Advanced Materials Research Vols. 433-440

Abstract: In this paper, a new method is introduced to derive the extended natural gradient, which was proposed by Lewicki and Sejnowski in [1]. However, they made their derivation under many approximations, and the proof is also very complicated. To give a more rigors mathematical proof for this gradient, the Lie group invariance property is introduced which makes the proof much easier and straightforward. In addition, an iterative algorithm through Newton's method is also given to estimate the sources efficiently. The results of the experiments confirm the efficiency of the proposed method.

Abstract: To overcome the defect of the classical ant colony algorithm’s slow convergence speed, and its vulnerability to local optimization, the authors propose Parallel Ant Colony Optimization Algorithm Based on Multiplicate Pheromon Declining to solve Traveling Salesman Problem according to the characteristics of natural ant colony multi-group and pheromone updating features of ant colony algorithm, combined with OpenMP parallel programming idea. The new algorithm combines three different pheromone updating methods to make a new declining pheromone updating method. It effectively reduces the impact of pheromone on the non-optimal path in the ants parade loop to subsequent ants and improves the parade quality of subsequent ants. It makes full use of multi-core CPU's computing power and improves the efficiency significantly. The new algorithm is compared with ACO through experiments. The results show that the new algorithm has faster convergence rate and better ability of global optimization than ACO.

Abstract: Support vector machine (SVM) is a general and powerful learning machine, which adopts supervised manner. However, for many practical machine learning and data mining applications, unlabeled training examples are readily available but labeled ones are very expensive to be obtained. Therefore, semi-supervised learning emerges as the times require. At present, the combination of SVM and semi-supervised learning (S3VM) has attracted more and more attentions. In general, S3VM deals with problems with small training sets and large working sets. When the training set is large relative to the working set, We propose a new SVM model to solve the above classification problem by introducing the fuzzy memberships to each unlabeled point. Simulation results demonstrate that the proposed method can exploit unlabeled data to yield good performance effectively.

Abstract: The quantity of information of digital Video image is huge and it is too difficult to be storaged and transmitted. Compression is a indispensable instrumentality .The text introduces MPEG-4, coding structure and expression of MPEG-4. A compression system of video is designed based on MAP-CA. The system can accomplish two-route-MPEG-4-video compression at the same time. It supports the video image previewed realtime. The video image can be straight carried to the storage. The code stream compressed is also straight carried to the memory. The CPU needn’t do anything in the process. It may no busy resource of CPU if it makes the image carried to the storage by PCI interface. The system is specially the same with DV, Tele-digital supervision, Web video applying, etc .

Abstract: The system of DES being issued by ISO is a data encryption standard . After analyzing the arithmetic of DES , this article brings forward realizing DES using C language . passing circulation or iteration this article uses ordinary base arithmetic (for example , shift left, shift right ,addition of mode 2,etc an) and commutation (choice function , replace function) constructing the nonlinear transform of data stream (encipher commutation or decrypt commutation) to realize cryptogrammic protect for the data of computer. This article emphasize to introduce primary encipher course which is realized by C language.

Abstract: Based on the classical Kalman filtering theory, the state estimation problem is considered for non-square descriptor discrete time stochastic systems. Under Assumptions 1~3, a fixed-Interval Kalman smoother for non-square descriptor systems with correlated noise is given. Some numerical examples illustrate the effectiveness of the proposed algorithm.

Abstract: The purpose of routing protocol is to achieve efficient data collection by making use of the limited resource in wireless sensor networks, which prolongs the lifetime of networks. Energy efficiency of the single hop or multi-hop communication mode is researched taking the uneven energy consumption caused by solely single hop or multi-hop mode into account. A hybrid communication mode in which nodes alternate between single hop and multi-hop mode with certain time proposed based on even cluster protocol. The simulation result verifies that hybrid mode provides the better energy efficiency as well as evener energy dissipation.

Abstract: Wireless sensor network (WSN) requires a wireless communication technology which is low in cost, low in power and easy to be implemented. IEEE802.15.4 standard is the standard specially designed for low-rate wireless personal area network (LR-WPAN) and aimed at introducing an unified standard for the low-rate interconnection between different devices used by a person or in a home, focusing on low power consumption, low-rate transmission and low cost. Since LR-WPAN which is defined on the basis of standard IEEE802.15.4 is of great similarity with WSN, it could be used as communication platform for wireless sensor. This paper, based on the analysis of LR-WPAN, puts forward the WSN networking scheme based on IEEE802.15.4.

Abstract: The Chinese off-line handwritten recognition is a multi-class classification problem. Since there exits certain relation between some classes, it is unnecessary to differentiate all types of the Chinese characters at the time of identifying one type of Chinese character. In this paper, an improved dynamic pruned binary tree SVM multi-classification algorithm is presents. Unvalued support vector is removed at each identification, then a new binary tree is reconstructed according to the features of the structure of the Chinese character, The number of the support vector machines is reduced and the identification speed is increased. Compared with different multi-class classification algorithms, the simulation results show that the method can improve the speed of identifying the classes of the Chinese characters. It is effective and accuracy high.

Abstract: In IEC61850 substation automation systems (SASs), substation configuration description language (SCL) is suggested to assure seamless information exchange among various intelligent electronic devices (IEDs) from different manufacturers in the form of unambiguous and standardized XML. In the paper, native XML database (NXD) is proposed to store and manage SCL files since NXD takes advantage over XML enabled database (XED) in maintaining information integrity that is the most important factor of SCL file management. And the detailed process of SCL file dynamic implementation in IEDs is demonstrated. According to information with different communication real-time requirements, configuration information from SCL files will be extracted in advance or only when the information is used. Finally, a testbed built by open source software Libxml2 under Linux system verifies the effectiveness of SCL file information extraction using Xpath.