Table of Content

19 August 2021, Volume 0 Issue 08

Previous Issue    Next Issue

Statistical Analysis of Wireless Sensor Network Monitoring Quality

BAI Xue, CHENG Zong-mao

2021, 0(08):  1-5. 

Asbtract ( 220 )   PDF (1008KB) ( 129 )  

References | Related Articles | Metrics

As a new network technology, wireless sensor network is a hot research field at home and abroad. It is a new network technology, which can collect the data information distributed on the network in real time, and transmit the information to the gateway node, and finally complete the complex network monitoring and tracking work. In order to solve the challenges faced by wireless sensor networks, for the problem of target coverage in wireless sensor networks, considering the exponential distribution of unknown random event parameters, the monitoring quality of random events is statistically analyzed, and the coverage problem is optimized under the background of wireless sensor networks. Firstly, the background and current situation of wireless sensor network are introduced. The purpose of this paper is to analyze the monitoring quality of wireless sensor network. Secondly, the coverage of wireless sensor network is optimized, the corresponding network model is established, and the optimized simulated annealing algorithm is designed. Combined with statistical knowledge analysis, the monitoring quality is studied by parameter estimation. Finally, the rationality and effectiveness of the method proposed in this paper are verified by the simulation experiment, and the purpose of prolonging the network lifetime is finally achieved.

Routing Protocol of Cluster Tree Wireless Sensor Network Based on Cloud Security Model 

ZHAO Ji-ye

2021, 0(08):  6-10. 

Asbtract ( 240 )   PDF (827KB) ( 108 )  

References | Related Articles | Metrics

Because of the large energy consumption variance of the cluster head of traditional cluster tree wireless sensor network routing protocol, the number of surviving nodes and the residual energy of the nodes are less, and the service life of the wireless sensor network is reduced. So a cluster tree wireless sensor network routing protocol based on cloud security model is designed. By calculating the transmission energy consumption of clustered tree wireless sensor network when receiving and sending data, the cloud security model is used to obtain the uncertain relationship between various elements of the cloud security situation, and the comprehensive trust value of nodes are predicted. Based on the predicted results, the ant colony algorithm is adopted to obtain the optimal path of the partition nodes, and the clustered tree-shaped wireless sensor network routing protocol is completed. The experimental results show that the designed routing protocol cluster head energy consumption variance is small, the number of surviving nodes and the remaining energy of the nodes are more, and the amount of received data packets is 48.1% and 22.6% higher than other 2 kinds of protocols respectively. It can be seen that the routing protocol designed by this paper extends the service life of clustered tree wireless sensor network.

Improved Discrete Firefly Optimization Algorithm to Solve Flexible Job Shop Scheduling Problem

ZHENG Jie , PAN Da-zhi ,

2021, 0(08):  11-15. 

Asbtract ( 261 )   PDF (965KB) ( 98 )  

References | Related Articles | Metrics

Aiming at the problem that when solving the flexible job shop scheduling problem (FJSP), the traditional swarm intelligence optimization algorithm has some disadvantages, such as insufficient optimization ability and easy to fall into local optimum, taking minimizing the maximum completion time as targets, firefly algorithm (FA) is applied to solve flexible job shop scheduling problem (FJSP), and an improved discrete firefly algorithm (DFA) is proposed. Firstly, the relationship between the FA continuous optimization problem and the FJSP discrete optimization problem is established through two-stage coding. Secondly, a population initialization method is designed to ensure the quality and diversity of initial solutions. Then, an improved discrete firefly optimization algorithm is proposed and a local search algorithm is introduced to enhance the global search ability and local search ability of the algorithm. Finally, the standard example is simulated and the validity of DFA algorithm for FJSP is verified. Through simulation comparison with genetic algorithm and particle swarm optimization algorithm, the superiority of DFA in solving FJSP is verified.

Intelligent Test Paper Generation Strategy Based on  Particle Swarm Optimization Genetic Algorithm

CHEN Chun-yan, LIU Meng-chi

2021, 0(08):  16-23. 

Asbtract ( 337 )   PDF (1928KB) ( 159 )  

References | Related Articles | Metrics

Online exams abandon the inherent shortcomings of traditional paper exams and have been widely used in the field of online education. The test paper of artificial intelligence is one of the important techniques for completing online examinations efficiently. The question of test paper is a multi-development goal combination and optimization problem, and generally has several solutions. Artificial intelligence algorithms have obvious advantages in finding the optimal solution of such problems. This paper first analyzes and studies the current mainstream intelligent test paper generation algorithm, combines the relevant principles of test paper generation and mathematical experiment models, and proposes an intelligent test paper generation strategy based on particle swarm genetic algorithm. The particles, individual extremes in the population and the extremes of the population are performed the crossover operation in the genetic algorithm and the mutation operation of the particle itself. At the same time, the algorithm performance is improved by adaptively adjusting the crossover probability and the mutation probability, and the segmented real number encoding. Finally, a comparative experiment is taken to prove the advantages of the algorithm.

Recipe Recommendation Algorithm Based on Knowledge Graph and Collaborative Filtering

GENG Hua-cong, LIANG Hong-tao, LIU Guo-zhu

2021, 0(08):  24-29. 

Asbtract ( 647 )   PDF (920KB) ( 210 )  

References | Related Articles | Metrics

In view of the traditional collaborative filtering-based recipe recommendation algorithm that only uses the user-item score matrix and does not consider the semantic information of the item itself resulting in low recommendation accuracy, this paper introduces the semantic information between recipes as an important recommendation basis by constructing a knowledge graph, and proposes a personalized diet recommendation algorithm based on knowledge graph embedding and collaborative filtering. By representing the recipe entity and relationship in two different low-dimensional continuous vector spaces, the semantic similarity between the dishes is calculated, and the semantic similarity is incorporated into the collaborative filtering recommendation for recommendation. The method in this paper alleviates the problems of data sparsity and cold start by strengthening the use of hidden information between dishes, and makes the recommendation result more reasonable. Experiments on the dataset show that the method has a significant effect on recipe recommendation, and it has a significant improvement in recall and AUC.

Review of Radar Imaging Simulation

ZHOU Xiu-zhi, CUI Yi-peng, SUN Zhong-yun

2021, 0(08):  30-34. 

Asbtract ( 494 )   PDF (3750KB) ( 180 )  

References | Related Articles | Metrics

Radar imaging for military flight simulator training plays an important role in simulation training because of its advantages of no limitation of time and space, controllability and strong security. In this paper, three main ground surrveying and mapping methods (RBM, DBS and SAR) for airborne fire control radar, domestic and foreign fighter radar models and their imaging systems are introduced. The research and development status of foreign radar imaging simulation software and domestic radar simulator are described. The advantages and disadvantages of existing products, their applicable scope and technologies are described in detail. The main methods and realization process of radar image simulation are summarized, which are divided into three categories: echo signal simulation, transfer function and fusion of radar image features. It is predicted that radar imaging simulation will develop in the direction of establishing general database, expanding imaging range, and improving real-time performance and fidelity.

Campus Fire Training Simulation Based on DI-Guy

ZENG Xu-yang, SUN Ka

2021, 0(08):  35-39. 

Asbtract ( 227 )   PDF (2899KB) ( 89 )  

References | Related Articles | Metrics

Aiming at the problems of high population density, difficulty of organizing fire training efficiently and lack of experience in dealing with emergency events in campus, a virtual simulation training method based on DI-Guy software is proposed. First, we build the three-dimensional model of campus using MultiGen Creator software and load it into the visualization window of DI-Guy for construction. Then we simulate the fire scene and the process of fire evacuation. Finally, we analyze the effect of crowd evacuation scheme. At the same time, some key technologies in the simulation process, including the establishment of user-defined model, the realization of special effects, the behavior control of characters and collision detection, etc, are researched and applied. The simulation results show that the proposed simulation model can reasonably reflect the crowd evacuation situation and play a good display role. It provides a feasible scheme for the simulation of campus emergency events on DI-Guy platform.

Dynamic Texture Synthesis Model Based on Self-correction Mechanism

YI Hui-min, ZHU Zi-qi

2021, 0(08):  40-45. 

Asbtract ( 203 )   PDF (3469KB) ( 66 )  

References | Related Articles | Metrics

Dynamic texture is one of the dynamic models in computer vision. It has statistical stationarity in the spatial range and random repetition in the time dimension. The goal of dynamic texture synthesis is to generate an image that is visually similar to a given texture. When performing dynamic texture synthesis, the accumulation of regression prediction errors is a key issue of leading to the degradation of texture quality. Therefore, this paper proposes a dynamic texture synthesis model based on self-correction mechanism. The model uses indicators such as clarity, structural similarity, and optical flow to determine the optimal data range and finds the optimal extreme point. Through the self-correction mechanism, the original data is replaced with optimized data, and the optimized data is used for regression prediction. Finally, a convolutional autoencoder is used to reconstruct the prediction data into high-dimensional dynamic texture video frames. Experiments  are conducted on the DynTex database and the model proposed in this paper is compared with several typical dynamic texture synthesis models. The experimental results show that, compared with other models, the Mean Square Error (MSE) value calculated by the dynamic texture video frame and the real video frame synthesized by this model is smaller, and the values of Peak Signal to Noise Ratio (PSNR) and Structural Similarity (SSIM) are larger. It solves the problems of residual image, blur, and noise in dynamic texture synthesis, so as to generate a better visual effect and longer dynamic texture sequence. At the same time, the effectiveness of the proposed modeling method is verified.

Non-intrusive GUI Input Item Recognition Based on Visual Feature

WANG Yan, QIAN Ju

2021, 0(08):  46-51. 

Asbtract ( 284 )   PDF (1171KB) ( 72 )  

References | Related Articles | Metrics

Recognizing input items in an application interface is a basis of automatic GUI test data generation. How to get information of input items in existing interface depends on the information provided by the system of the object under test. They cannot be used under non-intrusive scenarios. This paper proposes a non-intrusive GUI input item recognition method based on visual features. The method detects input items in a GUI interface from GUI images obtained from sources like external cameras. It first uses line detection and contour detection algorithms to obtain a list of candidate input items, then uses support vector machine (SVM) to determine which of these candidates are real input items based on visual features. The method does not rely on the underlying system of the target under test. Experimental results show that it can recognize the input items in a GUI interface with a good accuracy and hence is of effectiveness.

Clinical Electrocardiogram Classification Algorithm Based on Deep Learning

LIU Shou-hua , WANG Xiao-song , , LIU Yu ,

2021, 0(08):  52-57. 

Asbtract ( 463 )   PDF (1215KB) ( 187 )  

References | Related Articles | Metrics

Electrocardiogram (ECG) which can reflect the health state of human heart is widely used in clinical examination on heart diseases as an important basis. With the increasing number of ECG data, the demand of  the computer-assisted electrocardiogram analysis has become urgent. Electrocardiogram automatic classification as an indispensable technical means of computer aided electrocardiogram analysis has important medical value. However, because of the weakness and low anti-interference of ECG signal, the traditional ECG classification algorithms have the problems of good effect on test set and poor effect in clinical application. So, this paper introduces a ResNet network structure of one-dimensional convolution based on multi-lead two-dimensional structure, increases the diversity of training samples by means of data enhancement such as translation starting point and adding noise, and uses Focal Loss function to optimize the ECG classification model of individual patients. The model uses 20000 complete 8-lead ECG data and a total of 34 types of abnormal ECG events for classification experiments. The results obtained are: F1 score 0.91, accuracy 93.96%, recall rate 87.89%. Experiment results show the proposed algorithm has better ability of deep feature mining and classification, which verifies its effectiveness in arrhythmia classification.

Eye-movement Tracking Based on Deep Neural Network for Portable Devices

WANG Jian-hua, RAN Yu-kun

2021, 0(08):  58-63. 

Asbtract ( 166 )   PDF (1769KB) ( 89 )  

References | Related Articles | Metrics

Aiming at the problem that the current eye-movement tracking methods can not be applied to intelligent mobile phones, tablet computers and other portable devices, an eye-movement tracking method based on large-scale data sets is proposed. Firstly, a large-scale data set is constructed by crowd-sourcing method. Then a deep neural network is trained with the data set for end-to-end prediction. Finally, a smaller and faster network is trained to optimize, which makes the proposed method run in real-time on mobile devices. Experimental results show that the proposed method has better tracking robustness and data generalization ability than other similar methods. The speed of running in mobile devices can reach 10~15 frames per second. The prediction errors of this method are 1.71 cm and 2.53 cm respectively in mobile phone and tablet computer without correction. After calibration, the errors are reduced to 1.34 cm and 2.12 cm respectively.

Charging Socket Detection Algorithm Based on Region Proposal Net and LBP Feature

REN Chao-dong, ZHANG De-li

2021, 0(08):  64-69. 

Asbtract ( 343 )   PDF (5911KB) ( 102 )  

References | Related Articles | Metrics

With the large-scale promotion of electric vehicles in the world, people pay more and more attention to the automation of charging for electric vehicles. In the process of automatic charging, the key step is the detection and identification of charging socket, and the completion of docking and plugging between charging socket and charging gun. This paper proposes a charging socket detection and recognition algorithm based on Faster-RCNN, improves the RPN network part by combining the saliency image, strengthens the charging port area in the images, and uses the processed feature image as the input of RPN network. A multi-scale MB-LBP feature is designed to classify the candidate regions with neural network. The training and testing on the self-built dataset are conducted based on the Pytorch framework. The experimental results show that the proposed algorithm can meet the needs of the work scene and deal with the changes of illumination and seale.

Software Defect Prediction Based on Hybrid Sampling and Random_Stacking

YAN Ling-ling, JIANG Feng, DU Jun-wei, YANG Ai-guang

2021, 0(08):  70-76. 

Asbtract ( 243 )   PDF (1036KB) ( 92 )  

References | Related Articles | Metrics

The existing software defect prediction methods  face problems such as imbalance of data categories， high-dimensional data processing， and so on. How to effectively solve the above problems has become a research hotspot in related fields. Aiming at the problems of unbalanced categories and low prediction accuracy faced by software defect prediction, this paper proposes a software defect prediction algorithm DP_HSRS based on hybrid sampling and Random_Stacking. The DP_HSRS algorithm firstly uses a hybrid sampling algorithm to balance the unbalanced data, then uses the Random_Stacking algorithm to predict software defects on the balanced data set. The Random_Stacking algorithm is an effective improvement to the traditional Stacking algorithm. It constructs multiple Stacking classifiers by fusing multiple classic classification algorithms and the Bagging mechanism, votes multiple Stacking classifiers to obtain an integrated classifier, and finally uses the integrated classifier to predict software defects. The results of experiments on the NASA MDP data set show that the performance of the DP_HSRS algorithm is better than the existing algorithms, and it has better defect prediction performance.

A General Fast Design Method for Floating Point Arithmetic Instruction Based on TEC-XP16 Teaching Machine

ZONG De-cai , WANG Kang-kang

2021, 0(08):  77-84. 

Asbtract ( 270 )   PDF (827KB) ( 65 )  

References | Related Articles | Metrics

To solve the problem that TEC-XP 16 teaching machine does not have floating point arithmetic instruction, a general design and implementation method for 32-bit floating point arithmetic instruction is emphatically studied on TEC-XP16 microprogram controller. To solve the problem that manually designing microprogram and modifying the source program of microprogram controller is  inefficient and  error prone, a fast method which can automatically produce microprogram according to assembly program is put forward. And some database tables are designed and a program is written in Python language which can automatically modify the source program of the controller and generate a new source program file of ABEL language according to the contents of the database tables. The experimental results show that the design of 32-bit floating point arithmetic instruction is correct. The program written in Python language can automatically produce microprogram in less than five seconds and modify the source program of the controller in less than two seconds on average, which greatly improves the efficiency for designing floating point arithmetic instruction. The method can also be extended to the design of other complex instructions.

Information System Risk Assessment Model Based on Self-Adaptive Expert Weight

LU Sai, ZHUANG Yi

2021, 0(08):  85-93. 

Asbtract ( 249 )   PDF (2029KB) ( 90 )  

References | Related Articles | Metrics

Aiming to the problems that it is difficult to set the expert weight reasonably and the assessment result is greatly affected by the subjectivity of experts in the process of risk assessment, an information system risk assessment model SAEW-ISRA based on self-adaptive expert weight is proposed， and an adaptive adjustment method of fine-grained expert weight is presented. Firstly, triangular fuzzy number is introduced to score the attribute of risk indicators in the process of assessment. Secondly, the level of expert’s knowledge is described according to the fuzziness of expert score, and the posterior weight is constructed according to the distance from the average score of expert group, making the expert’s weight be adjusted adaptively; At the same time, the fuzzy analytic hierarchy process is used to construct the weight of risk indicators. Then, a risk quantification method of information system risk indicators is proposed, which can calculate the risk value. Finally, through an example of risk assessment of an information system, it is verified that the proposed method can achieve higher evaluation accuracy, and solve the problem of unreasonable weight in the evaluation process to a certain extent.

News Label Classification Based on BERT and Deep Equal Length Convolution 

YANG Wen-hao, LIU Guang-cong, LUO Ke-jing

2021, 0(08):  94-99. 

Asbtract ( 340 )   PDF (1036KB) ( 80 )  

References | Related Articles | Metrics

For the THUCNews’ Chinese news text label classification task, a news label classification model (DPCNN-BERT) that combines multi-layer equal-length convolution and residual connection based on BERT pre-training language model is proposed. Firstly, by querying the Chinese vector table, each word in the news text is converted into a vector and input into BERT model to get the full-text context of the text. Then, the local context relationship in the text is obtained through the initial semantic extraction layer and deep equal-length convolution. Finally, the predicted label of the entire news text is obtained through a single-layer fully connected neural network. The model proposed in this paper is compared with the convolutional Neural Network Classification Model (TextCNN), Recurrent Neural Network Classification Model (TextRNN) and other models. The experimental results show that the prediction accuracy of the model reaches 94.68%, and the F1 value reaches 94.67%, which is better than the comparison models. The performance of the model proposed in this paper is verified. 

Tourism Resource Information Database System Based on Geographic Information 

LI Wen-quan, XU Su-ping

2021, 0(08):  100-103. 

Asbtract ( 215 )   PDF (1221KB) ( 79 )  

References | Related Articles | Metrics

Strengthening management of tourism resource and building database of tourism resource information are very important for promoting the sustainable and healthy development of the tourism industry. Aiming at the problems of  tourism resource management, this paper proposes a tourist resources information database system based on geographic information, on the basis of detailed analysis of business process. The system adopts three-tier architecture to avoid direct access to data by users and improve the stability and scalability of the system. The system adopts role access control technology to realize authorized users to access system functions by roles, avoid illegal operations by illegal users and authorized users, and improve the security of the system. The system uses Gaode Map to realize the seamless association between single spatial data and attribute data of tourism resources and improve the user experience. The system can effectively overcome the shortcomings of traditional management methods and improve the efficiency of tourism resource management. 

An Attribute-based Multi-keyword Searchable Scheme Based on Bloom Filters

ZHANG Xiao-min

2021, 0(08):  104-111. 

Asbtract ( 428 )   PDF (1337KB) ( 210 )  

References | Related Articles | Metrics

Under the cloud computing environment, the attribute-based encryption multi-keyword searchable encryption can realize access control and searchability of encrypted data at the same time. For improving the retrieval efficiency of encrypted databases and reducing the cost of file keyword retrieval index storage, this article proposes an attribute-based multi-keyword search scheme based on Bloom filters. First, for the keywords set of a file, the Bloom filter is used to generate a fixed length index vector corresponding to the keyword set, so as to reduce the storage complexity of the keyword index. Besides, in order to prevent the adversary from obtaining the keywords information by means of statistical analysis of the index, this paper uses a permutation to randomize the elements of the index vector, and uses the attribute-based encryption to share the permutation. Thus, only legitimate users can obtain the permutation and construct a trapdoor for querying keywords with the permutation. Finally, through the security analysis and experimental analysis, the security, efficiency and practicability of this scheme are showed.

A Survey of Encrypted Traffic Classification Based on Deep Learning

LENG Tao ,

2021, 0(08):  112-120. 

Asbtract ( 927 )   PDF (1055KB) ( 372 )  

References | Related Articles | Metrics

In recent years, in order to protect the public privacy, a lot of traffic on the Internet is encrypted. The accuracy of traditional deep packet inspection and machine learning methods in the face of encrypted traffic has dropped significantly. With the application of deep learning automatic learning features, the encryption flow identification and classification technology based on deep learning algorithms have been rapidly developed. This article reviews these studies. First, this paper briefly introduces the application scenarios of encrypted traffic detection based on deep learning. Then, it summarizes and evaluates the existing works from three aspects: the use and construction of data sets, the detection model and the detection performance. The detection technology focuses on data preprocessing, unbalanced data set processing, neural network construction, real-time detection, etc. Finally, the problems in current research and future development directions and prospects are discussed, so as to provide some references for researchers in this field.

Design of Network-level Moving Target Defense System Based on Blockchain

DUAN Peng-fei, LAN Ru

2021, 0(08):  121-126. 

Asbtract ( 225 )   PDF (1205KB) ( 110 )  

References | Related Articles | Metrics

The network-level moving target defense is an effective approach to deal with the cyber attacks, like flooding attack. However, the existing network-level moving target defense systems mostly adopt the static central controller. This kind of centralized management architecture is prone to risks such as single point of failure or untrusted data. To address the above problems, this paper proposes a scheme of network-level moving target defense system based on blockchain, which realizes dynamically switching the central controller through the PoW consensus mechanism and overcomes the single point of failure of it and improves its robustness. In addition, based on the distributed trusted network environment constructed by blockchain, this paper establishes load balancing mechanism and disaster-tolerant backup mechanism for the dynamic central controller, making the system have good performance in dealing with the high concurrent service requests and recovering quickly from paralysis. Finally, this paper designs and implements the prototype system of network-level moving target defense system based on blockchain. The test results show that the designed system has good availability and robustness.