Volume 14 (2014): Issue 5 (December 2014)
Special Issue

This paper has designed a variable structure controller based on the nominal compensation of neural networks. The neural network input is the desired trajectory, which eliminates the strict assumptions of the control inputs in conventional neural networks. It also ensures the asymptotic stability of the system closed-loop global exponentials to introduce model compensation and continuous variable structure control rate. By means of Lyapunov stability theory, it is analyzed and researched how to guarantee good transient performance of the control system comprehensively and thoroughly. The theoretic analysis and simulation results demonstrate the efficiency of the method proposed.

Safety risk during the construction of a Large Scale Engineering (LSE) project can be avoided through safety education, following a correct procedure and using an engineering method. In order to ensure that LSE works in a safe state, some parameters need to be restricted to a certain range. Besides, in order to apply the former control for the predictive values, a Model Predictive Control (MPC) theory is suggested to be used for safety risk controlling. We share a MPC work flow and detail calculation steps of the rolling optimization, feedback correlation and constraint control, and a mean time gradient decent method is used during the optimization calculation iteration steps. At the end, regarding the resource allocation issue in the ship-lift engineering construction area of “Three Gorges” we verify how MPC works; and the numeric results show MPC’s efficiency. We can see that MPC might have important future in safety risk management of LSE, because as long as the model can be predictive, regardless of an accurate mathematical model of the system, MPC can control a dynamic and uncertain system very well, which is also a characteristic of the safety risk of LSE.

To have simple and efficient confidence machine learning is an important focus in confidence machine researches. Using one class classifier as a tool, the paper applies it twice for two-class classification problems. Setting reject options and a multi-layer ensemble learning method are used in this study. In this method there is no necessity to set up a specific threshold and the confidence computation is omitted. Realizing five experiments, the study proves it as efficient.

The Sparse Coding (SC) model has been proved to be among the best neural networks which are mainly used in unsupervised feature learning for many applications. Running a sparse coding algorithm is a time-consuming task due to its large scale and processing characteristics, which naturally leads to investigating FPGA acceleration. Fixed-point arithmetic can be used when implementing SC in FPGAs to reduce the execution time, but the implications for accuracy are not clear. Previous studies have focused only on accelerators using some fixed bitwidths on other neural networks models. Our work gives a comprehensive evaluation to demonstrate the bit-width effect on SCs, achieving the best performance and area efficiency. The method of data format conversion and the matrix blocking are the main factors considered according to the situation of hardware implementation. The simulation method of the simple truncation, the representation of the domain constraint and the matrix blocking with different parallelism were evaluated in this paper. The results have shown that the fixedpoint bit-width did have effect on the performance of SC. We must limit the representation domain of the data carefully and select an available bit-width according to the computation parallelism. The result has also shown that using a fixed-point arithmetic can guarantee the precision of the SC algorithm and get acceptable convergence speed.

Nowadays the development of Internet of Things (IoT) technology has witnessed great changes in the world. As it has often been mentioned, IoT Environment Monitoring Technologies and IOT Smart Home Technologies have been gradually accepted by people and have good prospects for development. Now we can research a networking based intelligent platform to monitor our forest environmental factors in time with the new IoT technology based on ZIGBEE protocol. ZIGBEE based networking technologies has the advantages of low power dissipation, low data rate and high-capacity transportation, which makes it more suitable for the design of the node of forest environmental factors collection platform. So, we are going to discuss a kind of IoT forest environment factors collection platform based on ZIGBEE protocol.

The traditional Self-Organize Map (SOM) method is used for the arrangement of seabed nodes in this paper. If the distance between the nodes and the events is long, these nodes cannot be victory nodes and they will be abandoned, because they cannot move to the direction of events, and as a result they are not being fully utilized and are destroying the balance of energy consumption in the network. Aiming at this problem, this paper proposes an improved self-organize map algorithm with the introduction of the probability-selection mechanism in Gibbs sampling to select victory nodes, thus optimizing the selection strategy for victory nodes. The simulation results show that the Improved Self-Organize Map (ISOM) algorithm can balance the energy consumption in the network and prolong the network lifetime. Compared with the traditional self-organize map algorithm, the adopting of the improved self-organize map algorithm can make the event driven coverage rate increase about 3%.

Multi-target tracking is a challenge due to the variable number of targets and the frequent interaction between targets in complex dynamic environments. This paper presents a multi-target tracking algorithm based on bipartite graph matching. Unlike previous approaches, the method proposed considers the target tracking as a bipartite graph matching problem where the nodes of the bipartite graph correspond to the targets in two neighboring frames, and the edges correspond to the degree of the similarity measure between the targets in different frames. Finding correspondence between the targets is formulated as a maximal matching problem which can be solved by the dynamic Hungarian algorithm. Then, merging and splitting of the targets detection is proposed, the candidate occlusion region is predicted according to the overlapping between the bounding boxes of the interacting targets to handle the mutual occlusion problem. The extensive experimental results show that the algorithm proposed can achieve good performance on dynamic target interactions compared to state-of-the-art methods.

In the paper presented the temperature of an oil-immersed transformer was measured, based on the principles of the fluorescence afterglow life. Three methods were used to calculate the fluorescence afterglow life τ by using the least squares method, the integral area ratio method and Prony algorithm. The leastsquare method, the integral area ratio method and the program of Prony algorithm are written using Matlab and C++. The Least-square fitting is susceptible to the influence of the DC component. When the DC location is different, the fluorescence afterglow life τ values vary widely. The integral area ratio method is not influenced by DC component, but it has low sensitivity. Prony algorithm is not affected by DC, it has high sensitivity. So Prony algorithm is selected as a way to obtain the fluorescence afterglow lifetime value τ.

In this paper a new wireless sensor network localization algorithm, based on a mobile beacon and TSVM (Transductive Support Vector Machines) is proposed, which is referred to as MTSVM. The new algorithm takes advantage of a mobile beacon to generate virtual beacon nodes and then utilizes the beacon vector produced by the communication between the nodes to transform the problem of localization into one of classification. TSVM helps to minimize the error of classification of unknown fixed nodes (unlabeled samples). An auxiliary mobile beacon is designed to save the large volumes of expensive sensor nodes with GPS devices. As shown by the simulation test, the algorithm achieves good localization performance.

Aiming at the problem of recommendation systems, this paper proposes a fuzzy clustering algorithm based on particle swarm optimization. This algorithm can find the best solution, using the capacity of global search in PSO algorithm with a powerful global and defining a proportion factor, which can adjust the position and reduce the search space automatically. Then using mutation particles it replaces the particles flying out the solution space by new particles during the searching process. In order to check the performance of the proposed algorithm, by testing with typical ZDT1, ZDT2, ZDT3 functions, the experimental results show that the improved method not only has a better ability to converge to the global point, but can also efficiently avoid premature convergence.

In connection with the problem that the food processing information system is poor due to the absence of knowledge acquisition and a knowledge selfupdating function, a knowledge acquisition approach, based on a Support Vector Machine (SVM) is proposed. First, the approach establishes a set of predicted samples for the relationship between the food processing parameters and product quality; then it uses discretization of the continuous attributes, attributes reduction and a rule extraction algorithm of SVM to automatically acquire predicted knowledge from a large number of predicted sample sets. After that it saves the predicted knowledge in the knowledge base of an expert system; finally, the method realizes extraction of the knowledge about the food processing process based on the inference engine, which greatly enhances the efficiency and applicability of the acquired knowledge in an online aided decision system of food processing quality and safety.

This paper firstly introduces the main content of the Unified Modeling Language (UML) and proves that it can transmit information among the users, the developers, the designers and the managers efficiently, which improves their collaboration capabilities and greatly increases the degree of industrialization in software development projects. Secondly, a library management system development and design is carried out, based on UML modeling mechanism to analyze a simple library management system. Thirdly, a demand analysis mode of the management system is built with the help of a case diagram and an analysis diagram after analysis of a simple library management system, using UML modeling mechanism. Then a book lending management subsystem has been designed in the library management system by a design class diagram and a sequence diagram. The design process indicates that as a modeling language of software engineering, UML has a very good application prospect.

In this paper, choosing highly frequent keywords from core journals in the field of 1992-2013 national knowledge discovery in CNKI database, counting the number of two frequent keywords co-occurrences in the same journal, then constructing the highly frequent keywords matrix, and transforming the highly frequent keywords matrix into a correlation matrix and a dissimilarity matrix, we analyze the dissimilarity matrix based on the use of factor analysis, cluster analysis. After discussing the results of the analysis, we found that the current hotspots in the field of domestic knowledge discovery have focused on the following six aspects, knowledge discovery based on data research, knowledge discovery algorithm optimization research, the model of knowledge discovery and references research, knowledge management based on domain ontology, expert system construction research, and applied research of the knowledge discovery. Finally, we summarized the research hotspots in the field of international knowledge discovery in the same way and suggested the domestic scholars to extend some directions of the research in the field of knowledge discovery.

Establishing cotton storage IoT can realize real-time, unified management of the national cotton storage, which has an important strategic significance. This paper describes the application system of cotton storage IoT and proposes an application platform architecture mixed with C/S and B/S. IoT development is still in its infancy, lacking a unified IoT system standard. Its system has some flaws, especially in its system security. This paper analyzes the security problems of the IoT system and offers some strategies for its research.

A Vehicular Ad hoc NETwork (VANET) is a special subset of multi-hop Mobile Ad hoc Networks, in which the vehicles wireless interfaces can communicate with each other, as well as with fixed equipments alongside city roads or highways. Vehicular mobility dynamic characteristics, including high speed, predictable, restricted mobility pattern significantly affect the performance of routing protocols in a real VANET. Based on the existing studies, here we propose a testing network according to the preferential attachment on the degree of nodes and analyze VANET model characteristics for finding out the dynamic topology from the instantaneous degree distribution, instantaneous clustering coefficient and average path length. Analysis and simulation results demonstrate that VANET has a small world network features and is characterized by a truncated scale-free degree distribution with power-law degree distribution. The dynamic topology analysis indicates a possible mechanism of VANET, which might be helpful in the traffic congestion, safety and management.