Table of Content

01 February 2007, Volume 27 Issue 2

Previous Issue    Next Issue

Network and communications

Comparison of two predictive algorithms for time delay in networked control systems

2007, 27(2):  257-260. 

Asbtract ( )   PDF (1222KB) ( )  

Related Articles | Metrics

In order to estimate random communication time delay in Networked Control Systems (NCS), two algorithms, a Least Mean Square (LMS) adaptive algorithm and a Least Square Support Vector Machines (LS-SVM) based online predictive algorithm, were proposed in this paper for the time delay estimation and prediction in NCS respectively. Based on more than 1.7 million real timedelay data, obtained from a six months' internet time-delay test, detailed performance evaluation and comparison were made for the two proposed algorithms. The results show that the LMS adaptive algorithm takes less calculation time in timedelay estimation, while the LS-SVM algorithm can achieve a more accurate estimation.

Application investigation of PN Even Balanced sequences in DS-SS Systems

2007, 27(2):  261-264. 

Asbtract ( )   PDF (712KB) ( )  

Related Articles | Metrics

Achieving processing gain as high as possible is one of the important factors that should be taken into account in design of a spread-spectrum system. In this paper, we analyze processing gain from a natural sampling point of view. The analysis has showed that very high processing gains can be achieved in band-limited direct sequence spread-spectrum systems by using PN Even Balanced sequences. Then the way to generate this sequence and its correlation levels has been introduced. Finally, simulation results are presented to verify and confirm the correctness of the claim made.

Study of QoS models in MANET

2007, 27(2):  265-268. 

Asbtract ( )   PDF (835KB) ( )  

Related Articles | Metrics

Quality of Service （QoS） in Mobile Ad hoc Networks （MANET） has been one of hot topics. However, it is very difficult to realize QoS in MANET due to the mobility, processing and power ability of mobile node. This paper summarized eight kinds of QoS models in MANET networks from the point of system and software engineering, and introduced mainly the approaches and characteristics in QoS implementation. In the last section, several QoS methods were compared and analyzed in scalability, level of QoS supporting, relaying on ability of node, flexibility and security. Directions for future study were pointed out.

Solution to the fairness of subscribers with different bandwidth

2007, 27(2):  269-271. 

Asbtract ( )   PDF (597KB) ( )  

Related Articles | Metrics

To solve the problem that subscribers with different bandwidths obtain fair bandwidth in shared uplink, a method called multi-thresholds Random Early Detection (RED) queue was proposed. It set a RED maximum threshold for each subscriber who may have different connecting bandwidths. In the same time, it measured the real bandwidth obtained by each subscriber and adjusted the RED maximum threshold values accordingly. Therefore, the packet loss rate for each subscriber can be controlled and their actual throughput can also be adjusted in proportion to their connecting bandwidth. Simulation results show that the improved queue inherits advantages of RED and can guarantee the fairness of subscribers.

Modeling and analysis of variable-sampling networked control systems

2007, 27(2):  272-274. 

Asbtract ( )   PDF (525KB) ( )  

Related Articles | Metrics

In this paper the means of event-time driven was used in variablesampling networked control systems. Using predictive control value and predictive feedback value， state transfer matrixes of the networked control system corresponding to variable network induced delays were given respectively. Finally, the conditions for the stability of the networked control systems were analyzed.

MAC protocol for wireless Ad hoc network based on IEEE802.11

2007, 27(2):  275-277. 

Asbtract ( )   PDF (786KB) ( )  

Related Articles | Metrics

Based on the IEEE802.11DCF, a Medium Access Control （MAC） protocol was proposed for wireless ad hoc network. It consisted of two key mechanisms: network adaptive backoff algorithm and congestion feedback mechanism. The main idea of proposed protocol was to schedule packet according to real contention status around the node and network congestion situation. The simulation results show that the proposed scheme can effectively reduce the average end-to-end delay, mitigate packet dropped ratio, and improve fairness of channel access.

Design for lwIP-based embedded gateway for CAN/Ethernet

2007, 27(2):  278-281. 

Asbtract ( )   PDF (835KB) ( )  

Related Articles | Metrics

To realize the interconnection between the nodes of Ethernet in the control center and the nodes of the CAN bus in the industry field, an embedded system based CAN/Ethernet was designed. The TCP/IP protocol was implemented by porting the lwIP and CAN application layer protocol was proposed and realized by redefining the content of the IDENTIFIER of the CAN2.0B extension data frame. A transformation protocol was proposed and implemented to realize the data transformation between the two protocols and the control of the communication. The communication between the server of the Ethernet and the CAN nodes and the communication between the CAN nodes with different gateways were realized.

Adaptive algorithm of improving TCP performance in wireless channel

2007, 27(2):  282-284. 

Asbtract ( )   PDF (582KB) ( )  

Related Articles | Metrics

The achieving process of TCP and the problems which TCP may encounter in wireless network were presented. Then an algorithm was presented to predict the next bit error rate of wireless link. Based on the prediction, the TCP was modified to adaptively select appropriate Maximum Segment Size （MSS） to transport data. By using NS2, a wireless channel with variable bit error rate was simulated. Last, a simulation about the modified TCP was done, and its result indicated the effectiveness of the modification.

Implementation of a simple cross domain single sign on system

2007, 27(2):  288-291. 

Asbtract ( )   PDF (747KB) ( )  

Related Articles | Metrics

Distributed structure based multi-sites corporation web applications need a universal authentication and resources access control mechanism, and Single Sign On （SSO） System is an indispensable module for such applications. In practice, we develop a simple single sign on solution that is a central authentication system based on HTTP Redirection and Tickets, with Cross Domain Cookie Sharing as key technology. This solution plays a vital role in distributed data sharing network construction, which realizes cross domain universal user login, user authentication and user authorization. By establishing standard control module, a simple configuration file may conveniently integrate distributed network nodes into one authentication system then accomplish Single Sign On and universal resources access control.

Articles

Security audit system based on LAN monitoring logs

2007, 27(2):  292-294. 

Asbtract ( )   PDF (902KB) ( )  

Related Articles | Metrics

To overcome the shortcomings of LAN monitoring system on security, a security audit system based on LAN monitoring logs was put forward and implemented. By gathering and analyzing LAN monitoring logs, this system kept watching on the running status and detects insecurity factors of LAN. Application results show that this system achieves the security audit function of LAN and provides basis for taking more security measures. Besides, it possesses good transportability.

Blind digital image watermark algorithm based on the second-generation wavelet and image-scrambling

2007, 27(2):  295-298. 

Asbtract ( )   PDF (995KB) ( )  

Related Articles | Metrics

In view of the firstgeneration wavelet’s deficiencies such as low operation speed and that the quality of reconstructed image has something to do with the way of extending the boundary, according to the characteristic of the second-generation wavelet transformation, using the traditional wavelet multi-resolution image watermark algorithm and combining the Amold scrambling algorithm， we proposed a multi-resolution image watermark algorithm which was based on wavelet multistagepromotion and watermark image scrambling. Haar, Db2, Db4 and 9/7 were used to do experiments. The simulation results show that this algorithm can not only make up for the quality deficiency resulted from watermark in the traditional wavelet multi-resolution image watermark algorithm when the filter becomes larger, but also greatly improve the processing speed.

Design and implementation of security gateway prototype system based on protocol translation

2007, 27(2):  299-302. 

Asbtract ( )   PDF (854KB) ( )  

Related Articles | Metrics

This paper described the design of a security gateway based on protocol translation in IPv4-IPv6 hybrid network, and implemented a security gateway prototype system based on Linux 2.6 kernel netfilter framework. The prototype system tracks up-layer UDP/TCP connections based on protocol translation and it performs hybrid end-to-end access control policies. Experimental testing results indicate that it has small latency during end-to-end packet transmission and may satisfy the needs of enterprise networking.

Formal model for cryptographic protocols based on problem-solving theory

2007, 27(2):  303-307. 

Asbtract ( )   PDF (914KB) ( )  

Related Articles | Metrics

This paper proposed a problem-solving theory based formal model, introduced the basic syntax and ρ-calculus based semantics of the model, and presented some pivotal concepts and propositions in the deduction of the model. The model has some properties as follows: can give accurate formal specifications for cryptographic protocols; has provable semantics which is reasonable and sound; can define the security properties precisely and reasonably; is easy to realize automatic deductions. All of those aforementioned make sure that the security analysis of cryptographic protocols based on this model is reasonable and efficient, and provide a dependable basis for analyzing the security of cryptographic protocols correctly.

P2P distributed digital signed system

2007, 27(2):  308-310. 

Asbtract ( )   PDF (741KB) ( )  

Related Articles | Metrics

A distributed digital signed system based on P2P Distributed Digital Signed System (PDDSS) was proposed in this paper. This system uses peer node to compute digital signature, which replaces the CA central server. The system has many advantages: low cost, high adaptation, high availability, and intrusiontolerance. And Verified Secret Share and Proactive Secret Share scheme are used to protect the secret key of CA.

Realization of SIP authentication scheme based on ECC

2007, 27(2):  311-313. 

Asbtract ( )   PDF (567KB) ( )  

Related Articles | Metrics

The Community of Session Initiation Protocol （SIP） entity is becoming an urgent problem to be solved with the continued enlargement of SIP application area. The security scheme, especially its security authentication was discussed and analyzed. Furthermore, a new authentication scheme based on Elliptic Curves Cryptography （ECC）was put forward, which assures the integrity and security during the transmission of SIP message.

DDoS attack detection based on global network properties of network traffic anomaly

2007, 27(2):  314-317. 

Asbtract ( )   PDF (739KB) ( )  

Related Articles | Metrics

Due to the invisibility and distributivity characteristics of Distributed Denial of Service （DDoS） attack, a new DDoS detection method based on global network was presented in this paper. Our method detects DDoS by analyzing OD traffic matrix, whereas the traditional methods detect it on single link or victim network. This method was carried out as follows: First, we need to get network traffic matrix in order to obtain the correlation character of attack traffic among multiple links. Then, traffic matrix was divided into normal space and abnormal space by K-L transformation. Finally, the correlation of abnormal space was achieved to detect DDoS attack. The simulation result shows that this proposed method is more accurate and faster than traditional methods. It is in favor of earlier detection of DDoS attack.

Research and implementation of security mechanism in electronic procurement system

2007, 27(2):  318-320. 

Asbtract ( )   PDF (783KB) ( )  

Related Articles | Metrics

This paper intended to solve the security problem of electronic procurement system. A model of security mechanism was put forward with the help of comprehensive application of the authentication technique combined with digital certificate and ASP.Net’s forms authentication, and of the data encryption technique and the digital signature technique. At the same time, concerning the problems of implementation of encryption and digital signature in B/S mode, this paper developed the electronic contract encryption and digital signature smart client program, and prepared the technical ground for the construction of secure, practical, and big electronic procurement system.

Study on theoretical security of information-hiding based on images

2007, 27(2):  321-323. 

Asbtract ( )   PDF (584KB) ( )  

Related Articles | Metrics

According to the characteristics of steganalysis, to restrict the visual and statistical, especially the high-level characteristic of the image carrier, the classical model called model based on the theory of hypothesis and proof-test was consequently improved and enriched. Then, the improved one was applied to the security analysis of specific hidden image information, thus testing of the security of information hiding system was accomplished.

Intelligence perception

Research into sampling measurement with two-computer cooperation and multiple sampling

2007, 27(2):  324-326. 

Asbtract ( )   PDF (526KB) ( )  

Related Articles | Metrics

This paper expounded the basic principle of sampling measurement based on mask matching, and provided the basic method and Top architecture for capturing the traffic of Intranet export. It also designed and implemented the co-operation algorithm between the two computers with java, so as to realize continuous sampling measurement for a long time. Finally, the authors proposed multi-sampling based on the mask matching to delete some traffic data gradually and smoothly without losing its character. Analysis of the traffic data shows that the length distribution and protocol distribution of message do not change after sampling.

Robust method for facial features localization

2007, 27(2):  327-329. 

Asbtract ( )   PDF (980KB) ( )  

Related Articles | Metrics

In this paper, a robust hierarchical approach based on AdaBoost algorithm and C-V method was presented for facial features localization. First, four kinds of detectors were trained by AdaBoost algorithm for detecting faces, eyes, noses and mouths. Second, face regions were detected using the face detector combined with a rule of face edge. Third, the eye regions, nose regions and mouth regions were detected using the facial feature detectors, and the feature contours and feature points were extracted from the feature regions by C-V method. The experiments on DTU_IMM face test set resulted in 94.6% accuracy rate on eyes, 95.3% on noses and 98.4% on mouths, and the positions of the extracted feature points were accurate. Results show that the proposed approach is efficient and robust.

Research on K-classification covering for PONN

2007, 27(2):  330-332. 

Asbtract ( )   PDF (575KB) ( )  

Related Articles | Metrics

From the aspect of coverage in the High-Dimensional Space (HDS), we discussed algorithms of Priority Ordered Neural Network (PONN), and promoted K-Partitioning Algorithm (KPA) based on anisotropy in HDS. Benchmark testing and application were made on KPA as well as the comparison with Center Adaptive Selection Algorithm （CASA）.The results of experiment prove that the constructive method of KPA is superior to CASA especially in continuous samples.

Grid method of data clustering based on RBF neural networks

2007, 27(2):  333-336. 

Asbtract ( )   PDF (687KB) ( )  

Related Articles | Metrics

Radial Basis Function （RBF） neural networks and grid partition were integrated into a whole in this paper, and an intelligent method of data clustering was proposed, which can process the data in parallel and make increments computation easy. This paper introduced the principle of grid clustering, and the choice of basis functions and the neurons number in RBF neural networks. And then, the clustering strategy and the clustering algorithm were deeply discussed with its aims to identify the locations of data clustering area, and to improve the resolution of data clustering area, and to accelerate the computing. The simulation verifies the validity of this method.

Pattern classification of brain computer interface based on support vector machine and selection of model parameters

2007, 27(2):  337-339. 

Asbtract ( )   PDF (763KB) ( )  

Related Articles | Metrics

Brain-Computer Interface（BCI） is a communication system that connects the brain with computer and peripheral equipments. Continuous Wavelet Transform (CWT) was adopted to decompose the acquired signals of the electroencephalogram (EEG). A multidimensional vector, composed of the variances corresponding to multi-scales, was constructed. Besides, Support Vector Machine (SVM) was applied to do the classified recognition and achieved good effects. The influences of the error penalty parameter C and the Gaussian kernel parameter σ on the generalization ability of support vector machine were studied on the basis of the structural risk minimization principle of the statistic learning theory. Simulation results show that the traditional experience risk minimization principle cannot guarantee good generalization ability. Therefore, the method of adjusting both the parameter C and the parameter σ was proposed to optimize the performance of support vector machine.

Improved fast method for eyes location on human face image

2007, 27(2):  340-341. 

Asbtract ( )   PDF (943KB) ( )  

Related Articles | Metrics

Based on generalized symmetry transform principle, an improved algorithm for eyes location was proposed: First, according to the characteristics of gray sharp changes on vertical, the candidate eyebrow and eye regions were selected rapidly by using integral image technique; then, a symmetry transform algorithm was applied to search eyes; finally, eyes were determined based on prior knowledge. Experimental results prove that the method is simple and easy to be realized, and higher accuracy and speed performance can be obtained.

Cluster-based energy-saving routing algorithm for wireless sensor network

2007, 27(2):  342-345. 

Asbtract ( )   PDF (956KB) ( )  

Related Articles | Metrics

A new Cluster-based Energy Saving Routing (CESR) for energy-limited wireless sensor network was proposed. In this protocol, the network was divided into overlapping clusters; each node only knew the node connectivity within its cluster and the cluster connectivity of the whole network. The global path was selected according to the cluster level topology. The node in the overlap region computed the local path and decided how to send the data to the next cluster. The algorithm tried to find the path with the minimum power consumption and can avoid the node with low energy. The storage and communication overhead was reduced.

Influences of perturbations of training pattern pairs on max-product fuzzy associative memory

2007, 27(2):  346-348. 

Asbtract ( )   PDF (569KB) ( )  

Related Articles | Metrics

A new concept was established in this paper that the robustness of a feed-forward fuzzy associative memory to perturbations of training pattern pair. Then a Max-Product-based Fuzzy Associative Memory (Max-Product FAM) was analyzed. Investigation reveals that such robustness of the memory is good when the fuzzy Hebbian learning algorithm is used, however is poor when another learning algorithm is employed. Finally, an experiment is given to testify the theoretical conclusion and illustrate practical application of Max-product FAM.

Drivers’ eyes state recognition based on fuzzy fusion

2007, 27(2):  349-350. 

Asbtract ( )   PDF (602KB) ( )  

Related Articles | Metrics

This paper proposed a method for drivers’ eyes state recognition based on fuzzy fusion by using multi-feature to recognize the states of eyes, which can overcome the disadvantage of using single feature which is half-baked, inexact and uncertain. The test result shows the recognition rate by using multi-feature fusion is much higher than that by using single feature.

Research of cluster-based multi-hop routing protocol

2007, 27(2):  351-354. 

Asbtract ( )   PDF (847KB) ( )  

Related Articles | Metrics

In order to adapt to the application of large-scale wireless sensor network and further balance the load of each node, a Cluster Based Multi-hop Routing Protocol （CBMRP） was proposed. This protocol was based on the concept of clustering. Cluster heads first fused data collected by member nodes. Then on the basis of a simple heuristic function and the residual energy of the gateways, they selected appropriate gateways, through which to route the fused data among clusters, and finally transmit to the base station. Because this protocol takes account of reducing redundant data and balancing power consumption during transmission, it can further prolong the lifetime of the whole network.

Research of the simplified method in point model based on data envelopment

2007, 27(2):  355-357. 

Asbtract ( )   PDF (602KB) ( )  

Related Articles | Metrics

To investigate the simplification of point-based model during the processing phase, a method based on data envelopment using the convergence of points to reduce the number of point elements was presented. The algorithm of dichotomy was applied to find the optimum declination value when the construction of inner and outer envelopment was designed. Meanwhile, space decomposition was implemented by eight-branch trees structure in the process of the merger of couple-point which was selected by greedy choice algorithm. The experiments have shown a highly effective and simple method maintaining the sharp features of the models.

Fast multiple reference frame selection algorithm based on H.264/AVC

2007, 27(2):  358-359. 

Asbtract ( )   PDF (562KB) ( )  

Related Articles | Metrics

H.264/AVC uses multi-reference frames to enhance the coding efficiency; and the quantity of increased computation is linearly proportional to the number of reference frames. On the study of the multi-macro-block motion estimation of H.264 standard, a low-complexity and fast multi-frame selection method was proposed. The experimental results show that the speed of the proposed method is over two times faster than that of the original scheme adopted in JVT reference software JM10.2 while keeping the similar video quality and bit-rate.

Preferential choice of service provider based on Pareto in migrating workflow system

Qiu-yun CHENG Fang-xi HAN

2007, 27(2):  360-362. 

Asbtract ( )   PDF (584KB) ( )  

Related Articles | Metrics

This paper combined Pareto optimization and genetic algorithm, built a mathematic model to describe the preferential choice of service provider, used the Pareto genetic algorithm to optimize the multi-objective problems, presented the suitable ways of initializing the population, selection, crossover, mutation for the model and got a Pareto optimal set. Also it gave the arithmetic for selecting the solution among optimal set according to the status of service provider. The results of experiments show that this algorithm can make the system resources more efficient than that introduced in reference [3].

Sample set shrinking strategy efficiently improving parameters seeking of support vector machines

2007, 27(2):  363-365. 

Asbtract ( )   PDF (535KB) ( )  

Related Articles | Metrics

The choice of kernel function and relative parameters plays an important role in Support Vector Machines (SVMs). It greatly influences the generalization performance of SVMs. It is time consuming to seek for optimal parameters when the training sample set is large. Concerning this problem, a sample set shrinking strategy was proposed. This method took some of the non-support-vector samples out of the training set; therefore efficiently reduced the set size. That is to say, with half the time consumed, a model can be constructed with testing accuracy just slightly changed.

Combined algorithms for training RBF neural networks based on genetic algorithms and gradient descent

2007, 27(2):  366-368. 

Asbtract ( )   PDF (672KB) ( )  

Related Articles | Metrics

The deficiencies of gradient descent method include the slow speed of convergence, the problem of local minima and the great influence of initial parameters on the performance of the network. Genetic Algorithm (GA) based methods can get rid of the problem of local optima, but they are not very effective to refine an existent good solution. For resolving these problems, in this paper we proposed a new algorithm. The experimental results show the algorithm performs well, and it is better than both Gradient descent algorithm and genetic algorithm.

Dynamic Bayesian networks inference algorithm based on adaptive particle filtering

2007, 27(2):  369-372. 

Asbtract ( )   PDF (695KB) ( )  

Related Articles | Metrics

A dynamic Bayesian networks inference algorithm based on adaptive particle filtering was presented in this paper. The algorithm can change the number of particles over time with dynamic Bayesian networks’ state uncertainty of dynamic Bayesian networks’ evolvement. The number of particles is determined based on statistical bounds on the sample-based approximation quality. The algorithm chooses a small number of particles when the state uncertainty is low, and while it chooses a large number of particles when the state uncertainty is high. The experimental results show that the algorithm can make good choice between inference precision and inference time, and it can outperforms particle filtering. Compared with RBPF algorithm, the algorithm has some advantages on stability and applicability.

Shape and texture-based image classification using wavelet

2007, 27(2):  373-375. 

Asbtract ( )   PDF (1031KB) ( )  

Related Articles | Metrics

This paper presented a method of image classification using wavelet based on shape and texture. Firstly, to use wavelet to transform the images and get edge images, then extract seven edge invariant moments as shape feature vectors. In experiments, finding that the background of images disturb the accuracy of classification in most cases, so to wipe off the background of images and account five second statistics as texture features on the basis of Gray Level Co-occurrence matrix. Lastly, to combine shape features and texture features as image feature vectors, and normalize them by Gauss method, and classify the images by SVM. Experimental results verify the superiority and strong practicability of this method.

Energy saving QoS routing algorithm for wireless sensor networks

2007, 27(2):  376-379. 

Asbtract ( )   PDF (923KB) ( )  

Related Articles | Metrics

Considering the different QoS requirements in different services, a new QoS Routing algorithm—EBQoS algorithm was proposed. The algorithm built multiple paths by memorizing the minimum hops to sink node of the neighbors. Then data was sent by the path that satisfied its QoS requirements and had enough residual energy. According to the comparisons with the SAR routing algorithm, this algorithm has obvious advantages in providing different services and prolonging the network lifetime.

Research of sampling and preprocessing on palmprint image

2007, 27(2):  380-383. 

Asbtract ( )   PDF (1387KB) ( )  

Related Articles | Metrics

In this paper, we focused on the essential technologies of palmprint identification including sampling, color adjustment, brightness normalization and direction alignment. We put forward the general conditions for getting an excellent sampling device and developed our sampling device accordingly. Based on the device, we proposed a statistical method to filter out splash, a method to overcome the difference of illumination, a method to adjust distribution of brightness based on a brightness distribution estimator and a morphology-based method to do direction alignment of hand geometry. Experimental data obtained through extensive simulations shows that our scheme is efficient.

Software process technology

Design of auto text categorization classifier based on Boosting algorithm

2007, 27(2):  384-386. 

Asbtract ( )   PDF (506KB) ( )  

Related Articles | Metrics

Boosting algorithm is a popular machine learning technology now. An auto text categorization classifier was designed by using an improved Boosting algorithm—Adaboost.MHKR. Then the evaluation method and results were given. The evaluation results show that the classifier is effective.

Multi-model synthesis dynamic prediction of software reliability based on inclusion degree theory

2007, 27(2):  387-388. 

Asbtract ( )   PDF (575KB) ( )  

Related Articles | Metrics

Different reliability models result in different reliability predictions for the same software, and failed process of the same software could not be described using a single model in software reliability project. Concerning these problems, a multi-model synthesis dynamic prediction method for software reliability based on inclusion degree theory was proposed. The definition of attribute importance degree based on inclusion degree theory and a select method of inclusion degree function were presented. Then the process of multi-model synthesis dynamic prediction for software reliability was put forward. This method integrates the strong points of various models and makes in-homogeneity model compensating each other through dynamic changing their weights, which can solve the practical problems in software engineering. A case study was done to demonstrate the feasibility of this method.

Goal and information driven software measurement model

2007, 27(2):  389-393. 

Asbtract ( )   PDF (932KB) ( )  

Related Articles | Metrics

At present, the enterprises especially the medium and small ones that adopt the CMMI model to perform their software process improvement often choose the metrics blindly, and are lack of the guidance of appropriate process framework when implementing the measurement. To solve this problem, the paper presented a goal and information driven measurement model (Goal Information Indicator Measure Dataset, for short GI2M-D), based on the characteristics of existing GQM model and its improved model as well as the Practical Software Measurement (PSM)model.Then a software measurement framework based on the GI2M-D measurement model was brought forward. The framework can help those who perform the measurement to choose metrics more relevantly and purposefully. It can also provide effective support and guidance for the decision-makers who are in charge of improving the software process, and especially for the enterprises that adopt the CMMI model to implement the measurement.

Design and implementation of virtual lab platform based on integration of Java and Matlab

2007, 27(2):  394-396. 

Asbtract ( )   PDF (879KB) ( )  

Related Articles | Metrics

This paper proposed a Virtual Lab Platform (VLP) that used JMatLink to realize the joint between Java and Matlab. This paper also introduced the design and implementation of VLP based on integration in detail, discussed the realization of Virtual Lab Server (VLS) and took digital communication virtual lab as an example to introduce the realization process and implementation mechanism of the whole platform.

SoC system-level specification based on PetriNet and systemC

2007, 27(2):  397-399. 

Asbtract ( )   PDF (541KB) ( )  

Related Articles | Metrics

An approach that utilizes Petri net and systemC language to establish system-level executable module was described. It can overcome the shortages that Petri net is short of information and systemC is not represented by graph. The comparability of the Petri Net and systemC was analyzed so as to transform Petri net to systemC simply. At the same time, hierarchical analytical method was adopted to avoid states blast. The system-level executable module was used to describe the states of software and hardware, as the golden modeling of the embedded SoC system.

Design of workflow model interoperation based on rule

2007, 27(2):  400-402. 

Asbtract ( )   PDF (610KB) ( )  

Related Articles | Metrics

In this paper, we mainly analyzed the structure and feature of workflow language, XML Process Definition Language （XPDL） and Business Process Execution Language For Web Services （BPEL），and then designed a strategy of workflow model interoperation in two aspects, control flow and data flow of workflow. Based on this idea, we developed a transformation algorithm and implemented this transformation by a Java-based open source rule engine named drools.

Discussion and research of a template based process definition language

2007, 27(2):  403-405. 

Asbtract ( )   PDF (783KB) ( )  

Related Articles | Metrics

By presenting the concept of workflow and XML Process Definition Language （XPDL）, this paper listed some drawbacks of XPDL in practice. In addition, the XPDL was improved, and another Process Definition Language based on template which had more practical value was put forward. Whats more， the meta model of the language and the way of separating the process data and business data were explained in order to acquire the reusability and expansibility。At last, a process definition tool using it was implemented.

Design and implementation of focused Web crawler based on semantic analysis

2007, 27(2):  406-408. 

Asbtract ( )   PDF (578KB) ( )  

Related Articles | Metrics

The design and implementation of a Semantic Analysis Focused Web Crawler （SAFWC） was introduced. In combination with HowNet，extended metadata semantic relevance algorithm for predicting the relativity between URL and top ic was applied. The result of experiments has shown that SAFC has higher efficiency and accuracy for Web pages relevant to a predefined set of topics.

Meta model and modeling language for workflow supporting service collaboration

2007, 27(2):  409-412. 

Asbtract ( )   PDF (782KB) ( )  

Related Articles | Metrics

There are limitations in currently existing modeling language supporting web services, such as high complexity, poor interoperability and lack of effective support for service collaboration relationship. Process definition meta model of service-oriented workflow was presented on the basis of the analysis of service collaboration requirements, and the corresponding formal model was introduced. A model language for Service-oriented XML Process Definition Language (SoXPDL) supporting service collaboration was defined by extending XPDL.

Database Technology

Study of SOA-based ERP system architecture

2007, 27(2):  413-414. 

Asbtract ( )   PDF (513KB) ( )  

Related Articles | Metrics

After analyzing the limitation of current ERP systems, this paper pointed out that the current ERP systems could not satisfy the needs of enterprises including the business process reorganization and the support for supply chain management function. Inspired by the property of Service Oriented Architecture (SOA), a new SOA, which was specialized for ERP system, was proposed in this paper. In the mean time, a J2EE-based implementation system was also illustrated in this paper.

The Design of Low-cost Query Middleware in Grid Database

2007, 27(2):  415-417. 

Asbtract ( )   PDF (534KB) ( )  

Related Articles | Metrics

Grid Database is a new research area in Grid and Database. There is new requirement for query plan optimization in this area. This paper described the data relation between sub-query and database nodes, and gave a query middleware Low-Cost Query Middleware （LCQM） based on low-cost model. An algorithm of low-cost query plan was proposed. Experimental results demonstrate the enhanced efficiency of the query plan.

Research of an improved Apriori algorithm in mining association rules

2007, 27(2):  418-420. 

Asbtract ( )   PDF (576KB) ( )  

Related Articles | Metrics

An enhanced Apriori algorithm which directly used the row vectors of boolean matrix for transaction databases to find out the frequent item sets was presented in this paper. It used the inner product and discriminant rule to concentrate the row vectors of boolean matrix step by step, so the frequent item sets of transaction databases can be inducted quickly and intuitively. Studies and analysis of the proposed algorithm show that it can not only scan the database once, but also has the virtues in high speed, less memory cost and handling with large item set dimensions. It can also be well applied to super transaction database and distributed transaction database.

Development of adaptive scheduling knowledge-based system based on UML

2007, 27(2):  421-424. 

Asbtract ( )   PDF (965KB) ( )  

Related Articles | Metrics

It is a fundamental and core issue that constructing better knowledge-based system to fulfill an efficient adaptive scheduling to complex manufacturing system. The adaptive scheduling Knowledge-Based System(KBS) analysis，design and implementation were described by utilizing Unified Modeling Language (UML), various views were built up to explore the static structure and dynamic behavior, and the semantics of class was accurately constrained by using Object Constraint Language (OCL). The high dependence on domain knowledge experts was alleviated through applying this modeling technology. The quality and efficiency of KBS were improved. Finally a practical application was given to validate the feasibility of the proposed methodology.

Time series pattern matching based on alternative covering algorithm

2007, 27(2):  425-427. 

Asbtract ( )   PDF (651KB) ( )  

Related Articles | Metrics

Time series pattern matching is an important issue in data mining. In this paper, a new matching method based on alternative covering algorithm was proposed. After pretreatment, time series were distributed over a spherical surface with its center at origin. Then a classifier was constructed through learning and pattern matching can be well and truly finished. All-pairs query was mainly discussed in this paper, that is to say, classifying similar time series into the same category was the primary task. Pattern matching discussed here can be effectively fulfilled by the matching method based on alternative covering algorithm, which has the function of classifying samples accurately.

Lexicon-based ontology integration system

2007, 27(2):  428-430. 

Asbtract ( )   PDF (522KB) ( )  

Related Articles | Metrics

To deal with the lack of semantic understanding on ontology integration, we used the lexicon technology, which can identify synonyms and adopt heuristic rules to compute similarity of entities that come from source ontology. Based on these approaches, we developed a semi-automatic system for ontology integration. The system can help user to integrate ontology and save labor, meanwhile improve the precision of ontology integration.

Technical study on the database of multiparameter state monitor and fault diagnosis system

2007, 27(2):  431-432. 

Asbtract ( )   PDF (628KB) ( )  

Related Articles | Metrics

Based on the data characteristics of large state inspecting and fault diagnosis system, comparing the applicability of WinCC configuration database，Microsoft Access database and file system in managing date to this system, file system was chosen to manage data and memory alignment was adopted to design storing format. Practical application shows that this method can build up a high real-time and large database so as to meet the needs of industrial spots.

Improved feature selection algorithm with conditional mutual information

2007, 27(2):  433-435. 

Asbtract ( )   PDF (584KB) ( )  

Related Articles | Metrics

Traditional feature selection algorithms have a common drawback, i.e. they do not consider the mutual relationships between features. It can result in that one feature’s predictive power is weakened by others and the lost of efficiency. In this paper, we proposed a new feature selection method called Conditional Mutual Information Maximin (CMIM).It can select a set of individually discriminating and weakly dependent features. Simulation results demonstrate that the proposed method can improve the precision of text classification.

Design of heterogeneity data source integration system based on B/S/S

2007, 27(2):  436-437. 

Asbtract ( )   PDF (502KB) ( )  

Related Articles | Metrics

As homogeneity data source structures are required in traditional tightly-coupled integration methods, these methods are only suitable for Intranet and difficult to extend to Internet. Therefore, study was carried out on heterogeneity data sources and Web database structures. A loosely-coupled heterogeneity data source integration solution based on dynamical database B/S/S architecture was proposed. XML, ASP and ADO techniques were used in this solution.

Typical applications

Blind separation of noisy image based on the curvelet transform and independent component analysis

2007, 27(2):  438-441. 

Asbtract ( )   PDF (1135KB) ( )  

Related Articles | Metrics

The curvelet transform is a new method developed from the wavelet transform, and it has good performance in denoising sophisticated image. This paper, based on the basic model and methods of Independent Component Analysis （ICA）, proposed a new blind separation method of noisy image using Fast Discrete Curvelet Transforms （FDCT） via UnequallySpaced Fast Fourier Transforms （USFFT）and the algorithm of Fast ICA. The simulation results show that for the mixed images with additive white Gaussian noise this method has effective performance in mixed image de-noising and separation.

Ant colony optimization algorithm for multiprocessor scheduling problem

2007, 27(2):  442-445. 

Asbtract ( )   PDF (687KB) ( )  

Related Articles | Metrics

The ant colony optimization (ACO) algorithm is an intelligent optimization algorithm which simulates the ants’ behaviors in searching for food. ACO is especially suitable for the discrete combination optimization problems. In this paper, we presented an ACO algorithm for multiprocessor scheduling problem. In the algorithm, each ant represented one processor and selected its tasks. The algorithm measured the urgency of each task by analyzing its critical path, the earliest and latest starting time. With the help of the obtained information, the ant selected the proper tasks for the processor it represented. Experimental results show our algorithm can obtain more efficient scheduling results than classical algorithms.

Recognition and positioning of hailstorm based on chain code information

2007, 27(2):  446-448. 

Asbtract ( )   PDF (629KB) ( )  

Related Articles | Metrics

In this paper, a new method of processing codes was proposed, which enabled the Minimize Sum Statistical Direction Code (MSSDC) based on Freeman chain code to be invariant to the origin and rotating, and also enabled the direction entropy to be invariant to scaling. In addition, a statistical algorithm was also proposed, which can detect the given shape effectively and supply important premise to position hailstorm area. The experiments show that the method can not only acquire high processing speed and high accuracy and robustness of the template matching, but also position the hail area accurately.

Traditional Chinese Medicine prescription mining based on attribute-oriented relevancy induction

2007, 27(2):  449-452. 

Asbtract ( )   PDF (1071KB) ( )  

Related Articles | Metrics

The traditional Attribute-Oriented Induction （AOI） technique is weak at efficiency and coarseness of generalization. In order to satisfy the complex requirements in Chinese medicine prescription mining, this study proposed a new algorithm based on attribute relevancy generalizing driven by Chinese medicine data, established concept-tree for relevancy-dimension, utilized pertinence of relevancy-attribute and benchmark-attribute to enhance efficiency of induction, and implemented a new data mining system TCMDBMiner based on attribute-oriented relevancy induction. Experimental results show that new algorithm is 23% faster than traditional method and the mining results are consistent with the Chinese medicine theory.

Image emotional semantic classification based on color features

2007, 27(2):  453-455. 

Asbtract ( )   PDF (794KB) ( )  

Related Articles | Metrics

This paper analyzed the strong relationship between color features and human sensations and showed how high-level emotional representation of painting can be inferred form perceptual level features suited for the particular classes (elegance vs. flaring). Classification was performed by Probabilistic Neural Network （PNN）.

Applied research of combinatorial prediction model on the GMDH

2007, 27(2):  456-458. 

Asbtract ( )   PDF (531KB) ( )  

Related Articles | Metrics

Making use of Group Method of Data Handle （GMDH） multilayer model, GMDH self-regression model, multi-dimension AC model and single dimension Analog Complexing （AC） model, a prediction model for industry increasing value based on the GMDH was established. Furthermore, an optimal linear combinatorial prediction model was also built up. Experimental results show that this prediction model can not only improve simulation precision of the model, but also evidently improve the prediction ability of the model.

Study on text summarization method based on users’ interests

2007, 27(2):  459-462. 

Asbtract ( )   PDF (964KB) ( )  

Related Articles | Metrics

An automatic text summarization method was proposed in this paper, which took sentence as the basic extraction unit and weighted each sentence by considering the topic-related phrases. The latent semantic structure of the document, which is very important to improve the quality of automatic text summarization, can be obtained by clustering the sentences based on Latent Semantic Index (LSI). Besides, a relatively objective and effective method for the evaluation of automatic text summarization was presented. Experiments have shown that this approach is effective.

Interactive segmentation method with SVM and region growing

2007, 27(2):  463-465. 

Asbtract ( )   PDF (670KB) ( )  

Related Articles | Metrics

The SVM segmentation method, as a kind of global thresholding, cannot segment the image finely. The results need to be post-processed by other segmentation approaches. In this paper, a new interactive segmentation method was proposed, which integrated the SVM and region growing. The proposed method can remove the regions of non-interest that have similar characteristics to regions of interest. Furthermore, it can combine the two procedures into one: the procedure of selecting samples for training SVM and the procedure of selecting seeds for region growing. As a result, the method improves the quality of segmentation and the performance of auto-segmentation.

Study on urban mixed network design problem based on genetic algorithm

2007, 27(2):  466-469. 

Asbtract ( )   PDF (759KB) ( )  

Related Articles | Metrics

In order to reduce the deficiency in urban road network design, a multi-objective optimization model of network design based on genetic algorithm was presented in this paper, which can calculate a lot of layouts of road network for decision-makers. A simulation test has been made in an urban road network design by using the model and algorithm. The results indicate the model and algorithm are feasible and effective, and they have provided a reference for urban road network design.

Curved surface modeling and texture acquisition from three images

2007, 27(2):  470-472. 

Asbtract ( )   PDF (757KB) ( )  

Related Articles | Metrics

3D curved object was represented by the standard parametric surface. The silhouettes of the images photographed from the frontal，lateral and rear positions were inputted as the linear constraints of the quadratic object function to find the smoothest curved surface of the object. The recovered surface was re-projected to the frontal, rear and lateral images of the object. Then, the pixels of the recovered texture were obtained as a barycentric combination of the colours of the near four pixels. The effectiveness of this method was demonstrated in modelling real curved objects.

Threshold abandoning strategy in hierarchical route-finding algorithm

2007, 27(2):  473-474. 

Asbtract ( )   PDF (637KB) ( )  

Related Articles | Metrics

In order to determine when to give up hierarchical route-finding algorithm, we put forward a kind of threshold abandoning strategy. When the estimated distance between original point and destination is shorter than the abandoning threshold, shortest route algorithm will be adopted, otherwise hierarchical route-finding algorithm will be adopted. Three parameters named estimate distance, abandoning threshold and abandoning factor were introduced to describe the new strategy, and abandoning threshold will be worked out by means of statistics. Experiments show that this new strategy can make more routes adopted by drivers.

Integrated simulation algorithm based on constrained TIN and CSG

2007, 27(2):  475-478. 

Asbtract ( )   PDF (905KB) ( )  

Related Articles | Metrics

In order to organize and manage 3D GIS spatial data effectively, and represent 3D spatial entity integrally, a new simulation modeling algorithm was put forward, which combined data structures of constrained Triangulated Irregular Network （TIN） and Constructive Solid Geometry （CSG）. In this algorithm, TIN model was used to represent terrain and CSG model was used to represent buildings. The two simulation models were integrated by extracting ground contour lines of buildings as constrained conditions when carrying out triangulating terrains. And the visualization for all parts of the integrated model can be processed synchronously. The results of simulation experiments demonstrate that the algorithm simplifies modeling process and to a certain extent, reduces computing complexity for visualization.

Framing and implementation of distributive population information system

2007, 27(2):  479-481. 

Asbtract ( )   PDF (856KB) ( )  

Related Articles | Metrics

To make full use of current population information resources and meet the demand of setting up a distributive population information management system, this paper put forward a three-layer structure information integration model of distribution population system by using Web services technology. The paper analyzed management mechanism and information processing process on all layers of the model, and solved the information integration problem between users such as the differences in system languages, platforms, protocol and data structures. Besides, based on the construction model, this paper discussed a programming technique, which developed a corresponding system by taking advantage of Java’s cross-platform function, the powerful exchanging ability of the Internet and the powerful linking capability of database.

Application of One-Class Support Vector Machine in Tobacco Leaves Selection

2007, 27(2):  482-485. 

Asbtract ( )   PDF (740KB) ( )  

Related Articles | Metrics

In order to solve the problem on substitute of tobacco leaves in cigarette blend, a model for selecting tobacco leaves based on one- class support vector machine is proposed. The principle of the model is introduced, and the primary factors influencing the veracity of the model are discussed, then an optimization strategy for the model is presented by joint optimizing of feature selection and parameters for support vector machine. It was proved by an applied experiment that the model could identify the comparability of tobacco leaves according to the measured data of the chemical components of tobacco leaves. This made it possible to carry out intelligent selection for tobacco leaves.

Application study of environment mapping in VR systems

2007, 27(2):  486-488. 

Asbtract ( )   PDF (713KB) ( )  

Related Articles | Metrics

To realize the realistic rendering of the scene in applying environment mapping to VR systems, an approach to fast calculating diffuse environment maps was proposed starting from the analysis of spherical harmonics. To simplify the filtering of environment maps for glossy reflections, we decided to use a cube map representation together with a simple box-filter instead of the Phong cosine kernel. In the end, we achieved fast rendering by generating a mip-map of the specular environment map. The experimental result shows that this technique realizes fast rendering of diffuse and glossy reflections and is well-suited for virtual reality applications.

Uncertainty analysis of resource scheduling in grid computation

2007, 27(2):  489-491. 

Asbtract ( )   PDF (576KB) ( )  

Related Articles | Metrics

Resource scheduling technology is a part of grid core service technology. In the process of the Grid system emulation experiment, many uncertain factors of resource scheduling were found. These factors would influence qualities of Grid system in various degrees. As a kind of learned disquisition based on experiment, the uncertain factors t found in experiment were analyzed, and resource scheduling strategy was improved accordingly.

Recognition of pigmented skin lesions based on fuzzy support vector machines

2007, 27(2):  492-493. 

Asbtract ( )   PDF (666KB) ( )  

Related Articles | Metrics

Unclassifiable regions exist in conventional SVM classification with the two-class problems being extended to the multi-class problems. In order to solve this problem, fuzzy membership function was introduced in conventional SVM. Fuzzy support vector machine was applied to solve the unclassifiable problem existing in multi-class recognition by conventional SVM. The experimental results demonstrate that the proposed method outperforms the conventional SVM in the recognition of pigmented skin lesions.

Improved contract-net cooperation strategy based on auction mechanism

2007, 27(2):  494-496. 

Asbtract ( )   PDF (592KB) ( )  

Related Articles | Metrics

Based on the present study of contract net model, the auction mechanism in Economics was introduced, and an improved contract net consultation strategy based on auction was proposed in this paper. It allocated task dynamically through independent competition. It can indirectly support the dynamically changing capability of each agent, and reduce the cost of communication and the difficulty of acquiring knowledge.

Algorithm to improve the rasterizational quality of small curve-outlined fonts

2007, 27(2):  497-498. 

Asbtract ( )   PDF (663KB) ( )  

Related Articles | Metrics

An algorithm was proposed to improve the rasterizational quality of small curve-outlined fonts. Three steps were involved in this algorithm. First, the curves should be separated into feature points according to the curvature of the typeface curve dynamic. Then, the redundancy operation according to the slope and the resolution size should be taken based on the characteristic that small font displayed in the low curvature. Finally, the typeface edge should be trimmed. The experimental results indicate that, this algorithm improves the rasterization quality of small curve-outlined fonts, greatly reduces the equipment memory demand, and enhances the typeface demonstration speed.

LOD algorithm based on quadtree for TIN

2007, 27(2):  499-502. 

Asbtract ( )   PDF (1009KB) ( )  

Related Articles | Metrics

Through building a quadtree structure on irregular point data, and combining the quadtree based Level-of-Detail （LOD） methods and the Triangulated Irregular Network （TIN） based triangle mesh simplification methods, we proposed and realized a LOD algorithm for TIN. The goal of the paper was to apply the method suitable for grid terrain data to the irregular point data, and to realize an efficient multi-resolution expression of the terrain.

Cube partitionbased rapid three-dimensional k-coverage decision algorithm for sensor networks

2007, 27(2):  507-509. 

Asbtract ( )   PDF (637KB) ( )  

Related Articles | Metrics

A Cube Partition based Rapid Three-dimensional k-Coverage Decision Algorithm （CP-RTCDA）was proposed, in which the region of interest was partitioned into triangles firstly, and then the complex area coverage problem was transformed into simple triangle coverage problem. Theoretical analysis and simulation results show that, for sensor networks with different sensors, the new algorithm can solve the k-coverage problem correctly for any given region of interest with time costs of O(n) only, which is far below the time costs O(n3logn) of other algorithms.

Algorithm based on local variance to enhance contrast of fog-degraded image

2007, 27(2):  510-512. 

Asbtract ( )   PDF (740KB) ( )  

Related Articles | Metrics

This paper presented an image contrast enhancement method based on local variance that can preserve brightness and gray levels. Because of the characteristic that local variance can show the detail of image, a method of contrast enhancement through local variance comparison to judge enhancement degree was proposed. It can not only enhance the image contrast and hold the details simultaneously, prevent gray levels losing, but also keep the contrast, brightness of the original image and suppress noise. Thus it is especially adapted to the fog-degraded images with scene depth variations．Experimental results show that the proposed algorithm is effective.