Search

The event response is a crucial feature in the real-time operating system RT-Thread, used for synchronizing threads. This paper aims to analyze the real-time performance of the event response and provide practical insights for its application. To achieve this, a low-latency piling timing analysis approach is proposed. The precise measurement of thread scheduling time is achieved by accurately recording timestamp of critical moments without external interruptions. Moreover, the theoretical time of thread scheduling under different compilation optimization levels is analyzed. And, a thread synchronization experiment based on event response is designed to measure the actual time of thread scheduling on Arm Cortex-M4 microcontroller STM32L431RCT6. The experimental results show that the low-latency piling timing analysis approach proposed in this paper has an error of about 1.5 microseconds compared to the measurement time of the logic analyzer and the theoretical analysis time. Finally, by comparing and analyzing theoretical time and actual measurement time of thread scheduling, it is evident that the compilation optimization level and PendSV exception are the key factors influencing the real-time performance of RT-Thread event response.

Effective leaf classification is crucial for ecological research and agricultural management, though handling diverse leaf shapes and environmental variations can be difficult. This study introduces an innovative approach that combines advanced feature extraction techniques with optimization and classification strategies. We employ Scale Invariant feature transform (SIFT), wavelet transform, and particle gradient motion for feature extraction, followed by bee colony optimization to enhance the process. A Gaussian distribution-based classifier is subsequently utilized, achieving an accuracy of 92%, while a Random forest classifier applied to the Grapevine Leave Dataset resulted in an accuracy of 84.63%. This approach highlights the potential for improved plant identification and disease management, providing a reliable solution for both small and large-scale agricultural applications. The resilience of this framework guarantees adaptability across various datasets, making it a valuable asset for improving crop quality and sustainability.

We consider the problem of sparse variable selection on high dimension heterogeneous data sets, which has been taking on renewed interest recently due to the growth of biological and medical data sets with complex, non-i.i.d. structures and huge quantities of response variables. The heterogeneity is likely to confound the association between explanatory variables and responses, resulting in enormous false discoveries when Lasso or its variants are naïvely applied. Therefore, developing effective confounder correction methods is a growing heat point among researchers. However, ordinarily employing recent confounder correction methods will result in undesirable performance due to the ignorance of the convoluted interdependency among response variables. To fully improve current variable selection methods, we introduce a model, the tree-guided sparse linear mixed model, that can utilize the dependency information from multiple responses to explore how specifically clusters are and select the active variables from heterogeneous data. Through extensive experiments on synthetic and real data sets, we show that our proposed model outperforms the existing methods and achieves the highest ROC area.

A slow-wave half-mode substrate integrated waveguide with a spoof surface plasmon polariton (SW-HMSIW-SSPP) transmission line structure is proposed, analyzed, and demonstrated. Results show longitudinal and lateral size reductions of more than 80% and 73%, respectively, as compared with an equivalent SIW at the same cutoff frequency while maintaining a low attenuation constant of 0.02 dB/mm. By altering the geometric parameters of the SW-HMSIW-SSPP the dispersion curve can be tailored. The dispersion and electromagnetic (EM) field distribution characteristics of the proposed SW-HMSIW-SSPP unit cell and overall structure are analyzed by HFSS simulations and an equivalent circuit, and pertinent performance parameters are compared to those of SW-HMSIW and HMSIW-SSPP structures. A proof-of-concept PCB SW-HMSIW-SSPP sample was fabricated, and measurement results showing a band-pass of 4.8-11.5 GHz are found consistent with simulation results. The SW-HMSIW-SSPP structure has applications in the design of miniaturized microwave circuits.

This paper studies a class of quasi-cyclic LDPC (QC-LDPC) codes, i.e., Tanner (3, 23)-regular QC-LDPC codes of code length $23p$ with $p$ being a prime and $p \equiv 1 (\mathrm {mod} 69)$ . We first analyze the cycle structure of Tanner (3, 23)-regular QC-LDPC codes, and divide their cycles of lengths 4, 6, 8, and 10 into five equivalent types. We propose the sufficient and necessary condition for the existence of these five types of cycles, i.e., the polynomial equations in a 69th unit root of the prime field $\mathbb {F}_{p}$ . We check the existence of solutions for such polynomial equations by using the Euclidean division algorithm and obtain the candidate girth values of Tanner (3, 23)-regular QC-LDPC codes. We summarize the results and determine the girth distribution of Tanner (3, 23)-regular QC-LDPC codes.

Integrating contextual information into sentence-level neural machine translation (NMT) systems has been proven to be effective in generating fluent and coherent translations. However, taking too much context into account slows down these systems, especially when context-aware models are applied to the decoder side. To improve efficiency, we propose a simple and fast method to encode all sentences in an arbitrary large context window. It makes contextual representations in the process of translating each sentence so that the overhead introduced by the context model is almost negligible. We experiment with our method on three widely used English-German document-level translation datasets, which obtain substantial improvements over the sentence-level baseline with almost no loss in efficiency. Moreover, our method also achieves comparable performance with previous strong context-aware baselines and speeds up the inference by $1.53\times $ . The speed-up is even larger when more contexts are taken into account. On the ContraPro pronoun translation dataset, it significantly outperforms the strong baseline.

In this paper, we proposed a broadband circularly polarized antenna with wide beamwidth. The main structure of the proposed antenna is a combined cross-polarized dipole antenna, four metal arms, a single coaxial cable and two feed structures with 90° phase delay. In order to attain a wide beamwidth antenna performance that can be used in a warehouse environment, we innovatively bent the metal arms and added a metal plane at a specific location to gain maximum beamwidth and the highest achievable gain. Especially, the application of metal bending arms not only reduce the space size of antenna, but also possess good axial ratio characteristic. The measurement illustrates that the antenna prototype achieves S11 < −10 dB impedance bandwidth (IBW) of 40% (820 – 1230 MHz), a 3-dB axial ratio bandwidth (ARBW) of 7.3% (895 – 970 MHz), and a maximum half-power beamwidth (HPBW) of 104° (@915 MHz). In addition, the good radiation characteristics of a gain of 6.33 dBic and an average radiation efficiency of 76.3% were measured in the operating bandwidth, which makes the proposed antenna a good candidate for the hasty logistics radio frequency identification (RFID) system.

The stator and rotor can are added to the permanent magnet canned motor (PMCM) to prevent the stator and rotor core corrosion. The can losses accounts for about 40% of the total losses, therewith the temperature problems can be brought in this kind of motors. Based on this, the accurate analytical calculation method of the magnetic field, can losses, and temperature of PMCM is the key to the rapid design of this kind of motor. The no-load magnetic field is solved by the subdomain method, and the harmonic magnetic density is decomposed by Fourier. Secondly, the calculation formula of can losses is proposed based on the law of electromagnetic induction. The derived calculation formula of can losses can separate the can losses into each harmonic can loss. Then, the results of the analytical solution are added into a new thermal network model (TNM) to predict the temperature distribution of the cans accurately. The new TNM of PMCM can consider the influence of uneven heat dissipation of cans and cooling channels. The key is to re-divide the structure of cans and winding and improves its thermal resistance calculation formula. Finally, the effectiveness of the proposed TNM is verified by comparing it with the results of experiment.

Recently, Wi-Fi-based human activity recognition using channel state information (CSI) signals has gained popularity due to its potential features, such as passive sensing and adequate privacy. The movement of various body parts in between Wi-Fi signals’ propagation path generates changes in the signal reflections and refraction, which is evident from the CSI variations. In this paper, we analyze the relationship between human activities and properties (amplitude and phase) of Wi-Fi CSI signals on multiple receiving antennas and discover the signal properties that vary remarkably in response to human movement. The variation in the signal received among multiple antennas shows different sensitivity to human activities, directly affecting recognition performance. Therefore, to recognize human activities with better efficiency, we proposed an adaptive antenna elimination algorithm that automatically eliminates the non-sensitive antenna and keeps the sensitive antennas following different human activities. Furthermore, the correlation of the statistical features extracted from the amplitude and phase of the selected antennas’ CSI signal was analyzed, and a sequential forward selection was utilized to find the best subset of features. Using such a subset, three machine learning algorithms were employed on two available online datasets to classify various human activities. The experimental results revealed that even when using easy-to-implement, non-deep machine learning, such as random forest, the recognition system based on the proposed adaptive antenna elimination algorithm achieved a superior classification accuracy of 99.84% (line of sight) on the StanWiFi dataset and 97.65% (line-of-sight) / 93.33% (non-line-of-sight) on another widely applied multi-environmental dataset at a fraction of the time cost, demonstrating the robustness of the proposed algorithm.

El Niño/La Niña events significantly impact human society, often resulting in considerable monetary losses. Accurate prediction has become crucial with triple La Niña events in this century. This study applied TimesNet to El Niño/La Niña event prediction for the first time. We proposed a hybrid prediction method based on extracting features from time series data and initially decomposing the time series data (Niño3.4) into several Intrinsic Mode Functions (IMFs) using the Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN). Based on the characteristics of each IMF, we used a hybrid method of TimesNet and ARIMA to make adaptive forecasts for them. We selected monthly data from 1950 to 2022, with the first 63 years used for training and shifted 12 periods (12 months) ahead to forecast the Niño3.4 index values for the next ten years. The experimental results of this study show that: 1) The pre-processing method using CEEMDAN can effectively extract the original time series data features and significantly improve the prediction performance; 2) proposed approach achieved good performance in predicting El Niño/La Niña events, particularly during the transition from El Niño to La Niña events (e.g., 2016, 2019-2020); 3) evaluation results indicate that the proposed model exhibits better predictive power (stability and accuracy of prediction results) than the current best single-order predictor, the ConvLSTM model, on the validation set of the last ten years.

The increasing amount of traffic in recent years has led to increasingly complex network problems. To be able to improve overall network performance and increase network utilization, it is valuable to take measures to capture future trends in network traffic. In traditional machine learning, to guarantee the accuracy and high reliability of the models obtained through training, there are two basic assumptions: (1) the training samples used for learning and the new test samples satisfy the condition of independent identical distribution; and (2) there must be enough training samples to learn a good model. However, time-series data are not easily accessible in real life, and even after putting in a lot of time and effort to collect them, the data may be unavailable due to confidentiality. In this paper, a neural network model based on long and short-term memory (LSTM) and transfer learning is proposed to address the problem of small sample size in network traffic prediction. Knowledge in the source domain is transferred to the target domain using transfer learning, and a prediction model with good performance is constructed with a small amount of target domain data. The results show that the performance of the transfer learning model improves by more than 40% over the direct training model when using the same samples for predicting 10,000 rows of data, resulting in better performance of the network traffic prediction task.

Load forecasting is critical for power system operation and market planning. With the increased penetration of renewable energy and the massive consumption of electric energy, improving load forecasting accuracy has become a difficult task. Recently, it was demonstrated that deep learning models perform well for short-term load forecasting (STLF). However, prior research has demonstrated that the hybrid deep learning model outperforms the single model. We propose a hybrid neural network in this article that combines elements of a convolutional neural network (1D-CNN) and a long short memory network (LSTM) in novel ways. Multiple independent 1D-CNNs are used to extract load, calendar, and weather features from the proposed hybrid model, while LSTM is used to learn time patterns. This architecture is referred to as a CNN-LSTM network with multiple heads (MCNN-LSTM). To demonstrate the proposed hybrid deep learning model’s superior performance, the proposed method is applied to Ireland’s load data for single-step and multi-step load forecasting. In comparison to the widely used CNN-LSTM hybrid model, the proposed model improved single-step prediction by 16.73% and 24-step load prediction by 20.33%. Additionally, we use the Maine dataset to verify the proposed model’s generalizability.

Massive high-resolution spaceborne synthetic aperture radar (SAR) data are available nowadays, effecting rapid development in the urban remote sensing technology. In this study, we use a set of 27 scenes from TerraSAR-X to reconstruct the three-dimensional (3D) point clouds of a skyscraper-North Star Times Tower in Beijing, China. We study in detail the derivation process of a 3D model of tomographic SAR (TomoSAR) from the two-dimensional (2D) imaging mathematical model of SAR. We observe the presence of a spatial shift in the elevation direction of TomoSAR, which yields poor and even incorrect results when greedy algorithms such as orthogonal matching pursuit (OMP) or regularized orthogonal matching pursuit (ROMP) algorithms are applied. Then, we propose a phase compensation method to eliminate the spatial shift. Both OMP and ROMP algorithms are used for TomoSAR imaging on high-resolution TerraSAR-X data. The experimental results with North Star Times Tower demonstrate the accuracy of the theoretical analysis and the effectiveness of the algorithms. Finally, we compare the ROMP and OMP algorithms and conclude that both algorithms can yield similar reconstruction results in TomoSAR imaging. The ROMP algorithm does not depend on sparsity and can suppress noise through regularization selection at the cost of time. The OMP algorithm introduces noise when the sparsity is inaccurate.

It is very important for the intuitionistic fuzzy decision making to rank the different intuitionistic fuzzy numbers (IFNs). Based on the full information of the membership degree and non-membership degree, an improved possibility degree method is defined for dealing with multi-attribute decision making (MADM) problems in this paper. Firstly, the limitations of the existing possibility degree are discussed, and then an improved one is defined to overcome these shortcomings. Subsequently, six properties of the proposed possibility degree, which are similar to that of the classic possibility degree, are presented and proved. Moreover, some numerical examples are provided to illustrate the superiority of the proposed method. And the calculation method of membership degree and non-membership degree from the statistical viewpoint is developed to construct the intuitionistic fuzzy evaluation matrix considering that the decision-makers are difficult to give exact values of the two variables directly. Furthermore, depending on the proposed possibility degree and the intuitionistic fuzzy generalized ordered weighted geometric average (IFGOWGA) operator, a new MADM approach is employed and used to solve the practical problem of aircraft cockpit display ergonomic evaluation under the intuitionistic fuzzy set environment. Finally, an evaluation example is examined to manifest the effectiveness and feasibility of the proposed decision-making method.

The Kalman-filter-based algorithms as the mainstream algorithms of dynamic state estimation of power systems have been extensively used to provide accurate data for power system applications. However, few comparisons are made to show their advantages and disadvantages. In this paper, four Kalman-filter-based algorithms (i.e., extended Kalman filter, unscented Kalman filter, cubature Kalman filter, and ensemble Kalman filter) are compared to show their differences from implementation complexity, estimation accuracy and calculation efficiency, the resistance to measurement errors, and the sensitivity to system scales. Finally, the simulation results on the 3-machine, 10-machine, and 48-machine power systems show their advantages and disadvantages.

A low-profile dual-band dual-polarized antenna with an artificial magnetic conductor (AMC) reflector is proposed for 5G communications. The antenna consists of a pair of crossed dual-polarized dual-band bowtie dipoles and a dual-band AMC reflector. By introducing trapezoidal slots and U-shaped slots on the bowtie dipoles, miniaturization and dual-band characteristics are achieved. Moreover, T-shaped feeding structures are utilized to broaden the bandwidth of the bowtie dipoles. By adopting a dual-band AMC reflector instead of a conventional perfect electric conductor (PEC) reflector, the distance between the radiator and the reflector can be reduced from 0.25λ0 to 0.08λ0 (where λ0 is the free-space wavelength at 3.5 GHz). The radiator can maintain the impedance bandwidth and a high gain can also be achieved, even if it is close to the AMC reflector. A geometrical optics model is used to explain the mechanism of the AMC. The optimal parameters of the AMC depend on the antenna operating frequency and the distance between the radiator and the reflector. Measurements show that the proposed dual-band antenna has an impedance bandwidth of 19.8% (3.14-3.83 GHz) and 13.2% (4.40-5.02 GHz), covering the sub-6 GHz frequency spectra of 5G mobile communications. The peak gain is 7.1 dBi in the lower band and 8.2 dBi in the upper band. Port isolation better than 20 dB is achieved. The proposed antenna can be used alone for 5G indoor base station, or as the element of an array for 5G outdoor base station.

Multi-agent reinforcement learning (MARL) for cooperative tasks has been extensively studied in recent years. The balance of exploration and exploitation is crucial to MARL algorithms' performance in terms of the learning speed and the quality of the obtained strategy. To this end, we propose an algorithm known as the weighted relative frequency of obtaining the maximal reward (WRFMR), which uses a weight parameter and the action probability to balance exploration and exploitation and accelerate convergence to the optimal joint action. For the WRFMR algorithm, each agent needs to share the state and the immediate reward and does not need to observe the actions of the other agents. Theoretical analysis on the model of WRFMR in cooperative repeated games shows that each optimal joint action is an asymptotically stable critical point if the component action of every optimal joint action is unique. The box-pushing task, the distributed sensor network (DSN) task, and a strategy game known as blood battlefield are used for empirical studies. Both the DSN task and the box-pushing task involve full cooperation, while blood battle comprises both cooperation and competition. The simulation results show that the WRFMR algorithm outperforms the other algorithms regarding the success rate and the learning speed.

The temperature-sensitive point is the input variable of the thermal error compensation model of computer numerical control (CNC) machine tools. At present, the most commonly used selection method is to measure the multipoint temperature and thermal error of the machine tool synchronously and select several temperature measurement points with the highest thermal error correlation as temperature-sensitive points according to the measurement data. This study reveals that this method sometimes has obvious mis-selection and causes the model to fail. The reason is that the weak correlation temperature measurement point away from the machine tool heat source amplifies the volatility of the correlation evaluation result, owing to the small overall change. If the calculated result exceeds the true strong correlation temperature measurement point, it will be incorrectly selected as the temperature-sensitive point. This scenario has been termed herein as the pseudocorrelation problem. With the gradual popularization of thermal error compensation technology for CNC machine tools, pseudocorrelation will seriously affect the mass production pass rate. Therefore, the study analyzes and rigorously proves the mathematical mechanism of this problem, and the temperature measurement point preselection algorithm based on the correlation coefficient volatility determination factor (CCVDF) is proposed to eliminate potential pseudocorrelation temperature points before selecting temperature-sensitive points. After 1000 random simulation experiments, the failure rate of Z-direction thermal error modeling after preselection decreased from 5.1% to 0.5%, and the Y-direction error decreased from 33.7% to 1.5%. This algorithm can greatly improve the qualification rate of large-scale equipment of thermal error compensation technology.

At present, the primary technical deterrent to the use of strawberry harvesting robots is the low harvest rate, and there is a need to improve the accuracy and real-time performance of the localization algorithms to detect the picking point on the strawberry stem. The pose estimation of the fruit target (the direction of the fruit axis) can improve the accuracy of the localization algorithm. This study proposes a novel harvesting robot for the ridge-planted strawberries as well as a fruit pose estimator called rotated YOLO (R-YOLO), which significantly improves the localization precision of the picking points. First, the lightweight network Mobilenet-V1 was used to replace the convolution neural network as the backbone network for feature extraction. The simplified network structure substantially increased the operating speed. Second, the rotation angle parameter $\alpha $ was used to label the training set and set the anchors; the rotation of the bounding boxes of the target fruits was predicted using logistic regression with the rotated anchors. The test results of a set of 100 strawberry images showed that the proposed model's average recognition rate to be 94.43% and the recall rate to be 93.46%. Eighteen frames per second (FPS) were processed on the embedded controller of the robot, demonstrating good real-time performance. Compared with several other target detection methods used for the fruit harvesting robots, the proposed model exhibited better performance in terms of real-time detection and localization accuracy of the picking points. Field test results showed that the harvesting success rate reached 84.35% in modified situations. The results of this study provide technical support for improving the target detection of the embedded controller of harvesting robots.

Compared to traditional images and 2D videos, 3D videos are more likely to cause distortion drift problems in data hiding. In this paper, an improved scheme for high capacity and efficient data hiding in 3D videos based on Multi-view coding (MVC) standard is proposed, which avoids the problems of distortion drift. To improve the visual quality of data hiding, two selection modes are provided to limit the distortion drift. By modifying the selected coefficients of 4 × 4 quantized discrete cosine transforms (QDCT) in macroblocks and hiding data into b4-frames, the proposed scheme will effectively prevent distortion drift which caused by intra prediction and inter prediction. Moreover, the proposed algorithm maintains great randomness by using two random sequences. Several benchmark 3D video sequences of different resolutions and diverse contents are used for experimental evaluation. It is experimentally proved that the algorithm has greater embedding capacity and higher efficient than the previous methods.

There is a certain energy loss in the process of wireless sensor network information collection. Moreover, the current network protocols and network coverage methods are not sufficient to effectively reduce system energy consumption. In order to improve the operating efficiency and service life of wireless sensor networks, this study analyzes the classic LEACH protocol, summarizes the advantages and disadvantages, and proposes a targeted clustering method based on the K-means algorithm. At the same time, in order to maximize the network coverage and minimize the energy consumption on the basis of ensuring the quality of service, a wireless sensor network coverage optimization method based on an improved artificial fish swarm algorithm was proposed. In addition, a controlled experiment is designed to analyze the effectiveness and practical effects of the proposed algorithm. The experimental results show that the method proposed in this paper has certain advantages over traditional methods and can provide theoretical references for subsequent related research.

We propose a bent corrugated substrate integrated waveguide (BCSIW) structure that can be used to design a high gain leaky wave antenna (LWA). The design removes the need for a metallic via fabrication process needed for standard substrate integrated waveguides (SIW) and displays superior performance to previously reported structures. We use simulations to compare the performance of the proposed BCSIW LWA to equivalent standard SIW and corrugated SIW (CSIW) structures, as well as experimentally characterize a fabricated BCSIW LWA. Simulation results show that the BCSIW structure can help improve the impedance bandwidth of a slotted LWA by about 14.7% while still maintaining high gain (about 13.2-17.4 dBi) as compared to an LWA based on a CSIW structure. Measurement results indicate that the proposed BCSIW LWA has a wide impedance bandwidth (32.6%) and a high peak gain (12-16.2 dBi) throughout a large frequency range from 22 to 29.2 GHz with a large beam angle range from -69° to -10°.

In this paper, the fractional-order modeling and analysis of a three-phase voltage source PWM rectifier (VSR) are researched considering the fractional-order characteristics of actual inductors and capacitors. While the Caputo fractional calculus operator is used to describe the fractional-order characteristics of an inductor and a capacitor, the fractional-order model of a three-phase VSR is established in the three-phase static coordinate system. With the coordinate transformation, the fractional-order model is expressed in the two-phase static coordinate system and in the synchronous rotation coordinate system, respectively. Simulation results verify the effectiveness of the established models, and indicate that the fractional-order models can describe the operating characteristics of a three-phase VSR more accurately. Besides, compared with the traditional integer-order three-phase VSRs, the three-phase VSRs which contain the fractional-order inductors and the fractional-order capacitor can present better static and dynamic performance indexes (such as smaller overshoot and shorter regulation time) by appropriately selecting the orders of the inductors and the capacitor.

A leaky-wave antenna (LWA) based on the double-layer substrate integrated waveguide (SIW) that has a wide scanning range within a narrow bandwidth is proposed. The LWA consists of two layers, each of which consists of multiple elemental SIWs arranged in parallel at each layer while the direction of the SIWs in the two layers are at an oblique angle to each other. The elemental top layer SIWs are connected to those in the bottom layer though parallelogram-shaped slots, so that the two layers together form a double-layer SIW. Thin slots etched on the upper surface of each elemental SIW of the top layer act as radiating structures. As the phase difference between the radiating slots of adjacent elemental SIWs can be enlarged by the waveguide length, wide-angle beam scanning is achieved within a specific frequency bandwidth. The proposed LWA has a scanning range from -58° to 49° in a relative bandwidth of 6.45%, from 12.75 GHz to 13.6 GHz. The scanning rate (SR), defined as the scanning range divided by the relative bandwidth, is up to 16.5. The size of structure is compact, which is 4.0λ0 × 3.6λ0 × 0.04 λ0. The proposed LWA is suitable for radar or communication system applications.

A new type of broadband circularly polarized leaky-wave antenna (LWA) with high-gain based on a corrugated substrate integrated waveguide (CSIW) structure is proposed and investigated. The CSIW structure that employs open-circuit stubs to replace metallic vias has the advantages of low-cost and easy fabrication as compared to conventional substrate integrated waveguides. Each unit-cell of the proposed LWA consists of two quarter-wavelength microstrip lines and a half-mode CSIW cell with three open-circuit stubs. Two M-shaped slots etched on the half-mode CSIW cell enables the generation of circularly polarized (CP) radiation. The full LWA, which consists of thirteen matched unit-cells cascaded along the direction of propagation enables continuous backward to forward beam scanning. The properties of the CSIW structure, including impedance matching and phase constant, are analyzed by Bloch impedance and dispersion simulations and the influence of the dimensions of the open-circuit stubs on the CP LWA performance is also investigated. A protype of the proposed LWA is fabricated and characterized. The measured results indicate that the proposed CSIW LWA has a high peak gain (9.3-12.5 dBi) throughout a large beam-scanning angle range from -28° to +25°, and the impedance bandwidth and 3-dB axial ratio (AR) bandwidth are over 40% covering the full $\textit {Ka}$ -band.

Deep neural networks of deep learning algorithms can be applied into regressions and classifications. While the regression performances and classification performances of the deep neural networks are depending on the hyper-parameters of the deep neural networks. To mitigate the adverse effect of the hyper-parameters for the deep learning algorithms, this paper proposes deep forest regression for the short-term load forecasting of power systems. Deep forest regression includes two procedures, i.e., multi-grained scanning procedure and cascade forest procedure. These two procedures can be effectively trained by two completely random forests and two random forests with the default configuration. Then, the deep forest regression is applied into the short-term load forecasting of power systems. The forecasting performances of deep forest regression are compared with that of numerous intelligent algorithms and conventional regression algorithms under the model with the data of previous 7-day, 21-day, and 40-day. Besides, the forecasting performances of deep forest regression with different parameters are compared. The numerical results show that the deep forest regression with default configured parameters can increase the accuracy of the short-term forecasting and mitigate the influences of the experiences for the configuration of the hyper-parameters of deep learning model.

The emergence of 4D heart images makes the data volume of the images multiply. It is more urgent to require an effective and fast segmentation algorithm. Therefore, a heart image can be accurately segmented from a large amount of image data and an area of interest can be extracted The segmentation algorithm is very necessary. Based on the segmentation and recognition of medical images, this paper proposes a neural network and image saliency based on the obvious difference between the heart image and other tissues in the slice, and the high similarity between adjacent slices in the CT image sequence. Fully automatic segmentation algorithm and 3D visual reconstruction is the segmented heart image. Convolutional neural network is a special deep neural network model of artificial intelligence. Its connections between neurons are not fully connected. The weights of connections between certain neurons in the same layer are shared, and the network model is reduced. The complexity reduces the number of weights. The use of visual saliency techniques to achieve cardiac segmentation based on CT images. An image saliency detection algorithm is adopted to introduce the image segmentation algorithm based on the saliency technique. In this paper, considering the PET image as grayscale image with low resolution, an improved Itti model and an improved GrabCut image segmentation algorithm are proposed to solve the problem of the original algorithm in grayscale image. At the same time, the operation steps of the user division area are cancelled, and the automatic processing is realized, and the running time of the algorithm is improved while optimizing the image segmentation effect. The convolutional neural network is constructed to realize the positioning function of the heart in the image. The original cardiac CT image is cropped by the positioning result, and some non-target areas are removed. A stacking noise reduction self-coding network is constructed, and the network is manually segmented. Training, realize the classification and recognition of the pixels belonging to the heart tissue in the CT image of the heart, and finally realize the segmentation of the heart image based on the classification result. The results of the above segmentation algorithm are quantitatively evaluated and analyzed with the artificial segmentation results, and the segmentation results are visually reconstructed by surface rendering and volume rendering. The algorithm has better accuracy, reliability and higher. The segmentation efficiency is more simplified for user operations.

Due to the high power consumption and hardware cost of radio frequency (RF) chains, the conventional fully-digital beamforming will be impractical for large-scale antenna systems (LSAS). To address this issue, hybrid beamforming has been proposed to reduce the number of RF chains. However, the fully-connected structure assumed in most hybrid beamforming schemes is still cost-intensive. Recently, the partially-connected structure employing notably fewer phase shifters has received considerable attention in both academia and industry. But the design of partially-connected hybrid beamforming has not been fully understood, especially in multi-user systems. In this article, we directly address the challenging non-convex non-smooth partially-connected hybrid beamforming design problem with individual signal-to-interference-plus-noise ratio (SINR) constraints and unit-modulus constraints in a multi-user massive multiple-input multiple-output (MIMO) system. An iterative alternating algorithm based on a penalty method is proposed to obtain a stationary point, which inevitably has relatively high computational complexity. Thus, two low-complexity algorithms are then proposed by utilizing matrix approximation. Numerical results demonstrate significant performance gains of the proposed algorithms over existing hybrid beamforming algorithms. Moreover, the proposed low-complexity algorithms can achieve near-optimal performance with dramatically reduced computational complexity.

Aiming at the problem of fault prognostics for the energy storage power station, this paper proposes a novel data-driven method named multiple elastic networks with time delays (MEN-TD). The proposed method can learn the status of the energy storage power station in advance and provide early detection of the fault. First, through the correlation analysis and the mechanism knowledge, the energy storage power station key parameter and corresponding key factors affecting the parameter are determined. Secondly, in order to predict the trend of the key parameter over a period of time and improve the prediction accuracy, the MEN-TD model is constructed. Then, based on the predicted values of the key parameter, compared with the control limit in the healthy status, the fault can be pre-warned in advance. Finally, through testing on the practical energy storage power station in Zhenjiang of China, the effectiveness and superiority of the proposed MEN-TD method are demonstrated.

Technology determines tactics, and tactics promote the development of technology. Artificial intelligence technology is a multiplier that accelerates the innovation and development of military theory. In this paper, we first list the different intelligence levels by introducing their corresponding applications. Then we review the technical classification based on the related concepts, Finally, we discuss technical and practical difficulties and give some solutions from the aspects of strengthening knowledge engineering, building simulation systems, and accumulating data engineering knowledge. The development of Artificial intelligence technology has a profound impact on military development trends, leading to major changes in the forms and modes of war.

Non-intrusive load monitoring (NILM) makes it possible for users to track the energy consumption of a household. In this paper, we present a new hybrid energy disaggregation approach named HAM. This event-based load disaggregation algorithm uses an improved multi-layer Hungarian algorithm to match appliances transient features and a supervised adaptive resonance theory mapping neural network (ARTMAP) to cluster steady features. This approach using a modified dual sliding window-based cumulative sum control chart algorithm (DSWC) to detect the transient event first, and we convert the multi-dimensional electrical features of appliances into a bipartite graph matching problem. This way, an improved Hungarian algorithm is proposed to find the perfect matching when appliances are in different states. For classification and identification, the ARTMAP network is used to learn and classify the steady features of appliances. Furthermore, we introduced a grey correlation evaluation and multiple matching strategies to increase the fitness and accuracy of the approach. Experimental results demonstrate that the proposed HAM outperforms the comparative algorithm in both our resident appliances dataset (RAD) and BLUED public dataset. The proposed approach could also facilitate NILM more applicable for common households.

Stock price volatility forecasting is a hot topic in time series prediction research, which plays an important role in reducing investment risk. However, the trend of stock price not only depends on its historical trend, but also on its related social factors. This paper proposes a hybrid time-series predictive neural network (HTPNN) that combines the effection of news. The features of news headlines are expressed as distributed word vectors which are dimensionally reduced to optimize the efficiency of the model by sparse automatic encoders. Then, according to the timeliness of stocks, the daily K-line data is combined with the news. HTPNN captures the potential law of stock price fluctuation by learning the fusion feature of news and time series, which not only retains the effective information of news and stock data, but also eliminates the redundant information of the text. Compared with the state-of-the-art methods, our method combines more abundant stock characteristics and has more advantages in running speed. Besides, the accuracy is averagely improved by nearly 5%.

To better grasp the status quo of urban rail transit (URT) safety research and explore the knowledge base and hotspot trends, 1249 URT safety research papers from 1983-2018 were collected from the SCIE and SSCI databases of the WOS search platform as data samples. Co-occurrence analysis, burst analysis of keywords and co-citation analysis were adopted to analyze the current research situation of URT safety in the world by means of information visualization. The study found that risk analysis in the construction/operational phase, methods for predicting ground movements, and cause analyses of tunnel deformation and settlement are the knowledge bases in the field of URT safety research. Tunneling and Underground Space Technology, Safety Science, Atmospheric Environment, Lancet and Transportation Research Record are the core journals in the field. At present, the basic theory and research system of URT safety have been basically completed, but the research directions are too concentrated, and the frontier branches are too few. The research themes of URT safety were roughly grouped into three core paths: the “health risk” path, the “risk analysis - safety management” path and the “tunnel displacement/deformation” path. In the process of research and evolution, the research paths are refined step-by-step, and the research directions are transformed from macro to micro. The emergency evacuation of personnel in subway stations, the effects of particulate matter in subway air on human health and the use of computer technology for engineering optimization are the current frontiers of URT safety. URT safety research also tends toward risk analysis in the operational phase.

Remote-sensing images contain lots of visual information about transportation systems, geomorphic conditions, and communal facilities. The secure storage and transmission of remote-sensing images are of great significance. This paper proposes a novel image encryption scheme-based DNA bases probability and applies it to remote-sensing images for data protection. First, the plain-image is randomly encoded in DNA rules and its result participates in DNA addition with DNA mask generated by 2-D logistic map. Second, the cryptosystem executes 2-D logistic map again with DNA bases probability against differential attacks. Third, the pixel-level rearrangement and the DNA base-level rearrangement are, respectively, implemented in order of chaos sequences for permutation and diffusion. The cipher-image has a uniform distribution, low coefficients, and ideal entropy. The cryptosystem has acceptable encryption speed ($speed=\textrm {0.651308Mbit/s}$ ) and high sensitivity to the plain-image and the security key. The experimental results show that the proposed algorithm can resist various existing attack schemes against remote-sensing images.

Maintaining power system stability is becoming urgent due to the large-scale interconnection of power grids and the high penetration of uncertain renewable energy sources. The excitation control and governor control of synchronous generators have been considered as two crucial measures for enhancing the power system stability. However, a major challenge is to simultaneously achieve global asymptotic stability (GAS), voltage regulation (VR), and power regulation (PR) in the excitation and governor control. In this paper, a Lyapunov-based decentralized control (LBC) is proposed to address this challenge. The time-derivative of the Lyapunov function is designed by the feedback control of synchronous generators in order to guarantee GAS. VR and PR are ensured by considering voltage and power deviations as the feedback variables. The simulation results on the New-England ten-machine power system validate the effectiveness of the proposed LBC in improving power system transient stability and simultaneously achieving VR and PR. Although the proportional-integral- and power system stabilizer-based control can also perform VR and PR, the proposed control has much better dynamic performance and can more significantly improve the system transient stability.

When faults occur in mechanical components, the faulty information is usually manifested as a series of periodic impulses which correspond to the faulty feature frequencies. However, due to the non-stationary characteristic of the raw vibration signals, the faulty feature frequencies are difficult extracted. In this paper, a novel strategy using variational mode decomposition (VMD), L-Kurtosis and minimum entropy deconvolution (MED) is proposed to detect mechanical faults. First, VMD is employed to decompose the raw vibration signal into a set of intrinsic mode functions (IMFs) to eliminate the interference of the noise. Second, the optimal intrinsic mode function (IMF) which contains the faulty information is determined using L-Kurtosis. Then, the impact characteristic of the periodic impulses in optimal IMF is enhanced through MED. Finally, a Hilbert envelope spectrum analysis is performed to the enhanced signal to extract the faulty feature frequency. In order to illustrate the performance of the proposed strategy, the simulation signal and real experimental signals collected from faulty rolling element bearings and gears are analyzed. The results show that the strategy using the VMD, L-Kurtosis, and MED can detect mechanical component faults effectively.

Crowdsourcing is gaining more and more popularity among the academic and industrial community. Organizations are adopting this technological advent and increasingly crowdsourcing their tasks to the unknown individuals. However, in the context of competitive crowdsourcing software development (CCSD), crowdsourcing is still unexplored. Too little is presently known about what intricate developers to participate in crowdsourcing software development competitions. Most importantly, what kind of developers are more likely to participate? Such open questions remain to be explored. To this end, in this paper, we present the results of an empirical study conducted to investigate what motivates software developers to participate in CCSD and what inhibits software developers to participate in such competitions. An online questionnaire is sent out to more than 300 crowdsource software participants, of which 113 return valid responses. It is also sent to more than 150 industry practitioners, of which 75 return valid responses. The results suggest that the monetary rewards are not significantly important to motivate software developers to participate in CCSD. Instead, learning, social contacts, and peer recognition are more important. Besides the survey, we also analyze the historical data collected from one of the most popular software crowdsourcing platforms. The analysis results reveal that the Pareto principle holds for CCSD as well, and 0.9% of the participants win 86% competitions. The results support the premise that CCSD market is still at an early stage. Most of the professional software engineers do not participate seriously in crowdsourcing software development. Therefore, many crowdsourced tasks, especially complex tasks, may fail to receive any satisfying submission. These findings are worthwhile for the crowdsourcing platforms and companies who want to outsource their software development tasks to the CCSD platforms.

An approach for the fast generation of a minimum-time, three-dimensional trajectory for a spacecraft propelled by a solar sail with reflection control devices (RCDs) is presented using the authors' previously proposed Bezier curve-based shaping approach. In this approach, the time variation of the position components of the spacecraft are assumed in advance to follow Bezier curve functions. By optimizing a finite set of unknown coefficients defining the shape of the Bezier curves used to approximate the generic state variables, the boundary constraints and equations of motion are satisfied simultaneously under a time-optimized performance index. Unlike the thrust vector of a conventional electric thruster, that of an RCD-equipped solar sail is constrained. To consider the thrust characteristics of such solar sails, the propulsive acceleration inequality constraints were numerically investigated for an asteroid rendezvous. The simulation results demonstrate that the presented approach is able to design the transfer trajectory of a spacecraft propelled by an RCD-equipped solar sail in about 1% of the time required by the conventional Gauss pseudospectral method with comparable accuracy by considering the actual characteristics of the thrust vector. This allows for quick feasibility assessments of different solar sail spacecraft mission profiles during the preliminary mission design stage.

A lot of cloud systems are adopted in industry and academia to face the explosion of the data volume and the arrival of the big data era. Meanwhile, energy efficiency and energy saving become major concerns for data centers where massive cloud systems are deployed. However, energy waste is quite common due to resource provisioning. In this paper, using dynamic voltage and frequency scaling (DVFS), a frequency selection approach is introduced to improve the energy efficiency of the cloud systems in terms of resource provisioning. In the approach, two algorithms, genetic algorithm (GA) and Monte Carlo tree search algorithm (MCTS), are proposed. A cloud database system is taken as an example to evaluate the approach. The results of the experiments show that both algorithms have its advantages. The algorithms have great scalability, in which the GA can be applied to thousands of nodes and the MCTS can be applied to hundreds of nodes. Both algorithms have high accuracy compared with optimal solutions (up to 99.9% and 99.6% for GA and MCTS, respectively). According to an optimality bound analysis, 26% of energy can be saved at most using our frequency selection approach.

Software refactoring has been extensively used to rectify the design flaws and improve software quality without affecting its observable behaviors. For a given code smell, it is common that there exist multiple refactoring solutions. However, it is challenging for developers to select the best one from such potential solutions. Consequently, a number of approaches have been proposed to facilitate the selection. Such approaches compare and select among alternative refactoring solutions based on their impact on metrics of source code. However, their impact on the traceability between source code and requirements is ignored although the importance of such traceability has been well recognized. To this end, we select among alternative refactoring solutions according to how they improve the traceability as well as source code design. To quantify the quality of traceability and source code design we leverage the use of entropy-based and traditional coupling and cohesion metrics respectively. We virtually apply alternative refactoring solutions and measure their effect on the traceability and source code design. The one leading to greatest improvement is recommended. The proposed approach has been evaluated on a well-known data set. The evaluation results suggest that on up to 71% of the cases, developers prefer our recommendation to the traditional recommendation based on code metrics.

The now ubiquitous use of information technology poses a crucial challenge to the cyber security of power grid operations, one that has aroused serious concerns from both industry and academia. The state-of-the-art research either focuses on the vulnerability assessment of particular types of components or concentrates on the prevention and mitigation of cyber attacks from the power grid’s perspective. Complete causal chains connecting component vulnerabilities to cyber attacks causing malicious system-wide effects are unclear, which hinders cyber-attack prevention and the consolidation of affected components. To bridge this research gap, this survey aims to study two issues that need further investigation. On one hand, the causal relationship between power grid component vulnerabilities and resulting cyber attacks has not been thoroughly explored. On the other hand, the evolvement of cyber attacks, from initial attempt to resulting serious consequences has not been resolved. To study these two issues, we first analyze various stakeholders and associated information flows in diverse applications in power grid operation. Second, we summarize the root cause of cyber attacks in the vulnerability of communication protocols. Third, a multi-stage model is proposed to describe the cyber vulnerabilities, which reveals potential attacks and their evolvement on power grid components at each stage, assesses the associated impact on the entire power grid, and elucidates possible countermeasures. With the above efforts, this survey establishes a complete causal chain from component-wise vulnerability to system-wide impact assessment for enhancing cyber security, and potential research directions for enhancing power grid cyber security are identified.

Identifiers chosen by programmers convey rich semantic information that could be exploited to enhance program analysis for various software engineering tasks, e.g., anomaly detection, argument recommendation, and code completion. However, such information has not yet been fully exploited, and little is known about its properties. In this paper, we carry out an empirical study to investigate the lexical similarity between the actual argument and the formal parameter names in method invocations. By analyzing actual arguments and formal parameters from 127 Java applications and 30 open-source C applications, we observe a number of interesting and valuable findings concerning the distribution of similarity, reasons for dissimilarity, and so on. Such findings have been exploited to facilitate argument-related software engineering tasks, i.e., anomaly detection and code completion.

A wideband single-fed, circularly-polarized patch antenna with enhanced axial ratio bandwidth for RFID reader application is proposed. The antenna consists of a modified half E-shaped patch, a near-field resonant parasitic (NFRP) patch, and four rotated vertical metallic plates on the ground plane. The metallic plates surrounding the main radiator of the modified half E-shaped patch antenna are fabricated vertically on the ground plane. The operating frequency of the antenna decreases substantially with the vertical metallic plates, which is important for miniaturization of RFID applications. With an elliptical slot and a cut on the modified half E-shaped patch, antenna impedance bandwidth and CP performance both improve. A truncated corner patch as an NFRP element can generate an additional CP radiation mode and further enhances the axial ratio and impedance bandwidth. The measured results show that the antenna has an impedance bandwidth of 14.2% (833-960 MHz), a 3-dB AR bandwidth of 9.0% (846-926 MHz), and an average gain of 7.3 dBic over a 3-dB AR bandwidth. The proposed antenna has a compact size and a low profile (0.45λ0 × 0.45λ0 × 0.074λ0).

Automatic vectorization is crucial for improving the performance of computationally intensive programs. Existing compilers use conservative optimization strategies for automatic vectorization, which, in many cases, lead to the loss of vectorization opportunity. Studies have shown that the use of machine learning algorithms to build a performance prediction model is beneficial to improve the program performance. The model input is program features, and the output is the predicted optimization strategies or the program performance related to the optimization. In this paper, we focus on a computational intensive loop structure-tensor contraction, which is common in quantum chemical simulations. Most existing machine learning methods rely on control and data flow graphs as features to represent programs, but different tensor contraction kernels have the same control and data flow graphs. In addition, the existing methods often use the same kind of learning algorithm to construct a learning model, which is prone to overfitting and low-precision problems. In this paper, we propose an automatic vectorization performance enhancement method based on ensemble learning. We construct an ensemble learning model to predict the performance of tensor contraction kernels with different vectorization strategies and select the best vectorization strategy for the kernels. According to the storage access patterns of the tensor contraction kernels, we propose a static method for features representation. Based on the multi-algorithm ensemble learning strategy, we obtain better learning results than the single learning algorithm. The experimental results show that the prediction model achieves 88% and 87% prediction efficiency on two different platforms with different instruction sets, data types, and compilers. Compared with the existing methods, the prediction efficiency is greatly improved. In addition, the average peak performance is 2.96× of Intel ICC 12.0 and 2.98× of GCC 4.6 compiler, respectively.

Issue tracking systems allow users to report bugs. Bug reports often contain product name, product component, description, and severity. Based on such information, triagers often manually prioritize the bug reports for investigation. However, manual prioritization is time consuming and cumbersome. DRONE is an automated state-of-the-art approach that recommends the priority level information of the bug reports. However, its performance for all levels of priorities is not uniform and may be improved. To this end, in this paper, we propose an emotion-based automatic approach to predict the priority for a report. First, we exploit natural language processing techniques to preprocess the bug report. Second, we identify the emotion-words that are involved in the description of the bug report and assign it an emotion value. Third, we create a feature vector for the bug report and predict its priority with a machine learning classifier that is trained with history data collected from the Internet. We evaluate the proposed approach on Eclipse open-source projects and the results of the cross-project evaluation suggest that the proposed approach outperforms the state-of-the-art. On average, it improves the F1 score by more than 6%.

Software refactoring is popular and thus most mainstream IDEs, e.g., Eclipse, provide a top level menu, especially for refactoring activities. The refactoring menu is designed to facilitate refactorings, and it has become one of the most commonly used menus. However, to support a large number of refactoring types, the refactoring menu contains a long list of menu items. As a result, it is tedious to select the intended menu item from the lengthy menu. To facilitate the menu selection, in this paper, we propose an approach to dynamic ranking of refactoring menu items for IDE. We put the most likely refactoring menu item on the top of the refactoring menu according to developers' source code selection and code smells associated with the selected source code. The ranking is dynamic because it changes frequently according to the context. First, we collect the refactoring history of the open source applications and detect the code smells. Based on the refactoring history, we design questionnaires and analyze the responses from developers to discover the source code selection patterns for different refactoring types. Subsequently, we analyze the relationship between code smells associated with the refactoring software entities and the corresponding refactoring types. Finally, based on the preceding analysis, we calculate the likelihood of different refactoring types to be applied when a specific part of source code is selected, and rank the menu items according to the resulting likelihood. We conduct a case study to evaluate the proposed approach. Evaluation results suggest that the proposed approach is accurate, and in most cases (95.69%), it can put the intended refactoring menu item on the top of the menu.

Linear power flow (LPF) is necessary for robust and fast centralized control of active distribution networks (ADNs). With penetration of distributed generators (DGs) into ADN, voltage-controlled nodes are becoming more common, to maintain a normal voltage profile in the distribution network. Existing three-phase LPF formulations under a rectangular coordinate system cannot cope with local voltage-controlled nodes, and does not consider multi-slack-node features of three-phase distribution network, detailed control characteristics and loss participation features of DGs. Here, a general three-phase LPF under a polar coordinate system is presented to address these issues. The proposed method can account for various connection ZIP loads, transformers, and single-phase or three-phase DGs. The detailed control model of the DGs and the distributed slack bus are taken into account. The effectiveness and advantages of the proposed method are validated with balanced 33, 70, 84, 119, and 874-node networks and modified IEEE 13, 34, 37, and 123 unbalanced networks.

In the cloud computing environment, the source-level energy consumption (EC) estimation is employed to approximately measure the EC of a cloud computing task before it is executed. The EC estimation on tasks is critical to task scheduling and source-code improvement in the aspect of EC optimization. The existing studies treat a task as a program, and EC of the task as the simple summation of each statement's EC. However, EC of two tasks consisting of the same statements with different structures is unequal; therefore, the code structure should be highlighted in source-level EC estimation. In this paper, an abstract energy consumption (AEC) model, which is static and runtime-independent, is proposed. For the model, the two quantitative measurements, “cross-degree”and “reuse-degree,”are proposed as the code structure features, and the relationship between EC and the measurements is formulated. Although AEC is not a precise EC measurement, it can properly represent the EC of a task, compare with other tasks, and verify the optimization effect. Experimental results show that the ratios between the EC and AEC with 50 test cases are stable; the standard deviation is 0.0002; and the mean value is 0.005. The regularities of EC and code structures, represented as “cross-degree”and “reuse-degree,”are also validated. Though AEC, it is easier to schedule the cloud computing tasks properly and further reduce the consumed energy.

In this paper, a novel time series heteroskedastic model is proposed for sea clutter modeling application. In the light of characteristics of the practical clutter at low grazing angle, the original generalized autoregressive conditional heteroskedasticity process, which has been widely used in various fields of economics, is extended from three aspects. First, the autoregressive moving-average terms are introduced for modeling the temporal correlation of both clutter returns and innovations. Second, the exponential of the conditional variance is generalized from one to arbitrary positive value, to capture the nonlinearity existing in the practical clutter. Third, the traditional Gaussian innovation is replaced by the Johnson Su random variable, which is a monotonic transformation of the Gaussian random variable and is capable of modeling the skewness and kurtosis. By systematically analyzing a large number of practical sea clutter data sets, we show that the proposed time series model fits the data better than some commonly used statistic-based distributions, such as the Weibull and compound Gaussian distributions.