperature adaptability, and generalization ability underwent comprehensive examination and evaluation using a public dataset under varying temperatures and complex operating conditions. Results indicate that: 1) The superior ability of the MLP-Mixer neural network to observe battery terminal voltage plays a significant role in the precise SOC estimation of MM-UKF; 2) The proposed method shows robustness and performs well under different driving cycles. The RMSE of SOC estimation does not exceed 0.01 under intense driving cycles, showcasing superior generalization ability; 3) While the estimator’s prediction accuracy deviates slightly from the reference value at low temperatures and SOC levels, it outperforms prediction methods based on MLP and LSTM neural networks in various temperatures and low SOC scenarios. Overall, the proposed SOC estimation method based on MM-UKF holds significant practical value for underwater vehicles operating under complex conditions.

In this study, we exploit the modified Jensen-Bregman LogDet (MJBLD) divergence to measure the dissimilarity between two region covariance descriptors extracted from an image, and design a target detection method based on this descriptor. In particular, MJBLD divergence, which considers the non-Euclidean geometric structure, is used as the measurement on the symmetric positive-definite (SPD) matrix manifold. The MJBLD divergence is a modified version of the Jensen-Bregman LogDet (JBLD) divergence which has many properties similar to the affine invariant Riemannian metric. Then, the MJBLD divergence is applied for the task of the image target detection where the image region of interest is represented as a covariance descriptor. The covariance descriptor is a SPD matrix which is constructed by the first and second gradients of intensity and the three-dimensional color information. Since the SPD matrix naturally resides on the non-Euclidean Riemannian manifold and the MJBLD divergence can be treated as a manifold metric, applying the non-Euclidean distance to SPD matrices can yield a better performance in comparison with the Euclidean distance. Experimental results show that our proposed method outperforms the state-of-the-art method.

Satellite remote sensing technology can achieve real-time observation of ships at sea, and the remote sensing images obtained have the advantages of high contrast and low noise and have become one of the important means of marine monitoring. For the satellite remote sensing image data, there are two main problems: first, remote sensing data class-imbalance problem, and second the existing target detector in the presence of clouds, islands, farmed nets, and other interferences on the small-target ship, and there is a leakage of detection and wrong detection problem. To address the above problems, first, a new dataset containing 3881 images of remotely sensed ships in a variety of complex environments is constructed, which contains a total of 8418 ship instances. Second, we propose CSDP-YOLO for the small-target ship detection method with remote sensing data class imbalance. In order to enhance the performance of neural networks for small-target ship detection in remote sensing images, the innovative CSDP module is proposed, which uses deep large kernel convolution to enhance the sensory field of shallow features and mixes the channel positions using point convolution to obtain a more excellent feature extraction performance. Finally, the MPDIoU loss function is introduced to solve the class-imbalance problem between remote sensing small target ships and the background. We compare the performance with other state-of-the-art algorithms. The experimental results show that the proposed CSDP-YOLO algorithm can significantly improve the performance of small-target ship detection for private datasets. Its average precision, recall, and $\text{AP}_{50}$ are improved to 90.1%, 86.6%, and 91.4%, respectively. For the SSDD public remote sensing dataset, its metrics can reach the highest 93.6%, 93.7%, and 96.8%, respectively.

Deep learning has shown remarkable proficiency in tasks related to synthetic aperture radar (SAR) interpretation. However, several studies have highlighted the inherent vulnerability of deep neural networks when faced with deliberately constructed adversarial examples (AEs). Current SAR adversarial attack research only focuses on generating AEs in the image domain, without considering SAR imaging systems. This approach can lead to two issues. First, the generated SAR AEs lack visual coherence. Second, they cannot be obtained as corresponding attack images through real SAR imaging systems. In this article, we propose a physically realizable SAR adversarial attack method, which includes two submodules based on optimization methods: Target perturbation generation and background perturbation generation. The former module utilizes attention mechanisms to extract the target regions from SAR images, followed by selecting the optimal small regions for perturbation placement. The latter module, on the other hand, utilizes scattering models to generate realistic scatterer images as perturbations, which are then optimized to identify the optimal position in the background regions. The individual attack performance of these two attack modules on five well-established SAR automatic target recognition models is demonstrated to be highly effective. Moreover, the combination of these attack modules achieves a fooling rate of approximately 90% and demonstrates superior adversarial transferability. The experimental results of this study provide an effective foundation for physical realization of SAR AEs.

Slow and deep breathing (SDB) is a relaxation technique that can increase vagal activity. Respiratory sinus arrhythmia (RSA) serves as an index of vagal function usually quantified by the high-frequency power of heart rate variability (HRV). However, the low breathing rate during SDB results in deviations when estimating RSA by HRV. Besides, the impact of the inspiration-expiration (I: E) ratio and guidelines ways (fixed breathing rate or intelligent guidance) on SDB is not yet clear. In our study, 30 healthy people (mean age = 26.5 years, 17 females) participated in three SDB modes, including 6 breaths per minute (bpm) with an I:E ratio of 1:1/ 1:2, and intelligent guidance mode (I:E ratio of 1:2 with guiding to gradually lower breathing rate to 6 bpm). Parameters derived from HRV, multimodal coupling analysis (MMCA), Poincaré plot, and detrended fluctuation analysis were introduced to examine the effects of SDB exercises. Besides, multiple machine learning methods were applied to classify breathing patterns (spontaneous breathing vs. SDB) after feature selection by max-relevance and min-redundancy. All vagal-activity markers, especially MMCA-derived RSA, statistically increased during SDB. Among all SDB modes, breathing at 6 bpm with a 1:1 I:E ratio activated the vagal function the most statistically, while the intelligent guidance mode had more indicators that still significantly increased after training, including SDRR and MMCA-derived RSA, etc. About the classification of breathing patterns, the Naive Bayes classifier has the highest accuracy (92.2%) with input features including LFn, CPercent, pNN50, $\alpha 2$ , SDRatio, $\alpha 1$ , and LF. Our study proposed a system that can be applied to medical devices for automatic SDB identification and real-time feedback on the training effect. We demonstrated that breathing at 6 bpm with an I:E ratio of 1:1 performed best during the training phase, while intelligent guidance mode had a more long-lasting effect.

In response to the issue of neglecting the inherent differences in samples in the traditional degradation analysis of long-life aerospace products, a degradation modeling method combining a frailty model is proposed to quantify the unobservable heterogeneity in random degradation processes. Firstly, the frailty term is described using a generalized inverse Gaussian distribution to more comprehensively capture random effects. Secondly, an inverse Gaussian degradation model is established, combined with the frailty model to consider the unobservable heterogeneity in practical use. Through maximum likelihood estimation, parameter estimation is carried out, and Bayesian theory is used to infer an independent frailty term in order to quantify differences between individual products. Finally, reliability analysis is conducted on a core chamber cooling control valve of a certain aerospace engine. The results indicate that neglecting the unobservable heterogeneity will lead to overly ideal reliability estimates, and quantifying random effects reasonably can make the model’s estimation closer to the actual situation.

Full-resolution quality evaluation model for pansharpened images is significant for remote sensing applications, yet presents a challenge of the absence of reference compared with the reduced-resolution approach. To predict the image quality accurately, it is necessary to consider the distortion during the pansharpening process. Based on an observation that the quality of pairwise images can more easily be ranked, we propose a rank learning based full-resolution quality evaluation method for pansharpened images. Our approach begins with the synthesizing of ranked distortion images in spatial and spectral domains. Then, we develop a pansharpening distortion-perceiving model. This model employs spatial and spectral Siamese networks to perceive distortions and applies a pair-wise learning strategy for ranked images. Consequently, we establish a distortion-guided full-resolution quality evaluation framework for pansharpening. This framework integrates the spatial and spectral distortion-perceiving network and is enhanced with a dimension alignment module and a discrepancy Rrpresentation module, enabling effective distortion extraction among high-resolution multispectral, panchromatic, and low-resolution multispectral images. We conducted a series of experiments on a large-scale public pansharpened database. The experimental results demonstrate the effectiveness of our proposed approach.

High-resolution, wide-range imaging is crucial for contemporary and future remote sensing surveys. Multichannel synthetic aperture radar (SAR) is an effective tool to fulfill this requirement. Among various multichannel technologies, null steering digital beamforming is commonly used for ambiguity suppression, SAR echo separation, and interference and clutter suppression. However, existing null steering beamforming algorithms lack robustness and are sensitive to channel errors. This article proposes an improved SAR echo separation scheme based on complex independent component analysis (ICA) to address the channel error issue. Initially, the problem caused by channel error leading to the failure of the linear constraint minimum variance (LCMV) beamformer is analyzed. Then, non-Gaussian and noncircular measurements of SAR echoes from different scatterers are conducted, concluding that multichannel SAR echoes satisfy the conditions for ICA application. For an elevated multichannel wide-swath SAR, an ambiguity signal separation scheme based on ICA is proposed. Block processing or pure-decimal-constraint LCMV (PDC-LCMV) beamforming is first employed to compensate for the spatial dependence of the mixing matrix. A complex entropy-bound maximum algorithm, based on source signal independence, is then used for further signal separation. Our proposed scheme effectively addresses the channel error issue in traditional null steering beamforming schemes and efficiently separates ambiguity echoes. It leverages the statistical characteristics of the signal itself and expands the processing dimension. Capable of handling multiple errors, it is suitable for systems with limited degrees of freedom and does not require additional hardware resources. Simulation experiments verify the effectiveness of this scheme.

Acupuncture is an important therapeutic method of traditional Chinese medicine and can effectively modulate brain disorders. The therapeutic efficacy of acupuncture is hard to evaluate due to lacking of effective measurements of brain activity. In this work, we design an EEG-based monitoring system to evaluate therapeutic effect of acupuncture on human brain by extracting periodic-aperiodic features. Power spectral density is estimated to compute the adjusted power of periodic oscillatory rhythm in EEG under acupuncture stimulation. It is exhibited that the brain activity in alpha band (8–12 Hz) is significantly enhanced during acupuncture, especially in parietal and occipital lobe regions. To probe the modulatory effect of acupuncture on aperiodic brain activity, we calculate the aperiodic exponent based on the parameterization of EEG power spectra. The aperiodic exponent decreases along with acupuncture process, which is more significant in central and frontal lobe regions. Furthermore, sensitivity of different brain regions to acupuncture is assessed by the integration of adjusted power and aperiodic exponent. Experimental results demonstrate the effectiveness of proposed periodic-aperiodic measurements of EEG signals, by which different effects of four acupuncture manipulations are precisely evaluated and a knowledge graph is established. The monitoring system provides a new perspective to quantitatively evaluate acupuncture effect on human brain and improve its therapeutic efficacy in clinical applications for neural disorders.

With the increasing prevalence of multimodal user-generated content on social media, Multimodal Aspect-Based Sentiment Analysis (MABSA) has garnered significant attention in recent years. MABSA aims to classify the sentiment polarity of aspects mentioned in textual content by leveraging both textual and visual modalities. Previous studies have primarily focused on using transformer-based models to fuse information from different modalities. However, the presence of irrelevant information in images and text related to specific aspects can negatively impact sentiment analysis results. Many studies still face challenges in handling the noise introduced during the multimodal fusion process. To address this issue, we propose a novel Position-Perceptive Multi-hop Fusion Network (PPMFN) in this paper. Our method includes a position-perceptive module and an aspect-guided multi-hop interactive attention fusion module, aimed at enhancing modal position perception and aspect-guided multi-hop interactions to filter out irrelevant noise information. Typically, parts closer to the target are more relevant. The position-perceptive module is used to capture positional features from both images and text. The aspect-guided fusion module leverages a multi-hop interactive attention network to merge multimodal features while eliminating irrelevant information from both images and text. Extensive experimental results demonstrate that our framework consistently surpasses strong baseline models across two public datasets.

The protection of dc systems in mobility applications, such as land transport, aircraft, and shipping, presents significant challenges due to the need for high-power-density equipment in confined spaces. This article focuses on dc systems onboard ships, for which diverse applications require different power levels, architectures, and protection strategies. Existing protection frameworks and regulations are often inadequate or outdated for the field, leading to certification issues and insufficient fault analysis. This research proposes a use-case-based categorization of short-circuit currents for primary systems. A reference scenario is created using a simulation model of a 5-MW system in a superyacht to provide a short-circuit inventory. The study proposes three contributions: a comprehensive fault inventory, a qualitative categorization, and relevant recommendations for power converter design. The research highlights the importance of fault categorization in understanding the impact of various short circuits on shipboard dc systems. The study emphasizes the importance of the evolution of materials and power converters in developing efficient protection technologies for ships. This work addresses some fundamental gaps in shipboard dc systems, providing a foundation for improved protection strategies and regulations, ultimately contributing to the advancement of protection of shipboard dc systems.

This study demonstrates the first work that achieves accurate modeling of Hydrogen plasmatreated (H-treated) p-GaN gate devices with the ASM-GaN model, facilitating simulations for applications in monolithic integrated circuit (IC) design. The workflow for ASM-GaN model parameter extraction and optimization using IC-CAP is proposed. The I-V characteristics of both Enhancement / Depletion (E/D) mode devices are modeled and fitted. The impact of device capacitance on the dynamic properties of monolithic IC is investigated through the ASM model. The results demonstrate that Cds, Cgd, and Cgs have different effects on the monolithic logic circuit performances. The high-level fitting of experimental data and circuit simulation of Inverter, NAND, and Comparator circuits proves the credibility of the modeling workflow and device capacitance modulation. This work provides a method to speed up the GaN monolithic IC design by accurate modeling with fast parameter extraction workflow regardless of the fabrication process. The reliable prediction of the circuit’s dynamic performance will lay the foundation for designing and scaling up the GaN monolithic IC application.

Many odors, such as poisonous and exhaust gases, are unsuitable for human detection. Therefore, techniques for predicting gas types and concentrations are essential. The metal oxide sensor utilized in this study is a standard gas sensor due to its excellent stability, affordability, and high sensitivity. However, it does have a drawback: lower selectivity towards different gases. To address this, we employ temperature modulation to enhance the sensor’s selectivity. By employing temperature modulation, we can observe the dynamic response of the sensor and capture more features. We utilize low-frequency square wave and triangular wave signals as heating voltages. The former represents rapid heating, while the latter embodies a slower heating process. The objective is to employ these two methods to classify and predict the concentrations of ethanol, methanol, MEK, and ethyl acetate, as well as mixtures of the four gases. Following feature extraction, a neural network is employed for classifying and predicting gas types and concentrations. The results demonstrate a 100% classification accuracy under two different heating voltages. Regarding concentration prediction, using the square wave alone yields an error range of +42.09 to -32.41 and a root mean square error (RMSE) of 4.68%; utilizing the triangular wave alone yields an error range of +13.86 to -18.10 and an RMSE of 4.33%. Employing both waveform types simultaneously results in an error range of +37.97 to -17.48 and an RMSE of 2.91%.