Search

Currently, some empirical formula methods, such as the Steinmetz equation, have been widely used in calculating high-frequency magnetic core loss. Although these methods may provide satisfactory accuracy under some certain conditions, they not only cannot present clear physical interpretation and fail to account for the origin and variation of losses, but also offer much lower accuracy for wide operation ranges. In this article, a new analytical model is proposed for the first time to predict core loss by the new vector magnetic circuit theory. In addition to reluctance, two new components, i.e., magductance and hysteretance are proposed in the vector magnetic circuit, which are used to describe the eddy-current loss and the hysteresis loss, respectively. Thus, the clear physical significance is available with the proposed loss model. Given the geometrical and physical parameters of cores, the loss can be predicted. Furthermore, the uneven distribution of magnetic flux inside the magnetic core is characterized by parallel vector magnetic circuits. By comparing with the available empirical formulas, it has been proven that the proposed analytical model not only offers clear physical concepts, but also has much higher accuracy.

Automatic coordinated control of power transformers is essential to stable operation of power systems. However, there still lack mature intelligent solutions for this purpose. As a result, this paper proposes a reinforcement learning-based automatic collaborative control approach for power transformers. Firstly, by establishing the state space and action of the transformer control system, deep reinforcement learning is used to optimize the control strategy. Then, deep Q-network is utilized to automatically learning and adjust the operating parameters of each transformer to achieve coordinated control of power systems. In the algorithm design, we consider multiple factors such as power grid load and voltage stability. And they are incorporated into the reward function of reinforcement learning to facilitate learning appropriate control strategies. The experimental results show that the new method not only improves the response speed of the control system, but also effectively enhances the stability and robustness. In addition, we conducted in-depth analysis on the convergence and learning efficiency of the system, verifying the feasibility and superiority of the proposed method compared with traditional methods.

In dynamic environments, the traditional relative positioning methods based on the Kalman filter model suffer from low accuracy and stability due to the influence of noise and outliers. This paper proposes a variational Bayesian filtering algorithm based on the combination of four-frequency observations from BDS (BeiDou Navigation Satellite System) and models the observation noise using the T-distribution to enhance the stability of filtering. Firstly, a geometrically correlated ambiguity resolution model is constructed based on the characteristics of the combined observations, effectively improving the precision of float ambiguity resolution and fixing rate. Moreover, considering the characteristics of outliers that are likely to occur in dynamic conditions, a T-distribution-based variational Bayesian filtering approach is employed to estimate the time-varying observation noise and system states. Experimental results demonstrate that the proposed method exhibits robustness and stability in dynamic short baseline scenarios, leading to further improvements in positioning accuracy, float ambiguity resolution precision, and fixing rate.

Leakage current is one of the important parameters reflecting the operation status of distribution network-type surge arresters. At this stage, the polymagnetic current sensor has the advantages of miniaturization and high accuracy for leakage current measurement, but the complexity of electromagnetic interference in the field easily introduces more noise interference signals, which limits the performance of the polymagnetic current sensor for field application. To this end, this paper proposes a leakage current measurement method for distribution network-type surge arresters based on EEMD-SVD and low-rank RBF neural network methods. Firstly, the measured leakage current is decomposed by Ensemble Empirical Mode Decomposition (EEMD) to decompose the modal components containing eigenfrequencies, and then Singular Value Decomposition (SVD) is used to extract the non-eigenfrequency signals. Finally, a low-rank Radical Basis Function (RBF) neural network is used to approximate the leakage current signal after de-interference and combined with the Gaussian window function to remove the white noise interference. The experimental results show that the signal-to-noise ratio of the polymagnetic current sensor is improved by about 20dB and the maximum average absolute error is only 2.62%, which can truly reflect the leakage current of the network-type lightning arrester.

Building a broadband equivalent circuit model of a transformer is highly significant for studying the characteristics of high-altitude electromagnetic pulse conduction environments in real scenarios and guiding the high-altitude electromagnetic pulse effect tests of power equipment. This paper utilized an impedance analyzer to obtain the port impedance characteristics of distribution transformers with two insulation forms and five capacities within the range of 20Hz to 30MHz, based on the concept of passive system circuit modelling. By analysing the impedance characteristics of passive circuit components and observing the trend of changes in the port impedance curve, broadband port circuit models were constructed for both the high-voltage and low-voltage sides of transformers. Model parameter acquisition methods using a combination of function fitting and resonance point characteristics were provided. A high-altitude electromagnetic pulse injection test platform was constructed to verify the accuracy of the transformer model through comparison. Finally, the influence of transformer capacity and insulation form on the parameters of the transformer broadband port model was summarized. The results demonstrate that the proposed broadband port equivalent circuit model is suitable for different types and capacities of distribution transformers and can be employed to investigate the characteristics of high-altitude electromagnetic pulse conduction environments in transmission and distribution lines.

Due to the fact that a high share of renewable energy sources (RESs) are connected to high-voltage direct current (HVDC) sending-end AC power systems, the voltage and frequency regulation capabilities of HVDC sending-end AC power systems have diminished. This has resulted in potential system operating problems such as overvoltage and overfrequency, which occur simultaneously when block faults exist in the HVDC link. In this study, a steady-state voltage security-constrained optimal frequency control method for weak HVDC sending-end AC power systems is proposed. The integrated virtual inertia control of RESs is employed for system frequency regulation. Additional dynamic reactive power compensation devices are utilized to control the voltage of all nodes meet voltage security constraints. Then, an optimization model that simultaneously considers the frequency and steady-state voltage security constraints for weak HVDC sending-end AC power systems is established. The optimal control scheme with the minimum total cost of generation tripping and additional dynamic reactive power compensation required is obtained through the optimization solution. Simulations are conducted on a modified IEEE 9-bus test system and practical Qing-Yu line commutated converter based HVDC (LCC-HVDC) sending-end AC power system to verify the effectiveness of the proposed method.

Power transformers are important equipment in power systems. Most medium and large transformers are oil-immersed transformers, apart from some small- and medium-capacity transformers and gas transformers with special purpose. Because the reliability and safety of oil-immersed transformers can cause significant implications for power systems, it is extremely critical to detect malfunction timely and accurately. Therefore, the fault diagnosis model based on the evidential reasoning (ER) rule with reference points in the Gaussian distribution form optimized by a constrained genetic optimization algorithm (GA), named GER-G model, is proposed in this paper. The GER-G fault diagnosis model can monitor and identify the degree of transformer faults in real time. First, the concentration of dissolved gases in the oil of oil-immersed transformers varies at different fault degrees, so the gases with high weight are selected using the maximum correlation-minimum redundancy (mRMR) to constitute the indicator system. Second, to solve the uncertainty existing in the fault information of transformers, the reference points of the ER rule are continuous probability distribution reference points described by Gaussian distribution form. Third, the constrained GA is proposed to adapt the Gaussian distribution. And the accuracy of fault diagnosis results can be enhanced by optimizing the grade parameters using constrained GA. Finally, the internal and external influences in the environment are quantified using perturbation analysis, the GER-G fault diagnosis model is applied to the dataset of dissolved gases in the oil, and the robustness and validity of the GER-G fault diagnosis model are validated. Meanwhile, the validation results also show that the GER-G fault diagnosis model possesses higher accuracy compared with other fault diagnosis methods.

Local field potential (LFP) recorded by sensing-enabled neurostimulators provided chronic observation of deep brain activities for the research of brain disorders. However, the contamination from the electrocardiogram (ECG) deteriorated the extraction of effective information from LFP. This study proposed a novel algorithm based on minimizing the variance combining template subtraction to improve the performance of ECG artifact removal for LFP. Four patients with implanted electrodes were recruited, and eight real LFP records were collected from their left and right hemispheres, respectively. The results showed that the proposed method improved the accuracy of artifact peak detection in LFP, and the subsequent signal quality after template subtraction compared to the traditional Pan-Tompkins (PT) method. The outcome of this study benefited the LFP-based brain research, promoting the application of sensing-enabled neurostimulators in more areas.

Networked robots have become crucial for unmanned applications since they can collaborate to complete complex tasks in remote/hazardous/depopulated areas. Due to the cost inefficiency of deploying cellular network infrastructure in these areas, hybrid satellite-UAV networks emerge as a promising solution. These networks provide seamless and on-demand connectivity for multiple robots with various task requirements, and support computation-intensive and latency-sensitive services through mobile edge computing (MEC)-based offloading. However, to complete tasks in limited times, the rapid collective movement of mobile robots may cause frequent service migration, and a large number of gathered robots may compete for limited bandwidth resources in satellite and UAV communications. As a result, offloading latency may increase significantly. To address this issue, the average completion time of multi-robot offloading in task-oriented satellite-UAV networks with MEC is formulated as an optimization problem. Unlike conventional mobility-aware MEC-based offloading schemes, joint optimization of mobility control, data offloading, and resource allocation is proposed using velocity control of multiple robots. According to Lyapunov optimization, the original optimization problem is simplified into minimizing the average completion time of offloading for all robots within UAV and satellite coverage. A multi-agent $Q$ -learning algorithm, including multi-group dual-agent $Q$ -learning, is proposed based on local state observation and global reward calculation. In each dual-agent $Q$ -learning, one agent is responsible for velocity control and communication resource allocation, while the other is responsible for data offloading and computational resource allocation. The convergence of the proposed multi-agent $Q$ -learning algorithm is also theoretically analyzed. Simulation results show that the proposed scheme can effectively reduce the offloading latency by up to 35% in the multi-robot environment over its conventional counterparts.

3D human pose estimation (HPE) has become increasingly important in baseball analytic, but there are several difficulties pertaining to pose estimation in real-world baseball pitching. First, in-the-wild baseball pitching lacks related 3D pose datasets and contains lots of joints occluded by other body parts. Second, baseball pitching contains dramatic velocity changes during arm acceleration phases. Due to these properties of pitching, it is difficult to use common filters to remove random noises while preserving high-frequency critical joint movements in pitching. To solve these problems, we propose joint-wise volumetric triangulation to reconstruct 3D human poses by utilizing the information of multi-view 2D joint heatmaps generated by 2D HPE methods. We also designed a baseball-customized filter system to remove noisy signal from pose movement while preserving the high-frequency pitching motion. Our proposed pose reconstruction scheme yields a 33.1 mm average position error and 0.35m/s (1.28 km/h) average velocity error on baseball pitching motion. Our work can be directly applied to estimate human poses either in indoor environment or real-world baseball field.

Phase picking is a critical task in seismic data processing, where deep learning methods have been applied to enhance its accuracy. While lightweight deep learning networks have been optimized for edge computing devices, there is a lack of networks developed explicitly for embedded systems. This paper presents a seismic phase picking model, a hybrid network integrating convolutional neural networks and Transformer, designed for embedded systems. Optimizing network parameters and computational resources, the model significantly reduces resource consumption while guaranteeing accuracy. It employs a multi-branch architecture. Specifically, the global branch employs a modified self-attention mechanism, effectively extracting global features through shared contextual information. The local branch retains local information from the input features. Such a multi-branch architecture facilitates effective interaction between global features and local details, thereby more efficiently capturing the relationships among features. The model can be configured into variants with different sizes to match various embedded systems. This research evaluated the model using the Stanford Earthquake Dataset, achieving a precision of 99.9% for the P-phase and 99.3% for the S-phase. On Raspberry Pi, the model reduced inference time by 58.1% compared to the earthquake transformer while maintaining comparable detection performance.

To overcome the challenges posed by the complex structure and large parameter requirements of existing classification models, the authors propose an improved extreme learning machine (ELM) classifier for human locomotion intent recognition in this study, resulting in enhanced classification accuracy. The structure of the ELM algorithm is enhanced using the logistic regression (LR) algorithm, significantly reducing the number of hidden layer nodes. Hence, this algorithm can be adopted for real-time human locomotion intent recognition on portable devices with only 234 parameters to store. Additionally, a hybrid grey wolf optimization and slime mould algorithm (GWO-SMA) is proposed to optimize the hidden layer bias of the improved ELM classifier. Numerical results demonstrate that the proposed model successfully recognizes nine daily motion modes including low-, mid-, and fast-speed level ground walking, ramp ascent/descent, sit/stand, and stair ascent/descent. Specifically, it achieves 96.75% accuracy with 5-fold cross-validation while maintaining a real-time prediction time of only 2 ms. These promising findings highlight the potential of onboard real-time recognition of continuous locomotion modes based on our model for the high-level control of powered knee prostheses.

Osteoporosis is a prevalent chronic disease worldwide, particularly affecting the aging population. The gold standard diagnostic tool for osteoporosis is Dual-energy X-ray Absorptiometry (DXA). However, the expensive cost of the DXA machine and the need for skilled professionals to operate it restrict its accessibility to the general public. This paper builds upon previous research and proposes a novel approach for rapidly screening bone density. The method involves utilizing near-infrared light to capture local body information within the human body. Deep learning techniques are employed to analyze the obtained data and extract meaningful insights related to bone density. Our initial prediction, utilizing multi-linear regression, demonstrated a strong correlation (r = 0.98, p-value = 0.003**) with the measured Bone Mineral Density (BMD) obtained from Dual-energy X-ray Absorptiometry (DXA). This indicates a highly significant relationship between the predicted values and the actual BMD measurements. A deep learning-based algorithm is applied to analyze the underlying information further to predict bone density at the wrist, hip, and spine. The prediction of bone densities in the hip and spine holds significant importance due to their status as gold-standard sites for assessing an individual’s bone density. Our prediction rate had an error margin below 10% for the wrist and below 20% for the hip and spine bone density.

The WiFi 6E extends the WiFi 6 operation in the 6 GHz band and features the $8\times8$ MIMO system. Recently, many countries have also made the lower 6 GHz band (5.925-6.425 GHz) available for the license-exempt wireless local area network, including the WiFi 6E. For such MIMO systems featuring 8 spatial streams in the lower 6 GHz band, we propose the compact $2\times2$ dual-polarized patch (DPP) antenna array with 8 isolated ports in a small size of 40 mm $\times40$ mm (about $0.66\lambda ^{2}$ at 5.9 GHz) to transmit 8 uncorrelated waves for the WiFi-6E MIMO access point featuring 8 spatial streams. In addition, over the operating band, the isolation of any two ports thereof is larger than 20 dB and the maximum port isolation is larger than 30 dB. The high port isolation of the compact $2\times2$ DPP antenna array is owing to each DPP antenna therein clad with an L-shape metal wall. We also test the fabricated compact DPP antenna array and its application in the $8\times8$ MIMO system in an on-campus public indoor space. The measured $8\times8$ MIMO system capacity reaches about 53 bps/Hz, even though the tested public indoor space is not an ideally rich multipath environment. The obtained results imply that for the maximum 160 MHz channel in the WiFi-6E, the system capacity based on using the compact DPP antenna array in the practical public indoor space can be about 8.48 Gbps (53 bps/Hz $\times160$ MHz), which reaches 88% of the WiFi-6E’s ideal projected system capacity of 9.6 Gbps. We present in this study the proposed compact DPP antenna array and its measured $8\times8$ MIMO system results.

Understanding the response of land surface temperature (LST) to urbanization is important for coping with urban warming. However, few studies have discussed the impact of twin-city integration on LST. Here, we explored the relationship between the twin-city integration and LST over the Xi'an-Xianyang Metropolitan Area (XXMA) using the local contour tree algorithm during 2000–2020. Results indicate that: 1) The urban integration of XXMA is obvious, and the spatial scope of XXMA converges toward the Weihe River between the two cities. Besides, its structure tends to be more complex, which develops from 3 levels and 8 centers to 10 levels and 33 centers from 2000 to 2020. 2) The heat islands of XXMA tend to connect and reach the maximum area of 547.43 km2 in 2015, while the LST evolution in two cities is not synchronized. The LST contour tree of Xi'an increased to 6 levels and 26 centers in 2000–2015 and simplified to 5 levels and 17 centers in 2020, while it increased from 1 to 4 levels in Xianyang in 2000–2020. 3) Synergistic effects exist between urban expansion and LST evolution. The correlation coefficient and bivariate Moran's I between nighttime light (NTL) and LST are greater than 0.35, indicating NTL and LST are correlated and have pleasurable spatial consistency. The high–high agglomeration of NTL and LST increases by 132% in area over the 20-year period, showing a trend toward integration. The study results can provide a scientific basis for thermal environment management and urban climate improvement in XXMA.

This work explores the application of hypothesis testing to enhance anomaly detection in time series data from the cigarette cutting production process, aiming to improve detection accuracy and robustness. Historical production data from the cigarette cutting workshop of the S Cigarette Factory are first collected. A deep learning-based anomaly detection model is developed using the Temporal Convolutional Network (TCN) algorithm, integrating TCN-Long Short-Term Memory (TCN-LSTM) and hypothesis testing. This model utilizes TCN-LSTM to extract features from the production data and applies the Wilcoxon signed-rank test for hypothesis testing. Results indicate that the model’s Mean Absolute Error, Mean Relative Square Error, and R2 values at a time step of 15 seconds are 11.54, 16.17, and 95.80%, respectively, improving by over 3% than other models. Additionally, the hypothesis testing further validates the model’s effectiveness and robustness at this optimal time step. Therefore, this model not only significantly enhances the accuracy of anomaly detection but also provides a new method and reference for intelligent monitoring and optimization of the cigarette cutting process.

Few-shot object detection (FSOD) aims to identify novel objects using only a limited number of samples. While impressive results have been achieved in natural scene images, object detection in remote sensing images (RSIs) presents unique challenges due to significant variations in orientation and size. Existing approaches often rely on horizontal bounding boxes, which may include a substantial quantity of irrelevant background because of object orientation, thus hindering accurate detection in RSIs. To address this limitation, we propose a metalearning-based method for arbitrary-oriented FSOD in RSIs, called AOFS. In our method, three key components are added to the YOLOv5-based one-stage detection architecture: a hierarchical metafeature encoder, an adaptive feature modulator, and an oriented bounding box decoder. We use features extracted from a small set of annotated samples to reweight the metafeatures of the objects. The oriented bounding box decoder provides outputs for object orientations. Experimental results on the benchmark datasets (NWPU and DIOR) indicate the effectiveness of our model. AOFS achieves 65.1% and 48.9% mean average precision in 3-shot setting on NWPU and DIOR, respectively, exceeding the second-best method by 4.4% and 5.4%.

Staggered synthetic aperture radar (SAR) is compatible with quadrature-polarimetric (quad-pol) SAR, enabling ultrawide continuous swath quad-pol SAR imaging with fine azimuth resolution. However, the quad-pol staggered spaceborne SAR is constrained by severe range ambiguity in the cross-pol channel. Moreover, azimuth phase coding (APC), as an excellent technique for range ambiguity suppression, is no longer exactly effective in the staggered imaging mode because the range ambiguities for different azimuth samples are located at different slant ranges. Given the above challenges presented in quad-pol staggered SAR, in this article, a novel pulse repetition interval (PRI) variation strategy based on APC is proposed to suppress range ambiguity in quad-pol staggered spaceborne SAR. First, the range ambiguity signal for quad-pol staggered SAR is modeled and analyzed. Then, the dedicated PRI variation strategy for quad-pol staggered SAR is developed, by which the near or far first-order range ambiguity is coherent in azimuth and can be subsequently suppressed using APC technique. Especially, a scale factor is developed in the novel PRI variation strategy, which can be flexibly designed to optimize the time interval of alternately transmitted polarized pulses. By this, most of the incoherent far or near first-order range ambiguity energy can be significantly reduced. Finally, simulations are conducted to verify the advancement of the proposed approach for range ambiguity suppression in quad-pol staggered SAR. The work in this article can be viewed as an important candidate for range ambiguity suppression in future quad-pol staggered spaceborne SAR.

Network neuroscience, especially causal brain network, has facilitated drug-resistant epilepsy (DRE) studies, while surgical success rate in patients with DRE is still limited, varying from 30% $\sim ~70$ %. Predicting surgical outcomes can provide additional guidance to adjust treatment plans in time for poorly predicted curative effects. In this retrospective study, we aim to systematically explore biomarkers for surgical outcomes by causal brain network methods and multicenter datasets. Electrocorticogram (ECoG) recordings from 17 DRE patients with 58 seizures were included. Ictal ECoG within clinically annotated epileptogenic zone (EZ) and non-epileptogenic zone (NEZ) were separately computed using six different algorithms to construct causal brain networks. All the brain network results were divided into two groups, successful and failed surgeries. Statistical results based on the Mann-Whitney-U-test show that: causal connectivity of $\alpha $ -frequency band ( $8~\sim ~13$ Hz) in EZ calculated by convergent cross mapping (CCM) gains the most significant differences between the surgical success and failure groups, with a P value of 7.85e-08 and Cohen’s d effect size of 0.77. CCM-defined EZ brain network can also distinguish the successful and failed surgeries considering clinical covariates (clinical centers, DRE types) with ${P}\lt {0}.{001}$ . Based on the brain network features, machine learning models were developed to predict the surgical outcomes. Among them, the SVM classifier with Gaussian kernel function and Bayesian optimization demonstrates the highest average accuracy of 84.48% by 5-fold cross-validation, further indicating that the CCM-defined EZ brain network is a reliable biomarker for predicting DRE surgical outcomes.

The successful launch of the first spaceborne CO2 IPDA LIDAR onboard Daqi-1 (DQ-1) in April 2022 marks a milestone in advancing scientific research. However, a notable discrepancy in the physical definitions of ${{\mathrm{X}}_{\mathrm{C}{{\mathrm{O}}_2}}}$ products between the IPDA LIDAR and TCCON presents a challenge for directly using TCCON for verifying and evaluating the performance of DQ-1’s ${{\mathrm{X}}_{\mathrm{C}{{\mathrm{O}}_2}}}$ products. To address this, we propose a method based on statistical hypothesis testing to globally validate the ${{\mathrm{X}}_{\mathrm{C}{{\mathrm{O}}_2}}}$ products derived from the spaceborne LIDAR. Our validation method does not compare DQ-1’s observations with the TCCON observations. We only utilize the useful information from TCCON to simulate the probability distribution of the true ${{\mathrm{X}}_{\mathrm{C}{{\mathrm{O}}_2}}}$ value in the LIDAR definition, and then hypothesis testing is adopted for deriving the systematic error. Our method improves the accuracy of computing the systematic error of DQ-1 by over 50% compared to the traditional approach. Up to now, we have produced the first four months (June 2022–September 2022) of DQ-1’s ${{\mathrm{X}}_{\mathrm{C}{{\mathrm{O}}_2}}}$ products. In this context, we present preliminary validation results based on the four months of data. Based on our method, we have validated that the accuracy of DQ-1’s ${{\mathrm{X}}_{\mathrm{C}{{\mathrm{O}}_2}}}$ products is 0.1 ± 1 ppm. This approach not only sets the stage for future official global validations of DQ-1’s ${{\mathrm{X}}_{\mathrm{C}{{\mathrm{O}}_2}}}$ products but also holds promise for application in upcoming similar missions such as MERLIN and DQ-2.

The power transmission network, which delivers power energy from generator to customers, plays an important role in the power grid. Insulator is a basic component in the power transmission network. Its defects may lead to the paralysis of the entire transmission network, resulting in serious electricity accidents. Therefore, how to use artificial intelligence and other emerging technologies to realize automatic detection of power line insulator defects has become an urgent problem to be solved. To accurately detect insulator defects in complex environment, this article proposes Attention Enhanced YOLOv5 (AE-YOLOv5) by inserting visual attention modules into original YOLOv5 model. In particular, we design a Channel-Spatial Attention module and plug it into the backbone of YOLOv5 to enhance its representation learning ability. Furthermore, a Multi-scale Attention module is also proposed to enhance the Feature Pyramid Network (FPN). To validate the efficacy of our proposed model, we conducted training and testing on a dataset collected from real-world scenarios. The experimental results demonstrate that our model can effectively and accurately detect defects of power line insulators in real-time.

Automatic seizure detection plays a crucial role in epilepsy diagnosis and treatment. Traditional machine learning based automatic seizure detection requires additional feature engineering and finding the optimal hand-crafted features is a challenging issue. Therefore, a novel end-to-end deep learning model that combines attention mechanism, temporal convolutional network(TCN), and bidirectional gated recurrent unit (BiGRU) is proposed for seizure detection in this work. Our model only requires filtering of the raw electroencephalogram(EEG) signals to remove artifacts, without the need for time-consuming feature extraction. Post-processing procedures including moving-average filtering, thresholding, and the collar technique are then applied to enhance the model’s performance. Experiments were conducted on the CHB-MIT dataset and the SH-SDU dataset. In patient-specific experiments, our model achieved average accuracies of 98.77% on the CHB-MIT dataset and 93.61% on the SH-SDU dataset. In cross-patient experiments, average accuracies of 93.78% and 91.37% were obtained, respectively. The total time required for the model to process 1-hour EEG signals is only 5.33s. These outstanding results indicate that our model achieves high accuracy and real-time performance in seizure detection tasks and could provide reference for clinical seizure diagnosis.

Realizing online monitoring of the hot spot temperature of grid mounted transformer windings is a challenge for the power industry. The heat exchange process between the internal heat source area and the external field area of the transformer is the direct cause of the temperature increase of the transformer shell. In this paper, based on the numerical coupling calculation of temperature - fluid field of 10kV oil immersed solid coiled iron core transformer, the typical heat flow streamline is extracted, and the measurable temperature characteristic points strongly correlated to the winding hot spot temperature are selected in the heat dissipation area of the shell, and the inversion model of the transformer winding hot spot temperature is established. This article verifies the effectiveness of the simulation results and the accuracy of the inversion model through the multi condition temperature rise test of transformer. The maximum error between the calculated steady-state temperature rise of each temperature measurement point in the simulation model and the test value is no more than 3°C, the simulation calculation results are in good agreement with the test results, and the average absolute error of the inversion model winding hot spot temperature inversion is 2.67 %, the maximum temperature difference is less than 3°C. This research provides an effective method for online monitoring of the hot spot temperature of the winding of a 10kV oil-immersed three-dimensional coil iron core transformer, and provides a reference for the state detection of grid mounted transformers.

Ship detection in synthetic aperture radar (SAR) images is crucial in both civilian and military fields, offering extensive application prospects. Nonetheless, owing to the distinctive characteristics of SAR imaging, this task confronts numerous challenges. Specifically, ships with high aspect ratios, dense arrangements and small sizes in complex environments frequently yield in suboptimal positioning effects, consequently impacting detection performance. In response to the challenges in ship target detection, this article introduces a novel approach, termed Inner-alpha-CIOU (IA-CIOU), that relies on an enhanced intersection over union (IOU). Primarily, the method introduces Inner IOU, which effectively regulates generation of auxiliary bounding boxes through scale factor r. This ensures a better fit for dimensions of ship target frames, thereby enhancing target detection performance as well as expediting model convergence. Subsequently, this method introduces Alpha IOU, enhancing robustness of small-size ship targets in complex backgrounds by adjusting α. This allows the detector to achieve greater flexibility in ship regression accuracy. Following numerous experimental validations, proposed algorithm consistently outperforms on both SAR-Ship-Dataset, MSAR-1.0 dataset, and SAR ship detection dataset (SSDD) dataset. This groundbreaking innovation not only possesses immeasurable practical worth, but also introduces a fresh perspective together with enlightening insights for future research efforts.

An adjustable gate driver is proposed based on the optimization of switching transient performances. The turn-on process is optimized through an adjustable charging current, while the turn-off process is optimized through an adjustable segmented drive current. The proposed gate drive circuit has the advantage of reducing switching loss, delay, and total switching time, while maintaining the switching stress and EMI noise level during both turn-on and turn-off transients. The proposed driver can be integrated into gate driver ICs, and the switching performance can be externally adjusted to match the power device for different applications. By combining the above techniques, an adjustable gate driver IC has been developed using the $0.25~\mu $ m BCD process. The simulated and experimental results validate the proposed technique. Compared to the conventional gate driver, the overshoot voltage has been reduced from 15.4% to 8.3%, and the overshoot current has decreased from 20.5% to 10.2% under a 10A load condition.

In arc array bistatic synthetic aperture radar (AA-BiSAR), an accelerated moving transmitter is helpful for enlarging the imaged scene and improving the flexibility of the imaging system, which is of great significance in the field of helicopter-assisted landing and emergency rescue. The motion state of the transmitter is closely related to the SAR echo model and the imaging quality. In this article, the imaging geometry of AA-BiSAR with an accelerated moving transmitter is established, and the impact of the transmitter acceleration on the imaging quality is analyzed. According to the echo signal model, a two-step compensation imaging algorithm based on the series reversion method is proposed. First, the first compensation for the residual phase caused by the movement of the transmitter is operated in the azimuth-time range-frequency domain by the transfer function, which is obtained from the approximated fourth-order high-precision Taylor series expansion of the double-root bistatic slant range. Then, the second compensation for the coupling terms between angle and range frequencies are carried out in the two-dimensional (2-D) frequency domain by uniform range migration and second range compression functions. The 2-D spectrum is derived by the series inversion method. Imaging results of simulated point and distributed targets are validated by the proposed imaging method.

The fractured-vuggy carbonate reservoir comprises various types of storage and seepage spaces, and is composed of multi-scale dissolution pores and fractures. The frequent changes to working systems make the characteristics of water breakthrough complex, and the production data nonlinear and non-stationary, resulting in great difficulty in real-time prediction. Traditional production forecasting methods only consider temporal correlations, neglecting the spatial correlations between production wells and local geological features. In this paper, adopts a modular design approach that comprehensively considers the spatiotemporal characteristics by abstracting each production well in the unit as a directed graph network node. We establish a graph attention network module based on the connectivity between wells to simulate fluid motion patterns and extract spatial features. To address the autocorrelation characteristics of the production sequences, we use a self-attention mechanism module to capture the temporal dependency relationships between production sequences. Finally, considering the fusion of spatiotemporal features, a gating mechanism is designed to adaptively aggregate spatiotemporal characteristics produced by the previous two modules, enabling dynamic production forecasting. We validate our proposed model using real-world production data from the Tarim Basin in China. Our experimental results demonstrate the superiority of the new model over existing production prediction models in fractured-vuggy carbonate reservoirs.

Predicting outage areas for affected customers in a disaster requires real-time meteorological data, which can be challenging to obtain and process with high quality. In particular, missing or inaccurate data can occur in certain regions or during extreme weather events. Therefore, a Convolutional Neural Network-Long Short-Term Memory (CNN-LSTM) based algorithm is proposed to improve the accuracy and reduce the prediction time in the outage region. Ground-based automated weather observatories are used to obtain real-time weather data, including extreme weather data. These data are pre-processed to enhance data quality and accuracy by removing duplicates, filling in missing values, handling anomalies, normalizing input variables, and reducing dimensionality. Based on the results of the data pre-processing, the outage rate was calculated for different types of meteorological disasters and the geographical characteristics of the outages were analyzed. This analysis provides insights into the impact of different types of meteorological disasters on power outages and helps improve the accuracy of predictive models. The proposed algorithm employs a CNN neural network to capture spatial and temporal information from raw meteorological data by stacking convolution and pooling layers. The extracted features are then organized and output by fully connected layers, laying the foundation for subsequent time series modeling. An LSTM network is further utilized to construct a prediction model for the outage area, which takes as input the feature extraction results of the meteorological data. By integrating the temporal dimension information of meteorological data, the model outputs accurate predictions for the outage area. The experimental results demonstrate a consistent outcome with the actual test results, achieving high prediction accuracy with a short prediction time of 4.3 s and a maximum non-outage detection rate of 2%. Therefore, the proposed algorithm proves to be significant for accurate and fast prediction of outage areas for affected customers in real world applications.

Zinc oxide arresters, a major type of overvoltage protection equipment in power systems, can suppress transient lightning overvoltage, operation overvoltage, and other types of power surges. However, the response characteristics and protection effects under nanosecond pulses, such as those in a high-altitude electromagnetic pulse conduction region, remain unclear. Therefore, the surge protection performance of arresters should be properly understood, and high-performance arresters should be developed. In this study, critical parameters, including overshoot peak voltage, residual voltage, response time, and voltammetry characteristics of a representative 10 kV ZnO arrester under nanosecond pulses, were obtained using a double exponential pulse current injection source. The overshoot peak voltage of this arrester under a nanosecond pulse is 2.19 times of lightning impulse residual voltage, and the residual voltage ratio under a nanosecond pulse is 4.31. The average response time is approximately 45 ns, and the response time is independent of the nanosecond pulse amplitude. Furthermore, an electromagnetic transient model of this type of lightning arrester is established based on the CIGRE guidelines, as well as a method to determine the key parameters. The model can simultaneously account for the response characteristics of the small current region and the conduction region of the arrester, which properly agrees with the experimental results.

Network Traffic Classification (NTC) plays a critical role in modern network security and management, used in various applications such as performance monitoring, anomaly detection, and bandwidth management. Traditional methods like port-based identification and deep packet inspection (DPI) face limitations due to evolving encryption techniques and dynamic port usage. Recent advances in machine learning have improved NTC but often struggle with data labeling, encrypted traffic, and model complexity. This paper introduces a novel Information Bottleneck Neural Network (IBNN) to address these challenges by efficiently extracting essential features from raw traffic data while minimizing irrelevant information. The IBNN framework balances model complexity and discriminative feature preservation, leading to enhanced classification accuracy and robustness, particularly for encrypted traffic. Through information-centric feature extraction and relevant information compression, the proposed method outperforms existing techniques about 5–7% in terms of accuracy, as well as F1 score on both ISCX 2012 and UNB CIS datasets, delivering superior resilience and adaptability across diverse network scenarios. Extensive experiments demonstrate the efficacy of IBNN, positioning it as a significant advancement in the field of NTC.

In recent years, the advancement in ranging accuracy and distance requirements in the ship industry has underscored the paramount importance of controlling ranging error. Beyond a reasonable range, any error may result in irreparable losses, including those pertaining to the safety of the ship and its personnel. Therefore, this study proposes a binocular vision ranging method based on sparse stereo matching and feature point extraction. The method accurately extracts the feature points in the image by optimizing the feature extraction algorithm and fast robust feature matching filtering algorithm. The experimental results showed that in data 1, the non-great suppression algorithm suppressed 5111 feature points when the suppression ratio was 20%. In data 2, the improved algorithm suppressed 1920 feature points when the suppression point feature ratio was 70%, while it was 3950 when the ratio was 20%. The total number of extracted feature points was the most in data 3. Moreover, the improved algorithm suppresses 4060 feature points in 70% ratio, compared to 89050 in 20% ratio. The results reveal that the ranging algorithm has higher accuracy in feature point extraction, which provides an important theoretical basis for the intelligent development of the ship industry.

The conversion of agricultural lands, termed “nonagriculturalization,” poses profound threats to food security and ecological stability. Remote sensing image change detection offers an invaluable tool for monitoring this phenomenon. However, most change detection techniques prioritize image comparison over exploiting accumulated vector datasets. Additionally, many current methods are not readily applicable in practical scenarios due to inadequate model generalization capabilities and a scarcity of samples, resulting in a continued reliance on manual intervention for nonagriculturalization detection. In response, this article introduces a novel change detection approach for nonagriculturalization based on the vector data and contrastive learning. Initially, the boundary-constrained simple noniterative clustering algorithm is applied to segment two-phase images under vector data guidance. Samples are then generated using an adaptive cropping method. For early phase image samples, a collaborative validation-based sample annotation framework is employed to optimize and annotate the samples, with the purified high-quality samples serving as the training set for subsequent classification. For later-phase image samples, only those within the cropland vector polygons are retained for prediction. Building on this, a semi-supervised cross-domain contrastive learning framework is proposed for remote sensing scene classification. Ultimately, by integrating nonagriculturalization rules and postprocessing techniques, areas undergoing nonagriculturalization are further detected. Validating our methodology on Wuxi and Yangzhou datasets yielded precision rates of 91.57% and 89.21%, with recall rates of 93.68% and 90.51%, respectively. These outcomes affirm the effectiveness of our method in nonagriculturalization detection, offering robust technical support for research in this domain.

The rapid increase in satellite imagery from Gaofen satellites has created an urgent need for efficient compression methods. Existing techniques are significantly limited in compression ratio when high reconstruction quality must be preserved, particularly with 16-bit multispectral images. In this article, we introduce a novel compression framework for Gaofen satellite multispectral images that leverages a lightweight bit-depth recovery network. Our approach divides the image into the most significant bits (MSB) image and the least significant bits (LSB) image. The MSB image is compressed losslessly using a conventional codec to preserve essential information. The LSB image, on the other hand, is compressed effectively by overfitting a lightweight neural network to predict the LSB image during encoding. The optimized network parameters are utilized as side information to ensure precise bit-depth recovery during decoding. Experimental results on three types of Gaofen multispectral images demonstrate that our approach significantly outperforms state-of-the-art methods, effectively addressing the demand for efficient Gaofen satellite image compression while preserving high reconstruction quality.

The conventional TO-247-3 packages with single-sided cooling limit the thermal and electrical performances of discrete SiC MOSFET devices. In this paper, a double-sided cooling press-pack (PP) packaging approach for the discrete SiC MOSFET device is proposed to optimize its thermal and electrical performances. First, a double-sided cooling PP structure for the discrete SiC MOSFET devices is designed with a copper foam gate pin and an embedded fixture. Then, based on finite element simulations, the steady-state thermal and electrical performances of the discrete SiC MOSFET device with the double-sided cooling PP package are analyzed, and the parasitic inductance of the designed SiC MOSFET device is extracted by the ANSYS Q3D software. Finally, a prototype of the double-sided cooling PP SiC MOSFET device is fabricated, and test platforms are established to verify its performance. The research findings demonstrate that the designed double-sided cooling PP SiC MOSFET device can reduce thermal resistance and switching loss by 47.4 % and 42.3%, respectively.

Electro-optic modulators with large bandwidth and low voltage are crucial for the high-baud-rate digital communication and high-carrier-frequency analog links. Recently, thin-film lithium niobate (TFLN) modulators with sub-1 V voltage and 140 GHz bandwidth have been emerged on the low permittivity substrates, such as quartz. However, on the high permittivity environment, such as silicon substrates, the bandwidth dramatically reduced to below 100 GHz even if the voltage was raised to around 3 V. We break the voltage–bandwidth trade-off limit in TFLN modulators on silicon substrates using low inductance thick metal traveling wave electrodes, which reduce RF phase index and microwave losses while preserving high EO modulation efficiency. We demonstrate a TFLN EO modulator on silicon substrate with 3 dB EO bandwidth $>$ 120 GHz and half-wave voltage $(V_\pi) < $ 2 V. Bandwidth/$V_\pi$ reach 60, which is significantly larger than traditional TFLN modulators. The proposed thin-film lithium niobate modulators offers a practical solution for the hybrid integration of silicon and lithium niobate.

Semantic change detection (SCD) aiming to localize changes and identify their types is a complex pixel-level classification task. In SCD, detecting changes of narrow and long objects (NLOs) common in various scenarios is more challenging than detecting these of typical small objects, primarily because NLOs' narrowness makes them carry local details, while their length gives them a global view. The fusion of semantic with spatial information and the interaction of semantic with spatiotemporal information are crucial to accurately detecting changed NLOs. Insufficient utilization of these pieces of information carried by multilayer semantic feature and change feature would limit the representation of the characteristics of changed NLOs, thereby leading to missed detection and false detection. We propose a new model that integrates semantic, spatial, and temporal information, named SSTNet, for SCD. This model utilizes a designed multilayer feature fusion module to fuse features from different layers in the encoder, leveraging global semantic information and local spatial information. In addition, the semantic change feature interaction module facilitates the interaction between semantic feature and change feature, enhancing the correlation between semantic information and spatial information. A series of experiments conducted on two public datasets demonstrate that SSTNet improves the detection accuracy of changed NLOs and outperforms state-of-the-art methods.

In this paper, we consider a system that integrates an unmanned aerial vehicle (UAV) for sensing and communication. The UAV is responsible for communication with multiple users and sensing targets for multiple users. Our objective is to minimize the age of information (AoI) of the sensory data by jointly optimizing the scheduling, transmit beamforming, trajectory, and transmit power of the UAV, while considering the communication quality constraints and physical constraints. To solve the optimization problem in a low-complexity manner, we first derive a closed-form solution for the transmit beamforming. Then, we decompose the remaining problem into two subproblems: one focuses on maximizing the total amount of the target sensing time, while the other aims to minimize the total amount of the sensing task completion time. For the former subproblem, we propose a low-complexity chain-based algorithm. For the latter subproblem, we utilize the Lagrange dual method and successive convex approximation technique. Simulation results show that our proposed algorithm can achieve a smaller average AoI with a low complexity compared to benchmark schemes.

Rapid and accurate monitoring of the chlorophyll content of ginseng leaves to reflect the growth condition of ginseng is of great significance in guiding the cultivation of ginseng. To address the problem that the existing means of detecting chlorophyll in ginseng leaves are time-consuming and disruptive and cannot meet the demand for rapid detection of chlorophyll in ginseng leaves, this study firstly establishes a variety of prediction models for chlorophyll in ginseng leaves based on the hyperspectral reflectance data and the vegetation index, respectively, and determines the strengths and weaknesses of the models by comparing the RMSE and the MAE of the test sets; Secondly, the analytical model construction process used for ginseng leaf chlorophyll content prediction was obtained through comparison and summary, and a fast and non-destructive ginseng leaf chlorophyll prediction method based on hyperspectral imaging technology and combining vegetation indices with machine learning algorithms was proposed; The experimental results showed that the final VI-SPA-RFR ginseng leaf chlorophyll prediction model had the best prediction performance, which had an RMSE of 1.1568 and an MAE of 0.9936 in the test set. In summary, this study provides a method to achieve rapid and accurate monitoring of the chlorophyll content of ginseng leaves, which is expected to provide a scientific means of monitoring for the cultivation of ginseng and expand the application field of hyperspectral imagers in the field of agriculture.

The stable operation of the flywheel system is crucial for the on-orbit safety of astronautic engineering, so it is very important to conduct a flywheel system health state assessment. In the assessment modeling process, the model not only needs to address uncertainty to ensure the accuracy of the assessment results but also needs to have a transparent and reasonable assessment process and interpretable assessment results. Therefore, based on the in-depth study of the belief rule base (BRB) modeling method, this paper proposes the interpretability criterion of the optimized model and constructs a new BRB model for assessing the health state of the flywheel system with interpretability with full parameter optimization (IP-BRB). Firstly, based on the traditional interpretability criterion, the interpretability criterion of the optimization model is defined. Then, based on the projection covariance matrix adaptive evolution strategy (P-CMA-ES), the IP-BRB is proposed. Finally, the effectiveness of the model in the health state assessment of the flywheel system is verified by a case study of the assessment of bearing components in a flywheel system. The comparative results show that the IP-BRB model has certain advantages in terms of the accuracy of the assessment results and interpretability of the assessment process.

Conventional iterative phase retrieval suffers from an inherent phase ambiguity due to limited measurement intensity. Multimodal amplitude modulation introduces physical constraints to tackle the underdetermination challenge. However, the time overhead caused by mask switching slows down the imaging speed. To increase imaging speed, we report an accelerated coded phase retrieval method by optimizing modulation masks. Compared to existing methods that require at least four patterns as inputs, the proposed method requires only three mask modulations to robustly reconstruct complex objects. The transparent pixels of the two masks partially overlap, constituting a strong constraint on the objective function. An additional random mask increases the difference between diffraction intensity patterns and ensures that the algorithm converges. The proposed method of efficient modulation using pure amplitude elements may open the door to short-wavelength high-speed complex amplitude imaging. Numerical simulations and proof-of-principle experiments have verified the feasibility of this method.

Snow depth monitoring is crucial for hydrology, climate research, and avalanche prediction. While traditional global navigation satellite system (GNSS) reflectometer methods offer cost-effective snow thickness retrieval, they suffer from poor accuracy and robustness, especially in complex terrains and extreme weather. This study proposes an innovative snow depth retrieval technique employing a time-series recurrent neural network with bidirectional gated recurrent units (Bi-GRUs). Unlike traditional methods using signal-to-noise ratio (SNR) features, our algorithm utilizes the detrended SNR as Bi-GRU input, aiming to enhance accuracy, particularly in low snow depths and complex terrains. SNR observations from GPS L1 carriers at stations P351 and AB33 were analyzed. The Bi-GRU algorithm demonstrated high consistency with true snow depths at station P351 (coefficient of determination: 0.9766), with the root-mean-square error (RMSE) and the mean absolute error (MAE) of 9.1559 and 6.4185 cm, respectively. Compared to traditional methods, the Bi-GRU model improved the RMSE by 30.9% and the MAE by 44.5%. At station AB33, where snow depth variations were significant, accuracy improvements of 65.6% (RMSE: 7.4905 cm) and 63.2% (MAE: 5.6074 cm) were observed. In addition, the Bi-GRU model exhibited greater robustness compared to long short-term memory. These findings highlight the efficacy of the Bi-GRU-based approach, suggesting its superiority and broader applicability.

The global navigation satellite system reflectometry (GNSS-R) technology has shown significant potential in retrieving snow depth using signal-to-noise ratio (SNR) data. However, compared to traditional in situ snow depth measurement techniques, we have observed that the accuracy and performance of GNSS-R can be significantly impacted under certain conditions, particularly when the elevation angle increases. This is due to the attenuation of the multipath effect, which is particularly evident during snow-free periods and under low-snow conditions where snow depths are below 50 cm. To address these limitations, we propose a snow depth inversion method that integrates SNR signals with the support vector regression algorithm, utilizing SNR sequences as feature inputs. We conducted studies at stations P351 and P030, covering elevation angles ranging from 5° to 20°, 5° to 25°, and 5° to 30°. The experimental results show that the root-mean-square error at both the stations decreased by 50% or more compared to traditional methods, demonstrating an improvement in inversion accuracy across different elevation angles. More importantly, the inversion accuracy of our method does not significantly lag behind that at lower elevation angles, indicating its excellent performance under challenging conditions. These findings highlight the contribution of our method in enhancing the accuracy of snow depth retrieval and its potential to drive further advancements in the field of GNSS-R snow depth inversion.

To address the problems of slow convergence, low search accuracy, and easy fall into local optimum, and generating a large number of infeasible solutions when solving the 0–1 Knapsack Problem, which makes it difficult to obtain the optimal solution scheme, in this paper, we present a greedy induced mutation method for locally optimal solutions, namely, an improved Shuffled Frog Leaping Algorithm (Shuffled Frog Leaping Algorithm based on Greedy Algorithm and Mutation, SFLA-GA-M). First, the proposed algorithm adjusts the population generation of individuals in the SFLA to avoid infeasible solutions, thereby optimizing the local update strategy. Secondly, an induced mutation mechanism that incorporates both greedy algorithms and genetic algorithms is introduced to enhance the search accuracy of the algorithm. Finally, ten classical 0–1 Knapsack Problem cases were selected to prove the feasibility and robustness of the proposed SFLA-GA-M by comparing with other two SFLA variants, which are MDSFLA, DSFLA. Meanwhile, to further verify the performance of the SFLA-GA-M, two classical 0–1 KPs with multi-dimensional test cases were introduced by compared with other five algorithm: DEA, PSO, GA, BMA and IFMA, respectively. The experimental results show that SFLA-GA-M has achieved stronger convergence and stability compared to DEA, PSO, and GA, and it has displayed better search efficiency and a superior global search capability than BMA and IFMA in solving large-sized 0–1 Knapsack Problem.

Federated learning (FL) is an emerging machine learning paradigm that enables the participants to train a global model without sharing the training data. Recently, FL has been widely deployed in industrial IoT scenarios because of its data privacy and scalability. However, the current FL architecture relies on a central server to orchestrate the FL process, thus incurring a risk of privacy leakage and single-point failure. To address this issue, we propose a fully decentralized FL architecture based on blockchain technology. Unlike existing blockchain-based FL systems that use blockchain for coordination or storage, we use blockchain as a trustable computing platform for model aggregation. Furthermore, we model the interaction between the FL task publisher and participants as a Stackelberg game and design a rewarding mechanism to incentivize participants to contribute to the FL task. We build a prototype system of the proposed decentralized FL architecture and implement an FL-based damaged package detection application to evaluate the proposed approach. Experimental results show that the blockchain-based decentralized FL is feasible in a practical industrial IoT scenario, and the incentive mechanism performs well with real application data.

A DC model is proposed for amorphous oxide semiconductor (AOS) thin-film transistors (TFTs) applicable to various active layer thicknesses. With the back surface potential and its coupling with the front surface potential being considered, an explicit potential solution is developed. Then, the analytical drain current and physical definition of threshold voltage are derived based on a non-chargesheet expression of free charge density. It is verified that in the previous models for AOS TFTs, typically ignoring the back surface potential and the active layer thickness effects could result in obvious deviations in the values of parameters during the characterization of DC performance, especially for scaled devices with low channel thicknesses. By comparing with numerical calculations and experimental data, this model is validated to be more suitable for AOS TFTs with decreased dimensions, which could give more realistic distributions of the density of states in the channel during parameter extraction.

There is often a lack of consistency in the footprints between satellite observation data, as well as between satellite observations and numerical models. This mismatch is typically caused by the different spatial resolutions of observed data in multiple frequency channels. The Backus–Gilbert inversion (BGI) algorithm is widely used for brightness temperature remapping in various microwave payloads to make the observation footprints more consistent. The microwave radiation imager for the rainfall mission onboard FengYun-3G introduces a novel integration of imaging and sounding channels for precipitation monitoring. However, this integration presents challenges in the matching of observation footprints, as they exhibit not only the conventional discrepancy in footprint sizes but also a misalignment between the center locations of the footprints of the sounding and imaging channels. This article proposes an improved BGI algorithm to match the observation footprints from sounding channels with those from imaging channels in terms of both spatial resolution and location. This study uses a distance-based fixed window for the filtering of source observations and employs an automatic noise parameter selection scheme to strike a balance between the fit error of footprints and noise. A coefficient reuse method is applied to improve computational efficiency. The experimental results using simulated data demonstrate that our improved BGI algorithm leads to a significant decrease in the overall root mean square error by 80% and 50% in the clear-sky and typhoon cases, respectively, compared to the raw ${{T}_A}$. The remapping results of real observations demonstrate that our algorithm effectively suppresses noise in each channel.

Lung cancer is the leading cause of cancer-related deaths worldwide. Early diagnosis significantly improves the 5-year survival rate from 6% in patients with metastatic cancer to 60% in those with localized cancer. Histopathological examination is the gold standard for lung cancer diagnosis, but analyzing whole slide images (WSI) is time-consuming and prone to error for pathologists. This study aims to enhance the classification accuracy of non-small cell lung cancer (NSCLC) histopathological images by proposing a novel deep-learning architecture that integrates convolutional neural networks (CNNs) and vision transformers (ViTs). The model classifies NSCLC into three categories: normal, adenocarcinoma, and squamous cell carcinoma. CNNs are employed to capture local features, while ViTs are used to understand long-range relationships between image patches. We trained and validated our model on the LC25000 dataset, a benchmark dataset for NSCLC histopathology image classification. Our proposed model demonstrated superior performance, achieving an accuracy of 0.988, an F-1 score of 0.980, specificity of 0.991, recall of 0.982, and precision of 0.980, outperforming existing state-of-the-art methods. Additionally, our model achieved a low inference time of 1.816 ms, highlighting its potential for real-world applications where both accuracy and speed are critical. Our code is now publicly available at https://github.com/ImranNust/LungCancerDetection to facilitate further research and validation of our findings.

Estimating coastal ground deformation caused by ocean tidal loading (OTL) through time-series interferometric synthetic aperture radar (InSAR) techniques is challenging but achievable. Based on the two-dimensional wavelet decomposition method to extract spatial long-wavelength interferometric phases, the OTL deformation, residual orbital error, and atmospheric delay error hidden in long-wavelength phases can be isolated by independent component analysis. The ground OTL deformation in time-series InSAR phases is then accurately estimated using tidal harmonic functions. Experiments conducted with Sentinel-1 measurements in coastal areas reveal the following findings. Leveraging the blind source separation method allows for the effective extraction of high spatial resolution and large-scale ground OTL deformation correlation signals from the time-series InSAR phase. The differences between parameter values of main constituents estimated by the modeled InSAR ground OTL deformation and the FES2014b global tidal model range from 1.1% to 5.8%. Additionally, the difference in their predicted OTL deformation values is within the range of 0.9–4.9 mm. Compared to the FES2014b model correction method, the proposed approach removes OTL deformation from the time-series InSAR and results in a 26.1% reduction in the residual standard deviation value of ground deformation, which leads to a significant enhancement in the accuracy of ground deformation acquired for coastal areas. This work provides a novel approach to the theoretical research on OTL deformations, contributing to the refinement of InSAR ground deformation monitoring accuracy in coastal areas.

Infrared small target detection has been widely used in military and civil fields. However, due to the insufficient feature integration capabilities of existing methods, effectively separating strong background clutter and targets in complex scenes remains difficult. To address this issue, we propose a two-stage feature complementary improved tensor low-rank sparse decomposition (TLRSD) method. The detection process is divided into two stages: tensor initialization and tensor decomposition, effectively integrating local and nonlocal features. In the tensor initialization stage, inspired by the local saliency of the target and the local consistency of the background, we design a three-layer directional filtering (TLDF) operator for preliminary clutter suppression and target enhancement. Then, to promote the complementary advantages of local and nonlocal features, we refer to the TLDF and the original image to provide a targeted initialization strategy for the TLRSD model. In the tensor decomposition stage, we develop a robust partial sum of the tubal nuclear norm as a nonconvex approximation of tensor rank, which can adaptively adjust the singular value distribution, thus adapting to diversity scenes. Meanwhile, we finely adjust the balance between low-rank and sparse components in the model-solving process through a nonlinear reweighting strategy, accelerating the optimization convergence speed and improving the model's background recovery ability. Extensive experiments on five practical datasets demonstrate that the proposed method is more effective and robust compared to ten state-of-the-art approaches.

To understand the mechanism of the effect of the cooling rate of the non-isothermal crystallization process on the crystallization morphology of polypropylene (PP) insulation and the relationship between the crystallization morphology and DC breakdown performance, this paper researched the crystallization morphology, physicochemical property, polarization current and DC breakdown strength of PP specimens with five cooling rates. As the cooling rate increases, due to the reduction of the diameter and perfection degree of PP spherulites, the density of amorphous regions increases, deep traps are introduced, and the DC breakdown strength is enhanced. The DC breakdown strength of the PP specimen with a cooling rate of $170~^{\circ }$ C/min is about 2.4 times higher than that of the PP specimen with a cooling rate of $10~^{\circ }$ C/min. Therefore, by controlling the cooling rates, the crystallization regulation of PP can be realized. That is, by affecting the development process of PP nucleation and spherulites development, resulting in changes in the crystallization morphology, which in turn changes the properties of PP such as glass transition temperature, melting range, and breaking elongation. Finally, the regulation of carrier transport in PP, i.e., insulation properties, is realized, which provides a basis for the improvement of PP insulating materials.