Recent Developments in Intelligent Computing, Communication and Devices

For cutting-edge protected agriculture, it is a leading trend to build an unmanned plant factory that requires robotic machinery as basic autonomous production units. However, the scheduling and cooperation among agricultural machines are rarely investigated. This paper reports on a modular logistics system based on the motion decomposition, and the objective is to develop a fully interrelating intelligent agricultural production system. Firstly, an unmanned ground vehicle is developed for transferring the intelligent agricultural machines such as agricultural robots. In consideration of the complex terrain features, both automatic mode and manual mode are available. The multifunctional smart remote-control mode is designed for users to control the logistics system accurately and the application of tablet PC provides an excellent user experience. Secondly, a control and mechanical system is designed to dock and transfer a harvesting robot. The safety and reliability of this system are jointly ensured by its movement pattern, the design of guide rails on the slope, and the precise positioning. Taking the performance on the test field into consideration, because the performance showed efficiency and accuracy of the system, future plans are highlighted to indicate the technical trends of unmanned plant factory.

From the past decade, data mining is becoming more important and tremendous amount of work is being explored in the healthcare industry, where most of the applications are introduced which could be classified into two branches: the bench of decision support and the development of policies. But still, it is hard to discover the valuable literature in the area of health care. This review paper initiates the current research overview which is carried out in the healthcare industry by using various DM techniques and algorithms to envisage different diseases such as cancer, diabetes, and HIV, and diseases that are related to skin and their accuracy ratio, and the abrupt discovery is presented to determine the paper.

From the tourists’ risk perception in the tourism supply chain, an evaluation approach that based on 2-tuple linguistic model and quality function deployment is proposed. House of quality using 2-tuple model is constructed to compute the preference of suppliers’ service factors. By using multi-granularity linguistic term aggregation approach, non-homogenous decision information problem can be solved successfully. The process of assessment criteria based on fuzzy linguistic term sets has been discussed, and at last a case study on a travel operator’s evaluation of suppliers with risk perception solved by the proposed approach is illustrated.

This paper firstly examines the total factor productivity on Nanshan economic growth by Cobb–Douglas production function in the background of Chinese supply-side structural reform. Then, we attempt to apply stochastic frontier production function to decompose total factor productivity into technological progress and technical efficiency and calculate the contribution to economic growth. Finally, the results of the analysis are used to comprehensively evaluate the quality of economic growth in Nanshan.

Spammers often embed text into images in order to avoid filtering by text-based spam filters, which result in a large number of advertisement spam images. Garbage image recognition has become one of the hotspots in the field of Internet spam filtering research. Its goal is to solve the problem that traditional spam information filtering methods encounter a sharp performance decline or even failure when filtering spam image information. In this article, the adaptive multipoint moment estimation (ADAMM) gradient optimization algorithm is first proposed and then applied to the convolution neural network to form a real-time garbage image recognition model weighted spatial pyramid networks (WSP-nets). Finally, MNIST, ImageNet and ImageSpam are used as the training and test data sets. Compared with other garbage image recognition models, our real-time image spam filtering model reaches a new state of the art.

Machine vision (computer vision) technology and image processing has been widely used for the automatic classification of fruit—representing an area for future research. In this chapter, we mainly use the method of machine vision to study color characteristics, size contours, and surface textures of fruit surfaces in the hue, saturation, and value (HSV) color space. The chapter goes on to realize the automatic classification of Hainan tropical fruit presenting typical characteristics.

This chapter summarizes ideas about thinking and mixed learning, summarizes common learning theories, and analyzes the characteristics of learners participating in basic computer courses at universities, whilst supported by standard university teaching practices. This chapter analyzes learning processes and the effects of learning for participants taking basic computer courses at university. It analyzes and summarizes the validity and feasibility of a hybrid learning model, based on computational thinking.

Recent years have witnessed the rapid progress of information technology, especially the high prevalence of mobile phones, which have penetrated into groups of various ages, especially the elderly who has never used mobile phones and the Internet before. Based on eight database platforms at home and abroad, this paper analyzes the affecting factors of the elderly’s use of mobile phones by literature retrieval. We identify that the influencing factors for the elderly’s use of mobile phones are mainly: characteristics of the elderly, individual motivation, social population, and technical convenience. Based on the above research, this paper also puts forward the basic framework that affects the elderly’s use of mobile phones in China, making people pay more attention to and understand the Chinese elderly’s use of mobile phones, providing theoretical and practical guidance for enterprises’ management of the elderly and corresponding product design and service.

Aiming at the problem of manual design, slow speed, and low efficiency in traditional face detection, this paper proposes a face detection algorithm based on Fast R-CNN. In this method, CNN algorithm, the Haar-Adaboost algorithm, and the candidate search algorithm are used to detect the candidate region of the face that may exist in the image. Then, the candidate region is input to the trained convolution neural network, and after a series of convolution and pooling operations, the final convolution feature (ROI) is based on the Fast R-CNN network structure, and finally, the feature is input to the two full connections (ROI), and then the characteristics of the input image are not specified. Branch, respectively, for feature classification and coordinate regression calculation. The experimental results of the proposed algorithm on the LFW test data set show that the proposed algorithm is compared with the traditional PCA + SVM algorithm, the traditional which avoids the manual design features and higher Identify the accuracy rate and speed up the detection time.

To solve the problem of “information overload,” personalized recommendation has played an important role; collaborative filtering algorithm in the personalized recommendation system occupies a pivotal position. Data sparse of users or projects limit the use of traditional collaborative filtering algorithms. In this paper, we optimize the recommendation of data sparseness to collaborative imprecision and use information entropy to optimize the similarity calculation in collaborative filtering algorithm. The two difference users concerned about the same number of items, the greater the weight of the parameter. The experimental results show that the improved algorithm can alleviate the problem of poor recommendation of data sparseness to a large extent.

Image segmentation is a fundamental component in image processing, object tracking, and clinical research. Many scholars have proposed various techniques; however, fuzzy c-means (FCM) algorithms have been proven effectively. In order to retain more image details as well as canceling image noise, a novel clustering algorithm based on the local coefficient of variation is presented in this paper, termed as LCVFCM. The major characteristics of LCVFCM are: firstly, the local gray similarity matrix is modified based on the differences between median pixel and neighborhood pixels to reduce the impact of noise on image pixels, and secondly, a new fuzzy factor is established by introducing local coefficient of variation to improve the performance of LCVFCM. The corresponding experiments with a large number of synthetic and real images prove our model achieves excellent segmentation capability and has stronger anti-noise ability, as compared with other segmentation methods.

The function plays a vital role in guaranteeing and improving performance, quality, and effectiveness of system task information. However, the health degree of resources directly influences function health. Comprehending the call relation of functions together with resources’ effectiveness from the data of functions’ resource effectiveness matrix will enable us to get a better understanding of the affiliation between resources and functions. Most methods proposed mainly focus on designing complex algorithms to extract dense-connected resources from one function’s resource effectiveness matrix. In this paper, we proposed a method to mine frequent dense resources across multiple real-value function’s resource effectiveness matrixes. The experimental results show our algorithm can extract frequent dense resource sets that satisfy the specified conditions in a more efficient way.

Recently, symbolic execution has been witnessed great success in applying symbolic execution to automatically generate high coverage test inputs for programs testing. While path explosion still a primary challenge in scaling concolic testing to large programs, meanwhile constraint solving is fundamental to concolic testing as a constraint solver is continuously invoked during analysis, while the main roadblock to performance of concolic testing engines is the time spent in constraint solving. To address this issue and achieve the purpose of selecting the path with simplest constraints to solve, this paper presents a new heuristic search strategy which determines the path priority based on the complexity of path constraints. We have implemented the strategy in angr, a binary-oriented concolic testing tools implemented by Shellphish, and evaluated this strategy on CTF binaries provided by famous institution. The results show that our strategy achieves significantly greater testing efficiency in larger programs while no apparent effect in minor programs.

In view of the direction finding problem based on time difference of arrival (TDOA), when the TDOA measurement noise was high, the direction finding accuracy of linear least squares (LLS) algorithm decreased. Iterative Taylor Series (Taylor) algorithm had high accuracy, but a dependency on the initial value. In order to compensate both shortcomings, a cooperative algorithm was proposed firstly, which combined the LLS with the Taylor. Then, the mean square error of direction finding about cooperative algorithm and the Cramer-Rao Lower Bound (CRLB) was derived. By comparing them, thus to prove that the cooperative algorithm can reach CRLB. Finally, the simulation results are shown that when the TDOA measurement noise is high, the cooperative direction finding algorithm based on LLS and Taylor can better improve the accuracy of direction finding and has strong robustness.

Aiming at the problem that the traditional kernel correlation filter (KCF) tracking algorithm cannot re-detect the target, when the target is missing due to illumination variation, severe occlusion, and out of view, this paper leads to the anomaly detection method as the target loss warning mechanism based on KCF, and at the same time, a target loss re-detection mechanism is proposed. This method detects the peak value of the response of each frame. If the abnormal peak value is found, the target is lost or will be lost. Then, the warning mechanism warns, the target template update is stopped, the target loss re-detection mechanism is started and tracks the target in full frame search. The experimental results show that the precision of the improved algorithm is 0.751, and the success rate is 0.579, which is 5.77% and 12.43% higher than that of the traditional KCF tracking algorithm, respectively. This solves the problem that the KCF tracker can recover the target to keep tracking after the target is lost, the performance of the tracking algorithm is improved, and the long-term tracking is realized.

For the spectral clustering algorithm in the large-scale samples, there are bottlenecks in the storage space and computing time, and the paper analyzes the current common solutions, namely based on the sparse t-nearest neighbor spectral clustering. In order to improve the accuracy of the spectral clustering algorithms, the Euclidean distance based on the attribute weight of information entropy is proposed to calculate the similarity between the samples. First, calculate the weight of each attribute of the sample and then calculate the similarity between the samples. The degree matrix is applied to the spectral clustering of sparse t-nearest neighbors in the last number of numbers. The experimental results show that the clustering accuracy of the method on some data sets is higher than that of the original spectral clustering algorithm.

Time Difference of Arrival (TDOA) is a popular interference location algorithm in the radio astronomy monitoring environment, where radio interference source may interfere with the performance of the receiver. However, with the geomagnetic white noise interference, interference detection is hard for a normal TDOA algorithm. In this paper, an improved interfere detection method is proposed. Single receiving antenna has been replaced by array antenna, and the adaptive beamforming algorithm is used in data preprocessing. A composite signal after beam-forming as an input of TDOA algorithm. The experiment shows that, comparing with the traditional TDOA algorithm and using beamforming, the location of the interference has been enhanced about 10 m for normal in the about 5 km.

In this paper, we propose a numerical method that is quasi-interpolation scheme for Korteweg–de Vries equation (KDV) equation, Camassa–Holm (C-H) equation, Degasperis–Procesi (D-P) equation, namely processing discontinuous points. The main idea is to separate the time and space separately. In order to consider accuracy, we use third-order Runge–Kutta for time. Numerical examples are shown to prove stability and accuracy of the schemes. The scheme in this paper can be applied to other area.

This paper proposes a novel nature-inspired algorithm called salmon migration optimization (SMO). The main inspirations of this algorithm are based on the navigation methods of the salmon migration activity in the nature. Three heuristics in SMO, respectively called water flow-oriented heuristic (WFOH), magnetic-oriented heuristic (MOH), and pheromone-oriented heuristic (POH), are developed to make SMO has the capability of exploitation and exploration. Optimization results illustrate that SMO always obtains competitive solutions on most of test functions.

Sentiment analysis of text has tremendous value in many fields. But verb expression is absent, while lots of researchers concentrate on identifying opinions from adjective, adverb, and noun expressions in recent years. In this paper, we find that verb expressions in a sentence can be more important because verb expressions not only imply opinions but also give a direct way for enterprise to improve their products. It is meaningful that the verb expressions are extracted and analyzed. In order to deal with this problem, we propose a new method of linear regression optimized by particle swarm optimization to analyze verb expression extracted from reviews. Since our training data is obtained from titles of reviews whose labels are automatically inferred from review ratings, our method is able to work without manual involvement. Experimental results demonstrate our approach has great performance in terms of both precision and efficiency.

This paper provides a Lyapunov inequality for one conformable fractional equation by using the Green’s function. The result of this paper extends the research field of similar problems and generalizes some early conclusions on this topic. Moreover, an application for this Lyapunov-type inequality is presented.

The process of feature selection and the extraction algorithm of dynamic and static gestures based on 3D skeleton data were mainly introduced, and ten customized gestures were classified according to the features above. At first, ten gestures were customized according to teaching demands and with Intel’s 3D camera SR300, a small gesture database based on classroom teaching was established; then, the characteristics of these gestures were analyzed so that features like the trajectory and the local area of some joints point and the extremum of adaptive Euclidean distance can be selected, and a packaged feature selection method (LVW-RF) which combines Las Vegas Wrapper (LVW) and Relief-F was proposed to extract the feature set. With these features input into support vector machine (SVM), an average accuracy of 98.67% was acquired for the ten gestures. The experimental results prove that the features extracted are reasonable and effective.

Acoustic positioning system, as a passive location, plays an important role in battlefield communication reconnaissance. It is an urgent problem to find a direction-finding method with high accuracy and robustness. This paper starts from the application background and studies significance of passive acoustic localization, summarizes the development of passive sound source direction, and briefly describes two basic direction-finding algorithms; after analyzing the current research situation of direction finding, this paper puts forward the deficiency of the current algorithm and the content this paper needs to study.

This paper takes Chinese college students’ English speech as research object. It improves the traditional computer-assisted evaluation methods for English pronunciation quality, considering multi-parameter evaluation indicators, such as accuracy, speed, rhythm, and intonation. The pronunciation evaluation includes accuracy evaluation based on Mel Frequency Cepstrum Coefficient, speed evaluation based on duration, rhythm evaluation based on short-term energy and pairwise variability index, and intonation evaluation based on fundamental frequency. Verified by experiments, the above evaluation indicators in this paper are reliable. Furthermore, considering the weights of the above multi-parameter evaluation indicators, the regression analysis method is used to construct a reasonable and objective English pronunciation evaluation model. It is also proved that the English pronunciation quality evaluation model in this paper is reliable, which can provide learners timely, accurate, objective feedback and guidance, help learners find out difference between their own pronunciation and the standard one, so as to correct their pronunciation errors and improve their learning efficiency.

An integrated simulation method of unmanned aerial vehicle (UAV) is proposed in this paper. Integrated simulation can reduce the risk effectively in UAV flight. Based on UAV wind tunnel data and MATLAB, simulation model is built and real-time simulation system is developed. With the help of the simulation system, an integrated simulation platform of a fixed-wing aircraft is constructed. Simulation and analysis are finished based on a type of UAV. The comparison between simulation results and flight data shows that the integrated simulation method is right and the simulation system has high confidence.

To address the flaws in clustering speed, this paper proposes a novel PSO-GGA clustering algorithm based on pattern reduction. To fully combine the pattern reduction method, the algorithm uses a generalized genetic algorithm in serial to improve the particle swarm optimization algorithm. This can increase the diversity of samples and protect patterns that need to be saved for compression. At the same time, to determine the number of particles needed to replace the poor particles an incremental strategy is employed. This fully embodies the PSO’s ability for rapid search optimization and the genetic algorithm’s advantage of a large search space. The experimental results show that the clustering time only required 20% compared to the original algorithm without showing any obvious decline in accuracy.

Effective and accurate detection of clothing categories in streets is imperative for the description and analysis of crowd clothing profile which is crucial for fashion designers and E-commerce. In this paper, we present a real-time tracking system from surveillance videos to detect and track the various clothing categories with a state-of-the-art deep learning approach, which proposes a combinational framework based on deep convolutional neural network (CNN). First, we take advantage of the mechanism of Focal Loss to improve the loss function of the one-stage detector YOLOv2. We adopt CREST as the visual tracker which largely overcomes the difficulties of occlusion and deformation for clothing detection. Additionally, we have collected and preprocessed a dataset including over 1,200,000 still images from previous works and thousands of video fragments for the training and validation of CNNs. Finally, we compare our work with the baseline and previous work, and our framework demonstrates its effectiveness and accuracy.

With the development of information technology, more and more data information uploaded to the cloud server, cloud computing encryption storage technology for data security has become particularly important. This paper mainly studies and analyzes the searchable encryption technology in cloud computing. Firstly, the research status of domestic and foreign research is analyzed. Secondly, the classification of searchable encryption technology is introduced from different aspects, and the classification model of searchable encryption technology A detailed elaboration. Finally, some problems encountered in searchable confidentiality technology and the prospect of this technology are introduced.

Aiming at the single-point vulnerability such as tampering attack of SDN controllers, current solutions have passive defense defects and cannot solve the security problem of control layer fundamentally. Combined with state-of-the-art moving target defense, cyber mimic defense, and other active defense technologies, we present a security scheduling mechanism with heterogeneous redundant structure, which provides system dynamics and diversity for improved security. We also consider the load factor while designing the scheduling algorithm in our model, thus transforming it into a dynamic optimization problem. Several solutions are compared in terms of anti-attack capability, and simulation results show that proactive scheduling strategies can obviously improve system security as dynamic and diversity properties block the adversary’s attack process.

Entity recognition of clinical documents is a primary task to extract information from unstructured clinical documents. Traditional entity recognition methods extract entities in a supervised learning framework which needs a large scale of labeled corpus as the training samples. However, clinical documents in real world are unlabeled. To construct a large scale of labeled corpus by manual is time-consuming. Semi-supervised learning that relies on small-scale corpus can solve such problem. Thus, this paper proposes an entity recognition model of clinical documents based on self-training framework. In such framework, we first establish partial annotation corpus through the way of dependency syntax analysis and the medical statement rule unifies. Then, a hybrid model of CNN-LSTM-CRF is proposed to label the unlabeled data in an end-to-end way. Specially, we will use CNN to embed characters in clinical document and use Bi-LSTM to extract the sentence-level feature. At the moment, we use CRF remedies the shortage of LSTM which further combined with the combination probability of CRF and the advantages of optimizing the whole sequence. Finally, the results of entity recognition with higher confidence level are fed back by self-training to expend size of corpus which improves the accuracy of the document entity recognition. The experiment result proves the availability and high efficiency of this model.

The standard SVM can be transformed into an unconstrained optimization problem, but the new objective function is not smooth, and lots of fast optimization algorithms cannot be applied to solve it. To overcome the problem, a new Padé33 approximation smooth function is put forward, based on rational approximation method. Then, a new SSVM based on Padé33 smooth function is established. Theoretical analysis proved that the smooth precision is significantly higher than existing smooth functions. Moreover, theorem proof is given to demonstrate the convergence of the new model. Finally, it is applied into the heart disease diagnosis. The experiment results indicate that the Padé33-SSVM model’s classification capability is much better.

First, four models are established to fit the data of the population for the year 2000–2016. The models include GM (1, 1) model, unary linear regression model, index model, and the logistic growth model. Second, simulation is conducted to demonstrate the four models. The results of the data from the statistical yearbook and website show that the fitting effect is good, and with high accuracy, therefore, we use four models to predict the population of 2017–2020. Then, a combined prediction model derived from the four models is constructed, which is more accurate than the single prediction model indicated from the fit results.

Due to there are many defects in the standard Symbiotic Organisms Search (SOS) algorithm, such as mature prematurely, fall into local optimum easily and insufficient convergence precision and so on, this paper adjusts the search strategy of SOS by adding asynchronous learning factor at the “Commensalism phase,” which enables the algorithm to update the cycle individual adaptively according to the current iterations, balance the local search and global exploration ability of the “Commensalism phase” automatically, and speed up the convergence rate and accuracy of the population ultimately. Simulation result shows that the proposed algorithm—Symbiotic Organisms Search algorithm—based on Asynchronous Change Learning Strategy (ACLSOS) in this paper has better performance compared with the basic SOS algorithm.

Diet-induced obesity is becoming more common all over the world, which is increasing the prevalence of obesity-induced chronic diseases such as diabetes, coronary heart disease, cancer, and sleep apnea. Many experimental results show that obesity is often associated with an elevated concentration of plasma leptin and triglycerides. Triglycerides inhibit the passage of leptin across the blood–brain barrier (BBB) to signal the hypothalamus to suppress appetite. However, it is still not clear how triglyceride concentration affects leptin transport across the BBB and energy balance. In this paper, we propose a novel ordinary differential equations model describing the role of leptin in the regulation of adipose tissue mass. Analytical and numerical results are analyzed using biologically relevant parameter values. Additionally, we perform sensitivity analysis of the equilibria and study the sensitivity of triglyceride production on leptin resistance. Equilibria analysis and simulation results show that triglyceride production plays an important role in determining the fat mass in an individual. As weight increases, the occurrence of leptin resistance increases. Obesity enhances the likelihood of creating a vicious circle, where more fat mass leads to greater leptin resistance. Thus, control of the triglyceride production may be effective in reducing the occurrence of leptin resistance.

Aiming at solving the problem that the feature of microseism signal is difficult to extract, a time–frequency analysis and classification method based on rearranged ST-NMF are proposed. Firstly, microseism signal is transformed into the time–frequency matrix by S transform, and the rearrangement is carried out in the frequency direction. Then, the decomposition vector of time and frequency domain is obtained by using the nonnegative matrix factorization technique, extracting the macro- and micro-statistics to construct the feature space of the signal. Finally, classify the feature space. The results of the experiment in the SAN Daogou show that the spectral resolution of the rearranged ST is significantly higher than that of the ST. The accuracy rate reaches 97% after multiclass classifier recognition.

There are some difficulties of knowledge acquisition in constructing knowledge base during the diagnostic reasoning. For this reason, a diagnostic knowledge base constructing model based on decision and RS theory is put forward in this paper. In this model, decision technology and RS theory are introduced to preprocess source data and construct decision table. Minimum reduction attribute set of fault diagnosis and diagnostic rules is obtained through attribute reduction and value reduction. Meanwhile, the diagnostic rules’ knowledge base is set up. It is proved through examples that the feature knowledge reflecting fault best can be extracted and the problems of knowledge redundancy or lack of knowledge in the construction of diagnostic knowledge base can be solved with diagnosis classification results kept in this method. And the accuracy and efficiency of fault diagnosis are improved significantly.

I present a translation approach based on a phrase-based maximum entropy model. To improve search performance, a beam search algorithm is exploited where the selection of the phrase candidate translation, the future probability calculation, and the pruning strategy are included. A neural network model with words and phrases is used. I obtain better translation results by the extensive experiments conducted on the real and synthetic datasets.

This paper proposes a robust local region-based active contour model for inhomogeneous image segmentation. In the model’s energy function, the gray intensity of each pixel within a region is fitted to a local weighted mean of the original image. To solve the local minimization problem of above energy, the consistency of local contrast between the inner and outer of curve at each pixel in the curve is imposed as a weighting for the local fitting energy. A normalization constraint for the curve is added to the energy, and the minimization of the energy is implemented with level set method. Experimental results reveal that compared to related methods the proposed method can improve the segmentation result, being robust to the initial contour.

The speech recognition on low-resource language is a research hotspot. In this paper, on the basis of transfer learning, we propose a cross-lingual speech recognition system based on time-delay recurrent neural network. This network is comprised of deep belief time-delay neural network layer interleaved with long short-term memory recurrent neural network layer. We first train the whole merged network with a large labeled corpus of source language and afterward retrain the hidden layers using small target language corpus with both cross-entropy and lattice-free MMI objective function to enhance the recognition performance on the target language. Experiments show that the proposed system outperforms LSTM and DBN-DNN baseline system on THCHS30 corpus.

This paper presents a new K-Medoids clustering algorithm based on gray relational degree. Analyze the gray incidence of each attribute and convert them into the weights of the attributes, and then apply these weights to the distance measure of the cluster; based on this measure, this paper proposed an improved clustering algorithm: Gray-K-Medoids clustering algorithm and applied it to the analysis of the aluminum electrolysis data. The paper introduces the gray relational degree and the basic principle based on the gray relational degree clustering and introduced the improved algorithm in detail. In order to test the effect of improving the algorithm, it was used to the production data of an aluminum plant, and the results show the effectiveness of the algorithm, has a certain promotional value.

With the rapid development of Internet technology, global data are growing at an exponential rate, and human computing has undergone profound changes. The early stand-alone mode has not met people’s needs, and the network-based collaborative distributed gained more and more attention and favor. Distributed computing system is one of the hottest Internet research directions in the era of big data era. It has the characteristics of high efficiency, high capacity, dynamic processing, and so on. It shows great application value in the commercial field and scientific research field of society. Based on the development and present situation of distributed computing system, this paper reviews and summarizes two popular key technologies: grid technology and cloud computing technology and expounds the differences between the two. At the same time, the design principles and system performance of Hadoop, Storm, Spark, and other typical distributed computing platforms are compared and analyzed in detail, which is of theoretical significance.

Learner’s learning states can be revealed from their spontaneous expressions while learning. The ubiquitous laptop camera is capable to provide learner’s facial image sequences. The combination of ubiquitous image perception and image processing technology can effectively compensate for the lack of feedback from learners on learning content in intelligent education. Therefore, this paper presents an affective state analysis algorithm, HowsLearning, and establishes the expression dataset in the actual teaching scene. On this dataset, the classification algorithm is verified. Experiments show that accuracy tested on real-world collected data reaches an acceptable level.

In the current entity relation extraction technology, more and more researchers focus on semi-supervised Bootstrapping method, because it does not require a large number of artificial tagging corpus, needs only a small amount of seed set, by self-iterative extended to obtain large-scale knowledge base. However, after a large number of iterations, there will be “semantic drift,” that is, the accuracy will reduce due to the accumulation of errors. In order to improve the accuracy of the relation instance the quality of the pattern, it is necessary to evaluate the reliability of instances and patterns. This paper uses large-scale news headline sentences set in the search engine, evaluates the reliability of instances by co-occurrence relation between description words and sentences set, then evaluates the reliability of patterns by the number of positive and negative instances in patterns historical matching record, and selects new patterns to extend and optimize. The experimental results show that the reliability evaluation of instances and patterns used in the iteration effectively improves the accuracy of relation extraction and improves the quality of the extracted pattern.

To solve the problem of Chinese text localization in natural scenes, a text localization method based on the feature of Chinese characters is proposed. Firstly, the Maximally Stable Extremal Regions (MSER) algorithm is adopted to extract candidate text regions. Then, the separated Chinese character strokes are fused by using mathematical morphological operations. Then, the non-text regions are filtered through heuristic rules and structural features. The Stroke Width Transform algorithm is used on the candidate text regions, and the non-text regions are removed by stroke width features. Finally, a text line construction algorithm in any direction is adopted to merge the localized text regions so as to achieve the final localization of Chinese text. The experimental results showed that the proposed method could provide better results for the localization of Chinese text in natural scenes.

The FAT file system was designed four decades ago, and its variants are still widely used today in embedded systems and removable devices. FAT is simple and robust, and it is supported by virtually all modern operating systems. However, FAT does not provide data redundancy per se, and bad sector(s) on storage media can incur permanent data loss. In this paper, we propose a block-based redundancy approach on FAT32 to tolerate sector errors. Parity data is stored in a reserved area within the file system. We use erasure coding rather than replication to minimize storage overhead. Test results indicate the effectiveness of the method.

In order to overcome the shortcoming of local convergence of simple particle swarm optimization algorithm (SPSO) in optimization of high-dimensional and complex function, a modified particle swarm optimization algorithm based on particle classification (PCPSO) is proposed. The particle classification is based on the particle fitness and the particle distance which is firstly proposed and defined in this paper. The coordinate system is established based on the mean value of particle distance and the mean value of particle fitness. Particle swarm is divided into four subclasses in four quadrants. The properties of the particles in each quadrant and their influence on the optimization process are analyzed, and different formulas of particle velocity updating are established to improve the particle update efficiency. In addition, the algorithm has the function of dynamically adjusting the origin of the classification coordinates as the iteration number increases to balance the algorithm’s global search performance and local search ability and improve the algorithm’s optimization efficiency. The optimization results of the typical test functions show that the accuracy and success rate of the modified algorithm in high-dimensional complex function optimization are greatly improved, and the premature convergence phenomenon can be overcome effectively.

In recent years, there has been an explosive growth in the volume of check-in data in location-based social network (LBSN). As we know, these check-in data imply temporal and spatial information about user behavior patterns. This paper proposed an overlapping community detection method based on similarity of user trajectories and reference spots similarity in LBSN. Based on traditional DBSCAN clustering algorithm, we divide the entire map into many atomic regions that contain density information. We measure the similarity of trajectories between users and combine the similarity of user’s reference spots, and then we can discover overlapping communities by edge clustering algorithm. The experimental results show that the proposed approach is more effective and more precise in discovering communities in LBSN.

Deep learning makes a great breakthrough in the field of artificial intelligence. The performance of robots on the uncertainty task can be enhanced using the deep learning. Due to the accumulative errors of the servomotors, the robot’s end-of-arm tooling (EOAT) could not grasp objects in proper position. It is worth to study robotic grasping detection with the deep learning while there has already been some successes practice in the robotics research. We propose a novel method for the robotic grasp detection that gives the grasp position of a parallel-plate robotic gripper based on the deep learning model with the RGBD image of the scene. The best model of our method archived an accuracy of 87.49% with an acceptable time speed. Our method introduces another way to solve the robotic grasping problem.

Test oracle is a mechanism that to determine whether the actual output value of the program is in line with expectations. It is an indispensable part of software testing process and also a weak area in software testing. The automation of test oracle not only effectively reduces the burden on testers, but also provides strong support for uninterrupted continuous testing. Heuristic-based oracle has the advantage of easy implementation, fast execution, and wider fault coverage. Heuristic-based oracle usually uses BP neural network as oracle information, but compared with probabilistic neural network, BP neural network has its limitation in classification. Aiming at the classification problem, this paper proposes a test oracle based on probabilistic neural network. Experiments show that it is better than BP neural network in prediction speed and accuracy.

I focus on mining opinion mining from news articles in this paper. Different from most of the previous opinion mining frameworks that focus on opinion mining from product reviews or news comment, opinion expressions in news articles always contain less emotional words and have no clearly defined opinion holder and opinion target. A new framework based on CRFs was proposed. It can extract the opinion holder, opinion target, and opinion polarity from news articles. My experience on news articles datasets shows that my method outperforms the traditional approaches designed for review mining.

The traditional sequence labeling systems rely on abundant specialized knowledge and handmade features. And state-of-the-art sequence tagging model requires no feature engineering and data processing by a combination of bidirectional LSTM, CNN, and CRF. In our paper, we propose a LSTM–CRF–DICT network by incorporating the word representation based on the dictionaries into the neutral network. The Trie tree is applied to the extraction of dictionary-based word features, which accelerates the retrieval efficiency. We evaluated our method on a Chinese corpus for sequence tagging with nine tags and five domain dictionaries, and the performance of sequence tagging is enhanced significantly. In addition, the introduction of dictionaries results in little increase of time assumption.

In this paper, authors have applied the ACO–SVM for the development of new-type urbanization in Jiangsu–Zhejiang and Shandong (i.e., JZS) Province which researchers believed to have embedded with many socially and practically significant factors. Based on the support vector machine optimized by ant colony algorithm (i.e., ACO–SVM), qualitative analysis should be combined with quantitative analysis, to assess the level of urbanization JZS Province in China. The result of assessing the status of urbanization JZS Province shows the method of ACO–SVM algorithm disappeared fast, and the precision was high. According to the empirical research, Kunshan City, Zhangjiagang, and Jiangyin, respectively, ranked the first, second and third. Kunshan City is the best quality to absolute advantages to become the current development of urbanization of county. As a new type of urbanization of Shandong Province pilot in most provinces, actively promote the development of new type of urbanization, compared with county in Jiangsu and Zhejiang, Shandong Zhangqiu, Laizhou City, Qufu rank in 14, 13, and 15.

The design of test case automatic generation technology is an important step in the implementation of software automated testing. It plays an important role in guiding the later testing work and is the fundamental guarantee for improving software quality and reliability. Ant colony optimization (ACA), as a robust optimization algorithm, has a strong ability of global optimization. In this paper, an automatic test case generation method of parallel multi-group adaptive ant colony algorithm is presented. This algorithm adopts multi-group parallel search, adopts adaptive migration rule based on population diversity, and updates pheromone strategy adaptively according to fitness function in population. Theoretical analysis and simulation experiments show that the application of this algorithm is superior to the generation of test cases.

In view of the difficulty of identifying the microseism signals, the present invention proposes a microseism signal recognition algorithm based on data mining. Firstly, ST is used to filter, and then using the space–time anomaly detection isolated noise signal and the seismic signal, using data mining to extract the feature information of seismic signals, finally using space–time forecast and classification model to recognize the seismic signal. Experiment result shows that the recognition accuracy of DM is slightly greater than that of LCD and LMD. However, the recognition rate of microseism signal based on DM is significantly smaller than LCD and LMD.

To research program design course teaching and practice based on the technique of SVN version control, we choose the right cases according to the program design course knowledge system. To split reasonably the function of the case according to the requirement of the program design course and knowledge system, the project case is broken up into teaching version. According to the teaching schedule for the project to version control to suit the needs of different students, the quality of the programming course teaching and students’ programming ability is improved, and the students’ team cooperation consciousness is enhanced by the practice teaching method reform. This enabled us to get the good teaching effect.

CSMA is carrier-sense multiple access and is one of the key protocols to ensure the efficient operation of the communication network. According to the characteristics of wireless sensor network nodes, the use of FPGA (field programmable array) for no-persistent CSMA access communication system provides a high-speed data transmission scheme to achieve gate. The design uses MATLAB and Quartus II as the experimental platform, which satisfies the Poisson distribution and the time slot length controllable source module for different events in the transmission process, using the method of hardware description language Verilog HDL and logic circuits to design the simulation, through calculation and analysis, the theory of throughput of the protocol are consistent with the simulation value, proved that the system has reliability design bureau, portability etc.

IEEE802.15.4/Zigbee is one of the most typical techniques for short-distance wireless communication. Due to the outstanding features of low cost and low power consumption, Zigbee has been widely applied into the field of smart home and security monitoring, etc. However, Zigbee signal is sensitive to obstacles in the communication channel. In the opposite view, the characteristics of signal degradation reflect the existence of obstacles in communication channel. In this light, a tentative study is made on indoor human intrusion detection based on the quality characteristics of received Zigbee signals. Experimental results show that the evidence of indoor human intrusion can be detected; furthermore, the information of human body characteristics can also be estimated.

In transport business is more important, maintenance is difficult, high fault probability of the optical fiber cable line section of the introduction of OLP fiber automatic protection switching system, can effectively achieve the result of failure prevention, failover, ensure smooth traffic, improve the reliability of optical fiber transmission. This paper mainly studies the related problems of the OLP system from the principle of technology and practical application.

The 3D NoC, which blends 3D IC with NoC, overcomes the performance bottleneck of 2D NoC. However, the conventional buffered routers in 3D NoC have the problem of consuming a large amount of resources. In this paper, we propose 3D minimally buffered router using priority deflection (3D MinBPD). This router is the 3D extension of MinBWD and introduces TSV Buffers at the same time. Then, a priority-based deflection routing algorithm is proposed, which only needs local information to compute the priority of flit. The experimental results show that the performance of the 3D MinBPD is better than the reference objects’ under the network of $$ 8 \times 8 \times 4 $$, which is similar to the traditional buffered routers’ in the low and medium packet injection rate.

In wireless sensor networks, there are many location-based applications. In the traditional greedy routing protocols, there are some defects, such as network lifetime is too short and packet delivery rate is low. A location-based multi-objective optimization routing (LMOR) protocol is proposed. By estimating the residual energy, the extent of promotion and link quality of each hop, a multi-objective optimization function is constructed, and the packet routing is based on the maximum of it. Test results show that, compared with the traditional greedy routing protocol (GREEDY) on non-ideal links, LMOR reduces the average energy consumption by more than 10% and extends network lifetime by 45.2%.

Network Function Virtualization (NFV) achieves the goal of decoupling network functions from dedicated hardware devices by software techniques, which improves the agility and scalability of network services thus reducing capital expenditures and operating expenditures. However, NFV orchestration (NFVO) still faces great challenges such as how to automatically generate an efficient orchestration in a user-friendly way. In this paper, we design and implement a platform based on OpenStack services which accomplishes a Draggable Topology Web UI to help administrators create and manage virtual networks (VNs) and orchestrations without too much expertise. We demonstrate the ease of use of the platform with experiments at the end.

This paper takes the logistics storage environment as an example and builds a temperature and humidity detection system based on 6LoWPAN wireless sensor network. It simulates and analyzes the situation of the routing when the nodes are moved. At last, it completes the actual environment test. The result shows that the system completes the interconnection and data exchange between wireless sensor network and IPv6 network, and realizes real-time detection of temperature and humidity in logistics environment. The system is stable and efficient.

We propose a modified PFNF routing algorithm based on the structure of WUDN, which is an adaptive routing algorithm. The algorithm has the characteristics of simple operation, low complexity, and fast response, and we prove that the algorithm is deadlocks free. The algorithm is simulated in a 16 × 16 network through the discrete time simulator NS-3. Compared with f-cube fault-tolerant routing algorithms, the results show that the modified PFNF algorithm can improve the throughput by 8% and reduce the average delay by 10% under saturated traffic.

Self-organizing maps (SOM) algorithm is a kind of tutorless learning method with good self-organization and visualization. It has been widely used and researched. SOM neural networks can classify unknown categories of data without supervision, but the same type of data in the classification result may correspond to different winning neurons. According to the principle of a category corresponding to a winning neuron, the SOM network classification may have more categories than the actual data categories. In order to solve this problem, this paper presents two improved methods: the improved SOM neural network based on system clustering method and the improved SOM neural network based on k-means algorithm. Two improved algorithms are used to cluster a total of 4000 sets of data from five kinds of network intrusion data, and the clustering results are compared with the fuzzy clustering and generalized neural network fuzzy clustering algorithms. The experimental results show that the SOM algorithm based on the system clustering method is better than the SOM algorithm based on k-means algorithm in the intrusion detection.

With the further smart grid, the research and application of network security situation awareness (NSSA) in smart substations are receiving more attention. In this paper, we propose a method of network safety forecasting for smart substations based on Gibbs-LDA and least square support vector machines. The project extracts all kinds of message information of intelligent substation network as message set, obtains sample characteristics and message set model, and establishes multi-dimensional prediction model of intelligent substation network security based on LDA. And use the least square support vector algorithm to get the prediction result. The experiment on the data set collected by smart substation shows that the method in this paper has the advantages of high prediction accuracy and short prediction time compared with other methods.

This paper proposes a novel approach to predict the link in OppNets. We utilize the deep belief network (DBN) algorithm for the link prediction. Katz Index, Resource Allocation, Adamic Adar, and Common Neighbor metrics are used separately as input vectors to the DBN algorithm, where the input vectors contain only unrepeated links. Two different time slot of 10 and 30 min used to divide the dataset into snapshots in a structure of matrix. Therefore, the results obtained from the experiment shows that the proposed method comes up with more precise prediction than applying each metrics individually. Katz-DBN has shown high precision with 98% depending on 30 min time slot, while in 10 min time slot the precision is 97%.

The AIS monitors and demodulates ship signals within the visible range and packs its demodulated information to the ground stations so as to provide a wide range of high-efficiency status data for both marine management and ship operations. The AIS subsystem receiver mainly brings about the real-time capture and data demodulation of the ship-borne AIS signal. In order to complete the comprehensive test of AIS subsystem receiver, the signal source of AIS ground test system is established. The primary mission of signal generator of the system is to generate simulated data in real time, according to the communication protocol stipulated in ITU-R Rev. 1371-5 of the ITU Radiocommunication Sector through simulation and maintenance of LEO satellites, ship status, etc. The communication frame simulates the generation of AIS signals so that the functionality and performance of AIS satellite receiver can be evaluated.

With rapid development and wide application of vehicular networks, data security is gradually valued. CP-ABE could be adopted based on the characteristics of vehicular networks data transmission. And short decryption keys generated in CP-ABE can reduce storage and transmission costs, as vehicles have limited storage capacity. In addition, the decryption keys are shared by multiple users whose attributes satisfy access policy in CP-ABE. So there may exist malicious key owners who have the intention to leak their decryption privilege for profit, especially when traitors cannot be traced. In this paper, we propose a traceable CP-ABE with short keys for vehicular networks. Our scheme not only provides short decryption keys and traceability of traitors, the cost of implementing traceability is also very low. Therefore, it is particularly appropriate for vehicular networks.

In this paper, a real-time virtual fitting technique based on Kinect is proposed to solve a series of problems in the existing virtual fitting technique, such as large amount of data, slow response speed, no reality, and popularity. This technique mainly includes acquisition of human pose based on Kinect, garment modeling, and garment deformation. Human pose information is obtained by body contour model and human feature points which are extracted from the human depth image, color image, and skeleton point data acquired by Kinect. Then garment model is established, and 2D garment deformation algorithm is used to reshape the garment model to adapt to the human body. Finally, the virtual fitting result is shown in the color image.

In the collaborative filtering algorithm, the user similarity and the selection of the nearest neighbor are the two most important links in the whole algorithm. The traditional collaborative filtering algorithm only relies on user ratings to calculate the similarity between users and then find each user’s K neighbors, and finally according to the K neighbor set recommended. However, in the face of large data processing, the traditional collaborative filtering algorithm is sparsely populated, resulting in the recommendation which is not obvious. A collaborative filtering algorithm for hybrid similarity is proposed for this problem. The algorithm is mainly from the user similarity degree of similarity, user attribute similarity and similarity of user purchase records, and many experiments to adjust the weight between them, the use of dynamic selection of the collection method to find the user’s nearest neighbor, so as to users Recommend more suitable items, improve the recommended quality of the algorithm. The experimental results show that the method proposed in this paper has lower MAE value than the traditional co-filtering algorithm, and the recommended quality is improved.

The army’s participation in building cyber security forces is an inevitable choice to link cyberspace with national security. This is also a common international practice. This article from the military field is the network battle “top” battlefield, the military force is to seize the “network power” of the main battle force, military applications are the most important direction of cutting-edge technology integration direction of three clear military dominance for the defense of national network sovereignty. Only by clearly defining the military’s role in defending the nation’s network infrastructure and vital information systems, it will be possible to build on the strength and operational mechanisms of its own military support that does not depend on other countries.

With the furtherance and progression of technology, secure communication got the paramountcy in every aspect of human’s life. Steganography is the science of techniques, concerned with concealing the existence of secret information in a cover file at the sending side and extracting back the hidden information at the receiving side in such a way that no one except the sender and the intended recipient knows about the presence of the secret message. In this research work, a novel audio steganographic technique is proposed which is based on additional audio channel. In the proposed technique, a new additional audio channel is generated on the basis of original audio. The secret information is embedded in the new additional audio channel according to the threshold level that works to be a secret key. As compared to the other existing techniques, the proposed technique achieves high embedding capacity and high peak signal-to-noise ratio (PSNR).

Based on the characteristics of radar signal and the algorithms’ applicability and performance of different algorithms, this paper selects scientifically the most suitable option for radar imaging among these four sparse fast Fourier transform (SFFT) algorithms for 2D ISAR imaging; evaluates each algorithm’s accuracy through simulation; and applies SFFT to inverse synthetic aperture radar (ISAR) imaging program for the stepped frequency radar. Through an experiment, author studies SFFT’s imaging efficiency and gets a conclusion that SFFT is able to reduce the run time into half at a low loss of image quality.

Convolutional neural networks (CNNs) have been widely applied in many computer vision tasks due to its high accuracy. Meanwhile, various CNN accelerators based on FPGA platform have been proposed because they have advantages of high performance, reconfigurability, low-power consumption, etc. Although current FPGA accelerators have demonstrated better performance over generic processors, it is challenging to deploy computation-intensive and memory-intensive CNNs on hardware implementations. In this paper, we have a comprehensive overview of existing FPGA technologies on accelerating CNNs and then outline underlying frameworks for mapping CNN models to energy-efficient FPGA. We adopted OpenCL hybrid systems to implement CNN accelerator using Xilinx’s KCU1500 board and demonstrated the better performance and resource utilization compared with existing work. This research traces the recent development trends and shows that the accelerator design space for CNNs on FPGAs can be further exploited.

Automatic expression recognition of human faces has been an active research area for decades. In this work, to improve the facial expression recognition effect, a new method based on SSVM algorithm and multi-source texture feature fusion using KECA is proposed. Multi-source texture features are introduced to describe the facial expression, containing GMCL, Gabor feature, and HOG feature. The results indicate that multi-source texture features are conducive to improve the recognition effect and make up for the deficiency of single texture feature in facial expression description. In addition, KECA and SSVM algorithms show better performance than traditional methods in feature extraction and classification. To further verify the effectiveness of the proposed method, three sets of comparative experiments are carried out: PCA+SVM (based on Gabor feature), PCA+SVM (based on GMCL+Gabor+HOG feature), KECA+SVM (based on GMCL+Gabor+HOG feature). The results of JAFFE database indicate that the accuracy of proposed method, equal to 93.04%, is at least 2.61% higher than conventional method. The results of JAFFE database demonstrate the validity of proposed method.

In order to improve the application effect of reversible watermarking algorithm, the chaos theory is applied in it. Firstly, the mathematical model of chaos theory is studied. Secondly, the reversible watermarking method based on blocking technology is analyzed. Thirdly, the procedure of digital reversible watermarking algorithm is designed. Fourth, the extraction algorithm of digital watermarking image is put forward. Finally, the simulation analysis is carried out, and results show that the novel algorithm can obtain good image quality.

The fractal property is considered as the third fundamental property of complex networks; its research has both rich theoretical value and practical significance. This paper reviews the research results of fractal network from three aspects: the origins of fractality, the algorithms for fractality, and the influence of fractality in complex networks; some of these algorithms are described in detail. Finally, the paper concludes and looks forward to the possible research focuses for fractal networks in the future.

Based on the group Diffie–Hellman technique, a contributory group key exchange protocol has been proposed by Biswas. Although Biswas claimed the protocol belongs to a contributory group key exchange, Tseng and Wu did not agree with. Then, they proposed an improved group key exchange protocol with a character of verifiably contributory and claimed the proposed protocol could not provide a secure scheme against passive attacks. However, Ling et al. pointed out that it is vulnerable to a man-in-the-middle attack with Tseng–Wu’s protocol. The attackers could intercept and modify the messages among the group members. In this article, we will propose an improvement of Tseng–Wu group key exchange protocol. Our protocol could against remedy the man-in-the-middle attack.

Recently, Li, Wang, et al. proposed a secure and efficient authentication protocol which enhances scalability function by using the simple index. This study further demonstrates that their scheme could not resist denial of service attack.

This paper presents some key technologies of differentiated anti-stealing electricity and its application. Collection of electricity data is implemented by the electric energy data acquisition system (EEDAS) firstly. The indicator evaluation system for power consumption and the evaluation model of user power consumption have been built. Then credit rating of electricity customer can be determined by the evaluation model. Online monitoring and video forensic system for anti-stealing electricity and comprehensive evaluation system of stealing electricity are introduced and applied. The application result which shows the comprehensive evaluation system can improve the efficiency of the anti-stealing electricity and reduce the workload of manual investigation for the power supply enterprise.

In this paper, the hybrid frequency synthesizer of PLL+DDS mixing in the outside of the feedback loop is researched and designed. The output signals of the PLL with low reference frequency and DDS with high reference clock are mixed directly, and then a narrowband band-pass filter is to get the required frequency signal. The hybrid frequency synthesizer is realized, and the measured results shown the phase noise and SFDR are −91.17 dBc/Hz @ 10 kHz and SFDR ≤ −33dBc (within ±25 MHz).

With the improvement of people’s living conditions, the applications of various TV display screens are becoming more and more common. Typically, the person viewing the display screen has a fixed angle and does not make full use of the view field, so the viewing quality is lowered. To improve the viewing efficiency, we propose a smart screen rotation platform for the TV Pan Tilt Zoom (PTZ). Through the Raspberry Pi microcomputer, image sensors in front of the screen are used to acquire the images of the viewer’s face and surroundings. Based on the machine vision technique, the human face is recognized, and the face orientation is found. The identification results are delivered to the lower machine–ATMEL328P microcontroller, which controls the PTZ actuator to rotate, so that the screen is able to align the viewer’s face and help to realize the high-quality watching.

As intelligent mobile devices become more popular, people have stronger and stronger intention to control the Entrance Guard System (EGS) with mobile devices (Abraham, Security Technology Research Report, 12, 2014 [1]). However, most of EGSs cannot connect to the mobile devices wirelessly and they also need to improve their performances such as security, practicability, and intelligence. Therefore, this research proposes a WiFi EGS with double authentication secure random dynamic code system based on WeChat. The user needs to send a certain command firstly and feedback the status of indicator lights on the controller before he gets the permission to open the door. If the user who does not belong to the accessible group, he needs to apply for the random dynamic access code from the owner with WeChat. The EGS can make sure that the user has been at the door by the feedback status of the indicator lights on the door controller. The random dynamic code applied by the visiting user can be used only once. Without installing any App, all the users can open, close, and check the lock with mobile phone just by following the WeChat public number on the lock. So, the EGS is very easy to use and can also help the owner to let his guests enter the door in security even he is out. Some experiments have verified that the EGS has many goodness such as convenience, security, and low cost, thus have a wide range of application.

Aiming at the problem of modularization, rapid prototyping, part feature identification, and assembly in intelligent design, using end-to-end integration design concepts in industrial Internet of things (IoT), taking the directional design and assembly of the safety fence equipment as the object of study, the intelligent design and assembly system of the safety fence is developed. It is based on embedded communication protocol (machine-to-machine—M2M), combining parts with rapid prototyping and assembly, the driver of parametric method and Object Linking and Embedding (OLE), the feature assembly technology (FAT) and data flow chain (DFC) combined with the determined parts assembly sequence, two-dimensional interface constraint, generating assembly model in SolidWorks, BOM list, to achieve security fence rapid design and assembly.

This study proposes a hardware Trojan horse (HTH) defense technology that uses a novel structure of physical unclonable function (PUF) which is based on resistive random access memory (RRAM). An application-specific integrated circuit (ASIC) combined with the technology can effectively prevent HTH embedded. Unlike the conventional silicon PUFs, the randomness of RRAM PUF is not based on process variation and is determined by the stochastic switching mechanism of the RRAM devices. It has some advantages over the conventional one of high security, high reliability, and low area. Experimental results show that the HfO2 RRAM devices have good performances about Hamming distance and reliability. This work demonstrates the RRAM PUF that has great potential for HTH defense.

In view of the need of automatic monitoring of oil field station, a joint algorithm for vibration monitoring of water injection pump in oil field is proposed. The algorithm combines the advantages of time domain algorithm and frequency domain algorithm, which filters the noise of original signals in time domain based on the digital signals collected by chip ADXL345, and gets the acceleration, velocity and displacement by Fourier series, to realize the purpose of double integral and optimization by using the relationship of Fourier series. Then combined with low-frequency cutoff algorithm in frequency domain, the accurate conversion between acceleration and displacement can be realized quickly. The algorithm can be effectively applied to vibration and displacement measurement of water injection pump.

The type of Marx high-voltage pulse source design of avalanche transistors is introduced in this paper. There are seven stages; connection of avalanche transistors is adopted. Via the component parameters selecting, the output pulse has several nanoseconds front edges and KV level peak voltage, and extremely small trailing ringing. The results show that this circuit is simple, stable, and reliable, with a good output signal which is good at the ultra-wideband ground-penetrating radar application.

In this paper, the batch scheduling of hot-rolled steel tubes with flexible maintenance is considered and abstracted into a single machine scheduling problem with flexible maintenance and sequence-dependent setup costs. A mathematical model is established to minimize the adjustment time and idle time of the machine, and a heuristic algorithm is designed. Finally, comparative experiments were carried out on the basis of actual production data, and the results show that the model and algorithm have an obvious effect on solving this kind of problem.

In this paper, a random number generator readout circuit with excellent performance is designed and implemented in combination with dual channel ADC acquisition circuit, FPGA, and host computer. The superlattice structure is used as the noise source of the random number generator; the high-speed ADC device is controlled by FPGA to sample and process the noise signal, and the random number is transferred from FIFO to the host computer by USB. The test results show that the system has high speed, good stability, and strong anti-interference ability, which shows that the random number generator can produce good random number.

For a vehicle, even the autopilot in future, the steering system is indispensable. From the analysis of the steering system, this paper provides an approach to evaluate and to optimize the steering system basing on the Jeantaud Epure (Ackermann steering geometry). At the same time, this analysis could also give the road planning for the autopilot vehicles. Then, we have constructed parameterized model under ADAMS to simulate and verify the result.

Large-scale multiple-input multiple-output orthogonal frequency-division multiplexing (MIMO-OFDM) has become a promising technology for 5G communication systems. However, existing precoder designs only aim at removing the multiuser interference (MUI) and do not consider the issue of high peak-to-average power ratio (PAPR) caused by multicarrier signals. In this letter, a linear precoding algorithm with low PAPR is proposed for large-scale multiuser MIMO-OFDM systems. The proposed algorithm can not only remove MUI but also reduce the PAPR significantly. Further, to reduce the computational complexity, a suboptimal precoding algorithm is also proposed. Simulation results validate the efficacy of proposed precoding algorithms.

A bus management strategy is designed for a 4-computer 4-bus system of a spacecraft. The strategy adopts the global optimization method using the scoring mechanism. It can find the best bus management mode, under the constraints of avoiding the bus conflict, balancing the computer load, and minimizing the influence of the mode switching. By adopting the designed strategy, the fault tolerance ability of the system can be maximized. The validity of the strategy is verified by simulation and experiments.

This paper presents a capability-based decentralized information flow control (DIFC) model and implements this model on a Barrelfish kernel. In this model, we use a uniform abstraction, capabilities, to describe the restriction rules and manipulate information labels. The structure of our model is concise and easy to implement on a real-world system. We have added several system calls in the Barrelfish kernel, including secure message passing, compartment allocation and capability transferring. The results show that the capability-based DIFC rules ensure the security and integrity of the system communication mechanism with a small amount of system overhead.

The similarity calculation of bilingual documents is an important link to cross-language information processing system, which is the necessary basis of cross-language news acquisition and cross-language documents sorting. The traditional bilingual documents association model based on parallel corpus is faced with the problem that high-quality parallel corpus is difficult to obtain. There are problems of semantic ambiguity and out of vocabulary based on bilingual dictionary. This paper presents a method of calculating the similarity in bilingual documents based on TF–IDF and bilingual word embedding. Firstly, the keywords of bilingual documents are extracted. The keywords of different languages are mapped into the same semantic space using the result of bilingual word embedding. And ultimately calculate the similarity in the documents based on the distance between the document keywords. The experimental results show that the proposed model is superior to the model based on parallel corpus and bilingual dictionary.

Integrated linear regulated voltage supply, with high precision, stable, and reliable operation, simple external circuit, small size, and lightweight, is widely used in a variety of power circuits. Because of its inefficiency and high loss of the regulator tube, protection measures are taken on the circuit in order to avoid the power supply damage caused by overload. Therefore, it is of great theoretical significance and practical value to design a linear regulated voltage power integrated circuit with overflow and safety operation area, overheat protection circuit. For this reason, a new circuit is adopted in this paper based on the basic linear power supply circuit. A linear regulated voltage power integrated circuit based on 2 μm 36 V bipolar technology is designed by adding overheating protection circuit, overcurrent and safe working area protection circuit module. The circuit is analyzed and simulated with Cadence. Through the adjustment of the circuit parameters, the final simulation parameters are met with the requirements of the proposed indicators and the correctness and feasibility of proposed circuit is verified.

Massive Online Open Courses (MOOCs) are popular these days as they provide a new way for learners. However, MOOC platforms are incapable of detecting learners’ emotions and cannot provide personalized assistance which is a key factor toward learners’ learning performance. In this paper, we propose SmartPlayer which is a framework of emotion recognition taking learners’ appearance and video interaction events into consideration and predicts learners’ emotion. Appearance features and interaction features are extracted from raw data. We applied Support Vector Machine algorithm to train a multi-class classifier. Data are collected, and the accuracy of the classifier reaches 71.89%. We also applied Principal Component Analysis method to reduce feature dimension and trained a new classifier with accuracy decreasing to 71.05%.

Power-to-gas (P2G) technology is an important way to accommodate wind power. This paper combines the P2G technology with smart microgrid and analyzes the effect of P2G on wind power accommodation in day-ahead scheduling. Firstly, theory of P2G technology is given, and the energy hub based on smart microgrid is modeled. Then, a day-ahead scheduling model of the smart microgrid is proposed. Finally, a case study is used to verify the effect of P2G on wind power accommodation. The results demonstrate the significant impact that P2G has on the wind power accommodation and can provide reference for the large-scale application of P2G.

The turbulent flow, which is generated by applying the alternating control voltage to the driving electrodes, can improve the mixing effect of micro-fluids in the Y-shaped micro-channel. In this paper, the simulation results of electroosmotic micro-mixing in Y-type micro-channel are numerically simulated by using COMSOL Multiphysics software. The mixing effect of free diffusion micro-mixing and electroosmotic micro-mixing was comparatively analyzed, and the relationship between the mixing effect and the number of electrodes, the electrode spacing, the electrode voltage, and the alternating electric field frequency was analyzed. The simulation results show that the mixing effect of electroosmotic micro-mixing is obviously better than the free diffusion micro-mixing.

We propose a method for indoor positioning of personnel based on volume spectrogram and neural network. This method has low requirement for indoor audio collection environment and realizes the original audio collection with low-cost equipment. The location method is based on deep learning. This paper compares and optimizes the network structure selection. Experimental results show that our proposed method can locate personnel with higher accuracy.

Combining with virtualization techniques, the paper proposes a novel architecture for fault diagnosis of weapon system based on test resource virtualization. Through collecting test data from the equipment, the paper applies the Condor and Hadoop techniques to process and achieve the virtualization of test resource. With the fault diagnosis methods applied to the architecture, the diagnosis results are obtained based on the test data and test resource, and artificial intelligence methods are inquired by the computational resource inquiring server to provide the fast computation for fault diagnosis. The original data and diagnosis results are stored in the data center for comparison and also for recalling to train the methods used. With the access between test equipment and the architecture, the architecture can fulfill rapid calculation, improve fault diagnosis efficiency and accuracy, and then reduce life cycle costs.

The rapid growth of air transportation demand led to numerous studies on the airport gate assignment problem (AGAP). As the problem scale gets larger, mathematical programming is no longer available, and heuristic methods like genetic algorithm (GA) have been applied. Taking into account of adding soft constraint to the model and ameliorate objective value of the AGAP, this paper proposes an improved GA considering structural properties to avoid GA’s prematurity. A dynamic topology integrated in crossover operator contributes to a better tradeoff between convergence speed and quality of solutions. Finally, experimental results illustrate the effectiveness of the proposed improved GA to solve AGAP in comparison with traditional GA and a reliable commercial software CPLEX CP Optimizer.

Unmanned aerial vehicles (UAVs) are widely applied in the civil domain. Meanwhile, the security issues concerning UAVs are emerging because the monitoring of small UAVs is difficult. In this paper, we designed a sound detection and alarm system for UAV monitoring. Firstly, the frequency domain feature extraction and Mel domain feature extraction are performed to obtain the features of UAV sound. The Mel frequency cepstral coefficient (MFCC) and support vector machine (SVM)-based sound detection method are designed. Besides, the real-time sound acquisition and detection are realized. The real test data has proved the efficiency of the designed UAV sound detection system. The advantages of the sound detection and alarm system in this paper are low cost and high accuracy, which may be widely applied in the area of UAV monitoring.

Quantitative relationships between land-use pattern and the water quality of river nets in the delta plains where the megacities lie need to be clarified. We selected Qingpu District of Shanghai City located in Yangtze Delta as a case study. Because the water flows in this area are almost static, the bad water quality might result from adjacent land uses. We calculated spatial patterns of urban, agricultural, and forest land uses related to the water quality in this area. In our method, remotely sensed large data were first registered into the same spatial resolution and radioactive level. Then, data were clipped via the boundary of study area. We used supervised classification algorithm (support vector machine) to obtain land-use classes with promising accuracy, followed by calculation of land-use pattern metrics. We correlated the water quality parameters with pattern metrics. The results showed that both the proportions of land use and the configurations of urban, agricultural, and forest areas were closely associated with the water quality. It is suggested that an appropriate land-use plan for this kind of region should be undominated regular simple shapes of urban or built-up zonings, aggregated agricultural fields, and large connected forests with complex edges.

Structured simulation environments are widely used in the research of robots. Combining with a kind of structured simulation environment, we proposed a new map building method based on image processing in this study. Firstly, the binary road network is extracted from the image by image processing, which is then thinned with improved Zhang parallel thinning algorithm. Then, Shi–Tomasi algorithm is adopted to detect corners from the obtained thinning road network. To get the connectivity between corners, we proposed a thinning-line searching method. Finally, we stored the obtained environment information in the form of adjacency list to represent the topology of the road network. The experiment was conducted to test our method: A topological diagram of the simulation environment was rebuilt. The experimental results showed that the proposed mapping method is effective and practical.

Androgenic hair which is also named as body hair can be used for identifying criminals and victims. It is a new soft biometric trait. Different with skin mark, tattoo, blood vessels, and so on, androgenic hair is suitable for low-resolution images. Recently, some researches proved that the androgenic hair pattern extracted from low-resolution images also has valuable discriminative power. In the features of androgenic hair pattern that used for identification, hair density which did not separately be described in the previous works is an important feature. This paper focuses on the hair density, describes, and extracts density feature explicitly. The different grade densities are classified, and the density difference is captured by the density descriptors. The three density grades which are called high density, medium density, and low density are classified by the hair density classifier. In this paper, the grades of different densities are tried to use for modifying the affect of the defect in dynamic grid system. The performance of hair density is tested on forensic skin image database. The experiment result shows that the method is efficient.

The development of Integrated Modular Avionics (IMA) has put forward higher and higher requirements for resource integration. Resource integration brings some security problems as well as high efficiency, and the security level of resources directly influences the function level of the system as well as the task level of the system. One of the practical applications in system security is to study the control theory especially in the changes of system security when regulatory relationships between resources vary. In this paper, we proposed a new method based on resource regulatory network to quickly determine which attractor that the affected states will eventually evolve into after perturbation. Compared to Hu’s algorithm, the new proposed algorithm is simpler and more efficient.

Although the BP neural network is recognized as an advanced method for predicting software quality, it still has some drawbacks, which is easy to fall into local minimum solution when training the data with complex structure. In this condition, the predict results are inconsistent with the actual software reliability. Deep learning is an improvement and extension for the neural networks, and deep belief networks (DBN) have a network model of deep structure. As a kind of deep learning model, deep belief networks (DBN) is not easy to fall into local minimum solution and has a higher Match Ratio than BP neural network than BP neural network when training the data with complex structure. Due to these good properties, we try to use DBN model to predict software safety based on the data of software code features, and it has been proved that, compared with BP neural network, DBN model is a better model to predict the software quality.

DoS attack is easy to implement, difficult to prevent, and is one of the most common ways of security threats nowadays. Its effect assessment is an important research content in safety evaluation. In this paper, an index system based on resource consumption type, service quality type, and attack attribute is established. Using the fuzzy synthesis evaluation model, the fuzzy vector and the single value result can be calculated, an evaluation method of quantitative attack effect is proposed. Finally, the feasibility and rationality of the method are proved by an example.

According to the vision of global energy interconnection (GEI), existing domestic and international information and communication key standards for supporting the GEI development are reviewed. The methodology of standard requirements through the “TOP-LEVEL DESIGN” and “UNDERLYING CASE” to find the GAP way is introduced. And the GEI information and communication conceptual model is built. Based on the analyses the standard requirement of GEI information and communication, the recent and long-term standards direction from the need of engineering construction, technical research and development, industrial development are analysed. It is able to provide reference to the GEI information and communication technology innovation and development and technical standard requirement.

The prediction method of service response time based on collaborative filtering technologies has some drawbacks; e.g., it is hard to find similar neighbors when user-item matrix is very sparse and the calculation complexity is too high. Aiming at these problems, a simplified method is presented to predict service response time of Internet of Things (IoT). The presented method is comprised of two sections. First, user-mean method is used to obtain essential prediction accuracy. Then, to improve the final prediction accuracy, the historical response time variation information of IoT services observed by different users is used to improve the prediction values. Furthermore, the amplitude ratios of average service response time between different users are considered to adapt to the IoT service response time prediction. Through the experiments on a real Web service response time dataset, the presented method is validated. By using the simplified prediction method with low calculation complexity, the satisfied prediction accuracy can still be obtained.

Streaming data also known as unbounded data are increasingly common in the big data era. Modern business collects massive amount of data that are never ending. Hence, there is an increasing need for continuous applications that can process streaming data from massive data ingestion pipelines. However, the majority of streaming system in existence remains relatively immature compare to batch process system, for it can be of great challenge for developers to overcome many obstacles, including: reliability, correctness guarantees, and handling out of order data. Fortunately, the latest Spark Structured Streaming provides the ability to tackle these obstacles. In this paper, we design an architecture that is build on top of Spark Structured Streaming to implement a big data processing system that can provide fast, scalable, fault-tolerant ability to process massive unbounded data.

The advent of the Internet of Things (IoT) significantly stimulates the development of context-aware applications. Complex event processing (CEP) is a technology for real-time data processing. However, a single node of CEP engine cannot keep up with the demand of high performance facing on the growing volume of sensor data. Many researches have focused on distributed complex event processing. In this paper, we propose a solution of distributed CEP based on pub/sub mechanism; by leveraging the loose coupling characteristics of pub/sub, it is easy to scale the system. The persistence storage and master–slave structure of the message broker also provide high reliability. Furthermore, we develop a configuration module for users to describe a complex event processing flow by a directed acyclic graph (DAG). The performance evaluation experiment demonstrates that our approach works well with large data set.

With the rapid development of Internet of Things (IoT) and in-depth study of Internet technology, the application of interactive system that depends on collaborative interaction of multiple business processes becomes more and more important in the IoT. In this paper, we present an event-driven multi-process collaborative interaction platform to improve the business processing capabilities of the interactive system and meet the growing business needs of the IoT environment and users. We integrate the IoT paradigm into the traditional business process to make a rapid and timely response to the IoT application. The platform provides visual business modeling (process and event), process management, and process real-time monitoring. Finally, we give an IoT application of the platform in the forest protection.

The issue of China’s economic structure is becoming more and more prominent, and the interactions between various departments have been increasingly strengthened. This paper takes into account the characteristics of the Chinese economy, introduces the structuralist characteristics, and on this basis integrates the financial sector to build a dynamic multi-sectoral CGE system to analyze the impact of the input–output structure on China’s macroeconomic influences. The results show that from 1997 to 2012, the Chinese economy saw changes in its technical structure (intermediate input coefficient) contributing to economic growth, social welfare, and employment in the informal sector.

This paper studies a coordinated production and distribution scheduling problem in flowshop with limited waiting times, which restricts the maximal waiting times in consecutive processing machines and between the production and distribution. This problem usually arises in the manufacturing of timeliness products, such as food and newspaper. In this problem, one order contains many jobs. Job is the smallest unit in the production, and order is the smallest unit of transportation. Moreover, there is one vehicle for transportation, and only one order can be shipped at one time. This problem is a joint decision of production scheduling and distribution scheduling, which is a strongly NP-hard problem. A mixed integer programming model is established for this problem with the objective of minimizing the tardiness, and a co-evolutionary genetic algorithm is presented. Computational experiments were carried out and experimental results verified the effectiveness of the model and algorithm.

Battlefield cyber warfare is different from strategic cyber warfare. It has the characteristics of good closeness, special agreement, complex structure, and good anti-interference immunity. However, cyber warfare and “Shute” systems are not the keys to battlefield cyber warfare. This article analyzes the characteristics of the battlefield network warfare mechanism implemented at the present stage and puts forward some countermeasures to clear up the confusion and accelerate the construction of battlefield cyber warfare strength. It provides some useful ideas to speed up the formation of battlefield cyber warfare system confrontation and system breakout ability.

In recent years, with the development of the network sector, more and more people pay attention to the online learning. However, the traditional method could hardly meet the needs of online learning. Online learning intervention attracts people’s attention. To improve the effectiveness of online learning, this paper calculates a new way which intervenes people’s online learning based on constraint theory. This new way focuses on the five-step method of constraint theory, tries to find out key activities, settles the timing of intervention and adjusts the intervention time according to the completion rate of learning activities. Based on the 50% rule, this paper calculates a new method—linear grading method—to calculate intervention time ensuring that the online learning can be successfully completed.

The utilization of recommendation system is becoming more and more common in our life. However, using user’s rating matrix to make recommendation (such as SVD, SVD++), traditional recommendation systems only focus on user’s rating information, yet the text information which contains more amount of information has not been effectively used. Meanwhile, previous recommendation algorithms did not consider the partial order relationship between items, which hampers the achievement of a better result. Therefore, we propose a combination algorithm which not only uses text information but also takes the partial order relationship into consideration. In this paper, we incorporate the text information into the recommendation learning algorithm Bayesian Personalized Ranking (BPR) as a hidden factor, which is learned from the topic model, in order to make the model more effective and more explainable and to solve the Cold Start problem to a certain extent. Meanwhile, the final experimental results prove that our model is feasible and effective.

E-learning systems offer new ways of learning and change the approaches of delivering learning materials. In the face of massive instructional data, efforts should be made to help learners choose more suitable materials based on their level and preferences. In this paper, we proposed a model of online learning style which offers a comprehensive description of learners’ cognitive style and other abilities. Then, we introduce an enhanced method for course recommendation based on our learner model. Experiment results show that the enhanced method evidently outperforms traditional collaborative filtering method, while maintaining a computational advantage.

Data centers provide services for various real-time applications, such as social networks, instance message, which produce a large number of bursty and urgent mice flow. However, the traditional flow queuing model cannot schedule these flow effectively, which leading the performance loss for some important applications. In this paper, a preemptive scheduling in data center for bursty flow (PSBF) was proposed. The scheduling leverage the preemptive scheduling for more critical flow and set the right queue for other data flow to keep the high aggregated throughput. The experiment and analysis show that our scheduling can handle bursty flow better than traditional flow scheduling for data center.

Cyber security visualization is becoming a hot research field. Visualization and interactive analysis can greatly help network managers and analysts monitor network, detect anomalies, and assess the situation of network. A large number of papers have been published in this field, and a lot of novel visual tools have been proposed. In this paper, we provide a comprehensive review of visualization systems for cyber security. Firstly, we introduce the advantage and importance of cyber security visualization. Then, we summarize the categorization of cyber security visualization systems. At last, we draw conclusion by evaluating these systems and prospect for future research.

This paper studies and explores a suitable blockchain solution for the IoT platform, enabling the platform user and physical devices or the virtual devices to participate in and make data or service transaction. We research on suitable blockchain-related technologies in use; consider the advantages when using blockchain to improve IoT platform’s ability, efficiency, and security, based on the different existing blockchain consensus; benchmark the performance in the IoT scenario; and make a comprehensive discuss.

Portal-based rendering as one of the visibility determination methods is widely used in large-scale indoor scene. The main idea behind portal-based rendering is that when a given portal is visible, the cell which is behind the portal is visible, and otherwise it is invisible. In order to simplify the complexity of visibility determination, they always define portal as a planar convex polygon and the usefulness of portals for scene composition is limited. A more useful portal is a general portal which can be an arbitrary polygon or be thick and connect any two independent cells, and we call it arbitrary portal. In this paper, we provide an improved scan-line algorithm, which is based on portal texture for rendering arbitrary portal. First, we simplify an arbitrary portal to rectangle portal. Then, we use ray-casting method to sample the visible area of rectangle portal and get the corresponding portal texture. Thirdly, an improved scan-line algorithm is created to render arbitrary portal correctly by using portal texture. Our focus is to discuss the details of the improved scan-line algorithm. We have devised some tests to verify the correctness of our algorithm. The testing results showed that our system could run smoothly on the general personal computer and render the arbitrary portals correctly.

At present, most RTU of the power plants adopt redundant configuration, host–standby switching mode in our country. Because of the two RTUs actually running in the virtual machine, the dispatching side cannot master the working state of the other machine in real time, which can have a great impact on the dispatcher in the actual operation. In this paper, a dual-machine real-time monitoring and maintenance scheme is proposed, which can not only guarantee the dual-machine monitoring of the plant and the station but also timely detect the plane failure of the dispatching data network. Then, the correctness and reliability of the monitoring information are improved effectively.

Image classification has constituted a “hot” issue in large-scale image data processing in recent years. In this context, a significant number of using deep learning theory has been proposed for machine learning. Some works had a great development; however, most works deal with these problems need training a lot of parameters, so this process cost a lot of time and resources. This chapter proposes a hierarchical learning method, based on distributed Hadoop, to solve these problems. Initially, input images are extracted randomly in image blocks and reduced, in terms of redundant information, to maintain key information. Next, a regularization algorithm and noise processes are used on the image blocks between multiple image nodes in parallel. The following steps are to extract the dictionary of image classification by k-means algorithm. Finally, in order to select the feature mapping function in above dictionary, and mapped to a new image feature, it will be as a vector. The experimental results show that the algorithm is both reasonable and effective compared with the results of K-means.

In recent years, the market of cloud solutions in various branches of science and industry has shown significant development. Cloud technologies and virtual services allow to reduce capital and maintenance costs, expenses for IT support and labour costs, and to better define investment plans in IT. At the same time, clouds are typically more secure and reliable, easier to use and better scaled than locally deployed solutions. They promote innovation in business, allow to respond more quickly to demand and provide the market with new products or services. In this paper, we present our work on applications of cloud technologies in oil production automation processes. We describe our project which uses of cloud computing to solve problems related to the processing of extra-massive volumes of data in the geophysical field.

Despite the current relatively low hydrocarbon prices, world oil reserves are dwindling and extracting hard-to-recover reserves and high-viscosity (heavy) products are becoming increasingly important. This paper presents the results of a study aimed at creating a new concept for a dynamic model of a heavy oil field-based artificial neural networks. The main idea is to replace current industrially used complex calculation schemes by neural network modelling. Thus, production models and formulae that are typically approved by an expert are replaced by an “intelligent core” that is formed on the basis of historical data on field processes and, in real time, determines the state of objects and processes, while giving recommendations to the operator for adequate management decisions in accordance with production regulations. In the course of the research, it was revealed that the task of recognising the states of oil field objects is very close to the task of pattern recognition. We therefore propose the use of convolutional neural networks (CNNs), whereas, to comply with the specifics of the domain, modification of established CNN methods has been implemented. This paper presents the main stages of the proposed CNN intelligent decision support system for digital oil field models.

The wide range of pumping units types, which are part of the complex for preparation and pumping of oil, limits the ability to accurately predict their conditions. While various neural network models can be employed for this problem, one of the most effective solutions is to use a combination of independent neural networks to obtain improved accuracy in comparison with a generalised complex network model. In this paper, we propose the application of fuzzy genetic algorithms to identify the optimal set of weights for such a network ensemble. In our approach, generations are randomly divided into two groups of individuals. Each member of the first group is marked on the basis of a fitness function. The corresponding second chromosome is then selected based on the fitness label and the population diversity of the previous generation using a set of fuzzy rules. To demonstrate its usefulness, we compare the results of our proposed method with conventional back-propagation neural networks.

The paper deals with the idea of modeling and simulating student’s abilities and the effect these abilities will have on his success in a university course. It is well known that new university students come from different schools, different backgrounds, and have different skills of their own. The experience and knowledge they acquired in secondary schools can help them to better understand the content of some university courses. Their skills can also help them to learn and to better understand the syllabus of these university courses. Of course, this is not the case for all students. Some come from middle schools that were not specialized in informatics or in some cases had nothing to do with it at all. To increase the chance to understand the content for all students, we want to make a model that would represent their knowledge and skills, and then we will use it to simulate their transition through a course. When we see that certain students have difficulty to deal with certain parts of these courses, we can adjust these parts to be easier to understand.

The influence of different types of districts on urban road capacity under different degrees of openness is discussed in this paper. Four evaluation indexes including the road capacity expansion degree, traffic instability variation coefficient, simplified OD (origin and destination) connectivity, and path betweenness are established, and simulation models of multi-objective decision making and traffic network vulnerability based on evaluation indexes are built. Finally, the influence of three different types of typical residential areas on the capacity of the surrounding roads before and after the opening is analyzed and evaluated under simulation models.

Apart from other common problems, judicial decisions are bound to an official and stiff format, lack of individualized considerations, and sticking to traditions. All these give the parties concerned, and the public an impression that judicial decisions are unreasonable. The origin of such a bad impression lies in the fact that judicial decisions are over-bound to a standardized template. Such a case breaks the organic relationship between all parts of a specific case. Moreover, the literal expressions also lower the authority of a court gravely, thus having an adverse impact to some extent. This article attempts to make a comparison of judicial decisions made on key cases in China based on including computerized keyword extraction and high-frequency word co-occurrence technology and by creating a database for judicial decisions. It takes the result of keyword extraction and data detection as a basic indicator for examining judicial decisions and then making an analysis on the linguistic features of judicial decisions. It is concluded that in view of pragmatics, juridical decisions shall be fuzzy and definite in language, reasonable and objective in language expression, as well as precise and concise in language expression effect, etc.

Accurate and timely prediction of safety stock is of great significance to reduce costs and improve economic efficiency. It is a problem that needs to be solved urgently in the manufacturing industry. Because there is a complex nonlinear relationship between factors that affect the safety inventory, it is difficult to obtain better prediction results with traditional methods. This paper presents an intelligent safety inventory prediction model based on artificial fish swarm algorithm and applies it to inventory management of Quanzhou chemical fiber enterprises. The research has important theoretical and practical significance for the inventory management of chemical fiber enterprises and even private enterprises.

Now, computer science has been applied to display, store the patient’s related information etc., condition diagnosis and prediction is largely dependent on the doctor. With the application of computer science to disease prediction and diagnosis, it will predict and deal with the disease, and it will reduce the wrong diagnosis and so on. This study will provide a method for chronic kidney disease (CKD),and will combine with BP neural network method and DS evidence theory algorithm to make predictions about the diagnosis of the disease; the influence of subjective factors on the prognosis of the disease was verified.

In order to promote the natural human–computer interaction in classroom, a set of general dynamic teaching gestures based on RealSense is designed and the gesture segmentation and recognition are realized. In this paper, a real-time gesture segmentation algorithm based on state automata is proposed to obtain gesture data. According to the hand joint data provided by RealSense SDK, the hand feature based on the gesture trajectory and joint is extracted, and then the recognition algorithm based on dynamic time warping is used to identify it. Finally, the error recognition rate is effectively reduced by the threshold training for the teaching application. Experiments show high accuracy and real-time performance of the method.

In recent years, the spreading of malicious social media messages about financial stocks has threatened the security of financial market. However, identifying these threats from noisy social media datasets challenge both research and practitioner communities. This paper describes a system named ASMODS for intelligent detection of abnormal stock price movements in response to social media postings. The system consists of intelligent modules to develop synchronized data collections of social media messages and stock prices, to identify abnormal price movements, and to balance the datasets for machine learning. Empirical findings show that ASMODS could enhance the effectiveness of logistic regression, artificial neural networks, and decision trees on detecting abnormal stock returns. The results have strong implication for cybersecurity in the financial market.

Dual-polarized antennas have became the main options of MIMO deployment, However, existing channel state information (CSI) feedback schemes ignore the polarization leakage between the polarization directions, which could result in significant performance degradation. To optimize limited feedback precoding scheme, this paper proposes a practical channel model for dual-polarized MIMO system and a novel CSI feedback schemes based on summarizing codebook design guidelines of precoding. The model formulates the channel as the sum of two components, i.e., the ideal polarization channel and the polarization leakage channel by taking polarization leakage into consideration. In the multi-components feedback scheme (MFS), the ideal polarization channel and the polarization leakage channel are fed back using predetermined cookbook. System simulation is conducted to validate the performance of the proposed schemes, and the results show they outperform existing CSI feedback schemes significantly.

Depth map acquisition is the key technology in free viewpoint system for multi-view video plus depth (MVD) video format, which can effectively improve the coding efficiency and the quality of synthesized virtual view in 3D video. Acquainting depth map can be accomplished by Kinect depth camera. But hole and noise problems are needed to repair and improve. In this paper, the depth map is inpainted and optimized. Firstly, depth map is clipped from the center of the image and then filled with foreground pixels. Finally, the proposed trilateral filter is used to optimize the image. The experimental results show that the better depth map can be obtained by our method.

The seasonal and diurnal variation characters of land surface temperature (LST) and their relationship with normalized difference vegetation index (NDVI) had been researched in Guangxi, with the support of GIS spatial analysis and statistical techniques based on MODIS LST products MOD11A2, MYD11A2, and MOD13Q1 in 2014. Results showed that: (1) spatial distribution of LST varied in different seasons, high LST was mainly scattered and distributed in the middle and south parts of the study area. (2) The diurnal LST variation was higher in nighttime than daytime, the nighttime LST at 22:30 p.m. was the highest, and the daytime LST at 22:30 a.m. was the lowest of all. (3) The LST and NDVI had no significant negative correlation in seasonal scale.

Dongba culture is the Naxi traditional culture. It is known as the “Encyclopedia of ancient society of Naxi nationality.” In 2003, Dongba scripture written by Dongba hieroglyphic is listed into Memory of the World Heritage List by UNESCO. In 2006, Dongba papermaking and Dongba painting are selected into the first batch of national non-material cultural heritage list. In this paper, through reviewing the protection and heritage of Dongba culture, as well as the opportunities and challenges brought about by tourist industry to the development of Dongba culture, we find that the digitizing approach is a viable way to protect the Dongba culture at present. However, we must continuously overcome the disadvantages existing in traditional digitizing approach and make the Dongba culture combine with digital technology perfectly through the measure, such as various digital forms, dynamic digital protection approaches, strengthening online communication and cooperation and the establishment of Dongba culture digital platform, to finally achieve the widespread, heritage, and protection of Dongba culture, promote cultural resource advantages to convert into tourist and economic advantages in Lijiang, and promote Dongba cultural protection and the better and faster economic and social development.

The talents with experimental and innovative spirit pay attention to the quick change of multimedia status every day. Since the beginning of modernism, the developing evolution of international culture and art has integrated and produced many achievements with general significance. By paying attention to and comprehensively absorbing the multimedia means and thoughts, we can get rid of the conventional education mode and we will not feel helpless to judge the new values. From the perspective of artistic education, the emerging pan art form, new medium courses and comprehensive material media are increasingly paid attention to in creation presentation. Therefore, it is supposed to look into the utilizability of interdisciplinary subjects with keen insight and extract new technologies and artistic rules to allow the forms of public art to make progress extensively and make more and more people imagine the beauty beyond the view.

[Objective] The paper was to formulate catchable size and total allowable catch of crucian carp in the Yellow River waters. [Method] Based on analysis of length–weight formula, body length and weight growth equations, and instantaneous mortality rate, the inflection age and critical age and weight growth were calculated, and the biomass of crucian carp in the Yellow River waters was estimated. [Result] The growth equation of body length was Lt = 278 * [1 − e−0.38 (t + 0.39197)], and the growth equation of the body of weight was Wt = 680.6 * [1 − e−0.38(t + 0.39197)]3.2049; the inflection age and critical age for weight growth of crucian carp were 2.67 and 3.01 a, respectively. The biomass in the Yellow River waters area was about 6088 t. [Conclusion] For the crucian carp flock in the Yellow River waters area, the catchable length was about 191.2–201.6 mm, the catchable weight was about 204.9–243.4 g, and the total allowable catch was about 7186.83 t.

Three commonly used robust statistical methods are introduced in this paper, namely quartile method, algorithm A, and Q/Hampel method. Two application examples are presented, and results of the three methods are compared. For data set of symmetrical distribution and small proportion of outliers, quartile and algorithm A are applicable. For non-symmetrical shape data set or a large proportion of outliers (>20%), Q/Hampel method is recommended. Code for Q/Hampel method is also provided to help quick calculation for this method.

Mutation testing is a fault-based software testing technique that can be used to measure the fault detection ability of a test set. However, high cost of the mutation testing has seriously hindered its wide application in the software testing. The fact shows that most of the mutants in mutation testing are redundant mutants. Recent researches have proposed using the subsuming mutants to reduce the number of redundant mutants. However, the existing algorithms are inefficient in identifying the subsuming mutants of real projects. This paper based on the concept of mutant utility, and researches on the relationship between mutant utility and the statement blocks. The result shows that the average utility of the mutants at the leaf node or near the leaf node in the block dominator graph is higher than that of the mutants at the other nodes. That is, the average mutant utility is higher in the dominated statement blocks generally.

Dependency parsing is a task of automatically parsing dependency grammar for a sentence in natural language, which sustains advanced applications such as machine translation, question answering. This paper thoroughly studies influential and state-of-the-art works of the two major classes of approaches: transition-based parsing and graph-based parsing. This in-depth comparative analytical study mostly focuses on fundamental concepts and current trends and has comprehensively analyzed state-of-the-art implementations of different approaches.

In recent years, as the structure of smart substation network changes, the prediction of high-quality network traffic becomes more and more important. Once the traffic is abnormal, it will affect the reliability and real-time relay protection device. In this paper, the wavelet transform and BP neural network algorithm are combined to construct and analyze the wavelet neural network model. Based on this, the improved particle swarm algorithm is used to replace the gradient descent training method of wavelet neural network to improve the wavelet neural network Easy to fall into the plight of the minimum, improve the convergence rate. Finally, taking the network traffic data of station-level switch in a smart substation as an example, the simulation is carried out based on the collected original frequency data. The experimental simulation shows that the improved particle swarm optimization wavelet neural network model prediction accuracy and convergence rate is better than the traditional model, thereby improving the accuracy and speed of smart substation network traffic forecasting.

According to the working condition and the common fault causes of the water-injection pump, this paper introduces a data acquisition and signal processing circuit of the oil field water-injection station pump vibration detection system. In terms of hardware, the scheme design consists of data acquisition module based on the three-axis acceleration sensor ADXL345 and microprocessor STM32F030, data processing and transmit module based on the 433 MHz wireless transparent data transceiver module YL-100IL and microprocessor STM32F407, and the PC module for the analysis and display of the vibration signal for the pump. For the processing and analysis of the vibration signal, the acquired acceleration signal is processed using the Empirical Mode Decomposition method to reduce noises, after which the frequency spectrum of the vibration signal can be obtained. Due to the fact that different faults of the pump can produce different characteristic frequencies, the frequency spectrum gained from using Empirical Mode Decomposition can be used to the analysis of the water-injection pump fault diagnosis. Finally, field tests prove that the circuit in the system works stably and meets the design requirements. It is able to achieve the function of vibration signal acquisition, processing, and diagnosis of the working condition.