Advanced Materials Research Vols. 791-793

Abstract: By mapping one dimension time series into two dimension graph, Horizontal Visibility Graph (short as HVG) method provides a new way to analyze time series structural properties, it has been used in economics, biology, sociology and other areas related. In this paper, we apply it to investigate the Low-rate Denial-of-Service (LDoS) flow packet sequence pattern variations. We establish the Packet Sequence Model and put new metrics to measure the pattern similarity. Using the data carried out on PlanetLab, we find out that the packet sequence time series of the receiver has a bigger HVG degree than the sender, indicating that the packet sequence patterns get more irregular under the impacts of end-to-end delay.

Abstract: In the virtual training simulation software, users through graphical interfaces to fullfill information exchange. By adopting the method of black box testing, treating the virtual training simulation software as a black box and input data to drive the software running, could test the interaction of graphical interface. The test case generation method based on data flow diagram (DFD), which is characterized to form a complete set of test cases and cover the entire path of the program, could improve the efficiency of the test and ensure the reliability of the test results.

Abstract: For many applications in wireless sensor network (WSN) node positioning by unmanned aerial vehicle (UAV) need a path planning algorithm optimizing positioning error. In this article, we present an algorithm combining Voronoi diagram and geometric dilution of precision (GDOP) to make track points, considering both the efficiency of flying and accuracy of positioning. Then, we connect all the track points using genetic algorithm and we get the minimized flight path.

Abstract: The paper introduces how to build a standard Chinese head to be used for numerical calculation of specific absorption rate of electromagnetic radiation using MAYA modeling software, including the demand background of building a standard Chinese head, international research status, as well as the urgency of relative research and study at present in China, and discusses the key steps of the standard Chinese head building using MAYA modeling software model, and the outlook for the numerical calculations after the modeling completed.

Abstract: The Linearized Auto-Localization (LAL) algorithm estimates the position of beacon nodes in Local Positioning Systems (LPSs) which are based on the transmission of ultrasonic signals and proposed for indoor positioning of mobile robots. In this paper we propose an improved auto-localization algorithm based on weighted least squares (WLS). The improved algorithm depends on the different error estimations which caused by the different relative positions of the beacons and the measurements nodes. Simulation results show that our WLS-based Linearized Auto-Localization Algorithm can provide improved accuracy in both distance estimation and position estimation.

Abstract: Virtualization technology has been widely used. Modern data center uses different equipments to support a large number of business systems. Whether the resources scheduling of virtual machine can be realized has become a very important issue, which will reflect the data centers management level of intelligence, intensiveness and automation. On the basis of comprehensive utilization of the physical machines various resources, the algorithm combines the advantages of genetic algorithm with particle swarm optimization algorithm, with reference to model of load balancing scheduling system, constantly adjusts to the load on the host, and consequently achieves the optimum balance of comprehensive utilization rate of different host resources.

Abstract: The present paper analyzes the phenomenon on spallation and failure of driven pile. Based on the nonlinear program LS-DYNA, the dynamical performance of driven pile into the soft soil foundation is studied by numerical simulation. Taking into account the stress-wave theory, some analyses are conducted to discuss the influence factors on spallation and failure of driven pile. The analysis results show that the compression and tension of stress wave plays an important role in the process of pile driven. Some guidance and reference is useful for the design, manufacturing and construction of piles.

Abstract: a visual simulation system of the high altitude gliding torpedo was researched, and the model was establish accurately. Because the attitude angle of the gliding torpedo changed quickly, and the background parameters were complicated, the problem of rapid change part and separation of the model always was exist, the visual simulation was hard to carried out. The method with combined using DOF and Switch node was proposed. The MATLAB Simulink model was used to calculate the 6 DOF data and save the data through the To Workspace module. Then Vega application program was written by VC++ 6.0, with reading the motion parameters and using LOD technology which couldnt reduce the display speed, the simulation visual effect was also improved. Simulation results show that the system can achieve the specific visual simulation functions and performance requirements for high altitude gliding torpedo, the visual simulation t will never be distorted because of rapid change part and separation of the model.

Abstract: The efficient data mining algorithm was researched in this paper, according to the massive data in the database, the efficiency and the fluency of the data mining should be attached much importance in the research. And yet at the same time, the precision of mining algorithm should be improved. Combing with the genetic algorithm and K-means clustering algorithm, an improved data mining algorithm was proposed. In the new algorithm, the slope factor was taken in advantage, then the phenomenon that the smaller classification caused the less optimum solution was avoided, and the defects of the two algorithms are offset. The mining simulation and experiment was taken based on the different databases with different sizes of data. Simulation result shows that the new algorithm based on the slope factor K-means clustering genetic method can solve the data mining problem for the large data base. The data mining result is much more precise than the traditional method. Research result shows the improved algorithm has predominant prospect in application, and it has good value in the engineering practice.

Abstract: With the development of digital information technology, speech coding enhances its development rapidly and plays an important role. Especially the speech coding algorithm, which has low bit rate and high quality, provides qualified multimedia services based on limited channel width. The paper studies the application of DSP technology of TMS320VC5509 in data speech coding algorithm, the overall plan of coding design, AIC23 interface design and coding algorithm. Based on these studies, it explores the optimization of coding algorithm and analyzes algorithm according to optimization results.