Search

Since control instructions are the fundamental component of thermal power generation, the quality and effectiveness of implementation directly affect the efficiency of the energy system. In order to improve the efficiency of internal combustion engine control, an optimization method of internal combustion engine control based on enhanced clustering algorithm and swarm intelligence optimization algorithm is proposed. The process simplifies the main structure of the gas turbine, divides the combustion engine model into multi-input single-output systems, and introduces the artificial bee colony algorithm to optimize the parameters. A new nectar search formula is constructed by using the global optimal nectar, and the control parameters are calculated by fuzzy logic clustering. The experimental results showed that the modeling error of the load model of internal combustion engine was in the range of −0.47 MW $\sim ~0.51$ MW. When the training iteration speed was tested, the loss value of the research method dropped rapidly in the first 10 iterations. When analyzing the change of the control quantity during load change, if the exhaust flow rate was taken as the control quantity, the control results of the research method was always kept within 5lbm/s error. It demonstrates that this research method can effectively improve the running quality of the internal combustion engine and has a good running efficiency. The research can provide certain technical reference for gas turbine control in thermal power generation.

With the rapid development of synthetic aperture radar (SAR) technology, SAR remote sensing has a wide range of applications in fields, such as marine surveillance and sea rescue. Currently, the SAR ship detection model based on deep learning suffers from the problems of low detection in real time and low detection accuracy. In order to solve the abovementioned problems, this article proposes a lightweight SAR ship detection network (EGTB-Net) based on transformer and feature enhancement. First, we design a novel Ghost-ECA model as the backbone network of EGTB-Net, which reduces the number of parameters of the model and enhances the ability to identify key feature information at the same time. Then, we incorporate the transformer block in the backbone network to capture long-range dependencies, enrich contextual information, and improve the network's ability to capture different types of local information. Finally, we adopt a new SIoU loss function, which is used to solve the direction problem of mismatch between the real frame and the predicted frame and improve the network's ability to localize ship targets. The experimental results on the SAR-ship-dataset show that the mean average precision of the method is 94.83%, the detection speed is 61 frames per second, and the model size is only 5.94 M, while the model has excellent anti-interference ability.

Emotion-Cause Pair Extraction (ECPE) aims to jointly extract emotion clauses and the corresponding cause clauses from a document, which is important for user evaluation or public opinion analysis. Existing research addresses the ECPE task through a two-step or an end-to-end approach. Although previous work shows promising performances, they suffer from two limitations: 1) they fail to take full advantage of emotion type information, which has advantages for modelling the dependencies between emotion and cause clauses from a semantic perspective; 2) they ignored the interaction between local and global information, which is important for ECPE. To address these issues, we propose an ECPE Pair Generator (ECPE-PG), with a Clause-Encoder layer, a Pre-Output layer and an Information Interaction-based Pair Generation (IIPG) Module embedded. This model first encodes clauses into vector representations through the Clause-Encoder layer and then preforms emotion clause extraction (EE), cause clause extraction (CE) and emotion type extraction (ETE), respectively, through the Pre-Output layer, on the basis of which the IIPG module analyzes the complex emotional logic of relationships between clauses and estimates the candidate pairs based on the interaction of global and local information. It should be noted that emotion type information is regarded as a crucial indication in the IIPG module to assist the identification of emotion-cause pairs. Experimental results show that our method outperforms the state-of-the-art methods on benchmark datasets.

Creep is a long-term property of concrete that poses a significant challenge to structural engineering. The research content of this paper concerns how to accurately predict the creep behavior of concrete. The overall purpose is to develop a robust predictive model based on the Northwestern University (NU) database that overcomes the challenges encountered when handling tabular creep data, including the lack of spatial correlations between features and the presence of noise. To achieve this goal, a novel deep learning methodology that leverages convolutional neural networks (CNNs) is introduced. First, the data preprocessing phase involves feature selection and K-nearest neighbor-based data filling to enhance the quality of the given data. A soft-sorted 1D CNN is introduced to transform tabular data into multichannel images, thereby taking advantage of spatial locality and feature sorting. Additionally, this model incorporates noise reduction mechanisms through residual network encoders, achieving enhanced data clarity and model accuracy. The mean absolute error (MAE) and root mean square deviation (RMSE) of the model increase from 7.48 and 10.76 to 5.83 and 8.88, respectively. The major findings of this research indicate a significant predictive accuracy improvement over the existing methods. The model effectively captures the complex relationships between concrete creep and a range of influencing factors, enabling precise and quantitative predictions to be obtained. These findings offer deep insights for the field of civil engineering, enabling enhanced structural design, safety, and performance prediction in concrete applications. This paper contributes to the ongoing exploration of advanced deep learning techniques for addressing complex challenges in the domains of concrete mechanics and structural engineering.

The information of cropland is obtained efficiently and accurately as the basis for achieving precision agriculture (PA). As the boundary between cropland and other types of land in remote sensing images with different resolutions is fuzzy, the characteristics of cropland are easily confused when extracting cropland, resulting in inaccurate identification and extraction of cropland under large and complex backgrounds and rough localization of marginal areas. We proposed a two-path multiscale attention self-supervised network with the perception of four directions in pixel space, called the multiscale bilateral spatial direction-aware network (MBSDANet), to solve these problems and improve the model’s ability to extract cropland in small samples. One path extracts attentional feature maps by spatial directions to preserve detailed direction-aware information and generate high-resolution features; the other path obtains local-to-global information through pyramid pooling and attention awareness to capture dense multiscale cropland features to separate targets in complex contexts. The features of the two branches at different levels are fused by weighting the multi-aware information. Triangular self-supervised and boundary-aware losses are used to achieve fine segmentation and extraction of cropland in small samples. We tested the extraction method on cropland in Denmark and the Hebei Province of China, demonstrating its effectiveness and generalization. Compared to other neural network models, MBSDANet achieves better accuracy with a precision of 0.9481, an IoU of 0.8937, and an F1 score of 0.9438.

Unmanned combat aerial vehicles (UCAVs) are preferred for regional electronic reconnaissance due to their versatility and stealth. This paper proposes a deep reinforcement learning (DRL) method to enable UCAVs to complete regional multi-target electronic reconnaissance (MER) tasks with continuous autonomous maneuvers. Distinguishing from traditional heuristic search algorithms, we first derive the objective function of MER and elucidate sufficient conditions to improve the success rate of reconnaissance recognition. Then, using the original cognitive electronic warfare framework, a three-dimensional MER simulator named Scouer-N is created to satisfy the requirements of dynamic environment training for DRL-based agents. To enable the processing of sequential situation awareness, a generative network is constructed by introducing a partially observable Markov decision process (POMDP) model, which assists the UCAV in filtering the observations from the sensor and predicting the actual states. Finally, we propose a priority-driven state reward shaping method that provides normalized state representation and dense rewards to the agent during training to improve the agent’s behavioral knowledge for MER. The experimental results demonstrate a considerable improvement in the task success rate of the trained UCAV relative to the benchmark, proving the efficacy of our approach in helping agents learn the optimal reconnaissance strategy from the potential state space.

To solve the global path planning problem of the ship in the static and dynamic environment, we propose an improved ant colony algorithm to plan the ship’s navigation path. We use the artificial potential field method to compute the force direction of the ship at the initial iteration stage. The attraction potential field function is modified to improve the iteration efficiency of the hybrid ant colony algorithm. We design the pseudo-random state transition rule and improve the convergence of the hybrid algorithm by strengthening the selection of good paths. When updating the pheromone, we consider the path’s length, safety, and smoothness to plan a safer navigation path. The simulation results show that the improved ant colony algorithm has a faster convergence speed than the original ant colony algorithm. The optimal solution quality is higher, which can realize global ship path planning in static and dynamic environments.

With the rapid development of information and sensor technology, the data-driven remaining useful lifetime (RUL) prediction methods have been acquired a successful development. Nowadays, the data-driven RUL methods are focused on estimating the RUL value. However, it is more important to quantify the uncertainty associated with the RUL value. This is because increasingly complex industrial systems would arise various sources of uncertainty. This article proposes a novel distributional RUL prediction method, which aims at quantifying the RUL uncertainty by identifying the confidence interval with the cumulative distribution function (CDF). The proposed learning method has been built based on quantile regression and implemented from a distributional perspective under the deep neural network framework. The results of the run-to-failure degradation experiments of rolling bearing demonstrate the effectiveness and good performance of the proposed method compared to other state-of-the-art methods. The visualization results obtained by $t{\text{-}}SNE$ technology have been investigated to further verify the effectiveness and generalization ability of the proposed method.

Detecting human actions in videos is crucial for human-computer interaction, intelligent security, etc. Temporal and spatial information is very important for action localization, as how we utilize them determines the performance of detection. In order to address the problems of poor real-time and complex models of human action localization, a real-time detection architecture with two branches is presented in this paper, which predicts bounding boxes and action probabilities directly from video clips in one evaluation. We adopt the similar guidelines of YOLO (You Only Look Once) for bounding box regression and classification. The difference is that our model completes two tasks through two branches, one branch is used to extract spatial information from a key frame for bounding box regression, and another branch extracts spatiotemporal information from consecutive video frames for classification. In this paper, our model provides 65 frames-per-second on 16-frames input clips. Results on UCF-Sports and JHMDB-21 datasets present comparable accuracy to State-of-the-Art approaches. We obtain F-mAP of 90.39% and 76.29% with gains of 2.19% and 1.89%, respectively. On the JHMDB-21dataset, V-mAP reaches 90.7%, 90.0% and 69.5% with gains of 2.9%, 4.3% and 11.4% at IoU thresholds of 0.2, 0.5, and 0.75.

Since the fingerprints of numerous crucial biologic materials can be identified by terahertz (THz) spectroscopy, THz sensing have become an important approach of biological and medical detections. Particularly, benefit from the excellent capability of metasurface, strong interactions between the metasurface and THz waves can be realized, thus the THz sensing with high sensitivity becomes reality. However, the common configuration of metasurface-based THz sensor is composed of metallic subwavelength structure. Due to the inherent resistive loss of metal, it is still a great challenge to further enhance the quality factor (Q-factor) of resonance and sensitivity of THz sensor. In this work, we designed an all-dielectric metasurface with high Q-factor for highly sensitive THz sensors. The metasurface is a windmill-like structure consisting of four cuboids, and every adjacent two cuboids are arranged alternately vertically and horizontally. The transmission spectrum of metasurface exhibits four polarization-independent and strong resonance peaks with high Q-factor in 0.1–2.5 THz, and all of them show high sensitivity related to ambient refractive index. The transmitted structure and polarization-independent resonances can relief the difficulty of measurement. We believe these studies will lay the theoretical and technical foundation for the design of high-sensitivity terahertz sensing.

Reverse conduction loss of GaN high electron mobility transistors (HEMTs) in very high frequency (VHF) converters is non-neglectable due to the absence of body diode. To reduce the reverse conduction time, this paper proposes a linear equivalent model for class $\Phi _{2}$ inverter to derive optimal duty cycles for different working conditions. The proposed model is derived from frequency-domain perspective, which simplifies the derivation process and provides an intuitive physical insight for class $\Phi _{2}$ inverter. Firstly, the power switch is modeled as a current source according to time-domain expressions of inductor current, which simplifies the circuit to a linear network. Based on quantitative analysis of the current source spectrum characteristics, the linear network response to the current excitation is derived, which forms the linear equivalent model. Then, drain voltage of the main switch under different working conditions is obtained. With the zero-crossing point of the drain voltage, numerical solutions of the optimal duty cycles are calculated. Finally, the optimal duty cycles are implemented with a high-resolution duty cycle generation circuit in a 27.12MHz prototype, which achieves not only a peak efficiency of 93.6% at full load, but also higher efficiency over the whole load and input voltage range compared to conventional fixed duty cycle.

Based on L-shaped array (LSA), a variety of algorithms for direction of arrival (DOA) estimation have been developed in recent years. However, conventional methods are sensitive to the mutual coupling which may cause the performance of the DOA estimation to degrade dramatically or even fail. In order to solve this problem, a new algorithm for DOA estimation is proposed in this paper. Different from the existing algorithms, the DOA estimation of LSA with mutual coupling is achieved from the perspective of sparse Bayesian learning. A hierarchical form of the Student $t$ prior is used to enhance the sparsity of source signal and achieve excellent performance for angle estimation. In the meanwhile, the mutual coupling effect is compensated blindly by sacrificing some sensors as auxiliary components and the hyperparameters and parameters updates are obtained by the expectation-maximization method. Simulation results verify that the performance of our method is superior than that of the state-of-art methods, particularly in the scenario of highly correlated and coherent signals.

Studies have shown that Sentinel-2 images have advantages over Landsat images in impervious surface area (ISA) extraction. The performance of index-based methods can be affected by different binary methods and subject to seasonal variation. This study marks the first attempt to assess the performance of different spectral indices for ISA extraction using multiseasonal Sentinel-2 images. Specifically, five indices (i.e., the biophysical composition index calculated using the Gram–Schmidt orthogonalization method, biophysical composition index calculated using a principal component-based Procrustes analysis, Normalized Built-up Area Index (NBAI), combinational build-up index, and perpendicular impervious surface index (PISI)) and three impervious surface binary methods (i.e., Otsu's method, manual method, and Iterative Self-Organizing Data Analysis Technique Algorithm (ISODATA) classification method) were tested on multi-seasonal Sentinel-2 images in the main urban area of Wuhan, China. Results showed that PISI combined with the ISODATA classification method achieved the highest accuracy with 92.64% OA, and 0.8410 Kappa coefficient, and NBAI combined with Otsu's method achieved the lowest accuracy with 35.37% OA, and 0.013 Kappa coefficient. Regarding the seasonal sensitivity, PISI is relatively more stable than the other methods. The superior performance of PISI is largely due to its capability in separating ISA from soil and vegetation. In addition, summer is the best season to map ISA from Sentinel-2 images when the impervious surface is generally less confused with bare soil. This study serves as a reference for the selection of spectral indices for ISA extraction from Sentinel-2 images in relation to binary methods and seasonal effects.

The burgeoning demand for enhanced power-system safety means that there is increased attention on technologies for the detection and location of cable faults. The double-terminal fault location method, which senses a fault current using a current transformer (CT), is the most effective and widely applied method in industrial practice. However, CT sensors may suffer from magnetic saturation when exposed to a high fault current, resulting in significant location errors, and their sensitivity cannot be adjusted. To address this problem, this paper describes an improved method for the acquisition of fault signal and the identification of fault-wave arrival time in power cables. First, a novel noncontact sensor (NCS) is developed for the detection of electric fields, and its sensitivity-adjustment curves are obtained by theoretical calculation and simulation. Then, based on feedback variational mode decomposition (FVMD) and the Wigner–Ville distribution (WVD), an improved method (the FVMD + WVD method) for the identification of fault-wave arrival time is devised, and its efficacy and accuracy are verified by simulations in Power Systems Computer-Aided Design/Electromagnetic Transients including DC (PSCAD/EMTDC) software. Finally, the fault-wave arrival-time detection performance of the NCS is examined in a series of on-site DC and AC experiments based on damped AC technology. The results show that the NCS exhibits feasible and effective performance, and that its response can be appropriately tuned based on its sensitivity-adjustment curves. The FVMD + WVD method is more accurate than several other current methods, as its error is only 0.48%, and thus this method enables the practical location of cable faults. In sum, these findings demonstrate that the NCS and the FVMD + WVD method comprise an improved system for the detection and location of cable faults, which will enhance cable safety in power systems.

A resonator is one of the important components in electronic systems. It is often used in an electronic system to give an accurate time reference. But it can also be used as a highly sensitive physical parameter sensor especially with the rapid development of fabrication technology for Micro Electro Mechanic System (MEMS) sensors. One of the parameters used to describe the performance of a resonator is its quality factor, also called Q-factor for short. Knowing the Q-factor of a resonator can make sure the performance of the system in which the resonator will be used. For applications in which a resonator is used as a sensing device, it is crucial for the Q-factor to be surveyed and measured in real time. In this paper, a review of a resonator’s quality factor (Q-factor) measurement is given. Three major approaches can be identified from the literature: frequency domain, time-frequency domain and time domain. Both advantages and limitations of each approach are presented. Based on the published results, a comparison of the three approaches is conducted. As the time-domain measurement is the only way to present the potential for an in-situ Q-factor measurement but it is relatively less exploited compared with others, a special focus on the time-domain measurement is granted with two time-domain circuit designs.

Serious mutual coupling effect must be taken into account for the miniaturized adaptive anti-jamming antenna arrays in global navigation satellite system (GNSS) receivers. In this paper, the effect of mutual coupling on the performance of GNSS antenna arrays has been comprehensively discussed. The nulling performance of the GNSS antenna array in presence of mutual coupling is efficiently evaluated by the proposed embedded element pattern method, which is further verified by a simulated miniaturized four-element GNSS antenna array in nonadaptive mode. In the space-time adaptive processing (STAP) array, a generalized wideband signal model based on the dispersion phenomenon is introduced to study the effect of mutual coupling on the optimal STAP array response, signal-to-interference-plus-noise ratio (SINR) loss, and the minimum mean square error (MMSE) of the Wiener filter. Employing the STAP normalized least mean square (NLMS) algorithm, the simulation results show that the mutual coupling leads to a decrease in convergence rate, an increase in misadjustment, and a loss of SINR, whether or not the compensation is considered.

The flight height of ambulance aircraft during aviation emergency rescue is low. Low-altitude wind fields and complex terrains may threaten the flight safety of rescue aircraft and thus limit rescue efficiency. The path planning of ambulance aircraft with consideration of low-altitude wind fields and terrains was presented in this study. First, a new wind field interpolation method was proposed to address the shortages of traditional reverse-distance wind field interpolation methods that neglect the effects of mountain slopes on wind vectors. This new interpolation method comprehensively considers the effects of terrain elevation, topographical slope, underlying surface roughness, and distance on wind vectors. The wind field interpolation results could accurately reflect the influence of topographic slopes on wind vectors and were close to practical results. Second, the effects of wind fields of different wind directions on flight paths were analyzed in view of the effects of low-altitude wind fields and restrictions of terrain obstacles. On the basis of the results, a method for correcting corresponding path nodes given aircraft performance constraints was proposed to ensure path safety. The connection matrix was solved by combining the A* algorithm and a 3D meshing technique. The safe path planning method for low-altitude rescue was concluded by combining a path-solving algorithm and the correction method for path nodes. Finally, the topographical features from Wenchuan and Chengdu to Mianyang during the Wenchuan earthquake were simulated with a constructed digital equation model. In the simulation environment, the safety of the path plan of the ambulance aircraft was verified under the influence of low-altitude wind fields.

A Spiking Neural Network (SNN) can be trained indirectly by first training an Artificial Neural Network (ANN) with the conventional backpropagation algorithm, then converting it into an equivalent SNN. To reduce the computational cost of the resulting SNN as measured by the number of spikes, we present Logarithmic Temporal Coding (LTC), where the number of spikes used to encode an activation grows logarithmically with the activation value; and the accompanying Exponentiate-and-Fire (EF) neuron model, which only involves efficient bit-shift and addition operations. Moreover, we improve the training process of ANN to compensate for approximation errors due to LTC. Experimental results indicate that the resulting SNN achieves competitive performance in terms of classification accuracy at significantly lower computational cost than related work.

Vehicular Ad Hoc Network (VANET) is an important foundation of intelligent transportation system and is widely used in traffic management, automatic driving, and road optimization. With the gradual popularization and further development of VANET, a large amount of VANET data has been produced. However, it poses huge challenges to the security and privacy when using VANET data provides services for users. In this paper, combining the technologies of blockchain and ciphertext-based attribute encryption (CP-ABE), we propose a fine-grained access control scheme for VANET data based on blockchain (FADB). In FADB, we employ the blockchain to replace the third-party service providers for user identity management and data storage. And different VANET data access rights can be established according to user attribute. By improving the CP-ABE, the lightweight VANET devices can outsource complex encryption and decryption operations to powerful RSUs and further improve the efficiency of data access. Final, we carry out a series of simulation tests and security analysis, proving that the FADB can provide effective data security and low performance overhead.

Wireless power transfer (WPT) technology has become a popular way to solve the battery charging problem of mobile devices. However, when designing a new WPT system, transfer power, power transfer efficiency (PTE), electromagnetic environment (EME) and other factors usually need to be considered. Firstly, for commonly used unshielded disk resonator, this paper establishes the theoretical analysis models of EME and transfer performance. Then, the multi-scheme evaluation method of disk resonators considering the WPT system transfer performance and the EME indicator is proposed. In addition, we also evaluate multi-resonator schemes made of the existing coil models in the laboratory using the above method. After considering the transfer power, PTE, coil overcurrent and EME, the optimal resonator scheme is obtained. Finally, the above evaluation method is verified by electromagnetic simulator COMSOL and experiment. The evaluation method proposed serves therefore for researchers and developers to significantly improve the design efficiency.

Reverberation chambers (RCs) have become a popular testing facility in antenna efficiency measurements. Unlike the standard reference antenna method, the non-reference antenna methods have been proposed to measure the antenna efficiency in the absence of a reference antenna. However, the statistical distributions of the measured (estimated) total antenna efficiencies using the non-reference antenna methods have not been derived before, making it difficult to perform a rigorous statistical analysis of the antenna efficiency measurement. In this paper, the distributions of the measured total antenna efficiencies using two non-reference antenna methods in an RC are derived, based on which the statistics (e.g., expectation and variance) of the measured total antenna efficiencies are also derived. It is shown that the original non-reference antenna methods (estimators) are only asymptotically unbiased and are biased with limited samples, thus corresponding unbiased estimators are proposed and analyzed. The derived analytical expressions of the statistics of the total antenna efficiency estimators are verified by simulations and measurements in an RC. The performances of different non-reference antenna methods are discussed.

Charged particles are widely used in the fields of defogging, haze removal and artificial rainfall. A novel transport strategy of charged particles based on the vortex rings is presented in this paper, to expand the effective action region and enhance the effect of charged particles. Transport distance and transport velocity (TV) of charged particles are the crucial two evaluation indexes. For the vortex rings generated by a piston-cylinder device, the transport velocity is closely related to Reynolds number and formation time. Based on the k-ω turbulence model, this paper verifies the relationship between vortex ring circulation, transport velocity and Formation Number in air by finite element methods, and studies the influence on the formation of vortex rings of an insulating bell mouth placed at the outlet of cylinder. The simulation results show that the Formation Number of vortex rings in the air is 4.0 ~ 5.0, and the growth rate of transport velocity of vortex ring slows down when the formation time is larger than Formation Number. The study also indicates that the bell mouth has little effect on the velocity of the vortex ring, when the bell mouth angle is greater than 45° and the length of generatrix is between 0.2 m and 1.0 m. Therefore, the research results can provide a reference for the design and optimization of a vortex ring generator for transporting charged particles to a certain target area.

How to apply data analysis algorithms to China's primary hospitals is still a problem that needs to be solved. In order to effectively explore the pathogenesis of myocardial infarction disease, this study collected a large amount of real data as a basis for data analysis through data survey, improved traditional cluster analysis and data mining methods, and proposed effective data mining methods for myocardial infarction. In addition, this study analyzes data sources by implementing clustering analysis algorithms and combines data mining algorithms to provide decision support information for disease research. Finally, this study uses experimental methods of image data analysis combined with data mining methods to record the results. The research shows that the algorithm of this study has certain feasibility and can provide theoretical reference for subsequent related research.

As one of the emerging application areas of the corona discharge technique, rain enhancement induced by atmospheric air ionization has made a great breakthrough. Although natural rainfall modified by artificial ionization was successfully observed via the statistical data of field experiments, deficient efforts have been devoted to the research on the near-field effect of corona discharge on the evolution of water droplets around the discharge device. In this article, we studied the electrostatic effects of corona discharge on the spectrum and density evolution of water droplets in a 96-m3 cloud chamber. An interesting phenomenon was found that two different electrostatic effects appeared when various high voltages were applied on the corona electrode. In detail, the growth of droplets larger than 10 μm was detected once the applied high voltage exceeds the critical value (e.g. -80 kV in our discharge device), and this is attributed to the electrostatic induction effect. By contrast, a narrow droplet size spectrum and reduced density caused by the strong electrostatic precipitation effect were shown below this critical voltage. To figure out the mechanisms underlying these phenomena, the charging status of water droplets with various sizes, forces of charged droplets under various electric field strength, and collision kernels between droplets, are analyzed through theoretical calculations. Results show that the electric field plays a dominant role in the charging process of droplets and the acceleration of collision-coalescence between charged and neutral droplets, which results in the different electrostatic effects on the evolution of droplets. This article may provide some enlightenment to fog elimination in industrial application and cloud-seeding in atmospheric water resources comprehensive utilization.

The equivalent resistance of the battery will change during charging in practical applications, so it is difficult to design a wireless power transfer (WPT) system with accurate constant current (CC) and constant voltage (CV) output characteristics. In addition, the WPT system efficiency is also affected by the change of equivalent resistance. Therefore, this paper studies the charging control and efficiency optimization strategy for the WPT system with dynamic loads. The WPT system with a structure of double-sided LCC compensation topology and containing controllable rectifier circuit at the secondary side is established, and the working process of full-bridge controllable rectifier circuit is analyzed. Then the output characteristics of the WPT system are analyzed. The expressions of the optimal transmission efficiency and the optimal equivalent impedance of the full-bridge rectifier are derived. The relationships between the equivalent impedance of the capacitor-filtered full-bridge controllable rectifier circuit and the phase shift angle, the charging current and the phase shift angle are also obtained. In this paper, the CC and CV charging strategies based on phase shift control of controllable rectifier circuit and efficiency optimization strategy based on dynamic equivalent impedance matching are proposed, which can improve the WPT system transmission efficiency while realizing CC and CV charging. Finally, the effectiveness of the proposed control strategy is verified by simulation and experiment. The WPT system realizes CC charging mode and CV charging mode in turn when the load resistance changes from $5~\Omega $ to $30~\Omega $ , and the overall efficiency is up to 90.71% at the transmission distance of 15cm.

Syslog parsing is of vital importance for the detection, diagnosis and prediction of network device failures in a datacenter. A common approach to syslog parsing is to extract templates from historical syslogs, after which syslogs are matched to these templates. To address the problems in the existing syslog parsing techniques, we propose a novel framework, Craftsman, which identifies frequent combinations of (syslog) words and then applies them as templates. Craftsman empirically extracts templates accurately, is extremely efficient in template matching, and naturally supports incremental learning. To compare the performance of Craftsman and three other template learning techniques designed for network devices, we experiment them on two-years' worth of syslogs collected from network devices deployed across 10+ datacenters of a tier-one service provider. The experiments demonstrate that Craftsman achieves a close-to-one accuracy (as measured by rand index), and improves the computational efficiency by 6.88 to 10.25 times in template matching, and by 730 to 6847 times in syslog parsing. It also improves the accuracy (as measured by F1 measure) of failure prediction by 13.07% to 188%. In addition, we demonstrate Craftsman's superior generality by comparing it with three widely-applied log parsing methods over five large log datasets collected from servers, distributed systems and applications.

Heavy metal pollution in soil has severe impact on the human health and global environment. There is an urgent need for cost-effective devices capable of recognizing and detecting heavy metallic matters in soils. A broadband photoacoustic spectrometric (PAS) system was established for the detection of heavy metal contaminants in soil. The heavy metal element lead (Pb) was selected as a preferred detection target. The near infrared photoacoustic spectra of contaminated soil samples with various concentrations of Pb were collected and the spectroscopic relationship between the soil absorption peaks and Pb concentrations was analyzed. Four advanced spectral preprocessing methods were explored to improve the robustness of the prediction model. Based on the maximal correlation coefficient and minimal root mean square error criteria of prediction, a two-layer feed-forward network with a continuum removing preprocessing method with a correlation coefficient as 0.96 was tested as the most appropriate method for Pb concentration prediction. This fact verifies that the broadband PAS evaluation methodology, as a nondestructive testing method, can be potentially used as an alternative quantification detection method of heavy metal contaminants in soil without the involvement of complicated sample pretreatment.

The ground-airborne frequency-domain electromagnetic (GAFDEM) survey is a fast method to detect the deep underground structure in areas with difficult access. However, the suspended coil sensor in the survey system experiences attitude changes, which inevitably yield attitude error on the measured magnetic field and result in false apparent resistivity imaging. Due to the separated system structure, the attitude error in GAFDEM survey is significantly different from that in traditional airborne helicopter frequency-domain electromagnetic survey. In order to improve the detection accuracy, a modified iterative algorithm is proposed to suppress the attitude error for apparent resistivity imaging. Both simulation and practical experiment are conducted to verify the usefulness of the correction method. The simulation results indicate that the correction procedure is important and effective to reduce the attitude error in layered earth model. The field experiment results show that the corrected apparent resistivity imaging section is more consistent with the high-density resistivity topography and more reasonable than the uncorrected apparent resistivity imaging result obtained by conventional method. Consequently, the proposed suppression approach can reduce the attitude error to obtain more accurate apparent resistivity imaging results in GAFDEM survey, which may effectively promote the developments and applications of GAFDEM method and systems.

Most existing studies and public datasets for handwritten Chinese text recognition are based on the regular documents with clean and blank background, lacking research reports for handwritten text recognition on challenging areas such as educational documents and financial bills. In this paper, we focus on examination paper text recognition and construct a challenging dataset named examination paper text (SCUT-EPT) dataset, which contains 50 000 text line images (40 000 for training and 10 000 for testing) selected from the examination papers of 2 986 volunteers. The proposed SCUT-EPT dataset presents numerous novel challenges, including character erasure, text line supplement, character/phrase switching, noised background, nonuniform word size, and unbalanced text length. In our experiments, the current advanced text recognition methods, such as convolutional recurrent neural network (CRNN) exhibits poor performance on the proposed SCUT-EPT dataset, proving the challenge and significance of the dataset. Nevertheless, through visualizing and error analysis, we observe that humans can avoid vast majority of the error predictions, which reveal the limitations and drawbacks of the current methods for handwritten Chinese text recognition (HCTR). Finally, three popular sequence transcription methods, connectionist temporal classification (CTC), attention mechanism, and cascaded attention-CTC are investigated for HCTR problem. It is interesting to observe that although the attention mechanism has been proved to be very effective in English scene text recognition, its performance is far inferior to the CTC method in the case of HCTR with large-scale character set.

Inspired from the idea that the contexts in which a word occurs are of different significance, this paper proposes a novel method, called word representation with Salient Features (SaFe), to represent words using salient features selected from the context words. The SaFe method employs the point-wise mutual information (PMI) method with scaled context window to measure word association between a target word and its context. Then, contexts having word associations will be selected as salient features, where the number of salient features for a given word is decided by the ratio between the number of unique contexts and the total counts of occurrences in the whole corpus. The SaFe method can be used with the positive PMI matrix (PPMI), with each row representing a word, hence the name SaFe-PPMI. Moreover, the SaFe-PPMI model can be further decomposed by using the truncated singular vector decomposition technique to obtain dense vectors. In addition to efficient computation, the new models can achieve remarkable improvements in seven semantic relatedness tasks, and they show superior performance when compared with the state-of-the-art models.

In recent years, intelligent fault diagnosis technology with the deep learning algorithm has been widely used in the manufacturing industry for substituting time-consuming human analysis method to enhance the efficiency of fault diagnosis. The rolling bearing as the connection between the rotor and support is the crucial component in rotating equipment. However, the working condition of the rolling bearing is under changing with complex operation demand, which will significantly degrade the performance of the intelligent fault diagnosis method. In this paper, a new deep transfer model based on Wasserstein distance guided multi-adversarial networks (WDMAN) is proposed to address this problem. The WDMAN model exploits complex feature space structures to enable the transfer of different data distributions based on multiple domain critic networks. The essence of our method is learning the shared feature representation by minimizing the Wasserstein distance between the source domain and target domain distribution in an adversarial training way. The experiment results demonstrate that our model outperforms the state-of-the-art methods on rolling bearing fault diagnosis under different working conditions. The t-distributed stochastic neighbor embedding (t-SNE) technology is used to visualize the learned domain invariant feature and investigate the transferability behind the great performance of our proposed model.

The China Fusion Engineering Test Reactor (CFETR) device is a milestone in the roadmap for the utilizing of fusion energy in China, whose final goal of plasma current is up to 10.02 MA. As the key system to generate a high enough voltage (20 kV) required for gas breakdown and plasma initiation, the Switch Network Unit (SNU) should have the ability to transfer 80 kA direct current by inserting resistor (0.25Ω ) to the superconducting coil circuit in 12 milliseconds. SNU topologies applied in other fusion devices are mostly based on the utilization of oscillation zero-point circuit, or combination of mechanical switch and semiconductor switch and so on, which bring stricter requirements for the components in the circuit and raise the cost due to the series and parallel connection of multiple devices if used in 80 kA/ 20 kV application. Based on the combination of solid-state switch (SSS) and auxiliary capacitor, this paper proposes a novel hybrid switch scheme. The SSS consists of insulated gate bipolar transistors (IGBT) group and thyristors group connected in series. The combination of added auxiliary capacitor and SSS can significantly reduce the number of series-connected IGBTs. This paper introduces the operation principle, design methods and simulation analysis of proposed scheme in detail. Coinciding with the analysis, the simulation results of SNU demonstrates the feasibility and reliability of the proposed scheme, which suggests its promising applications in fast switches with tens of kilovolts and hundreds of kilo-amperes.

To improve the stability of receiving power in the wireless power transmission system with unfixed receiving position, a flexible power control strategy without the measurement in secondary side and the wireless communication between primary and secondary sides is proposed. The circuit model of a multiphase resonant inverter (MPRI) with cascaded intercell transformers (ICTs) is developed. The output voltage control strategy of the MPRI is introduced based on the iterative summation method for trigonometric functions. The circuit model of the system is derived to deduce the expressions for the key parameters. Considering the cascaded ICTs and the inductor-capacitor-capacitor (LCC) compensation circuit in the primary side, the design method of the compensation parameters is proposed based on the circuit model of the MPRI. After synthesizing the online monitoring, parameter pre-calculation, prediction of several conditions, and online learning, the flexible power control strategy is put forward for the wireless power transmission system with unfixed receiver position. Without the auxiliary positioning methods, the measurement in the secondary side, or the wireless communication between primary and secondary sides, the recognition of the load access and exit, as well as the power stabilization control after access, are realized. Finally, the theoretical analyses and the control strategy are validated by experiments.

Recent work has shown that self-attention modules improve the performance of convolutional neural networks (CNNs), in which global operations are conventionally used to generate descriptors from feature context for attention calculation and characteristics recalibration. However, the performance gain is compromised due to sharing the same descriptor for different feature context. In this paper, we propose Pyramid Attention Mechanism (PAM) that incorporates contextual reasoning into self-attention module for enhancing the discriminative ability of descriptors. PAM is lightweight yet efficient and can be integrated with most self-attention modules. It consists of two operators: aggregation and distribution, which are used for assembling and synthesizing contextual information at different levels. Extensive experiments on different benchmarks (including CIFAR-100, ImageNet-1K, MS COCO, and VOC 2007) indicate that PAM can produce competitive performance gains. In classification tasks, by plugging PAM into self-attention modules, at most 2.18% accuracy improvement over various network structures can be obtained.

Broadband absorber in the terahertz region (0.1-10 THz) has attracted considerable attentions due to its important applications in detecting, imaging, and electromagnetic stealth. Recently, terahertz absorber with broadband features has been widely investigated, however, the achievement of ultrabroad bandwidth is still challenging due to the limitations of complex structural design and fabrication processes. In this paper, an ultrabroadband terahertz absorber covering the entire terahertz regime (0.1-10 THz) based on the heavily doped silicon has been designed and fabricated, which is composed of double-layer binary gratings filled with the SU-8 photoresist. Antireflection techniques (SU-8 layer) were utilized to further promote the performance of the terahertz absorber at high frequencies through matching the impedance between free space and doped-silicon substrate. The measured absorption exceeding 87% within the frequency range of 0.3-10 THz has verified the proposed approach in designing the ultrabroadband terahertz absorber. Furthermore, the designed absorber remains high performance in the case of wide-angle incidence even up to 60°. Benefiting from the simple structure, the absorber is easy to be fabricated by common optical lithography. We believe that the results of this paper could broaden the application areas of terahertz absorbers.

To achieve high-performance motion control accuracy for complex contouring tasks even under high-speed and large-curvature, a generalized global task coordinate frame (GGTCF)-based learning adaptive robust control (LARC) strategy is synthesized for the industrial biaxial mechatronic stage systems. Specifically, through the usually known desired trajectory information of each axis, GGTCF is firstly proposed and globally designed based on a synthesized equivalent shape function of the desired contour. Different from conventional GTCF just suitable for simple contours with explicit shape function, the proposed GGTCF can guarantee multi-axis coordination even under complex contouring tasks with highspeed and large-curvature characteristics. After transforming the system dynamics of an industrial biaxial mechatronic stage system into the proposed GGTCF, a LARC contouring controller is constructed for the strongly coupled nonlinearities in GGTCF to achieve great contouring motion performance. In LARC, adaptive compensation term and robust feedback term are, respectively, designed to deal with parametric variation and uncertain disturbances, whereas iterative learning term is designed to further suppress the unmodeled repetitive contouring errors. Comparative experiments under various complex contours are conducted on an industrial linear-motor-driven biaxial motion stage. The experimental results consistently demonstrate that the proposed GGTCF can successfully implement various complex contouring tasks. Furthermore, in comparison with the conventional cross-coupled control and adaptive robust control control algorithm, the proposed GGTCF-LARC simultaneously possesses strong coordination mechanism and excellent transient/steady-state contouring control performance, which actually provides a novel control framework with high-performance nature for contouring motion research field.

Bidirectional wireless power transmission (BD-WPT) systems are applicable to occasions that need bidirectional energy flows, such as vehicle-to-grid technology. The traditional BD-WPT systems work at a fixed frequency and the optimal transmission efficiency of the systems is fixed. Combining the frequency characteristics of AC resistance, the mathematical model of variable frequency BD-WPT system is built in this paper. Then, the characteristics of transmission efficiency and the characteristics of maximum active power are analyzed based on the model. In order to make the BD-WPT system have higher transmission efficiency, a BD-WPT system with variable structural parameters and the corresponding segmented powerefficiency coordinated control strategy are proposed in this paper. According to different reference power on the secondary side, the BD-WPT system can work in different resonant frequency segments by controlling the switching of corresponding relays in the capacitor arrays on the primary and secondary sides and changing working frequency of the H-bridge converters on the primary and secondary sides. By using the proposed control strategy, the transmission power, and direction, can be tunable segmentally while maintaining high transmission efficiency. The proposed concept is verified to be effective through experiment.

Direction-of-Arrival (DoA) estimation with Coarray can resolve O(N2) sources via only O(N) physical sensor elements. When it comes to model errors, i.e., manual coupling, gain and/or phase errors, and sensor location errors, whether Coarray is still effective for degree-of-freedom (DoF) enhancement has not been proved yet. In addition, calibration of the Coarray is also an open problem, which deserves more attentions. This paper formulates the error models of Coarray at first and then proves that the problem of Coarray calibration can be reformulated into an equivalent one of imperfect uniform linear array correction; meanwhile, DoF enhancement is promised. Based on these models, the problem of Coarray calibration, more specifically, joint DoA and error coefficients estimation, can be solved via the existing state-of-the-art array calibration schemes. An excellent scheme named sparse Bayesian array calibration is adopted as an example to estimate DoA and error coefficients jointly in this paper. Simulation results illustrate that for DoA estimation with imperfect Coarray, the calibrated estimator is effective for DoF enhancement and more robust than uncalibrated subspace method.

Classical Bose-Chaudhuri-Hocquenghem (BCH) codes over finite fields have been studied extensively. One can construct quantum stabilizer codes with good parameters using classical BCH codes. In this paper, our goal is to find such classical BCH codes. We study some properties of suitable cyclotomic cosets at first. These results make it possible to construct nonbinary quantum BCH codes with a given parameter set. Several new families of quantum BCH codes obtained are based on Steane's enlargement of nonbinary Calderbank-Shor-Steane codes and Hermitian construction, respectively. Meanwhile, we have shown that the cyclotomic cosets given in our schemes are optimal to design quantum BCH codes. The defining set contains most consecutive integers. Therefore, corresponding quantum stabilizer codes have better lower bound of minimum distance. Furthermore, it is convenient to compute the dimension of new quantum codes. Compared with the ones available in the literature, the quantum BCH codes in our schemes have good parameters. In particular, we extend known results to more general case.

Safety-critical embedded systems in application domains, such as aerospace, automotive, and industrial automation, must satisfy dual requirements of fault-tolerance and real-time predictability. Control flow checking is an effective technique for improving embedded systems' reliability by online monitoring and checking of software control flow to detect runtime deviations from the control flow graph. However, inserting instrumentation code in every basic block incurs significant execution time overhead, which may cause the program to violate its timing constraints. In this paper, we propose to selectively instrument a subset of code regions that are larger than basic blocks, called super-nodes, in order to make the program partially resilient to control flow errors while keeping the program worst-case execution time (WCET) below a given upper bound. WCET analysis is invoked to estimate the program WCET and to identify the corresponding worst-case execution path (WCEP). An ILP formulation is used to judiciously select a subset of super-nodes on the WCEP for instrumentation, so that the best fault detection coverage is achieved without violating the given WCET upper bound. The optimization is repeated for each identified WCEP until the program WCET satisfies the WCET upper bound. Experimental results demonstrate significant improvements of fault detection coverage compared with related work.

Quantum codes over finite rings have received a great deal of attention in recent years. Compared with quantum codes over finite fields, a notable advantage of quantum codes over finite rings is that they can adapt to quantum physical systems of arbitrary order. Moreover, operations are much easier to execute in finite rings than they are in fields. The modulo m residue class ring ℤm is the most common finite ring. This paper investigates stabilizer codes over ℤm and presents the Gilbert-Varshamov (GV) bound. The GV bound shows that surprisingly good quantum codes exist over ℤm, which makes quantum coding feasible for arbitrary quantum physical system. We also provide an enhanced version of the GV bound for non-degenerate stabilizer codes over ℤm. The enhanced GV bound has an asymptotical form and ensures the existence of asymptotically good stabilizer codes over ℤm. Finally, these two bounds are well suited for computer searching.

Vehicular social networks (VSNs) are moving from infancy to maturity. However, the characteristics of wireless channel availability and the high mobility of vehicles in the VSNs create new challenges for privacy preservation; the corresponding solutions require a combination of technologies. A common theme in this paper is the combination of crowdsensing (CS) and the VSNs. We first design a three-layer CS-based VSN architecture that includes a CS layer, mobile management layer, and core control layer. Then, we propose a two-stage privacy-preserving (TSPP) algorithm to protect identity and location privacy for the CS-based VSN. In TSPP algorithm, to protect identity privacy, we design a way of identity management. Simultaneously, by correlation analysis, we design a suppression strategy to satisfy ε-privacy with no information loss for location privacy protection. The simulation results show that our proposed TSPP algorithm better protects identity and location privacy than current algorithms. The results are based on four metrics: average processing time, identity privacy, location privacy, and information loss.

Cognitive radio technology is an important branch in the field of wireless communication, and automatic identification is a major part of cognitive radio technology. Convolutional neural network (CNN) is an advanced neural network, which is the forefront of application in the digital image recognition area. In this paper, we explore CNN in an automatic system to recognize the cognitive radio waveforms. Excitedly, it is a more effective model with high ratio of successful recognition (RSR) under high power background noise. The system can identify eight kinds of signals, including binary phase shift keying (Barker codes modulation) linear frequency modulation, Costas codes, Frank code, and polytime codes (T1, T2, T3, and T4). The recognition part includes a CNN classifier. First, we determine the appropriate architecture to make CNN effective for proposed system. Specifically, we focus on how many convolutional layers are needed, what appropriate number of hidden units is, and what the best pooling strategy is. Second, we research how to obtain the image features into CNN that based on Choi-Williams time-frequency distribution. Finally, by means of the simulations, the results of classification are demonstrated. Simulation results show the overall RSR is 93.7% when the signal-to-noise ratio is -2dB.

This paper proposes a carrier-sensing multiple access protocol for asynchronous multiple-packet reception (MPR), which enables wireless receivers to correctly decode partially time-overlapping packets. Unlike previous studies, which require back-off nodes to constantly monitor the channel, our design requires only that each node sense the number of transmitting nodes after the completion of the back-off period for determining whether to begin transmission in the next time slot. In addition, we develop an analytical model to evaluate relevant parameters and performance metrics of the proposed protocol. Our model relies on the channel-sensing probability of a node in a randomly chosen slot, rather than channel-accessing probability adopted in previous models for asynchronous MPR. The results are validated through numerical study under a variety of network conditions. We also show that the proposed protocol is quite robust to imperfect estimation in channel sensing, and is more energy-efficient than other similar threshold-based protocols.