基于光线投射的全GPU实现的地形渲染算法

地形渲染算法需要处理大量的地形及纹理数据,影响三维动画显示的流畅性和性能提高。随着GPU绘制能力提高,CPU与GPU的负载失衡逐渐成为制约性能提高的瓶颈。结合现代GPU体系结构,在GPU上实现了基于光线投射（Ray Casting）的地形渲染算法。算法简化了Ray Casting算法,把LOD策略和预裁剪统一到GPU中实现,保证了CPU和GPU之间的负载平衡,同时简化了应用程序的编制。为获得较好效果,还采用查找表（Lookup—Table）的实时纹理合成算法合成纹理,进一步降低了CPU处理纹理数据的开销。实验表明,本文算法不仅充分利用了GPU的处理能力,还降低了CPU负载,提高了动态三维重建的帧刷新率,并获得较逼真的渲染效果。     

基于谐波特征分析的时间调制阵列测向研究进展

针对近年来发展起来的基于谐波特征分析的时间调制阵列(time modulated array,TMA)测向这一新型无线电测向技术体制,全面总结了其目前的发展现状以及与传统阵列测向体制相比较的优缺点.首先分析了TMA的起源、发展与数学基础;其次总结了基于TMA的测向方法的研究现状;然后对二单元TMA测向、多单元TMA测向...     

图像处理技术检测微型冠状保持架形位误差

基于图像处理原理,提出了检测微型轴承冠状保持架多段冠状体圆度误差和同心度的方法。将显微视觉获取的保持架图像进行处理和边缘识别,对于识别得到的冠状体边缘点,运用最小二乘法得到拟合圆和圆心,从而计算出边缘点所形成的最大半径和最小半径,两者差值即为保持架冠状体的圆度误差;对于保持架各段冠状体圆弧,分别拟合出各自的圆心,进而求出保持架冠状体的同心度误差。     

基于全相位FFT的稳健型时间调制阵列测向

基于谐波特征分析的时间调制阵列测向方法的正确性与精度严重依赖接收谐波的估计精度. 传统的离散傅里叶变换(discrete Fourier transform, DFT)或快速傅里叶变换(fast Fourier transform, FFT)在估计谐波的幅相时,由于信号频率通常偏离采样频率的整数倍,会形成栅栏效应,从而引起基于谐波特征分析的时间调制阵列测向的精度降低甚至失效. 针对该问题,本文将全相位FFT引入二单元时间调制阵列接收谐波的分析中,通过提升谐波幅相估计的鲁棒性来提升时间调制阵列测向方法的稳健性. 仿真结果表明,当信号的载频为频谱分辨率的任意小数倍时,提出的全相位FFT时间调制阵列测向方法均能正确测向,且随着信噪比的增加,测向均方根误差收敛至0. 本文工作提升了基于谐波特征分析的单通道时间调制阵列测向方法的稳定性。     

Study of superparamagnetic nanoparticles as labels in the quantitative lateral flow immunoassay

The superparamagnetic nanoparticles (MNPs) with a series of sizes and magnetite contents are designed and prepared as labels used in lateral flow immunoassay (LFIA) to test if the size and magnetite content of MNPs can affect the performance of LFIA. The result showed that detection time was mainly determined by the size of MNPs, while the signal intensity was closely correlated to the magnetite content of the MNPs. In addition, the magnetic signals intensity remained stable over a long time period. Smaller MNPs with higher magnetite content are first choices as labels to construct a rapid and high-sensitive quantitative LFIA.     

基于CAN总线的AMT通信模块设计

AMT智能换档控制是当今自动变速器发展和研究的方向,如何获取可靠的车辆信息是其实现的关键。CAN总线是当前在汽车上应用最广的现场总线之一,本文介绍基于CAN总线的AMT通信模块设计,从硬件设计、数据定义到相关软件设计,实现与发动机、ABS及EPS等车辆部件之间的数据通信,为智能换档的最终实现提供基础。     

干涉仪测向技术现状与发展研究

基于时间调制理论,提出了一种单通道框架下的多基线相位干涉仪测向方法,其克服了传统相位干涉需要多个射频通道带来的系统复杂度高以及通道幅相不一致性问题. 所提方法利用单刀多掷开关周期性接通多基线相位干涉的各天线单元,并从接收信号产生的谐波特征中同时估计多组基线产生的相位差. 首先分析传统多基线相位干涉仪测向存在的局限性,然后提出基于时间调制的单通道多基线相位测量方法,最后结合2~8 GHz宽带测向需求,设计四基线相位干涉仪并进行仿真. 仿真结果显示,在2~8 GHz范围内,测向误差能达到0.1°以内,证明了本文方法的有效性.     

A general FORTRAN to C translator

Theorists in Edinburgh University Physics Department are currently using two ICL Distributed Array Processors (DAPs), programmed in a matrix and vector extension of FORTRAN called DAP FORTRAN, to perform a variety of numerical simulations. However, many of the next generation of array processors, in particular the GEC Rectangular Image and Data processor (GRID), will be programmed in parallel extensions of C, like GRID-extended C. In this paper software is described which translates DAP FORTRAN into GRID-extended C, as well as FORTRAN 77 into C, enabling DAP FORTRAN programs to be run on the GRID.     

Interdisciplinary study on popularity prediction of social classified hot online events in China

We offer an interdisciplinary study of computer science and social science, analyzing behavior surrounding three types of online events: political events, social events, and non-public events. Based on the intrinsic characteristics of the three event types, this paper creates an effective method to predict such events. We continuously followed and recorded data every 10 min for 10 months from September 14, 2012 to July 11, 2013, and collected over 14 million “hot” posts from Sina Weibo, the largest microblogging provider in China. After removing spammers and noises, we developed a database of 4180 hot online events and 7,761,395 threads. We found that people’s online behavior regarding event types varies in terms of follow-up statistics and the predictability of events. The Chinese are, typically, quite concerned with social affairs that relate most closely to their personal interests and preferences. People tend to cluster around political events more often than social events and non-public events. This is demonstrated by an algorithm embedded with a clustering growth pattern of events, which predicts the popularity of online political events above others. The statistical findings are justified by Habermas’ public sphere theory and the theory of vertical/horizontal collectivism/individualism. This research provides an interesting piece of computational social science work to assist in the analysis of incentives concerning China’s collective events.     

Improved Kernel PLS-based Fault Detection Approach for Nonlinear Chemical Processes

In this paper, an improved nonlinear process fault detection method is proposed based on modified ker-nel partial least squares （KPLS）. By integrating the statistical local approach （SLA） into the KPLS framework, two new statistics are established to monitor changes in the underlying model. The new modeling strategy can avoid the Gaussian distribution assumption of KPLS. Besides, advantage of the proposed method is that the kernel latent variables can be obtained directly through the eigen value decomposition instead of the iterative calculation, which can improve the computing speed. The new method is applied to fault detection in the simulation benchmark of the Tennessee Eastman process. The simulation results show superiority on detection sensitivity and accuracy in com-parison to KPLS monitoring.