Table of Content

03 April 2018, Volume 0 Issue 03

Previous Issue    Next Issue

A Servo Control System Based on Fuzzy Neural Network PID

BU Qing-wei, CHEN Xiong, CHAI Jin-bao, CUI Er-wei

2018, 0(03):  1.  doi:10.3969/j.issn.1006-2475.2018.03.001

Asbtract ( 149 )  

References | Related Articles | Metrics

This paper aims at optimizing the control system of servo in a certain missile. For a complex system with nonlinear and time-varying characteristics, an improved fuzzy neural network PID controller is proposed on the basis of the analysis of traditional PID control algorithm and fuzzy neural network control algorithm. Phased learning mode improves the network learning efficiency by the way of adopting self-organization learning and learning with a teacher. The mathematic model of DC brushless servo control system is established. Meanwhile, the simulation analysis is carried out by MATLAB as well. The experimental results show that the designed controller has faster step response with no overshoot basically and performs accurately with smaller phase shift under the rudder skew command.

A Method of Interactive Virtual Human Body Meridian Acupoints System Simulation #br# and Implementation Based on MFC and OpenGL

XU Yu-long1,2, ZHANG Pei-jiang1,3, WANG Zhong-yi1, SHENG Meng-yuan1, ZHAO Yu-mei1, WANG Bing-ying1

2018, 0(03):  6.  doi:10.3969/j.issn.1006-2475.2018.03.002

Asbtract ( 174 )  

References | Related Articles | Metrics

The existing models of human body meridian acupoints have less interactive function. This paper introduces an interactive virtual human body meridian system simulation and implementation based on MFC and OpenGL, and provides the method and process to simulate that. The 3DMAX is adopted to establish the human body model, acupuncture points and meridian lines. Based on OpenGL and MFC, the model, acupuncture point and meridians lines are loaded and rendered. The paper detailedly explains the framework and process of system implementation, and the key technologies involved in the development process, such as using dynamic direction adjustment to display the name of the acupuncture point, using the global function message to improve the efficiency of loading and responding for the model, the method of rapidly acquiring touch point coordinates. Finally, the model is deployed in a touch machine. The model system has a dynamic interactive function. Without mouse, only by fingers touching, the user can implement the model operations such as amplification, narrow, spin, acupuncture point, meridion cycle, etc. The model system can help users learn the knowledge of acupuncturist and main and collateral channels, its memory function can test the study situation for acupuncture of users, and it can provide a real acupoints environment of learning for the acupuncturist and learner.

Dynamic Spatial Reasoning Based on Action

XIE Yu-mei1, GAO Hai-yan2

2018, 0(03):  13.  doi:10.3969/j.issn.1006-2475.2018.03.003

Asbtract ( 111 )  

References | Related Articles | Metrics

Dynamic spatial relations representation and reasoning have always been a focus of research in qualitative spatial reasoning. In this paper, the qualitative spatial relation of spatial primitive area is represented by rectangle relation. Some elaboration and extension are made for dynamic spatial system defined by Bhatt. Eight moving actions are defined for spatial entity, and the state transition graph and the state transition table are given under different actions for interval relation in spatial situation with multiple spatial entities. A tuple is used to formalize spatial situation, a basic missions is defined in dynamic spatial system, what the new spatial situation is when an entity takes action in the history situation. The method to solve the missions is given, and the limitation and shortages of the method are also presented. Furthermore, an application situation is given in this paper.

Gait Optimization Research on RoboCup3D Simulation

HE Rong-yi1, LI Chun-guang2

2018, 0(03):  20.  doi:10.3969/j.issn.1006-2475.2018.03.004

Asbtract ( 177 )  

References | Related Articles | Metrics

In the RoboCup3D competition, a flexible and stable gait pattern is one of the key for the humanoid robot to win the match, to achieve such walk gait, a machine learning method of optimizing the vertical Center of Mass(Com) trajectory is presented. Firstly, to generate bipedal walk slowly and unsteadily, the vertical Com trajectory is planed by multiple polynomial, the inverted pendulum model(IPM) and a numerical method are utilized to control the Zero Moment Point(ZMP). Secondly, to get a fast and stable bipedal walk, a overlapping layered learning method is proposed to optimize the walking parameters which is based on Covariance Matrix Adaptation Evolution Strategy(CMA-ES). Finally, flocking control is applied to verify the flexibilty of the optimized gait in multi-robot environment. The experimental and competition results show that the proposed method is effective.

Human Activity Recognition Method Based on Neural Network

DONG Zhe-yu, WANG Qian-jun, LI Wan-jie, ZHOU Bo

2018, 0(03):  26.  doi:10.3969/j.issn.1006-2475.2018.03.005

Asbtract ( 176 )  

References | Related Articles | Metrics

Human activity recognition has always been paid attention to the field of computer vision. In this paper, a weighted recognition method based on neural network is presented to improve the accuracy of human activity recognition. Firstly, the ViBe algorithm is used to extract the foreground of human activity, and the center of gravity of the foreground is calculated. Secondly, the Fourier descriptor is obtained by the Fourier transform of the outline distance center of gravity. Finally, a weighted recognition method based on neural network is used to classify the Fourier descriptor. The experimental results show that the recognition rate of this method is more than 89%.

Outdoor Pedestrian Detection Method Based on Improved ViBe

CUI Ying, CHEN Sheng-dong, YUAN Feng, LI Yin

2018, 0(03):  30.  doi:10.3969/j.issn.1006-2475.2018.03.006

Asbtract ( 124 )  

References | Related Articles | Metrics

Outdoor pedestrian detection has always been research challenges because of the complex background, such as light mutations, shaking leaves, passing of small animals and other complex noises. Regarding this fact, this paper proposes a pedestrian detection method based on improved ViBe(Visual Background Extractor) method. This method first uses the region-based improved ViBe update strategy, which improves the completion rate of the extraction of pedestrian-target-region while accelerates the elimination rate of the ghost-region. Then, the background difference method is used to filter the false area which caused by complex background. Finally, on this basis, according to certain logical rules, the method presented in this paper performs pedestrian detection selectively. The experimental results show that this method can not only accelerate the elimination rate of the ghost-region, but also reduce the interference of complex variable background. This method has high accurateness and robustness for outdoor pedestrian detection.

Image Registration NLORB Based on Nonlinear Scale Space 

DONG Hao1, LYU Dong-yue2

2018, 0(03):  38.  doi:10.3969/j.issn.1006-2475.2018.03.007

Asbtract ( 166 )  

References | Related Articles | Metrics

 An image stitch method in video scene need to perform well in real time, scale and rotation invariance. Aiming at the issue that the ORB algorithm does not have scale invariance and SIFT algorithm with high time complexity does not respect the natural boundaries of objects, we put forward an image registration algorithm based on the nonlinear scale space and improved the ORB and a mean coordinate blend method based on YUV brightness template. Firstly, the nonlinear scale space is used to smooth noise and retain object boundaries, to set scale parameters based on image entropy and to set proper distance between feature points, so that the ORB detectors are stable. Then, maxima is searched for in scale and spatial location and descriptors are compute. Finally, the feature points are matched with hamming distance and RANSAC algorithm. The experimental results show that the improved algorithm can improve robustness, shorten time of registration and improve greatly the matching accuracy. What’s more, the mean coordinate method with brightness template enhances blend.

Moving Background Subtraction of Video Image Based on 1-D Maximum Entropy

LI Ya, WANG Ying

2018, 0(03):  44.  doi: 10.3969/j.issn.1006-2475.2018.03.008

Asbtract ( 82 )  

References | Related Articles | Metrics

Background subtraction of video image is critical for detecting and tracking moving objects. Because of variety of video image and complicated background, it is difficult for a method to detect moving objects for all video images. This paper proposes the 1-D maximum entropy of image to accomplish the object detection of video images with dynamical and complicated background. The algorithm obtains the gray value which makes the entropy of object and background maximum, retains most information of moving object. Compared with frame difference method and Gaussian mixture background modeling method, 1-D maximum entropy method obtains the accurate object with shorter computational time, and much information of object. It is suitable for detecting objects of the videos with dynamical and variable background in real time.

An Optimal Virtual Machine Placement Algorithm for Data Center Based on Orbital Shrinking

SUN Qing-gong, CHEN Hai-biao, LIN Nan, WU Jin-ming

2018, 0(03):  48.  doi:10.3969/j.issn.1006-2475.2018.03.009

Asbtract ( 181 )  

References | Related Articles | Metrics

 Virtual machine placement can control the efficiency of physical server in data center of smart grid. The main idea is to make the balance of time, space, computing resources and energy consumption. It has met the bottlenecks, which are the contradiction of robustness and flexibility, and the problem of non-equilibrium allocation of finite resources. This paper proposes an optimal virtual machine placement algorithm for data center based on Orbital Shrinking. First, the applicability of virtual machine placement for data center is analyzed, to determine objective optimization and dynamic boundary constraint resources effectively. Then, based on the Orbital Shrinking model, a multidimensional knapsack model with computing resources, temporal and spatial conditions and energy consumption conditions is established. The overall balance of the placement strategy of the virtual machine is realized. The simulation results show that the new algorithm can effectively improve the computing resources by 9.8% and reduce data processing latency by 10.3 s.

Research on TSP Problem Based on Crossover and Mutation Combination

LIU Zhi-yong 1,2, ZHOU Jie 1, ZHANG Lin3,ZHANG Jia-xin1, ZHANG Qian1, SHA Ren1

2018, 0(03):  54.  doi:10.3969/j.issn.1006-2475.2018.03.010

Asbtract ( 160 )  

References | Related Articles | Metrics

Genetic algorithm is a kind of random search algorithm based on natural selection and natural genetic mechanism, which is a relatively common algorithm for solving TSP problem. However, when the algorithm solves the TSP problem, there is a problem that the convergence speed is slow and easy to get premature. This paper proposes an algorithm design that combines five kinds of crossover algorithms and three kinds of mutation algorithms, then achieves 15 kinds of combination methods, and then uses Java language programming experiment, and finally through the Chinese 144 (CHN144), it is proved that the genetic algorithm combined with the THGA algorithm and the reverse order mutation algorithm can solve the traveling salesman problem by using the combination of the crossover algorithm and the mutation algorithm to achieve the best results.

Research on Named Entity Recognition Method in Specific Fields

ZHANG Lei

2018, 0(03):  60.  doi:10.3969/j.issn.1006-2475.2018.03.011

Asbtract ( 179 )  

References | Related Articles | Metrics

For named entity recognition technology in a specific domain, there are various identification methods corresponding to different fields. Different fileds of texts have their own unique textual features, which leads to the existing identification method is difficult to adapt to new specific domain. In order to solve this problem, this paper proposes a method based on conditional random field, semi-supervised learning and active learning, which forms a unified technical framework to adapt to the named entity recognition in each specific domain. This method constructs the feature set based on characteristics of rail transit text, then trains CRF to recognize named-entity of rail traffic text, and selects the samples with lower confidence level than the selected threshold, and then manually extends the training samples to achieve high goals. In order to validate the method, this paper carries on the experiment in the field of rail transit. The experimental results show that the method is effective and has a good recognition effect in the field of rail transit.

EM Algorithm Amendment of Zero-and-one Inflated Binomial Mixture Regression Model

LYU Min-hong1, YAN Yi-rong2, WU Cheng-jing1

2018, 0(03):  65.  doi:10.3969/j.issn.1006-2475.2018.03.012

Asbtract ( 143 )  

References | Related Articles | Metrics

This paper proposes a zero-and-one inflated regression model with “heterogeneity”. Since standard EM algorithm will make the parameter estimates converge to local optimum, this paper introduces the MCEM algorithm to address this shortcoming, then the estimation of parameter is setted up for the zero-and-one inflated binomial mixture regression model. Finally, the corresponding simulation is given to illustrate the proposed method.

Recommendation Algorithm Based on Difference Value Matrix Factorization

CHENG Peng, LIU Wen-bin

2018, 0(03):  69.  doi:10.3969/j.issn.1006-2475.2018.03.013

Asbtract ( 139 )  

References | Related Articles | Metrics

Matrix factorization has become a common way to predict user ratings of items. Traditional matrix factorization algorithms do not take account of the differences between users. To address this problem, a difference value(D-value) matrix factorization model is proposed. First, for each user, the difference between his/her rating score and the average rating score from users with similar social attributes is calculated, which finally results in a matrix called D-value matrix. Then the D-value matrix is factorized to calculate the predicted ratings. Experimental results on the Movielens 1M dataset show that the proposed method significantly outperforms Bayesian probabilistic matrix factorization, matrix factorization and the latent factor model fused with user attributes in terms of prediction accuracy.


A Novel Underwater Dam Crack Extraction Algorithm Based on Lorentz Information Measure

FAN Xin-nan, WU Jing-jing, SHI Peng-fei, ZHANG Xue-wu

2018, 0(03):  73.  doi:10.3969/j.issn.1006-2475.2018.03.014

Asbtract ( 136 )  

References | Related Articles | Metrics

Because underwater images generally have many problems, such as uneven illumination, low contrast, blurring and much noise, a novel underwater dam crack extraction algorithm based on Lorentz Information Measure is proposed to deal with the problems. First, the image blocks without crack information are removed according to Lorentz Information Measure. Second, features of connected domains, i.e., circularity and area, are extracted. Then k-means clustering method is used to achieve final cracks by 2-D feature space. The experimental results show that the proposed algorithm can extract cracks from underwater dam images precisely and effectively.

A Recommendation Algorithm Based on Denoising Autoencoders

WU Ling-mei, LU Jian-bo, LIU Chun-xia

2018, 0(03):  78.  doi:10.3969/j.issn.1006-2475.2018.03.015

Asbtract ( 202 )  

References | Related Articles | Metrics

The traditional recommendation algorithm generally uses the user project score matrix to learn the potential factors, understand the user’s personal preferences and make recommendations, but in practice, the score matrix is usually very sparse. Aiming at the shortcomings of traditional recommendation algorithm, a recommendation model based on noise reduction automatic encoder is proposed. First, two automatic encoders are used to train the potential factor matrix of the user and the project, and then the implied feature vector is input into a neural network to carry out the scoring prediction. Finally, the recommendation is made according to the new score matrix. The experimental results show that the proposed algorithm improves the recall rate of the recommended results and reduces the reconstruction error.

Research on Classification of Improved Smote Algorithm on Imbalanced Datasets

YI Wei1, MAO Li1, SUN Jun1, WU Lin-hai2,3

2018, 0(03):  83.  doi:10.3969/j.issn.1006-2475.2018.03.016

Asbtract ( 131 )  

References | Related Articles | Metrics

In imbalanced datasets, the oversampling algorithm, such as Smote (Synthetic Minority Oversampling) algorithm, R-Smote algorithm and SD-ISmote algorithm, may blur the boundary between the majority and the minority and use noisy data to synthesize new samples. The ImprovedSmote algorithm proposed in this paper uses cluster center of minority set and their corresponding minority set to generate new samples. The Smote, the R-Smote, the SD-ISmote and the ImprovedSmote algorithm combined C4.5 decision tree and neural network algorithm are used on the experimental datasets. The results show that the ImprovedSmote algorithm is better than other algorithms in classification and can effectively improve classifier performance.

An Improved Cause-effect Graph Method and Its Implementation

ZHANG Zhi, WANG Min

2018, 0(03):  89.  doi:10.3969/j.issn.1006-2475.2018.03.017

Asbtract ( 134 )  

References | Related Articles | Metrics

To improve the efficiency of test cases design for multi-condition combination object, this paper proposes a method that uses the cause-effect expression to replace the cause-effect graph. Firstly, the syntax rules of cause-effect expression are made. Secondly, the data structure and algorithm of decision tables are determined. Finally, the test case design tool for multi-condition combination object is realized by visual design and programming. The functions of the tool include: test input editing, automatic test cases generation, test case maintenance and design process view. To compare with cause-effect graph, this method doesnt have to draw cause-effect graph, and easily generates test cases automatically, so can improve the efficiency of test case design and maintenance.


A Safe and Feasible Digital Watermarking Protocol

ZHAN Hu, MAO Li

2018, 0(03):  96.  doi:10.3969/j.issn.1006-2475.2018.03.018

Asbtract ( 146 )  

References | Related Articles | Metrics

Considering the problems of the frequent participation of buyers and the involvement of too many third parties in current digital copyright protection protocols, a secure and efficient watermarking protocol is proposed. This protocol utilizes the homomorphic Paillier encryption algorithm and the iris recognition technology to solve the anonymity problem, unbinding problem, piracy tracing problem and some other relevant secure problems. In addition, to ensure the confidentiality and completeness of the communication data, the communication data between the participants in the transaction are expressed in the forms of encryption, digital signature authentication and so on. The buyers select the trading application to finish the operations of information transmission and purchase, which makes this protocol closed to the daily trading model and easy to realize as well.

Identification System for Enterprise Users Based on

XU Rui1,2, YOU Jia1,2, LIU Kun1,2, MA Feng1,2, DUAN Ke2, ZHONG Yan-tao3

2018, 0(03):  102.  doi:10.3969/j.issn.1006-2475.2018.03.019

Asbtract ( 169 )  

References | Related Articles | Metrics

In order to overcome the disadvantages of login methods of current enterprise information systems on security and flexibility, an identification system for enterprise users based on national commercial cryptographic algorithms and physical un-clone-able function is designed and implemented. The frame of FIDO U2F is used in this system with identity tokens as the second factors in user authentication. Moreover, by using national commercial cryptographic algorithms, autonomous and manageable security is obtained. Thus the system achieves high security and flexibility. The security analysis shows that the system is evidently more secure than current other enterprise identification systems. The experimental results show that the system takes advantage of low overhead and high reliability thus can be implemented in enterprise information systems easily and quickly.

Design of Hyperchaotic Lorenz Encryption Algorithm Based on Hadoop

WEN He-ping1, BAO Jing-jing2, KE Ju-xin1, LIU Shu-wei1

2018, 0(03):  108.  doi:10.3969/j.issn.1006-2475.2018.03.020

Asbtract ( 120 )  

References | Related Articles | Metrics

Aiming at the problems of data security and privacy protection in big data environment, a hyperchaotic data encryption algorithm based on Hadoop big data platform is proposed. The dynamic behavior of hyperchaotic Lorenz system is more complex and the generated sequence has better characteristics of randomness, combined with MapReduce parallel programming model, an cipher algorithm with higher efficient and safety is designed. Experimental results show that compared with the AES algorithm, the efficiency of the proposed algorithm is improved by nearly 40%. In security aspect, the algorithm has the characteristics of large key space and good key sensitivity.

Palmprint Recognition System on Mobile Devices Based on Dual-point Assistance

WANG Tian-hua1, LENG Lu2, YUAN Ming-wen3

2018, 0(03):  112.  doi:10.3969/j.issn.1006-2475.2018.03.021

Asbtract ( 133 )  

References | Related Articles | Metrics

Palmprint recognition on mobile devices suffers from several severe technical challenges, including non-uniform palm location and gesture, complex background, diverse illumination, limited hardware resource, etc. A novel palmprint recognition system based on dual-point assistance acquirement technique is designed and implemented independently on Android platform in order to overcome the aforementioned problems. In addition, several technological difficulties are addressed in entire engineering implementation. Both assistant points and restricting blocks assist users to place their hands at appropriate position with correct pose during acquirement. The two assistant points are located on the two key-points, i.e., the two finger valley points between index finger and middle finger as well as ring finger and little finger. Then a palmprint image is rotated so that the line connecting two key-points is horizontal. Finally, the region of interest is cropped for the following feature extraction and recognition. The proposed assistant localization scheme enhances the robustness of palmprint preprocessing against interference factors and real-time property.

Urban Traffic Flows Analysis and Visualization Based on Mobile Phone Signaling Data

CAO Zhong, LI Fu-chen, YANG Hao-fei

2018, 0(03):  116.  doi:10.3969/j.issn.1006-2475.2018.03.022

Asbtract ( 150 )  

References | Related Articles | Metrics

In recent years, the research based on mobile phone signaling data which contains time and space information to solve various types of traffic problems is becoming a hot spot. In this paper, a technical scheme combining Spark with MongoDB is proposed to process the mobile phone signaling data, and to generate the traffic trajectories of mobile users, and then the traffic flow between urban areas is analyzed. Taking the Haidian Administrative District and Sanlitun Commercial District in Beijing as examples, the paper analyzes the attractiveness and occurrence of inter-regional traffic flow, and finally uses the JavaScript library provided by OpenLayers to visualize the research results.

Design and Implementation of Software with Mobile Study in English Listening #br# and Speaking Based on Augmented Reality

HUANG Min, LAN Hong

2018, 0(03):  122.  doi:10.3969/j.issn.1006-2475.2018.03.023

Asbtract ( 111 )  

References | Related Articles | Metrics

In the modern age, most of the laptops are without CD-ROM, but a huge amount of teaching material like video, audio and text information are still available on Compact Disk (CD). Exclusion of CD-ROM creates problems for the students to get video lectures and other electronic information from Compact Disk. Considering the above problems, this paper solves it with the implementation of the mobile augmented reality application, which is called VBook. The proposed solution is based on Unity3D development tools and Vuforia SDK expansion of real-world software development kit to achieve the objective. First, the application accesses the mobile camera to get the pictures and performs image recognition with Vuforia SDK. The recognition maps are generated and saved as a database, and meanwhile named as videos accordingly. Secondly, the application designs scenarios using Unity3D and downloads Vuforia toolkits package integrated in Unit3D, and after that, programs the recognition codes to access data. The picture from the maps database is transferred to our proposed system and requested the local server to get the related video. Finally, Android application is designed. This paper uses virtual reality (Unity3D) function to access the mobile camera and video player to play the retrieved video in the focused area of the book page. The experiment shows the impressive result. Using augmented reality with Unity3D, the paper achieves the goal to provide efficient data (english learning video) availability technique.