Advances in Intelligent Systems and Interactive Applications

Aiming at the rolling force model of hot strip rolling mill, the forecasting technique based on ELM (extreme learning machine) is presented in this paper. Initially, the variables associated with control rolling are identified by the analysis of the traditional formula of rolling force, in order to ensure the effectiveness of the model, and then apply ELM network to predict models. Production data is applied to train and test the above network, and compare with the modified calculated value of rolling force, which got from the self-learning model of rolling force. The results show that the thickness can be predicted more rapidly and exactly, which can meet the actual demand of production, when this model and the rolling force learning model are integrated.

The big data of electric power industry contains the information of users’ values, credits, and behavior preferences. On the basis of the data, electric power companies can provide personal services as well as increasing profits. In this paper, we proposed a labeling system based on the clustering algorithms and Gradient Boost Decision Tree (GBDT) algorithm to establish user behavior profiles for State Grid Group of China, including basic information labels, behavior labels, behavior description labels, behavior prediction, and user classification. The experimental results showed that the approach can describe the behavior features of users in the electric power industry effectively.

In order to maximize the pan number placed in the oven and the even distribution of heat in a pan when baking brownie, we propose two mathematical models, named as “edge cutting model” and “weight evaluation model”, to get the solutions. The “edge cutting model” mathematically proves how the heat diffusion distributes over the outer edge of the pans with various geometric shapes by cutting the edge into small components and then deducing the functional relation between the heat distribution respect to the ratio of the surface area to the mass of each component. And we prove why the heat concentrates in the four corners of the rectangular pan and why the heat evenly distributes on the outer edge of the round pan. Further, we simulate the heat distribution results of different shapes of the pan when it is placed in the oven by using the software of the ANSYS. And the results are consistent with the theory analysis, which demonstrate the feasibility of the proposed model.

To solve the problem of safety avoidance of fishing boat in open water, the mathematical motion models of collision avoidance is established on the basis of geometry of collision avoidance. Taking into account that alteration of course alone is the most commonly used action to avoid collision in sufficient sea-room. Therefore, the intelligent steering control based on Radial Basis Function (RBF) neural networks is proposed and added in the above models. The simulation results verify the effectiveness of the models.

The purpose of this paper is to study the dynamical behavior of fractional order hyper-chaotic complex systems based on the bifurcation theorem. The variation of the system parameters and fractional order can induce the bifurcation by the simulation results.

The remarkable development of the logistics industry motivates the optimization researches on the logistics park planning. After the construction period of the logistics parks when the relative location and layout problems are highlighted by the scholars, the performance evaluation on the logistics parks during their operation period is of great significance. On one hand, the well-performed logistics parks can be determined and stand out from their peers through performance evaluation, and can hence provide a benchmark for others to learn and further make progress. On the other hand, logistics parks can analyze their SWOT based on such evaluation and realize their sustainable development. Consequently, the performance evaluation on the logistics parks has been attached great importance in recent years. In this study, we present a systematical review on the logistics park performance evaluation problem from two aspects, including evaluation index system and quantitative evaluation methods (e.g., AHP, TOPSIS and DEA, etc.). We wish this review can help the readers clearly understand the research progress of this problem and also draw more colleagues to this research field.

This study explores a multi-objective optimization for loading containers on double-stack cars that considers safety issues including lowering the center-of-gravity height and balancing the wheelsets’ load of the loaded cars. To effectively solve this problem, a Tabu Search algorithm with 2-opt and Tabu list techniques is designed to obtain the close-to-optimal solutions to the problem. In this study, two experimental cases are presented to demonstrate the efficiency of the proposed algorithm and its computational accuracy by comparing with the exact solution strategy proposed in our previous study. The experimental results indicate that the Tabu Search algorithm can obtain the optimal solutions to the double-stack car loading problem more efficiently.

During the prefabricated component (PC) production stage, for many kinds of PCs, production technologies are not the same. It is an important reason for high cost of PC production that reasonable group batch plan is lacked. By group technology and planning theory, an optimization model of grouping batch production for a variety of PCs is established. The model regards the minimum of the cost of production of PCs as the objective function. The size of the bottom mold, the type of embedded parts and connectors and the number of molds in the production process of PCs are constrains. By comparative analysis of application instance, the production cost under grouping batch for PCs is significantly reduced. It proves that the optimization model has a reference for production cost control under grouping batch for PCs.

Air quality reflects the extent of air pollution, and it is evaluated based on the concentration of air pollutants. Air quality level is traditionally assessed by mathematical formula, which cannot precisely represent the level of air pollution in some circumstances. This paper introduces a novel Air Quality Evaluation system based on Stacked Auto-Encoder (named AQES-SAE) to tackle these problems. The data of air pollutant concentrations and the air quality index are collected and clustered with improved K-means algorithm. The labeled clusters as training samples are then transferred to a deep neural networks and trained on the SAE model, which can be used as the classifier in the air quality evaluation. The proposed method is compared with traditional mathematical formula, and it is verified that the AQES-SAE system is prior to the traditional method because it can identify air quality level more accurately and more reasonably.

Taking Chengdu plain as the research object, using 2014 TM image and the data of social economic development, from the perspective of urban and suburban, studied the differences of the coupling relationships between landscape pattern and economic development in the studying area, the results are as follows: the coupling degree show that the coupling degree Chengdu plain is 0.411, the coupling coordination degree is 0.332, which is high strength and low coordination. The coupling degree is 0.498 and the coupling coordination degree is 0.436 in the urban, which is high strength and normal coordination. The coupling degree is 0.342 and the coupling coordination degree is 0.325 in the suburban, which is low strength and low coordination. The coupling differences in urban and rural between the economic development and landscape pattern is because the result by natural conditions and human activities, the human economic activity is the leading cause of urban and rural landscape pattern differences.

Clustering is a unsupervised pattern recognition method and an important research content in data mining and artificial intelligence. A new mathematical model of clustering is proposed using graph theory. It is proved theoretically that the model is a submodular function. According to it, a corresponding algorithm is proposed.

In this paper, we consider that a Gibbs phenomenon exists for the biorthogonal wavelets expansions of a discontinuous function. Firstly, we study properties of the biorthogonal wavelets kernel. Based on these properties of the kernel, we discuss existence of this phenomenon for the biorthogonal wavelets expansions of a discontinuous function. Also, we present the necessary condition of this phenomenon existence for the biorthogonal wavelets expansions of a discontinuous function.

In this paper, we give $$ {\varvec{\upalpha}} $$α-convergence theory of nets, ideals and filters by means of the concept of $$ {\varvec{\upalpha}} $$α-closed L-sets. Its applications are presented.

Software product line (SPL) engineering is an effective method to improve the software development process in terms of development costs and time-to-market by using comprehensive software reuse technology. The feature model is a demand model that describes the common and variability of software product family and the relationship between features in SPL engineering. The difficulty of product configuration based on the feature model is how to choose the optimal combination of features from the complex feature model to satisfy the constraints. In order to achieve the problem of constrained feature selection optimization, we propose a method based on atomic set and a genetic algorithm to optimize feature selection. Firstly, the feature model is optimized by using the atomic set algorithm. Then, the whole constraints of the model are modeled as the evaluation function of the effective and invalid configuration in the genetic algorithm. Finally, by the genetic operations of combining the effective configuration and the invalid configuration, such as crossover, selection and mutation, it selects the best effective configuration for output.

The approaches of product configuration based on feature are about how to select features from a feature model based on specific domain requirements and stakeholders’ goals. Although the literature on this topic has gained most importance in Academic and industrial fields, only little effort is dedicated to compare and analyze them. In order to address this shortcoming and to provide a basis for more structured research on feature modeling in the future, we first build a framework model to describe model structure, dependency management, automated support, configuration approaches and so on shared in the approaches family. Secondly, we understand and classify different configuration method based on the framework. Meanwhile we analyze the commonalities and variabilities among different approaches.

The development and application of image data has the characteristics of high resolution, large data volume and so on. Research on how to take advantage of MapReduce to distributed processing efficient and fast is one of the focuses and hotspots in the field of massive image data management. To solve the above problems, combing efficient distributed programming and running frame provided by MapReduce model and image pyramid algorithm, proposing and realizing a distributed model for massive image data. Experiment expresses that this model is good performance in massive image’s browsing and rooming.

QRS complex is with significant use in electrocardiogram (ECG) components. This paper uses a special algorithm to recognize QRS complexes in the ECG, which is based on relevance vector machines (RVM). RVM is used as a classifier to detect QRS areas. Performance of the adopted approach is validated by cross-validation. The experimental result has showed that, the RVM method has better classification effect, along with fewer parameter settings and fewer kernel functions than that of support vector machine (SVM). It can perform a satisfying result under the poor condition of the ECG signal.

The increasing amount of the automobile has led to many problems. How to carry out parking management work efficiently and reliably has become a vital part of the smart city development. In this paper, an intelligent parking system based on RFID technology is proposed, which identifies a vehicle by a reader and a tag connected to the car and achieves the management functions like non-conduct and cashless charging. The system could effectively deal with many problems of the traditional charging work such as inefficiencies, imbalance of resource allocation and loss of fare.

The application of robot is expanded unceasingly followed by the development of artificial intelligence, such as household and nursing, education and other services. The human–machine interactive mode changes fundamentally as the relation within them is closer, while emotional interaction is the prevailing trend. And thus there is a need of the guidance of cognitive psychology in robotics. This paper reviews the robotic emotion recognition, and then introduces the cognitive theory applied in robotics, including robot autism intervention, robot education and household service robot.

This paper presents a visual navigation method based on Improved ORB mobile robot. In view of the fact that ORB algorithm does not have scale invariance at feature point matching, an improved ORB algorithm is proposed based on the idea of SIFT algorithm. Firstly, the multi-scale space of the image is generated, and the stable extremum is detected in the multi-scale space, so that the extracted feature points have the scale invariant information. Then, the feature points are described by the ORB descriptor to generate the binary invariant descriptor. Using Improved ORB to extract the ORB features of the input scene, and combined with the robot odometer information to achieve robot navigation. The experimental results show that the method can accomplish the navigation task of the robot well, and it has some robustness to the dynamic information in the environment.

RelieF algorithm is a series of feature selection methods, including the first proposed RelifF algorithm and later extended ReliefF algorithm. The core idea is to give greater weight to the features which have great contribution to the classification problem. The algorithm is simple and has high efficiency. However, it is not ideal to apply the Relief algorithm directly to unbalanced multi-label datasets. Based on the Relief algorithm, this paper proposes an improved multi-label ReliefF feature selection algorithm for unbalanced datasets, called UBML-ReliefF algorithm, which effectively eliminates the problem of datasets unbalance. Finally, we use the stroke cases to do experiments, and the experimental results show that the multi-label classification effect in UBML-ReliefF algorithm is better than the original ReliefF algorithm.

Location-based topic modeling is an emerging domain to capture topical trends over the area dimension, which makes it possible to conduct further analysis on the various preference of users in different areas, such as investigating users’ social opinion tendencies, generating personalized recommendation and so on. In this paper, we proposed a novel topic model called Area-LDA to discover the latent area-specific topic. Previous works which based on pre-discretized or post-hoc analysis of area topics had no ability to help generate topic for a document. Different from previous works, our model extends the original LDA by associating each document with area factor and introducing a new area distribution over topics into the model. Therefore, for each generated document, the distribution over topics is influenced by both word co-occurrences of the document and word co-occurrences of the corresponding area. We present the experimental analysis over the real-world dataset and the results demonstrate the effectiveness of the proposed method to mine interpretable topic trends on different areas.

Interaction behaviors of users on different platforms on the Internet make user-generated content spread widely and become popular. How to model and predict the popularity of topic concerned by users is vital for many fields. Aiming at the problem of topic popularity prediction on microblog platform, a popularity prediction method based on similar topics and co-occurrence topics is proposed. The method is further evaluated with the Sina Weibo dataset. The experimental results show that our method can have relatively better performance in predicting topic popularity than the baseline methods.

Recently, data mining developed fast and attracted a lot of attention. When using data mining in real world, privacy protection is an important problem. Over the past ten years, many researchers study this problem and propose a lot of PPDM (privacy preserving data mining) methods. These methods can complete data mining task when protecting privacy. This paper gives a new Bayesian-based PPDM method, which is designed for classification. This method is a data perturbation method and is algorithm-independent, which means the perturbed data can be used by normal classification methods directly. Experiments show that comparing with existing methods, this new method perform better for protecting privacy, when they keeping data utility both.

In order to overcome the influence of traditional gesture recognition on the surrounding environment, illumination change and background, a gesture recognition method based on fingertip detection is proposed. Firstly, the depth image of Kinect is collected, and the threshold segmentation and color space are combined to complete the hand segmentation. And then calculate the curvature value of the palm of your hand, and obtain the fingertip and depression points by defining the value of curvature. Finally, the fingertip detection algorithm with the center of gravity distance is set, the distance from the palm to the fingertip and the depression is set, Fingertip mark points, complete gesture recognition. The experimental results were verified by 100 experiments, and the recognition rate was 97%. Experimental results show that this method can accurately carry out hand segmentation and fingertip recognition.

Bioinformatics applications have become more data-intensive and compute-intensive, which requires an effective method to implement parallel computing and get a high-throughput. Although there exists some tools to realize parallelization of BLAST, but most of them depend on complex platforms or software. A parallel BLAST is implemented using Spark, which is called Parka. The parallel execution time and speedup of Parka are evaluated in a cluster environment. Then, it is compared with Hadoop-based parallelization method. Results show that it is a scalable and effective parallelization approach for sequence alignment.

Although the target tracking is a difficult problem in the field of computer vision, it has a bright development prospect with the continuous development of video surveillance technology, and it adopts many kinds of algorithms whose effects are various. Based on the bibliography, the thesis introduces four kinds of tracking algorithms and analyzes their advantages and limitations. Adopting two or more algorithms, it can take the tracking research and also can improve the accuracy and efficiency of target tracking. Finally, it summarizes the problems and research trends of target tracking which need to be improved and also raise the prospective future of tracking algorithms.

Under the background of global ageing and empty nest families, the senior tumble behavior has becomes a focus problem in today’s society. In order to provide timely help for seniors, and relieve the injury of tumble to them, a judgment method of senior tumble behavior based on potential behavior association mining is proposed in this paper. Using clustering algorithm and the mining algorithm for potential behavior association rule, it calculates the similarity of the senior behavior feature, then extracts the senior behavior features, calculates correlations between behavior features in seniors, which can complete data mining in senior tumble behavior. The experimental results show that the proposed algorithm could greatly improve the accuracy of the senior tumble identification, so as to provide security for senior trips.

For the purpose of image distortion caused by the oblique photography of Android CMOS camera, a auto-rectification approach was proposed in this thesis. According to the camera stance information took into geometric correction model and provided by the phone built-in acceleration sensor and geomagnetism sensor, the relationship of pixel coordinates between distortion image and standard image was obtained. The pixel coordinates was resampled to obtain orthorectified images without the squint distortion by OpenCV4Android perspective projection transformation. The experiment results show that the approach is efficient, real-time and has great significance for implementing precise position of the target.

Fish-eye camera has been widely used in computer vision, mobile robots, photogrammetry and other fields for its wide-angle imaging. Fish-eye camera should be calibrated before being used, and in this paper, we first introduce several projection models for wide angle cameras, and choose the most commonly used equidistant projection, combined with the lens distortion model, to establish a calibration model for fish-eye camera. Then, by the linearization of the model, the least square solution of both the calibration parameters and the exterior orientation elements is given. Finally, through the experiment, it is concluded that the fish-eye camera calibration model can reach the ideal calibration precision with 0.3 pixels.

Algorithm for digital location and recognition of digital instruments in complex scenes is proposed. First of all, the digital instrument images collected during the inspection process of the substation inspection robot are preprocessed by gray scale, filtering and morphology. Secondly, looking for extreme area, and using extreme values for coarse positioning, and then extract the ROI area of the number. In the ROI region, the preprocessing operation such as filtering and morphology are performed again, and the extreme region is searched, the error window is eliminated which can be achieve accurate positioning and segmentation of digital. Finally, the digital feature vectors are extracted and sent into the SVM classifier in sequence. The determination of significant figure of result based on the first nonzero value, and the correct reading of the meter is exported. The experimental results show that the algorithm can realize the digital location and recognition of digital instrument in complex scene. The algorithm has high accuracy and strong robustness, which can be met the requirement of digital instrument for substation inspection robot.

As the pillar industry in the national economy, energy is facing a serious problem of overcapacity. Therefore the visualization study is demonstrated to more clearly understand the situation of tackling overcapacity of energy. The data source is based on the 4527 articles included in China National Knowledge Infrastructure database from 2000 to 2016, whose topics are all tackling overcapacity of energy. The tool SATI for statistical analysis of articles information is used to have frequency statistics of keyword and literature source, and to generate keywords co-occurrence matrix. Then the softwares of Ucinet and Netdraw are utilized to further draw the keyword co-occurrence of knowledge mapping for articles visualization, and to analyze the hotspots and tendencies in the area of tackling overcapacity of energy. From the research trend, tackling overcapacity of energy will be the hotspot for a period in the future. The paper is expected to provide the corresponding theoretical reference for the researchers in the field of tackling overcapacity of energy.

Key links and main influencing factors of battlefield rescue are analyzed and studied. The index system of prediction of battlefield rescue capability of brigade medical aid station is established. Through integrating queuing theory, fuzzy comprehensive evaluation, analytic hierarchy process and mathematical definition with exponential method, a hybrid algorithm for prediction of battlefield rescue capability is proposed. A simulation example of evaluation is presented. The validity and practicability of the comprehensive method is verified, and the logistics command decision and system optimum design of the battlefield rescue health service support are strongly supported by the prediction algorithm.

A new structure of synthesis filter bank with optimal synthesis filter and a receiver based on carrier frequency offset (CFO) are designed for the multicarrier (FBMC) system. Firstly, A synthesis filter bank (SFB) of a FBMC system is designed with dimensions. but the complexity will increase. An optimal dimensionality reduction method is proposed to overcome it. Secondly a receiver of FBMC system is proposed based on frequency-domain compensation technology (RDCT). Finally, the simulations are given to prove the advantages and the effectiveness for designed FBMC system.

It is of a great importance for ensuring the correctness of system reconfiguration information and the satisfiability of partition time requirement in safety and reliability of critical systems such as integrated modular avionics (IMA). This paper considers a configuration information model transformation and verification approach and scheduling validation of IMA systems in the model-driven architecture with ARINC653 specification. Considering the features of IMA systems such as time or space multi-partition, this paper firstly defines a semantic mapping from the core elements of reconfiguration information (e.g. modules, partitions, memory, process and correspondence, etc.) to the MARTE model elements, and proposes a transformation approach between system configuration information and MARTE models. Then, design a scheduling validation framework of IMA partition system and then use MAST tool to make simulation for the MARTE model to verify the schedulability. Finally, a case study is illustrated to show the effectiveness of above proposed approach.

Barcode has been widely adopted in many aspects, it is the unique identification and contains important information of goods. Regular barcode scanning device usually requires human being ¡¯s aids and is not suitable for multiple barcode scanning, especially in a complex background. In this paper, a cascaded strategy is proposed for accurate detection of 1D barcode with deep convolutional neural network. The work contains three parts: Firstly, a faster Region based Convolutional Neural Net (Faster R-CNN) framework is used to train a barcode detection model. Secondly, a powerful lo-level detector called Maximally Stable Extremal Regions (MSERs) is developed to eliminate the background noisy and detect the direction of the barcode. Thirdly, a postprocessing with like bilateral filter, called Adaptive Manifold (AM) filter, is applied when the image is blurred. We have carried out experiments on both Muenster Barcode Database and ArTe-Lab Barcode Database and compared with the previous barcode detection methods, the result shows that our method not only can get a higher barcode detection rate but also more robustness.

Real-time positioning system, circuit system, data transferring and remote monitor and control used for farmland operation are studied with Beidou satellite navigation system signal, and Locomotive operation management system used for crop harvesting is designed. Precise Beidou satellite positioning system, accurate area calculation method, Intelligent mobile phone APP software measurement are integrated in the system. Farm area, soil temperature and humidity and locomotive speed can be measured in real time. It also has the function of locomotive management.

In order to detect radial velocity ship in complex sea state, a novel detection algorithm is proposed. Considering that the frequencies of the ship and first-order sea clutters meet the requirements of sparse in the Doppler domain, certain sparseness constraints are imposed on Non-negative Matrix Factorization (NMF) to improve the efficiency of extracting the frequency spectrum of ship from that of the sea clutter. The following two topics are discussed: (1) the detection algorithm based on NMF (NMFSCC); (2) the detection results of NMFSCC, compared with Singular Value Decomposition (SVD) algorithm. Simulation results demonstrate the usefulness of the proposed approach.

Support Vector Machine (SVM) is an effective tool for classification problems. With the advent of the information age, tensor data problems are common in pattern recognition field. However, SVMs may lead to structural information loss and the curse of dimensionality when encounter into tensor data. In this paper, we propose a novel tensor-based classifier called the v-Twin Bounded Tensor Machine $$ \left( {\nu {\text{-TBSTM}}} \right) $$ν-TBSTM. It is an extension of $$ \nu {\text{-TBSVM}} $$ν-TBSVM. $$ \nu {\text{-TBSVM}} $$ν-TBSVM solves two smaller Quadratic Programming Problems (QPPs) instead of a larger one, meanwhile, it adopts the structural risk minimization principle. Compared with existing SVMs, $$ \nu {\text{-TBSVM}} $$ν-TBSVM has certain advantages. $$ \nu {\text{-TBSTM}} $$ν-TBSTM inherits all the advantages of $$ \nu {\text{-TBSVM}} $$ν-TBSVM, moreover, it utilizes the structural information of tensor data more directly and effectively, thus it gains better performance. The experimental results indicate the effectiveness and superiority of the new algorithm.

A Face Image Super-Resolution (SR) reconstruction based on Convolution Neural Networks is constructed. Firstly, extract two-level feature map by multiple convolution kernel images. Secondly, after each feature map is extracted, mapping the extracted features to another plane by means of a non-linear mapping method. Lastly, rebuilding the final SR images through adding all the second-level feature maps and plus a constant. The experimental results show that our method can get a better result in single face image SR reconstruction.

According to the image retrieval process the user need to submit many times and feedback the query, based on semantic correlation self-feedback algorithm, a user submits the demand, active feedback query, finally get to meet the needs of the results, simplify the query process, Enhanced the “fool” feature of the image retrieval system. The test results show that the algorithm has higher accuracy and higher flexibility.

Down syndrome (DS) is a genetic disorder caused by trisomy of all or part of the human chromosome 21. Since currently there is no cure for Down syndrome, the screening tests became the most efficient ways for DS prevention. Here, we used various supervised learning algorithms to build DS classification/screening models based on the protein/protein modification expression level of mice DS model Ts65Dn. Furthermore, we applied an adaptive boosted Decision Tree method to identify the most correlated and informative proteins factors that were associated with DS biological processes and pathways. Moreover, we improved the DS classification/screening models by using these selected DS related critical proteins. Finally, we used unsupervised learning algorithms to confirm the results we obtained above. These selected DS related proteins could be further used for protein-related and coding gene-related drugs developments.

We introduce a new visual analytics approach to the vehicle experiment data with multi-dimensional and time-series characteristics. Our approach integrates four visualization techniques to explore potential patterns hidden in huge amounts of actual experiment data. Overview view is the main view, which is used to show high level data characteristics of the experiment process and provide an intuitive and compact view for analyzing temporal and multi-dimensional patterns. Other three visualization techniques can be viewed as the complementary views, specialized in analyzing temporal features, relationships between different attributes and vehicle status variation, which are also important domain tasks. A case study has been conducted using a real-world dataset to verify the effectiveness and scalability of our approach.

This paper presents the motion characteristic and power response of a multi-section wave energy converter (WEC) consisting of two hinged cylindrical buoys and two power take-off (PTO) damping units at the joint according to a heaving and pitching motion model. The influences of geometry parameters of buoy (length, radius and draft) on the pitching motion and energy conversion ability are analyzed and summarized. The analysis results show that the relative pitching angle of two buoys depends on the pitching angles and the phase difference. The resonance frequency of pitching motion of buoy is mainly related to the radius and draft, but the phase difference generally depends on the length of buoy. The resonance of buoy and phase difference of odd multiples of 180° are necessary conditions for attaining maximum relative capture width.

The paper generalizes PinSVM to multiclass version by using binary tree structure. At each internal node, all inherited classes are first divided into two groups via farthest centroid selection. Then, PinSVM is constructed between two groups. When each group contains only one class, the leaf node can be identified. The experimental results show that binary-tree multiclass PinSVM is very competitive with one-versus-one PinSVM and one-versus-one SVM. Especially in terms of computational time, it has clear superiority than them.

Fractal feature can measure the complexity of radar signal. Radar signal is a kind of time series, which can be represented by fractal dimension. Wavelet transform is the signal processing microscope, which can observe the general situation and details of the signal and reduce the influence of noise. In this paper, the fractal characteristics of radar signals are extracted by combining the wavelet transform and fractal theory with the advantage of the two. Firstly, the wavelet transform is used to decompose the radar signals, and then the fractal dimension of radar signals under different decomposition levels is calculated. The radar signals with different complexity are obtained according to the difference of the fractal dimension. Finally, the validity of this method is verified by Matlab platform.

Moving Object Detection Based on Global Motion Video Sequence, Background’s movement mainly behave in the translation and scaling. Due to camera movement caused the background produces a zoom. This paper present an improved gray-scale projection algorithm. This method can quickly estimate the scaling parameters between the two images before and after, and then refer to the background of the reference frame. The background is mapped to the current frame image according to the scaling parameters. Using this algorithm to improve the three-frame difference method can accurately detect the moving target. Experimental results show, this method can effectively extract the target in the video sequence.

Facial expression recognition (FER) enables computers to understand human emotions and is the basis and prerequisite for quantitative analysis of human emotions. As a challenging interdisciplinary in biometrics and emotional computing, FER has become a research hotspot in the field of pattern recognition, computer vision and artificial intelligence both at home and abroad. As a new machine learning theory, deep learning not only emphasizes the depth of learning model, but also highlights the importance of feature learning for network model, and has made some research achievements in facial expression recognition. In this paper, the current research states are analyzed mostly from the latest facial expression extraction algorithm and the FER algorithm based on deep learning a comparison is made of these methods. Finally, the research challenges are generally concluded, and the possible trends are outlined.

When the traditional connected components labeling algorithm is used to label the connected components, it is necessary to scan the image multiple times, resulting in unnecessary repetition. In addition, additional analysis is required to obtain the characteristic information of the connected components. Aiming at the above problems, this paper proposes a connected components labeling algorithm for one scan based on the idea of “linked list” and “Multi-tree”. It only needs one scan to complete the merging process of the connected components, and can get the connected components feature information. The experimental results show that the algorithm has strong anti-noise interference performance and fast processing speed, and can be applied to the actual target extraction.

In this paper, we used the methods of bibliometrics and information visualization to quantitatively analyze papers in the field of metallurgy in the past 20 years. Methods of bibliometrics and information visualization can make scholars intuitively understand the research hot spots in the academic field and the trend of disciplines. This paper analyzed the overall study of the domestic and foreign technical development, we adopted keyword coexistence and the co-citation network map of literature. We revealed the distribution of metallurgical papers and the new hot spot on the field of metallurgy in recent years both here and abroad, and mastered the development trend of metallurgy field.

Motion related human activity recognition using wearable sensors can potentially enable various useful daily applications. In this study, we start from a deep analysis on natural physical properties of human motions, and then extract the implied commonly prior knowledge. With the prior knowledge, a hierarchical decision tree (H-DT) model has been proposed to recognize human motions and activities. H-DT has a multi-layer heuristic structure that is easy to understand. Support Vector Machine (SVM) has been selected as sub-classifier of each layer in H-DT. The experiment results indicate that the proposed H-DT methods performs superior to those adopted in related works, such as decision tree, k-NN, SVM, neural networks, and the H-DT has achieved a general classification rate of 96.4% ± 0.025.

Human motion monitoring widely used in medical rehabilitation, health surveillance, video effects, virtual reality and natural human–computer interaction. Existing human motion monitoring can be divided into two ways, non-wearable and wearable. Non-wearable human motion monitoring system can be used when it is not contact with user, so it has no influence on users’ daily life, but its monitoring range is constrained. Wearable human motion monitoring system can solve this problem well. But as the dataset is not complete, it is inconvenient for the research of related algorithms, and many researchers choose to build their own data acquisition platform. On the other hand, the traditional motion recognition platform which is based on inertial features, is difficult to identify static pose. This limits the application of wearable motion recognition technology. In this paper, we design a human motion recognition platform based on inertial features and positional relationship, which can provide a platform for data acquisition and dataset for the researchers who study algorithm of wearable motion recognition.

This paper mainly studies the characteristics of third-order computational GI whose imaging system is a single-arm system using a spatial light modulator (SLM) instead of a rotating ground glass in the classic GI system. To simplify the theoretical analysis of three-order computational GI, we assume that the two reference beams are equal and find that it is very similar to the theoretical results of the two-order GI. We analyze the impact of signal intensity fluctuation of three light beams on imaging. The numerical simulation shows that the intensity fluctuation of the light is the source of imaging information. We propose an inverse algorithm, which can improve the visibility of imaging effectively.

In this paper we study face recognition using convolutional neural network. First, we introduced the basic CNN neural network architecture. Second, we modify the traditional neural network and adapt it to another database by fine tuning its parameters. Third, the network architecture is extended to the cross database problem. The CNN is first trained on a large dataset and then tested on another. Experimental results show that the proposed algorithm is suitable for building various real world applications.

In this paper, S-system model is first presented to forecast stock market index. An improved additive tree model named restricted additive tree (RAT) is proposed to represent S-system model. A hybrid evolutionary algorithm based on structure-based evolutionary method and cuckoo search (CS) is used to evolve the structure and parameters of RAT model. Shanghai stock exchange composite index is used as example to evaluate the performances of S-system model. Results show that S-system model outperforms other traditional models, including neural network, wavelet neural network, flexible neural tree and ordinary differential equation.

Aiming at the detection requirements for a certain type of auto parts production and processing, an online detection approach to defect auto parts internal defect based on industrial endoscope, is studied in this paper. In this method, auto parts internal detection device is designed based on hard tube industrial endoscope and industrial camera, and the structure and working principle of the device are described detailedly, ultimately proceed in sliding defect detection making use of the texture block.

With the continuous development of new technologies such as global data, the internet, machine translation and speech recognition, language service industry, language education industry and even relevant vertical areas have undergone a series of profound changes. How to integrate innovation and breakthrough in translation teaching, scientific research and practice has becomes an urgent problem to be solved. Big data and artificial intelligence has brought new opportunities for the development of education, and language services play the core values of language for the “going out” of Chinese culture.

Amid global competition, the environments of industry businesses are being tested. With the development of new technologies, which can be created anytime and anywhere, talents have become the most important resources for enterprises, and are the driving force behind the growth and profits of enterprises. The quality of human resources determines the operational performance of enterprises, and the success or failure of enterprise education and training will affect the quality of human resources. This study is designed to explore the impact of education and training, as implemented by enterprises, on performance evaluation, and develops a model for a quality education and training system in order to analyze the overall structural equation modeling (SEM), as based on the relationship between Kirkpatrick’s four levels, organizational performance, and relevant literature. This study modifies Kirkpatrick’s L1-L4 four-level training effectiveness evaluation model and Phillips’ fifth level evaluation ROI model into a framework of “Education and Training Evaluation Scale” suitable for this case study. A questionnaire survey is conducted on the employees and directors of a multinational group, and verification and analysis are conducted based on AMOS SEM.On the basis of the data analysis: (1) The model relationship of education and training effectiveness, organizational commitment, and employee productivity is verified, and the impacts of education and training effectiveness and organizational commitment on employee productivity and related models are offered; (2) The resulting level of education and training produces significantly positive impact on the improvement of organizational performance; (3) The more trainees are satisfied with training, the better the trainees perform in improving their personal abilities and practical work application effectiveness. The resulting levels also demonstrate that education and training can enhance the skills and productivity of employees, increase employee loyalty, and reduce turnover and absenteeism, thereby, making a practical contribution to organizational performance.

The character recognition technology has been widely developed during these years, but the character recognition study for the new Tai Lue has lagged behind. As a result, the digital processes of new Tai Lue have been in troubles. To solve this problem, this paper proposed an online handwritten character recognition method of new Tai Lue based on online-random forests in on-line training model. Firstly, the handwritten new Tai Lue characters are preprocessed, and then the eigenvectors of these characters are extracted. Finally, the appropriate training samples and test samples are selected, online-random forests algorithm is used to train and test them. When it iterates ten times, the recognition rate reaches 87.86%. The experimental results show that online-random forests algorithm is effective in online handwritten character recognition for new Tai Lue.

To improve the transfer efficiencies when information is sent from Android to web server or vice versa, the information transmit method between them is studied. And to save information access time in Android, how to store the information in Android is proposed. How to send and extract the information between Android and web server is presented. JSON technology is put forward to achieve transfer information, the information storage way is analyzed and SQLite is used to store information in Android in this study.

Since the traditional annual physical fitness test for adolescents in schools lacks of real-time dynamic data collection and deep analysis, we propose a data analysis model mapping from raw dynamic data to physical fitness evaluation, based on a self-developed wearable smartband by which we collect dynamic data from middle school students in Beijing. Firstly, the model presents a preprocessing algorithm which consists of a smoothness priors approach (SPA) and a median filter (MF), aiming for preprocessing of both photo plethysmo graphy(PPG) signals and three-axis acceleration data collected from the wearable smartband. Secondly, the model implements physiological and physical index estimation to acquire heart rate (HR), blood oxygen saturation (SpO ₂ ) and exercise amount estimates from the preprocessed data. Thirdly, the model extracts several key features closely related to physical fitness evaluation from the estimated HR, SpO ₂ and acceleration data. Finally, a support vector machine (SVM) algorithm is employed for classification of physical fitness level of different smartband users. An application testing of our self-developed wearable smartband has been implemented in Tongzhou No. 6 middle school of Beijing. Experimental evaluation results demonstrate feasibility and effectiveness of both the smartband hardware/software and the proposed data analysis model.

In the “Internet +” times, the E-commerce has rapidly developed. But it still remains a problem whether consumers who are accustomed to the traditional face-to-face shopping way are willing to cooperate with this rhythm to accept the brand new E-commerce consumption pattern. So it is critical to improve consumers’ trust towards the E-commerce patterns and let consumers have the courage to try. This paper will take fresh produce as the breakthrough point and from two aspects of product characteristics and quality of service to analyze the influence of the quality of fresh agricultural products, the value of products, the quality of logistics services, the quality of interface design and the quality of after-sales service on customers’ trust. And we carry out practical investigation and then use statistical software to analyze the results of the valid survey and reach a conclusion. These conclusions show that these five factors all have an impact on consumers’ trust. The research results of this paper have guide meaning to the troubles how enterprises of E-commerce of fresh agricultural products improve trust in the minds of consumers and develop better.

Concerning the business architecture as the center of the enterprise architecture development, the TOGAF business architecture deserve to be modeled with a simple and definite model suit for the subsequent modeling phases to follow a clear requirement. We propose a modeling method named PROUD based on tailoring & mapping of the content meta-model of TOGAF business architecture. Furthermore, a 2-dimensional iterative modeling matrix is defined to ensure the derivation of the PROUD model instances from meta-models in gradient granularity. A simplified use case of gift exchange in Disneyland with MagicBand is illustrated for PROUD modeling example.

Information entropy and SVM can be applied to the intrusion detection system, the combination of the two, the user measured the inherent nature of the audit data or the implementation of appropriate data deformation, so that it can be applied to the model when the training data, experimental results prove that the two combination can be more efficient detection of abnormal intrusion.

Aiming at the recommendation accuracy, diversity and timeliness of music recommendation system, this paper puts forward the music recommendation based on user behavior analysis and user emotion extraction. User behavior analysis can analyze the music preferences, and establish user’s interest model. Collaborative filtering algorithm and user similarity calculation can be a good way to explore the user’s new interests. In addition, by analyzing user’s real-time text information in the social network. Using word2vec and clustering can help achieve the user’s real-time feeling. Combining user’s interests with user’s emotional needs, filter out the user’s current emotional music recommendation-list from the recommended music list. By this way, users get a better experience.

It is one of the basic and essential ability of university students to use computers and Internet to solve the problems of learning, living and working. University students are becoming the main victims of all kinds of security incidents due to the long time online, the large absolute number of Internet users, less social experience and the weak information security awareness. Some of them lose their money and some even lose their lives. University students should be taught the knowledge of information security so that they can deal with the increasingly serious forms of security to protect individual, social and national security. This paper discusses practical experience and exploration in the process of education from educational targets, forms of education, educational contents and assessment methods in information security.

The Wireless Multimedia Sensor network (~WMSNs~) differs from the traditional Wireless sensor network (~WSNs~) because of the direction and rotation of the multimedia sensor sensing area. The perceptual model was established, and the path monitoring capability of wireless multimedia sensor network was researched, and the path monitoring conclusions adapted to the multimedia network were obtained, and the simulation experiments were carried out, and the simulation results verified the correctness of the conclusion.

In this paper, we shall give concepts of fuzzy preopen set, fuzzy preclosed set,fuzzy preinterior, fuzzy preclosure on intuitive fuzzy topological space and we shall introduce precontinuou, preirresolute, preopen (preclosed) mapping between two intuitive fuzzy topological spaces. Moreover, we shall give their equivalent characterizations.

By using approximate smooth method of $$ l_{1} $$l1 norm, a new smoothed model to approximate traditional compressed sensing problem with small dense noise is established. Utilizing projection technique, two discrete-time projection neural networks are proposed for the optimization solution of concerned compressed sensing problem. Applying the definition of exponential stability, two exponentially stable criteria are also given to guarantee the state vector convergent to optimization solution.

Owing to its wide coverage and high observation density, remote sensing satellite microwave observation data has become the most data used in numerical weather prediction system. However, due to the influence of the uncertainty of the surface emissivity, most of the data collected from remote sensing satellites are the observations at higher-peaking channels over land. A large number of these observations at lower-peaking channels are discarded, so the accurate surface emissivity is the key of assimilation of remote-sensing satellite microwave observation data over land. This paper introduces several commonly retrieval methods used for passive microwave surface emissivity and discusses their advantages and shortcomings. At last, an evaluation of each retrieval method and a conclusion is made.

In order to enhance the damping adjustable working mode and the adjustment range, a new multi-mode switching adjustable shock absorber was designed based on a traditional hydraulic damper shock absorber. The structure characteristics and working principle were analyzed. The adjustable damper was controlled in the soft compression and soft rebound mode, the hard compression and soft rebound mode, the soft compression and hard rebound mode and the hard compression and hard rebound mode. The strut assembly and electromagnetic valve assembly was designed. The ride comfort test of full vehicle under random input and pulse input were conducted respectively. The results shown that the structure design scheme of multi-mode switching adjustable shock absorber was feasible, and the mean square root value of the weighted acceleration of the driver’s seat had a significant increase trend with the increase of the speed. The root mean square value of weighted acceleration increased with the increase of damping value under random road input. The adjustment range and comprehensive properties of multi-mode switching adjustable shock absorber for active suspension system are improved. It has important theory value and engineering application prospect for active suspension system and its control strategy.

This paper studies the system of patrol unmanned aerial vehicle (UAV) autonomous identifying, tracking the charging pile and distance calculation in the process of achieving innovative autonomous charging target. This paper firstly proposes an SRDCF-based (Spatially Regularized the Correlation Filters) identification and tracking algorithm. The algorithm extracts sift features within the scope of real-time image to match the existing template. Then, a minimum circumscribed rectangle around the target is created as the initial tracking box to determine target area. Afterwards, the target is tracked through training and detecting the location of area. Lastly, camera ranging module with monocular camera measures the distance between the camera and the target. Thus UAV body position relative to charging pile label target can be obtained. The real UAV experimental results shows that the target detection and tracking algorithm can accurately recognize and track the charging pile docking label under the condition of camera movement in the UAV normal flight. Compared with the STC (Spatio—Temporal Context) tracking algorithm, the improved SRDCF algorithm has improved accuracy and robustness obviously. In addition, camera ranging module can accurately measure the distance between camera to the target and the requirement of real-time performance and reliability is reached.

The forest fire is one of the great threats to the operation of transmission line. The reasons and prevent measures of forest fire induced breakdown of the transmission line were investigated, and the application of data analysis in this field was discussed in the present study. According to the data statistics on the forest fire induced breakdown of the transmission line, the occurrence of transmission line trip has apparent time and space rules. The flow chart of big data analysis on transmission line breakdown induced by forest fires with the data analysis was proposed. By using big data analysis, important information can be derived, which is helpful to the design of early-warning system, as well as the proper strategic decisions to control the damage of transmission line in forest fires.

Current detection means are faced with the single test parameters and are unable to detect the internal hidden defects; also, the point detection method adopted is of a low detection efficiency. In this paper, with the establishment of the three-dimensional model of barrel defects through the finite element simulation and the adoption of the line detection method based on the L (0, 2) mode guided wave, the quantitative relationship between the axial and circumferential size of crack defects and the echo reflection coefficient could be drawn, and a kind of nondestructive testing method of cannon gun tube crack defects based on ultrasonic guided wave will be put forward. Simulation experiments show that this method can realize high, quantitative detection of cannon gun tube crack defects.

It is one the hot topics that how to improve the success rate of target recognition in the field of target recognition. Aiming at the change of the appearance of the target in the multi-camera surveillance videos, proposed a new algorithm of target re-identification that based on the sparse representation and LCC dictionary learning. Firstly, we constructs a joint learning dictionary, then training the joint dictionary by using the target data in the surveillance images. Then extracting the main features and carry on sparse coding of the re-identification target. Finally, using the features and sparse coding to match all the target data in another camera. In the simulation experiments, the results proved the superiority of ours algorithm, reduced the computational complexity, and improve the success rate of the re-identification.

Due to high impedance when the distribution network fault over current relay can’t generate action, and therefore more difficult to detect its time and frequency information. At present, all kinds of methods in signal processing, S transformation is the most effective method used to extract the frequency distribution. In this paper, S transform the distribution network of high impedance fault detection and simulation results show that the distribution network fault detection method based on S transform high impedance can accurately identify the distribution network of high impedance fault event, the results can be of practical engineering has a certain value.

Nowadays, the scale of enterprises expands continually, and its use of the network is constantly changing. Most enterprises in the country and even abroad are equipped with branches of the organization, want to share resources, work together to improve the efficiency by linking the various branches of the network. How can these branches communicate safely on the Internet? Virtual Private Network (VPN) can realize the secure data transmission equal to the traditional private network, while cutting great costs of network’s establishment and maintenance. In this paper, the basic principles and several protocols of VPN technology are analyzed, and the IKE process is described in detail.

Small hydropower station, connecting with the power system through the distribution network, can lead to the changes of the detected current for protection relay and the protection coverage. This paper firstly analyzes the effects of the small hydropower station connecting with the IEEE7 power system in current protection theoretically, and calculates the corresponding parameters of current section I and section II protection of the original power distribution without small hydropower station in PSCAD/EMTDC environment. Lastly, it analyzes the maximum allowable capacity of the small hydropower station in the distribution network.

It is of great significance, especially for power quality, to realize the smooth switching and conversion between the two operation modes of Micro-Grid. In this paper, for the control of the two operation modes of Micro-Grid, constant power control strategy and Constant-Voltage Constant-Frequency control strategy are adopted, and analysis of these two kinds of control strategies. In order to solve the transient oscillation problem in the switching process between the two operation modes of Micro-Grid, a smooth switching control strategy is proposed. And then the strategy is applied to realize the smooth switching to ensure power quality in this paper. Lastly, a mode is constructed by PSCAD/EMTDC software to test whether the strategy is effective and practicable or not.

FTA is a NP-hard problem. When using the Binary Decision Diagram (BDD) to analyze a fault tree, the size of the BDD transformed by the fault tree increases as the number of basic events increases. When using the BDD method to analyze a large-scale fault tree, there is not enough memory to store the resulting BDD structure, and the process of obtaining the minimum cut set is time consuming. This paper introduces a simplified BDD (SimBDD) structure to solve these problems. SimBDD is a BDD whose scale linearly relates to the number of basic events in the fault tree. Therefore, when using this structure to deal with large fault tree, the structure is small and easy to store. The number of paths in SimBDD is small and the minimum cut set gets faster.

For dealing with the problem of sparse nonlinear signal reconstruction of wireless sensor networks (WSNs), a distributed sparse nonlinear signal reconstruction algorithm which is on account of square root unscented kalman consensus filter (SRUKCF) with embedded pseudo-measurement (PM) is proposed in this paper. The pseudo-measurement embedded square root unscented kalman consensus filter (SRUKCF-PM) is reestablished in the information form. By introducing pseudo-measurement technology into SRUKCF, a kind of distributed reconstruction algorithm is exploited to integrate the random linear measurements from various nodes of the WSNs, so that all filters are able to achieve agreement about the calculation of sparse nonlinear signals. The simulation results demonstrate the effectiveness of the proposed algorithm.

Traditional Chinese Part-of-speech tagging (POS tagging) typically use statistical approach, which makes it rely on hand-crafted feature templates. In this paper, we propose an effective attention based bidirectional Long Short-Term Memory model for Chinese POS tagging. In our model, first, the word distributed embedding features are used as the input of the network to go through network transformation. Then, the bidirectional LSTM network is used to encode the input tokens. What’s more, the attention mechanism is introduced into the hidden layer to make the model focus more on the related part. Finally, the state transition probability matric is explored and the log likelihood loss function is used for model training. In the experiment, we test the tagging model in Penn Chinese Treebank to observe its performance. The results show that the proposed model achieves competitive accuracies compared with state of the art works.

Sound source localization is a challenging task in adverse environments with high reverberation and low signal-to-noise ratio. To accurately localize the source through classification methods, the number of sub-spaces for classification decision should be large. However, this also causes high misclassification rate, leading to larger localization error, especially in adverse environments. In this paper, we propose a new weighted decision making method (WDMM), which can effectively improve the localization accuracy of the likelihood based classification algorithms, by revisiting and accessing the probabilities of the adjacent sub-spaces. The synthetic experimental results have shown that the average mean and average standard deviation of the localization errors from the 20 different acoustic environments by the proposed WDMM are only 0.42 and 0.21 m respectively in a 4.0 m × 4.0 m × 4.0 m room. The 20 different acoustic environments include the high reverberation T₆₀ up to 0.6 s and low signal-to-noise ratio to −10 dB. Compared to localization results without WDMM, the proposed method has reduced averages of mean and standard deviation localization errors by 35.8 and 55.2% respectively.

The research on the radio-frequency wireless energy transmission in the cavity has just started at present, which is limited to the feasibility and loss analysis of fixed size and working frequency. This paper deeply studies the point-to-point wireless energy transmission model in the cavity, analyzes the field and the energy loss energy of the transmission in the cavity. Based on the electromagnetic simulation software HFSS, the cavity simulation model is used to obtain the different influencing factors of energy transmission. The results show that the greater transmission efficiency and working bandwidth can be obtained when the transmitting and the receiving antenna are in the diagonal of the rectangular cavity when the TM220 wave is transmitted in the cavity. The longer the receiving antenna length is, the lower the transmission efficiency is. In addition, when the cavity size is larger and the resonant frequency is low, the loss is not great.

The development of location-based social networks (LBSNs) generates large volume of check-in data. Point-of-interest recommendation (POI) is important for users to find some attractive venues, sometimes when users are in some places far away from their living cities. However, POI recommendation is so difficult compared to the classical recommender system. Users may access only a small portion of POIs, with a sparse user-POI matrix. The bulk of the POIs accessed by users located in a near close to users’ residences, which leads it difficult to put in a good word for POIs when the user travels to a faraway region. Meanwhile, uses’ preferences may be different in various geographical regions. Different users may prefer to go to different venues at different time. From our paper, we present a novel model represented as probabilistic graphical model to describe users’ check-in behaviors, which can overcome the data sparsity for the users far away from their living cities. To demonstrate our proposed model can recommend effectively, we do experiments to calculate the precision. The results show our model can do recommendation effectively and efficiently.

With the development of robot systems, distributed control systems become more and more widespread, which leads to much more complex communication specifications and hardware structure designs. Since the communication structure has important influence on robot security and real-time property, we proposed using formal method to verify the repeater with xMAS and ACL2 in this paper. Giving out the data flow logic information can also keep the low level properties while avoiding state explosion. The formalization method and model logic can also be used in many areas.

Nowadays, Network Traffic Classification has got pivotal significance owing to high growth in the number of internet users. People use a variety of applications while browsing the pages of internet. It is very crucial for internet service providers (ISPs) to keep an eye on the network traffic. Most of the researches made on Network Traffic Classification, using Machine Learning Based Traffic Identification to collect data set from one campus network, don’t provide far better results. In this paper, we attempt to achieve highly precise results using different kinds of data sets and Machine Learning (ML) algorithms. We use two data sets, HIT and NIMS data sets for this work. Firstly, we capture online internet traffic of seven different kinds of applications such as DNS, FTP, TELNET, P2P, WWW, IM and MAIL to make data sets. Then, we extract the features of captured packets using NetMate tool. Thereafter, we apply three ML algorithms Artificial Neural Network, C4.5 Decision Tree and Support Vector Machine to compare the results of each algorithm. Experimental results show that all the algorithms give highly accurate results. But C4.5 decision tree algorithm provides 97.57% highly precise results when compared to other two machine learning algorithms.

This paper analyses the current situation and researches the fault cause and the common fault type in active distribution network (ADN). The fault diagnosis methods are summarized and the advantage of the common methods are compared. The general problems consist in the current fault diagnosis methods of ADN are researched in this paper. Combining the development trend of ADN and the power supply reliability demand, the further research direction of ADN fault diagnosis method is proposed, which lays the foundation of the AND self healing and protection control.

Networked control, especially wireless one, on account of the flexible and effective application it facilitates, has been increasing scholars with diverse background in various fields participate into such a hot issue. Due to the advance of new technologies and applications, relevant classic issues, such as delay, queuing and discussion on stability under a certain of effectiveness, have been brought up on a new height to research with new challenges. This survey paper presents an epitome of NWCS, which is short for Networked Wireless Control System, and related research in the past few years covered with the discussions range from characteristics to materializations.

With the rapid development of Internet of Things (IoT) technology and cloud computing technology, smart home received more and more attention, which focuses on integrating with the home life-related facilities and building efficient residential facilities and family affairs management system to get safe, convenient, comfortable and artistic home life. Firstly, this paper introduces some problems of traditional smart home, which analyzes the advantages of ‘Internet plus’ and cloud computing combining with smart home and proposes a new kind of smart home system based on cloud computing. At last, The simulation experiment proves that the new system can solve and improve the problems existing in traditional smart home system.

Applied the results of user’s behaviour analysis to optimize website structure, this essay accordingly refines the session on behaviour cluster. Thus the solution model for optimizing website structure based upon BS (Behaviour-Service) matrix has been put forward. This model demonstrates application solutions respectively involving page navigator recommendation and link relations adjustment.

The transfer of huge amounts of funds over international financial markets from commercial activities involving the banking industry must provide a reliable operating environment to ensure data security. This study explores the relationship between bank employees’ information security awareness and performance of information security governance by employing empirical methods and collected data through a questionnaire with the subjects being bank employees. The goal is to understand the effects of the subjects’ information security awareness on the banks’ performance of information security governance from the analysis results. The results demonstrate that information security awareness produces a significant positive impact on the performance of information security governance - that is, information security awareness significantly affects the performance of information security governance. Moreover, the better the bank employees understand information security, the more significantly the information security governance influences information security awareness.

The central city drives the development of the surrounding cities through the agglomeration effect of producer services. This paper measures the clustering abilities of producer services in central city of the three urban agglomerations through the location quotient coefficient. Conclusions can be drawn as follows: the agglomeration effect of the central cities of “Beijing-Tianjin-Hebei” is the largest, followed by the “Yangtze River Delta” and the “Pearl River Delta”. Based on this, this paper puts forward some policy suggestions for the healthy development of the three major urban agglomerations.

As the basic unit of Intelligent Transportation System (ITS), Vehicular Ad Hoc Network (VANET) is responsible for transmitting traffic related messages between vehicle nodes, so it is important to ensure communication security of VANET. However, because of openness of wireless channels, self-organizing of network topology and mobility of vehicle nodes, detecting OBU’s trust level in real time traffic environments is very difficult. This paper presents a dynamic node’s trust level detection scheme, which is based on uncertain reasoning algorithm. Furthermore, the proposed scheme is integrated with asymmetric encryption to implement dynamic OBU’s identity authentication.

Aimed at the problem of slow convergence of genetic algorithms, an improved genetic algorithm is given. The improved algorithm is based on the traditional NSGA, and uses crossover limited and elitist strategy to solve multiple objective network optimization. By these operations, the algorithm can effectively improve the speed of convergence. At the same times, the algorithm also uses objective layer method to select the non-dominated individuals. The current algorithm is used to solve the network optimum design problem with multiple objectives. Two objectives, the cost of path and delay of path are considered. A few examples for the network optimization generated by randomly are used test the algorithm. The result shows the algorithm can find better Pareto solutions.

Physics problem understanding in natural language (NL) is always a challenge for machine. To address this challenge, this paper proposes a general approach to extract physics objects of middle school physics problems in Chinese. According to the appearing of object name and unit in a given physics problem, this approach use unit semantic model to extract physics objects. In the process of physics object extraction, an unit semantic model described by the components of semantics, unit and output is proposed to identify physics objects. The effectiveness of our proposed approach was examined in a dataset collected from the textbook. Experimental results demonstrated that the proposed method is very effective in identifying physics objects.

Paper fragments reassembly has been playing an important role in many places such as public security and even archaeology. Combined with the travelling salesman problem, a novel approach based on the matching of greyscale difference matrix is adopted. Experimental results demonstrate its potential in speed, accuracy and less human intervention for double-sided paper fragments reassembly. The study may provide a new direction for the automatic stitching or image mosaic technique.

Memetic algorithms combining with effective constraints handling techniques have been becoming the focus of concern as well as the subject of substantial research issue in decision-making in complex systems. As one of the most effective techniques for constraints handling, level comparison based method has been used to solve various constraint optimization problems. However, in most of the existing research, the constraint satisfaction level keeps constant or follows a fixed regulation during search, which may weaken the efficiency of relaxation of constraint violation. In this paper, we propose a TLBO based constrained optimization method, labeled as TLBO–LCSQP by incorporating a novel level comparison technique to effectively handle the constraints, and the sequential quadratic program (SQP) to enhance the searching performance. According to the simulation on a well-known constrained benchmark, the proposed TLBO–LCSQP could effectively enhance the constraints handling efficiency and greatly improve the searching ability.

A field programmable gate array (FPGA)-based real-time processing architecture for recurrent neural network (RNN) is proposed and presented; the proposed FPGA processing architecture is based on echo state network (ESN) and can get the output weights of RNN in real-time. The proposed architecture and the performance have been verified on an Altera FPGA chip. Experimental results show that the real-time hardware RNN can be trained to recognize different duty cycles of the input signal. We also performed experiments to investigate the ESN demand for resources and systems convergence in FPGA.

In view of the problem that the result accuracy of the research publication recommendations of manuscripts is often low. It is very difficult for many contributors to find the suitable journals from many recommendation results. To solve this problem, a novel method of publication recommendations for authors’ manuscripts is proposed using our improved C4.5 decision tree algorithm. For the shortcomings of too many values of data samples during dealing with the traditional C4.5 methods, the five-value logic ideas are adopted to improve this C4.5 algorithm to apply in the research fields of publication recommendations. The experimental results show that the related publications are easier to obtain for authors via the decision trees with the improved C4.5 method than before, and this method has higher accuracy than ever. It is of great significance to help the researchers, especially for those who have no enough experience to choose the most suitable journals from a lot of publications.

The multiple data centers storage system of clouding SCADA is proposed and the replica consistency maintenance policy is designed in paper. The oil and gas pipeline clouding SCADA multiple data centers storage system can support storage enormous capacity data, expanding in data center lever and independent reading and writing for each data center. The two stage lock replica consistency maintenance policy can ensure the consistency of data copy of multiple data centers.

This research aims to investigating the language learning strategies use of foreign language learners. The computer program SPSS (Statistical Package for Social Science) was adopted, and a questionnaire with 24 items about learning strategies was made to collect data from 32 Chinese university freshmen. The study of the collected data shows a positive relationship between learning strategy use and efficiency, which also verifies that the deep influence of the conventional learning strategies prevents those students from employing many other new ones to improve their foreign language skills.

This paper discusses the design and implementation of a transformer intelligent monitoring system based on the FPGA Stratix IV. The monitoring system utilizes the chip resources of FPGA, and uses precise rectification circuit and amplifying circuit to collect data of voltage and current. Pulse Value is converted from the electric energy with the sensor and the pulse conversion circuit, and is sent into the FPGA for processing. The software part of the system can complete the monitoring of power transformer parameters and running state, including the record of voltage, current, reactive power, apparent power, frequency, power factor, and automatic formation of various power curve analysis report, monthly management curve report, etc. The system adopts double anti-interference measures, and can be used on transformers of self closing railway lines and various substations.

Successful prediction of mutual fund with maximum accuracy is a great challenge because of highly fluctuating behaviour of the financial market. The prediction of Net Asset Value (NAV) of mutual fund helps investors to wisely plunge money into profitable mutual funds. This survey covers more than 40 related published articles in the field of mutual fund and gone through a systematic review of the various nature inspired techniques, used for NAV prediction with its performance analysis. The performance of mutual fund is highly correlated with the stock market, hence some of the stock market prediction analysis is also well throughout. Through this survey it is found that very few works is done in the field of NAV prediction using optimization techniques while for performance analysis of mutual fund many nature inspired optimization (NIO) techniques have been employed. On the whole, this paper gives a comprehensive review of the literature in the field of mutual fund.

In order to design quantum circuits for LNN, in this work, we propose a kind of synthesis algorithm based on the pre-evaluation of quantum circuits for LNN. Through pre-evaluation, not only can the algorithm accurately calculate if there are deletable redundant SWAP gates and remove them, but also convert every non-adjacent quantum gate to adjacent quantum gate, with inserting a minimal number of SWAP gates, and therefore get quantum circuits of minimal quantum cost (qc). As for quantum circuits of n lines and m quantum gates, the time and space complexity of the algorithm and optimized algorithm is O(m^3) and O(n^n+m), respectively. The results present that, with fewer average gates of quantum circuits and higher improvement efficiency of quantum cost, the algorithm has a wider range of application compared to the existing algorithm.

An SIR epidemic disease model with special nonlinear incidence was dealt with in this paper. By constructing the Lyapunov function, the global stability of the infection free equilibrium is proved when one is greater than or equal to the basic reproduction number; and the global stability of the unique infection equilibrium of the system is also proved when one is less than the basic reproduction number. At last, the results were verified by computer simulation.

For structural reliability analysis, the response surface method is popularly used to reduce the computational efforts of numerical analysis. The general method in the response surface method is to use the least square regression method. To give higher weight to the points closer to the failure curve, an improvement response surface method based on weighted regression for reliability analysis is presented, the new weight function is built based on parallel circuit theory, because the closer sample points are to the failure curve (the smaller the branch resistance is), the higher weights are given (the greater the branch current is). Numerical applications are provided to indicate the significance of the presented method.

Maximizing or minimizing submodular function is widely used in combinatorial optimization problems. In this paper; we present an approximation algorithm for maximizing submodular function subject to independence system that be represented as the intersection of a limited number of matroids, and discuss its performance guarantee.

According to the factors of affecting the employment quality for the university graduates, we established the index system of evaluation on the employment quality for the university graduates. The factor analysis was used to establish the model for evaluating the employment quality for the university graduates. Several data sets were analyzed to illustrate the proposed method with the SPSS software. The results not only included the employment quality for the graduates from different colleges, but also included the employment quality for the graduates from different departments in one university. The results coincide with the statistical data about the employment on the official website. Some advice and countermeasures were given on the personnel training and the employment promotion to the universities and education authorities, which can provide the regulatory on employment quality and services of the graduate employment.

The overall efficiency of multi-robot system is closely related to the cooperative control algorithm, so the research on cooperative control algorithm has been a hot topic in the field of multi-robot. In order to improve the overall efficiency of the multi-robots cooperative system, this paper proposes a Distributed Collaborative Control Model based on improved contract net (DCCM). According to the characteristics of the robot cooperative system, the negotiation and evaluation rules of the multi-robots cooperative system are proposed. This model combines the traditional contract net model with the broadcast algorithm to solve the shortcomings of the traditional contract net. This model solves the bottleneck problem of the traditional contract net model in the bidding evaluation stage, and also reduces the waiting time of the robot. With the help of the supervisory mechanism, the problem of the loss of the traditional contract net model in the task execution stage has been solved, and the stability of the system has been improved. Through the experiment of multiple trackless robots, it proves that this model can effectively reduce the execution time of the whole multi-robots system.

As technology continues to progress and develop, the electromagnetic environment becomes more and more complex. So in the research field of modern radar, it is very important to model clutter accurately. In this paper, in order to overcome the spectral broadening matter, we set up a new architecture based on Chebyshev filter and propose a novel clutter modeling method for K distribution clutter. In simulations, curves demonstrate that the estimated value of the improved Chebyshev filter method is more close to the theoretical value than the original zero memory nonlinearity method in the aspects of probability density and power spectral density, and the improved Chebyshev filter method is valid. In the end, the conclusions are given.

Due to the low viscosity of seawater, it is difficult to form the seawater-lubricated film, and the bearing capacity and stiffness of the seawater-lubricated film is very small. It is easily to cause the “overload” and “burning” phenomenon of the seawater-lubrication sliding bearing, and the operation stability and service life can be shorted. The magnetic bearing takes the bearing form of non-contact suspension, and it is more suitable as an auxiliary support to the seawater-lubrication sliding bearing. Therefore, the paper introduces the electromagnetic suspension support into the seawater-lubricated sliding bearing. And then a novel Magnetic-Liquid Double Suspension Bearing with the advantages of electromagnetic suspension and hydrostatic supporting can be formed. The structural characteristics, supporting mechanism, hydrostatic self-adjustment and electromagnetic-adjustment processes of Magnetic-Liquid Double Suspension Bearing can be analyzed in the paper. Based on force balance equation, electromagnetic equation and flow equation, the transfer function of different adjustment processes under constant-flow supply model are deduced. Then adjusting time, dynamic stiffness and phase margin are selected as dynamic indexes. The influence rule of the liquid film on capacity property of single degree freedom bearing system of Magnetic-Liquid Double Suspension Bearing can be analyzed. The results show that as liquid film thickness increases, dynamic stiffness decrease and adjusting time increase, and phase margin remain the same during the hydrostatic self-adjustment process. The proposed research provided a basis for the design of Magnetic-Liquid Double Suspension Bearing in the engineering practice.

Design of unequal-area oil pocket can decrease required pressure and flow of primary oil-supply oil pocket in open-type self-adaption oil-supply liquid hydrostatic slide. But anti-vertical-loads and anti-overturning-loads ability is quite different in composition from hydrostatic slide with equal-area oil pocket design. Taking single-row oil pocket group as research subject, ma matics relationship expression between length of primary oil-supply oil pocket and bearing capacity and stiffness of oil pocket group under vertical and overturning load is presented. Results indicate that with length of rimary oil-pocket increases, vertical bearing capacity and stiffness of unequal area oil pocket group is greater than of which equal area oil pocket group. But overturning bearing capacity and stiffness is smaller and anti overturning ability gets worse.

Due to the low viscosity of seawater, it is difficult to form the seawater-lubricated film, and the bearing capacity and stiffness of the seawater-lubricated film is very small. It is easily to cause the “overload” and “burning” phenomenon of the seawater-lubrication sliding bearing, and the operation stability and service life can be shorted. The magnetic bearing takes the bearing form of non-contact suspension, and it is more suitable as an auxiliary support to the seawater-lubrication sliding bearing. Therefore, the paper introduces the electromagnetic suspension support into the seawater-lubricated sliding bearing. And then a novel Magnetic-Liquid Double Suspension Bearing with the advantages of electromagnetic suspension and hydrostatic supporting can be formed. The structural characteristics, supporting mechanism, hydrostatic self-adjustment and electromagnetic-adjustment processes of Magnetic-Liquid Double Suspension Bearing can be analyzed in the paper. Based on force balance equation, electromagnetic equation and flow equation, the transfer function of different adjustment processes under constant-flow supply model are deduced. Then bearing capacity, static stiffness and total power loss are selected as static indexes. The influence rule of the liquid film on capacity property of single degree freedom bearing system of Magnetic-Liquid Double Suspension Bearing can be analyzed. The results show that as liquid film thickness increases, static stiffness decrease, carrying capacity and total power loss remain the same during the hydrostatic self-adjustment process. The proposed research provided a basis for the design of Magnetic-Liquid Double Suspension Bearing in the engineering practice.

Deep excavation of large data provides immeasurable value in agricultural application, such as agricultural market consumption. This paper initially outlines the application status of consumption data for Beijing agricultural products market. Then, the analytical ideas of agricultural market consumption data are explored and the difficulties for building application architecture are investigated. Ultimately, the corresponding counter measures are proposed based on the aforementioned analysis.

The issue of path selection in workflow can be resolved by establishing a model of decision tree. The solution presented in this paper is based on analysis of workflow application in real world and probability theory. The solution can provide accurate and reliable runtime data for workflow engine to determine the future execution paths. The data provided by our solution has significant impact on the decision making of workflow, which contributes to optimization of workflow in many aspects. Finally, the efficiency of business process in real world can be greatly improved.

In this paper, a novel formal language Z is adopted to describe the state spaces of workflow process models. We construct the state space for two aspects of workflow process: control flow and resource management. The formulation proposed in this paper can guarantee the correctness of process models to some extent. At the same time, it provides a useful tool for checking the consistence of workflow process models and real processes.

The emergence of PCBA heavy disc soldering tin bad phenomena in two times of furnace process, the failure of the pad, a furnace a pad, not a furnace of solder surface observation, analysis of the FIB sample preparation section, the reason to search the AES failure surface composition analysis. The results showed that: the the failure pad at second times in front of the furnace has been oxidized, and the pad surface tin thickness dramatically thinned, resulting in a pad of tin.

No matter what the welding technology, should ensure that meet the basic requirements of welding, welding to ensure good results. High quality welding should have the following 5 basic requirements: 1. appropriate heat; 2. good wetting; 3. appropriate solder joint size and shape; 4. controlled tin flow direction in the welding process; 5. the welding surface does not move to have enough solder joint life, we must ensure that the shape and size of solder joint with welding end structure. The mechanical strength of the solder joint is too small, unable to withstand the stress in use, even after the welding stress is unable to bear. But once in use began to appear fatigue or creep cracking, the fracture speed is rapid. The shape of the solder joints will cause bad homes from the phenomenon of light, life expectancy shortened the solder joint.

This paper alms to raise a instruction set simulator prototype based on ARM7. The simulator complete the application-simulate through execute the application instructions’ fetch, decode and perform. In this paper, we realize the simulation of the ARM7 3-stage pipeline, decoding logic, the implementation of logic, a variety of abnormal patterns of switching and other basic functions.

This paper based on the Vector Control theory uses the MATLAB/SIMULINK simulation software to simulate the rotor flux current model and rotor flux voltage model in the asynchronous motor. It comes to a conclusion that the two waveforms produced by simulating the rotor flux current model and voltage model are credible under the ideal conditions through comparing the two simulation waveforms. After the waveform simulated by asynchronous motor model, which has been transformation in different degree, is compared with the other simulated by original asynchronous motor model, it finds that the changes of motor parameters have a great influence on the rotor flux current model and that the integral part of the motor model affects the rotor flux voltage model. Thus, the following conclusion can be drawn that the current model is suitable for low speed operation of the motor and the voltage model for high speed operation.

A fuzzy PID algorithm is proposed and implemented for the temperature control of the insulating oil of the test cup in the overvoltage withstand tester. Due to the specific heating of the insulating oil and the high viscosity, it is difficult to control the temperature. It is necessary to design other method to control the temperature rather than the conventional control method. Combined with some characteristics of temperature control, a method of fuzzy PID control is proposed and validated by MATLAB for the adaptive control ability. The simulation results show that the control strategy has good robustness and stability, and the control precision is higher than conventional PID control.

According to the impact principle, the response characteristics of the newly designed anti-high overloaded steering gear under the half sine wave impulse excitation are studied, and the effect of the buffer structure is verified by the combination of the finite element simulation. At the same time, the establishment of a finite element under impact load model of the ultrasonic actuator, using the explicit algorithm in the simulation environment in the transient process of impact load.

This paper investigates the design and analysis of resonator for the resonant accelerometer. The difference in the resonator resonant frequency in driving and sensing mode and the resonator sensitivity were found to be key factors in determining the performance of the resonant accelerometer. The methodology suggests a simple way of designing and analyzing the resonator. Firstly, the three structural parameters that the width b, the thickness h and the length L of the resonant beam were optimized. And the sensitivity of the resonator to the axial force is related to the three structural parameters. Moreover, simulations of resonator performance are achieved using ANSYS finite element software. The simulation results show that the sensitivity of the resonator 1 is 3.35 Hz/μN, the sensitivity of the resonator 2 is 4.06 Hz/μN. At the same time, the linear degree of sensitivity of the resonator 1 is better than that of the resonator 2.

The damage scenario of toys has a high degree of uncertainty. The parents with children at the age of 0–14 have been investigated by the method of stratified random sampling. The investigation contents include basic information, the possibility of the usage scenarios of children’s toys, the possibility of injury under the usage scenarios, the severity of the injury, and recommendations and measures to reduce the safety risk of children’s toys. This paper analyzes the factors related to the quality and safety of children’s toys, including the correlation of parents’ age, education background, occupation and toy purchase channels, and the correlation between the purchase channel and product quality safety level. By using stata12.0, this paper analyzes the relationship between parental heterogeneity and its safety awareness, as well as the relationship between parents’ safety consciousness and purchase channel.