基于神经网络的仿真输出数据的分布识别

分布识别在仿真输出数据的分析中有着重要的理论和实际价值，但长期以来一直没有很好解决，在研究和改进神经网络的基础上，提出用多层感知机ＢＰ网格进行分布识别，给出了基于ＮＮ的分布识别模型，并进行了实验检验，实验结果表明本方法识别效率高，并具有很高的可靠性，可以辅助分析者作出准确的判断。     

光纤捷联惯组四点减振隔振模式分析研究

光纤陀螺捷联惯组减振设计中,经常采用平面四点减振隔振模式。本文结合减振系统对角振动固有频率的特殊要求,建立减振系统动力学模型,对角线振动固有频率比进行分析,通过改变惯组本体组件的质量分布提高角线振动固有频率比,可为光纤捷联惯组减振设计提供一定的指导。     

An experimental method for squealer tip flow field considering relative casing motion

Squealer tip is widely used in turbines to reduce tip leakage loss. In typical turbine environment, the squealer tip leakage flow is affected by multiple factors such as the relative casing motion and the wide range of variable incidence angles. The development of experimental methods which can accurately model the real turbine environment and influencing factors is of great significance to study the squealer tip leakage flow mechanism. In the present paper, a low-speed turbine cascade test facility which can model the relative casing motion and wide range of variable incidence angles (−25° to 55°) is built. Based on the similarity criteria, a high-low speed similarity transformation method of the turbine cascade is established by considering the thickness of the turbine blade. A combined testing method of Particle Image Velocimetry (PIV) and local pressure measurement is proposed to obtain the complex flow structures within the tip cavity. The results show that the experimental method can successfully model the relative casing motion and the wide range of variable incidence angles. The low-speed cascade obtained by the similarity transformation can model the high-speed flow accurately. The measurement technique developed can obtain the complex flow field and successfully capture the scraping vortex within the squealer tip.     

A star identification algorithm based on radial and dynamic cyclic features of star pattern

A full-sky star identification algorithm based on radial and dynamic cyclic patterns is presented with the aim of solving the “lost-in-space” problem. The dynamic cyclic pattern match is applied with a maximum cumulate comparison method to identify sensor-catalog pairings in initial match, which substantially eliminates the effects of the star position noise, magnitude noise, and false stars. After initial match pairings of stars are obtained, a chain part extension technique is employed to quickly search for the longest match chain as the final result. Experimental results indicate that the proposed algorithm is highly robust to star position noise, magnitude noise and false stars. In a series of simulations, the identification rate of the algorithm is 97.50% with 2.0 pixels star position noise, 96.90% with 0.4 Mv star magnitude noise and 95.30% with four false stars respectively. Moreover, the algorithm achieves an identification rate of 58.08% when only six stars are in the field of view.     

证言广告法律制度立法模式探析

在借鉴国外立法之经验并结合我国实际国情,提出对我国证言广告立法模式的构想。     

设计模式探析及其在测控软件中的应用

针对测控领域中复用程度低和开发周期长的问题,提出在该领域软件进一步推广和使用设计模式的思路。分析当前测控软件在实际开发过程中遇到的突出症结,并结合具体工程实践描述在测控软件领域引入设计模式的作用和影响,以单态模式为例分析引入设计模式对软件体系架构造成的深刻改变和潜在风险,给出测控软件领域中运用设计模式的建议。通过分析不同应用场景表明,灵活运用设计模式能够解决航天测控软件中常见的突出问题。     

基于聚类的多平台协同数据关联算法研究

从多平台协同工作的角度出发,针对目标和传感器观测数据之间的关联问题进行了研究。给出了模糊聚类目标函数、FCM算法步骤和量测目标归属判定方法,基于此,提出了模糊C-均值聚类和量测目标归属判定算法,然后提出了基于聚类算法的多平台协同多目标数据关联算法。仿真实验结果表明,该算法数据关联正确率较高,能够提升跟踪精度。     

三体问题地面实验的相似性分析

研究了三体问题相似性的分析流程,给出了三体问题实验近似相似性程度的公式化度量方法。首先,明确了三体问题地面实验的构建原理与方法;其次,建立圆形限制性三体问题的动力学模型,并利用相似性分析方法建立三体问题的相似性度量函数;再次,建立存在各种干扰因素情况下实验系统动力学模型,并利用相似性度量函数衡量在干扰因素作用下实验满足相似性的程度;最后,利用数值仿真法进行仿真分析从统计的角度得出各种干扰对于三体问题实验相似性的影响大小,并给出各种干扰因素对于三体问题相似性的影响排序。仿真表明这种分析方法对于实验系统的提高与实验方案的选择具有指导意义。     

空间非编码标记特征点中心定位方法与比较

归纳总结了应用于可见光测量的空间标记非编码设计原则,例举了棋盘格型、栅格型、圆型等3种典型的非编码标记图案设计,分别阐述了各标记特征点中心定位方法和计算过程,开展了对比试验,结果表明：圆型非编码标记图案设计及其特征点中心定位精度明显高于其他两种图案,其图案设计简捷、特征点突出易于识别、在像平面中呈现各向同性、点中心定位精度高、计算速度快、噪声抑制能力强。