DOM tree browsing of a very large XML document: Design and implementation

Browsing the DOM tree of an XML document is an act of following the links among the nodes of the DOM tree to find some desired nodes without any knowledge for search. When the structure of the XML document is not known to a user, browsing is the basic operation performed for referring the contents of the XML document. If the size of the XML document is very large, however, using a general-purpose XML parser for browsing the DOM tree of the XML document to access arbitrary node may suffer from the lack of memory space for constructing the large DOM tree. To alleviate this problem, we suggest a method to browse the DOM tree of a very large XML document by splitting the XML document into n small XML documents and generating sequentially the DOM tree of each of those small n XML documents. For later reference, the information of some nodes accessed from the DOM tree already generated has been also kept using the concept of their virtual nodes. With our suggested approach, the memory space necessary for browsing the DOM tree of a very large XML document is reduced such that it can be managed by a personal computer.     

One-step t-fault diagnosis for hypermesh optical interconnection multiprocessor systems

To maintain high reliability and availability, system-level diagnosis should be considered for the multiprocessor systems. The self-diagnosis problem of hypermesh, emerging potential optical interconnection networks for multiprocessor systems, is solved in this paper. We derive that the precise one-step diagnosability of kⁿ-hypermesh is n(k − 1). Based on the principle of cycle decomposition, a one-step t-fault diagnosis algorithm for kⁿ-hypermesh which runs in O(kⁿn(k − 1)) time also is described.     

Magnetoelectric transducer with high quality factor for wireless power receiving

To achieve strong power coupling, a resonance-type magnetoelectric (ME) transducer with high quality factor is developed to achieve strong ME coupling. The ME transducer employs a type of iron–nickel-based ferromagnetic alloy with constant elasticity and piezoelectric Pb(Zr,Ti)O₃ (PZT8) material. The dynamic magnetomechanical behavior of the ferromagnetic alloy is investigated. The result indicates that the strain coefficient of the ferromagnetic alloy at resonance achieves 557.07 nm/A due to the high effective mechanical quality factor of the alloy. The transducer is designed to operate as a half-wavelength, longitudinal resonator. The dynamic performance of the transducer is evaluated by measuring its electrical and vibrational characteristics. The results reveal that (1) the resonance of the transducer occurs at the frequency of 26.9336 kHz with a strain coefficient of 314.74 nm/A, an effective mechanical quality factor of 1600; (2) the ME voltage coefficient achieves 30.07 V/Oe (i.e., 375.875 V/cm Oe) at resonance; (3) the ME output power density at optimal load resistance of 25 kΩ achieves 0.956 mW/cm³ under 0.3 Oe root-mean-square AC magnetic field. The performances indicate that the transducer is promising for ME energy conversion application.     

A nodal analysis method for electromechanical behavior simulation of bow-tie shaped microbeams

In order to perform mixed-domain simulation of electrically actuated bow-tie shaped doubly-fixed microbeams, a nodal model for the trapeziform beam and corresponding equivalent circuit are developed by HSPICE. The nonlinearities including mid-plane stretching and electrostatic forcing are considered. In the modeling, the Galerkin residual method is used to discretize the partial differential equations of the trapeziform beam. It proves accurate in comparison with FEA simulation results and available experimental data. In particular, as a lumped behavioral model, it is executed much faster than FEM program. The proposed model could be expected to be useful in the optimum design of related MEMS devices.     

一种提取人脸特征3D信息的方法

针对头部姿态大角度偏转时,传统的2D人脸模型(主动表观模型,Active Appearance Model)定位人脸特征点的精度会急剧下降.引入与MPEG-4兼容的3D人脸形状模型Candide,并为其建立2D的形状模型和表观模型.模型拟合人脸图像的过程中,在保持3D形状模型与2D形状模型形变一致的基础之上,应用反向组合算法,提取人脸特征点的位置和头部姿态等3D信息.实验结果表明:在C环境下,算法的处理速度为250帧/秒以上,能达到实时处理;当头部姿态的偏转角度为0°-60°时,算法的标准误差值介于0.01-0.1之间.     

密集多径信道下的IR—UWB跟踪环路的研究与设计

在脉冲体制超宽带(IR-UWB,Impulse Radio Ultra Wide-band)无线通信系统中,极小的定时误差会引起系统性能的极大衰减.为了解决这一难题,提出一种数模混合同步采样的适用于脉冲体制超宽带无线通信系统的锁时跟踪环路来减少定时误差,提高系统性能.采用最大似然估计的方法,得到跟踪环路中最重要环节一时间误差检测器的结构,基于该时间误差检测器完成整个跟踪环路的设计并通过分析该环路的S-Curve和定时误差方差得到了定性跟踪性能的数学分析和仿真.最后,在密集多径信道下面的系统仿真实验结果表明,当采用选择性Rake(SRake)接收机时,该跟踪环路能有效跟踪半脉冲周期左右的时间抖动,使系统的误码率性能有近三个数量级的提高.     

基于混合DSm模型的多机器人地图构建

针对多移动机器人探测静态未知环境,提出了多机器人的一类新的体系结构和机器人内部控制结构,采用分布式融合框架,引入了最近提出的一种在贝叶斯和DST扩展而来的信息融合理论DSmT,结合限制传播算法,建立了不精确传感器(声纳)的混合DSm模型.构造了基本信度赋值函数,计算每个栅格的基本信度值(gbba),有效地融合了多个移动机器人使用声纳获取到的不精确、不确定和高冲突环境信息.最后,以Pioneer 2-DXe机器人作为实验平台,将由混合DSm模型构建出的静态环境地图与实际环境布局做比较,并利用openGL绘制出三维置信度分布图,充分验证了所提出的算法和基于通信的多机器人系统的有效性.     

在静态编译器中实现Java异常机制的算法

将Java程序静态编译成可执行程序是使用Java虚拟机动态编译／解释执行Java程序的另一种运行Java程序的方式。针对Java异常机制的特点和静态编译的需求,在介绍Java异常处理逻辑的基础上,提出一种在静态编译器中实现Java异常机制的算法,结合Open64开源编译器,给出该算法的具体步骤以及实现方式,以SPECjvm98为测试集,验证该算法的有效性。