Advanced Materials Research Vols. 546-547

Abstract: Data acquisition begins with the physical phenomenon or physical property to be measured. Examples of this include temperature, gas pressure, and light intensity, and force, fluid flow, regardless of the type of physical property to be measured. Physical property converted into digital, and then by the computer for storage, processing, display or printing process, the corresponding system is called data acquisition system. With the rapid development of computer technology, data acquisition systems quickly gained popularity. A variety of products based on digital technology have been created. Digital System spread quickly; it’s mainly the following two advantages: the first is the digital processing flexible and convenient; the second is a digital system is very reliable. The main idea of Reconfigurable computing technology [1] is using the FPGA [2][3] allows the system has a dynamically configurable capacity, suitable for harsh environment applications, improve the speed of data processing. By the use of dynamic reconfigurable FPGA devices can be realized on the hardware logic function modification, application of reconfigurable computing technology can improve the speed of data processing. Data acquisition system is widely applied in many fields, and often used the abominable working environment place. The reconfigurable computing technology, can greatly improve the data acquisition system reliability and safety. The paper introduces a kind of multi-channel data acquisition system based on USB bus and FPGA, the factors affecting the performance of system are discussed, and describes how to use reconfigurable computing technology to improve the efficiency of data acquisition system while reduce energy consumption. The system in this paper uses AD's AD9220, ALTERA's EP1C6-8 and IDT's IDT70V24, Cypress’s CY7C68013.

Abstract: It makes a series of improvements of the traditional oriented way of human face, which is based on ASM. First, we orientate some outstanding feature points on the basis of the results of face test, then make use of these feature points to initialize model. Second, we get some statistics information of the edge in the training process and take a vertical range of the edge in a feature point. According to the scope of positioning, we looked at the local search results on its position before reconstruction the local adjust, thus preventing the search results from departing from the feature pointed have orientated.

Abstract: Aiming at improving the disvantage of single attribute analysis in rough set, the problem of combined attributes analysis is researched in this paper. The single attribute ayalysis only consider the effect of the factor itself on the decison-making but neglect the interaction between multy-attributes influencing the results of the decision-making. So a new combined attributes weighting is presented in this paper.On this basis,a novel analysis method of combined attributes in rough set is proposed. The experiment and theory analysis shows that the new approch is feasible. In addition,it can mine the latent effect on decision making of combined attribute and make up for the shortage brought by single attribute analysis .

Abstract: The complexity is one important index for Galois Field multiplier. This paper presents one low complexity GF(2^m) multiplier based on iterative Karatsuba algorithm. The multiplication is replaced iteratively by three ones of half-length operands which are performed in parallel. The operands are divided into different width such as 64-bit, 32-bit, 16-bit and so on. For the 2^m*2^m multiplier, we take 128 bit-widthGF(2¹²⁸) multipliers as an example. We implement them on FPGA and count the number of the used LUTs and the used registers. Through analyzing the statistic, we find that, when the width of the two multiplication operands is divided to 8 bit, the multiplier consumes the least resources. Compared with the FPGA implementation of the other previous multiplier, this optimum multiplier can save 50% resources in LUTs and the registers.

Abstract: Mobile communication technologies have been widely utilized in daily lives, many low-computing-power and weakly-structured-storage devices have emerged, such as PDA, cell phones and smart cards, etc. How to solve the security problems in such devices has become a key problem in secure mobile communication. In this paper, we would like to propose an efficient signature-encryption scheme. The security of the signature part is not loosely related to Discrete Logarithm Problem (DLP) assumption as most of the traditional schemes but tightly related to the Decisional Diffie-Hellman Problem (DDHP) assumption in the Random Oracle Models. Different from the existing solutions, our scheme introduces a trusted agent of the receiver who can filter the “rubbish” messages beforehand. Thus, with high efficiency in computation and storage, it is particularly suitable for the above mobile devices with severely constrained resources and can satisfy the security requirements of mobile computations.

Abstract: This thesis presents a parallel scheduling algorithm of multi-core processor based on task clustering and duplication. This algorithm, using the strategy of task clustering, gives priority to the operation of thread nodes of the same process on the same processor and effectively reduces time complexity of task scheduling. In order to avoid the unbalanced task load on the processors, it will employ their ultimate values to control the load. Finally, for achieving the optimal time of task operations, this algorithm, with the adoption of task duplication strategy, looks for the key tasks and duplicates them so as to fully utilize the resources of each core on the processor and improve the efficiency of task scheduling. The analysis of the experiment shows that, with the increasing number of task scheduling, the time of task operation of this algorithm is always the least.

Abstract: This paper proposes the concept of Multi-address-transport (MAT) communication model in IPv6 which allows one server to transport a packet to several nodes at the same time without increasing the router’s burden. This model can achieve dynamic join and leave of this group without restrictions of physical location and all happened on servers and clients．In addition this model can overcomes multicast’s shortcoming and retains its advantage. This model need not any new hardware, just need modify software a little, and it have good Compatibility, it can run well even if some machine don’t modify its software, because it works on router that has been modified and passes through router not modified like the original. And, it is particularly suited to the increasingly popular BT transmission.

Abstract: On-demand service is one of the most important characteristics of cloud computing. Cloud-computing services should dynamically and timely deliver computing resources, storage resources and network resources etc. to consumers according to user needs, and service level should be able to meet quality of service. Users only need to pay on demand; otherwise they have to maintain sufficient resources just in order to meet peak requirements, which can be costly. In this paper, we present the design and implementation of Auto-Scaling system and illustrate its system architecture, components and scaling algorithm. Finally, we test the system and the results show that it can be capable of handling sudden load surges, delivering resources to users on demand, saving cost for users and improving resource utilization.

Abstract: Traditional evaluation methods do not consider that e-commerce credit evaluation is inherently uncertain and ambiguity .This kind of method makes the evaluation results unnatural. In this paper, a model of e-commerce credit evaluation based on cloud theory is presented. It can transfer quantitative data to qualitative data by cloud transform and cloud maker and matches the index for rules by cloud reasoning engine. If this model is used to explain the result of evaluation, the result could be easier to understand.

Abstract: Take the SPOT remote sensing images for a district in June 2009 as study object, using the texture feature methods based on the wavelet transform to the extraction for image water information, the results show that: The introduction of the wavelet transform and the edge of the restrictions to the the split of waters , can effectively extract and classify different colors, types of water bodies .