Current Issue

• 2013, No. 7　Published:25 July 2013 Last issue    Next issue
  Columns in this issue: 论文

• 论文

  Dissembler design of the innovative DSP
  for high computable density of soft base station           

  WU Jiazhu,TIAN Xi,ZHAO Chuanjun,LIU Hengzhu,CHEN Shuming

  2013, 35(7): 1-5. doi:

  Abstract ( 169 )   PDF (1437KB) ( 312 )   　　

  Aiming at high computable density, the innovative DSP’s research includes the chip design and the development of related tools  such as Compiler, Assembler, Dissembler, Linker, Loader, and Integrated Development Environment (IDE). This paper designs the Dissembler based on the vector processor with variable length instructions, and it allows the programmer to use macro assemble pretreatment instruction.

  Novel way of syndrome testing 

  TONG Chunchun,XU Shiyi,WU Yue

  2013, 35(7): 6-10. doi:

  Abstract ( 136 )   PDF (1468KB) ( 235 )   　　

  Syndrome testing and parity testing are exhaustive input based fixed fault testing methods. They have been employed for many years. The paper proposes a novel syndrome testing method, named third order  syndrome testing, which combines syndrome testing and parity testing. Its main feature is that both the testing efficiency and the fault coverage are improved so that syndrome testing can be performed on the circuits whose syndrome could not be tested. Based on the traditional syndrome testing, parity testing is introduced to preprocess the circuits, thus improving the testing efficiency. Then, syndrome testing is enhanced to be the second and the third order syndrome testing, thus improving the fault coverage. Experiments on reference circuits and commonly used circuits demonstrate the practicability and effectiveness of the proposal.

  Prototype construction of a dual-channel
  3D many-core NoC based on KILL rule

  TAN Hai1,2

  2013, 35(7): 11-15. doi:

  Abstract ( 133 )   PDF (1014KB) ( 233 )   　　

  In order to design a lowdelay and lowoverhead manycore NetworkonChip (NoC), the paper proposes an Architecture Utilizing 3D Stack Hierarchical Dualchannel NoC (AUSHDN) and KILL（Kill If Less Linear）rule is used to determine the number of processing cores in each group of AUSHDN. What's more, the prototype of the AUSHDN system is established based on the Graphite simulator from MIT. In AUSGHN system, 3D stack Hierarchical multilevel internetonchip is employed and different communication link is used to transfer control and data signal according to different content of communication. The results of simulation test in prototype system show that: compared with the traditional 2D NOC, the power consumption has reduced by 20% and the time delay has shortened by 30%. Meanwhile, the hierarchy feature of the AUSGHN system guarantees its good scalability.

  Dynamic grid technology for aircraft conceptual design  

  WANG Man,LIU Yi

  2013, 35(7): 16-22. doi:

  Abstract ( 102 )   PDF (2362KB) ( 227 )   　　

  Based on the two dynamic grid methods (spring smooth approximation method and local remeshing method) and the function/shape function method, according to the facts that the program and the parameters change frequently at the phase of aircraft conceptual design, a dynamic grid method is proposed for the shape of the 2D airfoil. The method passes the changed boundary information to the grid, so that the grid is changed accordingly. Compared with the commonly used dynamic grid method, this method has better strength and resilience of the parameters performance, so the grid is generated fast. NACA 2415 airfoil, which adopts the “class function” and “shape function” translation technique (CST), is simulated in the case of lowspeed turbulent flow under different angles of attack. The results show that the proposed dynamic grid technique can better simulate the flow characteristics of the airfoil and draw airfoil aerodynamic coefficients more accurately. The technique has good application value for parameter determination at the phase of aircraft conceptual design.

  Optimization of RPC based on InfiniBand            

  MING Liang,FENG Dan,WANG Fang,CHEN Biyan

  2013, 35(7): 23-30. doi:

  Abstract ( 163 )   PDF (1099KB) ( 236 )   　　

  Along with the development of InfiniBand (IB) network, IB is becoming an important component in the data center due to its characteristics such as high performance, low latency, low power consumption and low price. As an important communication protocol, it is a significant problem that how the Remote Procedure Call (RPC) utilizes the RDMA feature of IB. NFS over RDMA solves the problem that the NFS transfers data by RDMA protocol in the kernelspace. However, there is no RPC optimization on InifiniBand in the userspace communication and the communication between userspace and kernelspace. In this paper, we integrate the Socket Direct Protocol into the RPC so as to achieve the RDMA feature of IB in both userspace and kernelspace. Through the experiments, we can obtain 10% and 35% peak performance improvement in userspace and kernelspace communication respectively.

  Brain  storage model based on
  single fuzzy cognitive unit event

  FENG Kang

  2013, 35(7): 31-38. doi:

  Abstract ( 118 )   PDF (741KB) ( 203 )   　　

  Aiming at the flaw of the former storage model of the brain, a storage model of the brain based on single fuzzy cognitive unit event is proposed. The model processes the external cognitive information as fuzzy cognitive unit, and the storage of the model is triggered by a single fuzzy cognitive unit event. The storage is defined as storage computation which is made of a series of storage subcomputation. The storage data is cognition and the memory of the model is memory band which had limited volume, variable rows and columns, so it is able to store more cognitions. Experimental results demonstrate that the model can simulate storage functions of the brain, such as concluding cognition, remembering cognition, recalling cognition and losing cognition. Therefore, it can be conclude that the storage model of the brain based on single fuzzy cognitive unit event is optimal.

  Modeling and simulation of routerlevel
  topology based on scalefree network  

  CHEN Peng

  2013, 35(7): 39-45. doi:

  Abstract ( 129 )   PDF (1580KB) ( 205 )   　　

  Although the powerlaw distributions of nodes degree in scalefree networks are similar, the topologies of them are distinct in all probability. Considering the hierarchy and the quadrant feature, the distribution with weight between nodes and the statistical eigenvalues of complex networks, three restrictions, which are classification of nodes， distribution of dynamic increase and rules of attachment, are added, and a routerlevel topology model and algorithm are proposed. According to the quantitative data of simulation and theoretical analysis, the model is proved valuable and applicable.

  Synchronization of heterogeneous chaotic CNN
  systems and its application in secret communication          

  ZHANG Xiaohong，LI Deyin

  2013, 35(7): 46-52. doi:

  Abstract ( 101 )   PDF (977KB) ( 196 )   　　

  The traditional signal encryption methods are easily attacked so as to be destroyed because of the lack of keys. Owning a well pseudorandomness and a wealth of system parameters, chaotic systems make up for the deficiencies of the traditional encryption. The complex dynamics spatial characteristics of chaotic system with delays make it ideal for signal encryption. Synchronization control of chaotic system with delays is used to achieve ideal secret communication. Cellular neural networks’ structures are more complex than usual chaotic systems and they have higher system dimensions. Chaotic cellular neural networks systems with heterogeneous structures, respectively, act as the drive and response of the synchronization system. The error system between the drive and response is obtained to substitute into a suitable LyapunovKrasovskii functional and this functional is analyzed on stability. An effective synchronous controller is designed and numerical simulations verify its effectiveness. The synchronization system is further used in signal and image secret communications, and its results are very good.

  An enhanced LAMBDA description language
  oriented to combined attack effectiveness evaluation               

  PENG Zimei1，ZHAO Wentao2，ZOU Rongnian1

  2013, 35(7): 53-59. doi:

  Abstract ( 117 )   PDF (979KB) ( 208 )   　　

  Modeling attack actions with description language can depict the details of the attack more effectively. By evaluating the effectiveness of network attacks, the qualitative and quantitative evaluation of the effect of network attacks is concluded, and it can test the effectiveness of attack actions and help to establish effective security policy of network. Based on the LAMBDA description language, this paper expands its expressiveness on time constraints and effectiveness constraints, proposes an enhanced attack description language, the enhanced LAMBDA, and then gives an application example of the enhanced LAMBDA. In the end, this paper uses DARPA data set, the LLDOS1.0, to construct a test scenario of combined attack of network and evaluates its effectiveness based on the enhanced LAMBDA. The experimental results indicate that Enhanced LAMBDA can effectively support the evaluation of effectiveness of combined attacks.

  Security protocols  analysis
  of improved colored Petri nets

  ZHANG Hui,LI Xuwu,ZHAO Yuanli,XIAO Yunchao

  2013, 35(7): 60-63. doi:

  Abstract ( 97 )   PDF (619KB) ( 232 )   　　

  According to the deficiency that space exploration may occur when the security protocols of colored Petri nets are analyzed, the paper proposes to construct the knowledge RI that is required by intruders to successfully attack the security protocols and the knowledge KI that can be obtained by intruders. And the improved colored Petri net of successful attack functions. The improved colored Petri net is used to analyze Helsinki protocol and TMN protocol. Experiments show that the proposal can simplify the procedure of constructing the Petri net with intruders, and efficiently solve the space exploration problem.

  Research on dynamic spectrum allocation algorithm based on spatial congestion games with adaptive load balancing

  LIU Juefu,CHEN Xiao

  2013, 35(7): 64-70. doi:

  Abstract ( 119 )   PDF (1301KB) ( 228 )   　　

  In the future of cognitive wireless networks, spectrum owners in specific region will control a large number of spectrums in different bandwidth and use them to provide different cognitive users with communication services, then how to improve the reliability of the dynamic spectrum access in order to achieve load balancing of spectrum access has become a key issue in the spectrumallocating environment with the coexistence of multiple spectrum providers and cognitive users. This paper simplifies the aforementioned manytomany allocation problem and proposes a dynamic spectrum allocation algorithm based on spatial congestion games with adaptive load balancing by using the extended analysis of Space Congestion Game (SCG) model. The theoretical analysis of existence of its Nash equilibrium is given, and then the utility function added with load regulating is elaborated. Simulations show that this algorithm can not only accelerate the convergence rate, but also improve the spectrum utilization efficiency, thus ameliorating the reliability of spectrum access and finally achieving a load balancing.

  A selfadaptive text visible watermarking algorithm
  based on uniform distribution of gray level           

  LI Xiang，DING Wenxia

  2013, 35(7): 71-76. doi:

  Abstract ( 90 )   PDF (2783KB) ( 234 )   　　

  Text visible watermarking is mainly used for determining copyright of the text digital media in order to prevent illegal usage. In order to overcome the drawback that the existing text visible watermarking algorithms have weak binaryzation attackresisting capability, the paper proposes a selfadaptive text visible watermarking algorithm based on uniform distribution of gray level, according to the characteristics of gray level and texture of text image. Firstly, the rowoutline of the cover text image is extracted, and the embedding strength of watermark is determined by the outline; Secondly, the embedding strength is used to select the black pixels of watermark. Finally, the cover text image and the watermark are mapped to the same graylevel range. The algorithm has advantages of adaptive embedding strength, random embedding position and strong binaryzation attackresisting capability. The experimental results show that the algorithm has good visibility and robustness and it can apply to any forms of text and graphics watermark.

  Improvement and security analysis about
  signcryption scheme in electronic commerce     

  XU Dewu

  2013, 35(7): 77-81. doi:

  Abstract ( 107 )   PDF (805KB) ( 277 )   　　

  Firstly, the paper presents the advantage of digital signcryption over the traditional “signature+encryption” method, and describes the four subalgorithms which compose the digital signcryption scheme. Secondly, it briefly introduces the mathematical basis of the digital signcryption, reviews several existing identitybased digital signcryption schemes under the standard model. It analyses the security of the digital signcryption scheme designed by Yu, and points out that it is insecure when it is attacked by select ciphertext. Thirdly, it gives an improved digital signcryption scheme, and respectively analyses the digital signcryption scheme consisting of four subalgorithmsthe system establish algorithm, the key extraction algorithm, the digital signcryption algorithm and the unsigncryption algorithm. The scheme's security is briefly described through the two aspects of unforgeability under adaptive select message and ciphertext indistinguishability under select ciphertext attack. Finally, the paper compares the efficiency of the improved scheme with other schemes, concluding that the improved scheme is superior at ciphertext size etc, then puts forward plans for future research.

  Asymptotic stability of delayed cellular neural networks   

  LIU Xiaoqun

  2013, 35(7): 82-86. doi:

  Abstract ( 127 )   PDF (384KB) ( 203 )   　　

  By constructing the new Lyapunov functional, ingeniously importing real parameters which can be adjusted, and making use of the inequalities, the global asymptotic stability of delayed cellular neural networks with variable coefficient was discussed and new sufficient conditions were obtained. The results of this paper improve, extend, unify and complement a number of existing results. They also handle a number of cases not covered by known criteria. Theoretical analysis and mathematical derivation show that the global asymptotic stability is related to the delay. The results we obtained highlight the impact of delay on the global asymptotic stability of cellular neural networks. This is of important guiding significance for the design of cellular neural networks with delay. Interesting examples were included to show the versatility of our results. And they greatly enlarge the area of designing neural networks. This has important significance in both theory and application.

  Stock networks community structure division algorithm
  based on multigene families encoding         

  LI Kangshun1,2,CHEN Guihua1

  2013, 35(7): 87-94. doi:

  Abstract ( 122 )   PDF (2639KB) ( 218 )   　　

  In order to solve the problems of the traditional stock networks community structure division, such as low searching accuracy, high time complexity and ease of going into a local optimal solution, this paper proposes a new community division algorithm based on MultiGene Families (MGF) encoding GEP to research the stock market complex networks community structure division phenomenon. The algorithm takes advantage of the MGF property to respectively encode the stock node ID and the community type into two different MGFs, and then implicitly encodes the relationship of the two MGF into the chromosome through a mapping function. Meanwhile, the elite migration strategy is applied to the whole hereditary stage e.g. gene selection, chromosome crossover, chromosome inversion, restricted permutation and so on, which can prevent premature and speed the convergence. Experimental analysis shows that this algorithm implements the stock complex network division accurately and efficiently, and the community structure division result can provide the stock investors with profound information.

  Evolutionary algorithm for constrained optimization
  problem and its engineering applications          

  ZHU Gaofeng1,WU Tiebin2,3,ZHANG Yanlei1,CHENG Yun2,LIU Yunlian2

  2013, 35(7): 95-101. doi:

  Abstract ( 104 )   PDF (1195KB) ( 191 )   　　

  A modified evolutionary algorithm (MEA) is proposed to solve constrained optimization problems. Chaotic sequence method is introduced to construct the initialization population that is scattered uniformly over the entirely search space in order to maintain the diversity. In the evolution process, our algorithm is based on individual feasibility; the population is divided into feasible subpopulation and infeasible subpopulation, which evolve with different crossover operator and different mutation operator, respectively. Numerical simulation results on four benchmark problems demonstrate the effectiveness and robustness of the proposed algorithm. Several engineering optimization problems are designed to test the MEA, and the results show that the MEA can solve different constrained optimization problems.

  Improved particle swarm optimization algorithm
  and its application in modulation recognition     

  QIN Lilong1，WANG Zhenyu2

  2013, 35(7): 102-107. doi:

  Abstract ( 111 )   PDF (723KB) ( 318 )   　　

  In order to resolve the problems that the standard PSO algorithm is apt to be easily trapped in local optima and the LDWPSO algorithm cannot adapt to the complex and nonlinear optimization, the paper proposes a modified particle swarm optimization algorithm based on the information entropy theory, named EPSO. The information entropy value is used by EPSO to determine the inertia weights, which make the algorithm have the ability of “explore” and “exploit” adaptively. The new algorithm is realized for the parameter selection of support vector machine. The simulation results prove that the proposed EPSO is stable. Compared with PSO and LDWPSO, EPSO enhances the ability of escaping from local optimal solution, and becomes more feasible in engineering application.

  Optimization algorithm based on HS
  to solve blocking flow shop scheduling problems 

  BAO Yun,JIANG Hua,ZHENG Liping

  2013, 35(7): 108-112. doi:

  Abstract ( 124 )   PDF (601KB) ( 346 )   　　

  An improved GA scheduling algorithm (IHSGA) base on HS is  firstly proposed to solve the blocking flow shop scheduling problems(BFSP) with the objective of minimizing the makespan of the jobs in the literature. The BFSP scheduling principle figure is constructed and the process of operation method is described. The application steps and the key problems of hybrid algorithm are discussed in detail and NEH & local search methods are used to improve the hybrid. Simulation results show that IHSGA algorithms can obtain a better solution than other algorithms.

  Pixeloriented visualization techniques
  based on the scan line seed fill algorithm 

  GUO Jingfeng,DU Shanshan

  2013, 35(7): 113-118. doi:

  Abstract ( 138 )   PDF (1278KB) ( 341 )   　　

  With the maturity and popularization of the database techniques, datasets which are stored in computers become larger and larger. People are already not only satisfied with the data management, but also hope to analyze data, finding knowledge and information from large numbers of datasets. Pixeloriented visualization techniques stand out from other visualization techniques with advantages of its effectively analyzing and dealing with large scale database. With the comprehensive contrasts of several classic visualization techniques from the time efficiency and the display results, the paper proposes a method to realize the data visualization which based on the pixel techniques. This new method applies the highefficiency scan line seed fill algorithm to making the large scale datasets switch into different color pixels and then displaying them in the structured pixel mapping model. It provides guidance to a macroscopical analysis of the changing regularity to large scale datasets, and greatly improves the efficiency of the algorithm.

  Logo recognition based on grid feature and fuzzy matching        

  WANG Gang1,JIN Yanqing2,CHU Ruilai3

  2013, 35(7): 119-123. doi:

  Abstract ( 113 )   PDF (1041KB) ( 237 )   　　

  Aim at the application requirement of logo recognition in image intelligent processing, a logo recognition algorithm based on grid feature and fuzzy matching is proposed. The grid feature of the logo is extracted, and then transformed into the membership degree of fuzzy sets. At last, closeness degree is used to accomplish logo recognition. This significantly enhances adaptability and antiinterference to the poor quality image, and effectively improves the flexible processing capability of logo recognition system. Experimental results show that the recognition rate of the logo recognition algorithm can reach as high as 95.5%.