Current Issue

• 2011, No. 5　Published:25 May 2011 Last issue    Next issue
  Columns in this issue: 论文

• 论文

  A Cluster Head Election Algorithm Based on the Coverage of Nodes

  GAO Tiegang,NIU Weiwei

  2011, 33(5): 1-8. doi:

  Abstract ( 553 )   PDF (1760KB) ( 2139 )   　　

  A wireless sensor network aims to get the useful information of an area. However, the amount of information obtained depends on the ratio of coverage. Based on some cluster head election algorithms, a new cluster head election algorithm is proposed in this paper. In the scheme, the selected cluster heads do not sense data, and do not affect the coverage of the network even it dies earlier. If the consumption of the transmission is firstly distributed on this kind of nodes, then, the other nodes whose death can affect the coverage of the network can be used to be cluster heads later, and this may postpone the death of nodes, and accordingly, the lifetime of the network can be prolonged. Experiments show that compared with other current algorithms, the proposed algorithm based on the coverage of nodes has a higher ratio of node coverage and lower energy consumption, and thus has a longer network life cycle.

  A Comprehensive Weight Method Based on the Game Theory for Information Security Risk Assessment

  WU Yeke1,SONG Rushun1,CHEN Bo2

  2011, 33(5): 9-13. doi:

  Abstract ( 622 )   PDF (414KB) ( 1179 )   　　

  In order to determine the weights of risk factors reasonably in information security risk assessment and scientifically evaluate information security risk, the weights of risk factors are  calculated by a comprehensive weight method based on the game theory. Finally, an instance of risk assessment is given to show the rationality and feasibility of the results by analyzing an information system. It provides a new method for information system risk assessment.

  Research on ECard Data Synchronization Based on a Shared Data Center

  NIE Ruihua,LIN Huaigong,ZHENG Kai,WANG Juan

  2011, 33(5): 14-17. doi:

  Abstract ( 530 )   PDF (486KB) ( 480 )   　　

  A digital compus requires the intergration of  the data of application systems.As for the Ecard system as an important part of the digital compus, its data derives from ralative application systems.How to keep the data synchronization with that in the ralated application systems is an important question when constructing an Ecard system.This paper analyses the collecting mode of the Ecard data,puts forward an Ecard data synchronization solution by the data center based on SOA,and describes the process of data synchroization.The rusult proves the efficiency of the solution.

  A Dynamic Graph Software Watermarking Scheme Based on Variable TamperProofing

  LI Shuzhi,LIU Meng

  2011, 33(5): 18-21. doi:

  Abstract ( 424 )   PDF (396KB) ( 467 )   　　

  In order to improve the robustness of software watermarking and the quantity of hidden information, a dynamic graph software watermarking scheme based on variable tamperproofing is  put forward. The big watermark is  split into a set of small watermarks and they are represented as a set of graphs, which are  embedded into the program dynamically. The paper encodes the watermark with a parent pointer of the dynamic graph node, and combines graph nodes and the original grogram variables for tamperproofing. The results of experiment show that the watermark is not obviously effective in changing the program speed and results, and can  resist many types of attacks.

  An Optimal Strategy on Malware Attack and Defense in EnergyConstrained Wireless Sensor Networks

  YANG Xiong1,2,ZHA Zhiqin1,ZHU Yuguang1,XU Zezhong1,2

  2011, 33(5): 22-26. doi:

  Abstract ( 456 )   PDF (431KB) ( 463 )   　　

  As for the shortcomings of  the research on malware modeling in wireless sensor networks the paper starts from the attack and defense perspectives, and considers the impact under the scanning rate on the optimal strategy of attack or defense, and takes the energyconstrained property in wireless sensor networks into account. Then an paper novelly proposes an optimal model on malware attack and defense in WSNs based on the traditional malware propagation models. The model is extended from the  traditional SIR propagation model and the immunization and healing factors are also concerned. Finally, an optimal solution to evaluating the effectiveness of attack and defense is given. The simulation experiment  indicates that this optimal evaluation strategy is able to effectively depict the scanning rate’s key impact on the maximal results for both attack’s and defense’s point of view.

  Evaluation of the IDS Capabilities Based on Return Expectations

  YANG Zhongming1,WU Yugang1,2,QIN Yong1,CAI Zhaoquan3

  2011, 33(5): 27-31. doi:

  Abstract ( 481 )   PDF (522KB) ( 396 )   　　

  It has become a hot topic  in the current research of network security to evaluate the effectiveness and reliability of intrusion detection systems effectively with the development of technology and products of intrusion detection. This paper indicates the advantages and disadvantages of some intrusion detection models which are commonly used at present. An optimal reference value is given through analyzing and deriving an intrusion expectation model, which is different from the current problem of velocity and accuracy of the assessment of intrusion detection systems. Meanwhile, the relations of false alarm rate, false negative rate and the product of intrusion rate and intrusion number are given. Therefore, the product of intrusion rate and intrusion number can be given through the relation which is significant for the IDS capabilities evaluation. The feasibility of the return expectationsbased IDS access model in assessing IDS is proved by the experiments of the Matlab software.

  Research on  a  2DMeshBased NetworkOnChip Routing Algorithm

  SHI Shuai，DOU Wenhua，LU Jia

  2011, 33(5): 32-37. doi:

  Abstract ( 460 )   PDF (584KB) ( 468 )   　　

  Networkonchip and network coding both are quickly developed research fields in recent years.NOC will play an important role in the VLSI field. Network coding has an incomparable advantage compared to other methods in improving the network throughput and network load balancing has.This article uses the main idea of network coding in 2DMesh NOC to do some exploratory research,and proposes a framework routing strategy based on the 2DMesh structure.The conclusion demonstrates that this strategy has good compatibility, and  will bring a good performance improvement in multicast applications,and this framework routing strategy has strong portability and can be deployed in other structures of NOC.

  An ObjectBased Storage System Based on CustomBuilt HighPerformance Interconnects

  GUO Yufeng,LI Qiong,LUO Li

  2011, 33(5): 38-43. doi:

  Abstract ( 441 )   PDF (750KB) ( 457 )   　　

  The trend as such makes it very promising to build a scalable I/O system for highperformance computing through the implementation of an objectbased storage system based on highperformance interconnects. We design and implement a custombuilt highperformance, low latency interconnects chip HSNI, which provides a  very good performance of communication, and can be used to build objectbased storage systems. Firstly, we introduce the hardware architecture, the software structure and the communication mechanism of HSNI. Then we implement an objectbased storage system based on HSNI. The experimental results show HSNI can provide high bandwidth, low latency data transfer, and is very suitable for building objectbased storage systems, and the storage systems based on HSNI can exert the performance of Lustre effectively with good scalability, which satisfies the demands of the I/O system for highperformance computing perfectly.

  Random Number Generation Based on the Galois Type LFSR

  GU Xiaochen,ZHANG Minxuan

  2011, 33(5): 44-47. doi:

  Abstract ( 482 )   PDF (557KB) ( 530 )   　　

  With the development of FPGA, more and more attention has been paid to FPGA based computing acceleration. And random number generators are frequently used in many of these applications. In this paper, a hardware structure for random number generation based on the Galois type the LFSR and Leap Forward method is proposed. By analyzing the characteristics of the proposed structure, some methods to improve the clock frequency and to save the hardware resources are put forward. A random number generator with 16bit outputs based on the above structure is implemented on the Xilinx Vertex 6 FPGA. This generator occupies only 6 slices, while the frequency and the throughput are as much as 951 MHz and 15.2 Gbps. The KS method is used to test the quality of the generated random numbers, and the CDF curve of the generated 105 random numbers is compared with the theoretical one.

  A Dynamic Optimization Framework for MultiThreads Based on MultiCore Platforms

  XU Fan,SHEN Li,WANG Zhiying

  2011, 33(5): 48-53. doi:

  Abstract ( 484 )   PDF (906KB) ( 543 )   　　

  Dynamic optimization is one of the most important techniques for computing system virtualization. Based on the analysis of some existing dynamic optimization systems, such as DynamoRIO and Jrpm, this paper concludes the problems and challenges of traditional dynamic optimization mechanisms on multicore platforms first. Then, it proposes a dynamic optimization framework for multithreads based on multicore platforms, analyzes its structure and working principles, and proves its feasibility by experiments.

  A Mechanism of Extending the Generic Snapshot Capability for the Virtual Block Devices in Virtual Machine Monitors

  CHEN Bin1,2,CAI Zhiping1,XIAO Nong1,CHU Fuyong1

  2011, 33(5): 54-58. doi:

  Abstract ( 474 )   PDF (665KB) ( 443 )   　　

  Based on the characteristics of diversity of virtual block devices and requirements for the capabilities of high availability and fault tolerance, we propose a mechanism of extending the generic snapshot capability for the virtual block devices(VBD) in virtual machine monitors for the mainstream virtual machine(VM) architectures. Based on the mechanism, we can provide uniform and extensible snapshot services for the multiple VBDs in the VM environments while avoiding taking much efforts in modifying the lowlevel multiple specific VBD drivers. Based on a Linux system, we have extended the generic snapshot capability for QEMU to prove the feasibility of the extending mechanism, and have measured and analyzed the effects that snapshot brings to the VM performance through our experiments.

  An Improved ModelBased Recognition and Localization Method for Articulated Objects

  LI Qianqian

  2011, 33(5): 59-62. doi:

  Abstract ( 411 )   PDF (610KB) ( 594 )   　　

  Aiming at the complexity of articulated object detection, this paper proposes a novel casting of visual inference, based on a standard object model, makes full use of the edge feature of the image(it does not rely on local features), as well as the space information of objects, and performs an iterative parsing process, where we can sequentially handle the articulated object detection and estimate the local object components. Experiments show that this method has a strong resistance to the background interference, and can also greatly improve the results of articulated object detection by human vision.

  Progress in 3D Video Compression

  CUI Dawei,DOU Wenhua,JI Yongxing

  2011, 33(5): 63-68. doi:

  Abstract ( 446 )   PDF (571KB) ( 512 )   　　

  Information compression is one of the most important technologies of information transmission. With the development of 3D display and 3D video technology, video transmission will not only be limited in 2D video in the future. However, the large amounts of data of the 3D video have hindered its transmission. Therefore, 3D video compression with a high compression ratio and high compression efficiency must be designed, which can not only solve the transmission of 3D video but is also conducive to promoting the 3D video and 3D display for further development. This paper describes the progress in 3D video compression, and gives some advices for further study, based on summarizing the technologies of video compression.

  A Trigonometric Spline Curve with Shape Parameters

  YAN Lanlan1,2,LIANG Jiongfeng3

  2011, 33(5): 69-73. doi:

  Abstract ( 651 )   PDF (456KB) ( 441 )   　　

  In order to improve the disadvantages of the cubic Bspline curve,its  shape is fixed to its control polygon. It cannot describe the conic section except the parabola, and the shape parameters and the trigonometric functions are combined. A trigonometric curve with shape parameters is defined. Its structure is similar to the cubic Bspline curve. The new curve not only inherits the major advantages of the cubic Bspline curve, but also can express ellipse with shape adjustability. Besides, the continuity and approximability of the new curve is better than the cubic Bspline curve. For equidistant knots, the new curve is C3continuous, and it is C5 continuous under special conditions. Using the tensor product method, the curve can be extended to the surface. The surface has properties similar to the curve. The way to represent the ellipse surface using the new surface is given.

  Application of the BP Neural Network Based on PSO in Dynamic Gesture Recognition

  LI Wensheng,YAO Qiong,DENG Chunjian

  2011, 33(5): 74-79. doi:

  Abstract ( 512 )   PDF (551KB) ( 501 )   　　

  In order to improve the training speed and identification accuracy of dynamic gesture, a method of gesture recognition based on the particle swarm optimization(PSO) BP neural network is  put forward. First, a set of dynamic gestures is defined for HumanMachine Interaction (HMI). The engenvectors vectors of dynamic gestures are extracted as the input of the BP neural network on the basis of obtaining the trajectories of moving fingertips. An improved PSO algorithm is used to train the BP neural network and get the weights/thresholds of the network. Finally, the gestures based on machine vision are  recognized through the trained BP neural network. The experimental results show that the proposed PSO algorithm can enhance the speed and precision of network training, and improve the accuracy of dynamic gesture recognition.

  Design of a Virtual and Reality AirForce Integrative Combat System

  HUANG Zhenyu,YANG Rennong,BAI Juan,WU Meng

  2011, 33(5): 80-84. doi:

  Abstract ( 550 )   PDF (1050KB) ( 462 )   　　

  The PLA air force is devoid of other Air Forces’ primary battleplans, which cumbers the upgrade of combat ability. Aiming at this disadvantage, we propose a scheme which puts the simulator s’ virtual force and reality force into the same battlefield and design an airforce combat system based on HLA. A realtime datacollecting subsystem based on Lua is realized to implement the communication between simulation federates and simulators. An approach to multilevel network communication comes up, and the development of the interfaces of different levels is implemented. The simulator s’ data should be sent to battleplans during the training to upgrade the combat ability.

  MultiView Face Detection Based on the AdaBoost Algorithm and the Skin Color Model

  ZHAO Nannan

  2011, 33(5): 85-90. doi:

  Abstract ( 528 )   PDF (702KB) ( 461 )   　　

  The AdaBoost algorithm is not ideal in multiview face detection,and the skin color model has a high false alarm rate under the complex background.This paper presents a method of rotating images form flank to frontal, which combines face detection based with the AdaBoost algorithm and the Skin Color Model. This method is mainly to rotate images, and makes different angle faces detected, then makes eyes detected, if the position satisfies the condition by calculating the position between the two eyes. A final verification is performed  by the skin color model. The final key is inversion calculation, which draws the face location. Experiments prove that this method is adaptive to different views.

  A Particle Swarm Optimizer and Its Application Based on Dynamic Population and Comprehensive Learning

  LIU Yanmin1,2,ZHAO Qingzhen1

  2011, 33(5): 91-96. doi:

  Abstract ( 515 )   PDF (827KB) ( 421 )   　　

  In order to improve the ability to escape from local optima, we present an improved particle swarm optimizer based on dynamic population and comprehensive learning (DCPSO for short). In DCPSO, the swarm population growing and declining strategies are introduced to increase the swarm diversity, further improve the ability to escape from local optima; a comprehensive learning strategy also is used to improve the probability of flying to the global best position. In the benchmark function, the results demonstrate good performance of the DCPSO algorithm in solving complex multimodal problems when compared with other PSO variants. In the optimization design for the box grider of portal gantry, the experimental results show that the DCPSO algorithm can achieve better solutions that other PSOs.

  A Cloud Adaptive Particle Swarm Optimization Algorithm Based on Mean

  LIU Hongxia,ZHOU Yongquan

  2011, 33(5): 97-101. doi:

  Abstract ( 511 )   PDF (584KB) ( 421 )   　　

  Based on the cloud adaptive theory, the particle swarm optimization algorithm is improved and the particle swarm is divided into three populations. It  modifies the inertia weight using a cloud method, and meanwhile modifies the “social” and “cognitive” sections, and introduces the notion of mean, and proposes an improved cloud adaptive theory particle swarm optimization algorithm named MCAPSO. The greatest advantage of the method is that the algorithm in the later iteration, when the different value between an individual optimal to some particle corresponding of the fitness value and a global optimal corresponding to the fitness value is significant, overcomes the shortcoming that the algorithm does not benefit from converges to the optimal solution. Numerical experience shows that, MCAPSO runs  less iteration to find the optimal solution, and the average time is lower. The average time cost is reduced accordingly.

  The 01 Knapsack Problem Based on the BeeSwarm Genetic Algorithm

  WU Di,JIANG Yongzeng,SONG Guangjun

  2011, 33(5): 102-105. doi:

  Abstract ( 522 )   PDF (455KB) ( 554 )   　　

  This paper presents a beeswarm genetic algorithm for the 0-1 knapsack problem. There are two populations, one for global search, and the other for local search. Each individual adopts the binary code. Only the best one can crossover. The strategy of managing the feasible solution is to enclose the goods which is out of the knapsack and costeffective, until no goods can be put into. The solution which does not accord with the constraint condition mutates under the instruction of mutagens. The genetic operators include order crossover operator, twoblockexchange mutation operator and restraint operator. The method sufficiently takes the advantage of the genetic algorithm such as group search and global convergence in order to have a quick parallel search, which efficiently overcomes the problem of local optimization. The experimental results show that the bee swarm genetic algorithm is efficient in solving the 0-1Knapsack  problem, and is also suitable for other combinatorial optimization problems.