Fault reconstruction for Takagi–Sugeno fuzzy systems via learning observers

This paper addresses the problem of observer-based fault reconstruction for Takagi–Sugeno fuzzy systems. Two types of fuzzy learning observers are constructed to achieve simultaneous reconstruction of system states and actuator faults. Stability and convergence of the proposed observers are proved using Lyapunov stability theory, and necessary conditions for the existence of the observers are further discussed. The design of fuzzy learning observers can be formulated in terms of a series of linear matrix inequalities that can be conveniently solved using convex optimisation technique. A single-link flexible manipulator is employed to verify the effectiveness of the proposed fault-reconstructing approaches.     

接触式多关节三维扫描仪研究

论述了拥有自主知识产权的接触式多关节三维扫描仪3DLCS-400的硬件系统和软件系统的基本构成、主要特点和实现方法。在此基础上,研究了接触式三维信息获取中的一些关键技术———数据采集与三维显示的同步、坐标计算和误差校正、三维构型重建。采用基于API函数的异步多线程串行通信和OpenGL三维显示技术,提出了一种修正最小二乘法和基因可变限改进遗传算法相结合的参数标定方法和运用轮廓线的三维散乱数据构型方法。该产品的研制成功填补了国产三维信息获取系统的一个空白。     

RELATIONSHIP BETWEEN TRUE AND ESTIMATED POSSIBILISTIC SYSTEMS AND THEIR RECONSTRUCTIONS†

This paper is a theoretical contribution to reconstructability analysis of possibilistic systems. A method to estimate a given possibility distribution from the associated observations is derived based on the one-to-one correspondence between possibility distributions and their basic probability assignments. It is shown that in this estimation the possibility value for a state is proportional to the frequency of the state in the observations which are drawn from nested state subsets. The order preservation relationship between a given (true) possibility distribution and its estimation is derived. Based on this preservation property, it is shown that the order in the distance from an original possibilistic overall system of their unbiased reconstructions from two alternative reconstruction hypotheses (structures) is preserved in the estimated systems domain. Particularly, it is proved that systems conceptualized in terms of possibility theory possess an important property: if a possibilistic system is perfectly reconstructable from a specific reconstruction hypothesis, then a system estimated from any data produced by the former system is also perfectly reconstructable from the same hypothesis.     

基于康普顿散射的能谱数据重建电子密度图像

提出一种基于能谱和行扫描的康普顿散射成像方法。该方法通过逐行扫描，把一个较大的图像重建问题分解成几个独立的小而容易解决的问题进行重建。对于正向方程中源于非扫描行的衰减作用，利用已重建行的像素值来计算；对源于扫描行自身的衰减作用，用逐步近似(SAP)的方法来计算。为了减小说差在不同扫描行之间的传播，利用途射数据对重建结果进行校正。实验证明，该方法速度快、可靠性高，易于实现。     

Fault Reconstruction for Continuous‐Time Systems Via Learning Observers

This paper addresses the problem of learning observer (LO)‐based fault reconstruction in continuous‐time systems. An LO is constructed such that simultaneous reconstruction of system states and actuator faults can be guaranteed. Then, existence conditions of LOs are first introduced to verify whether or not the LOs for fault reconstruction exist. An effective LO‐design algorithm is explored such that observer gain matrices can be conveniently solved using linear matrix inequality technique. An extension to sensor‐fault reconstruction is also investigated. Finally, two simulation examples are employed to demonstrate the effectiveness of the proposed LO‐based fault‐reconstructing approaches.     

一种三维准欧氏重建方法

研究在不必精确知道相机内参数的情况下进行场景的三维重建，充分利用建筑场景的平面信息，假设场景由一些基本平面元素组成，从而自动生成场景的拓扑结构．虽然重建结果从严格意义上仍是射影重建，但已经非常接近欧氏重建结果，文中称之为准欧氏重建，该结果还可以为其它优化算法提供很好的初值．真实图像的实验结果证实了本文算法，重建模型达到了很好的可视效果．     

逆向工程中基于语义的测量信息传递方法研究

针对逆向工程中曲面数字化与模型重建的信息交互方式单一,一些对后续建模有用的辅助信息（隐式信息）流失（目前的数据文档格式无法保存与传递）,造成后续模型重建困难的问题,提出一种基于语义的测量信息传递方法,分析测量过程中有利于模型重建的隐式信息,建立语义信息模型,采用类IGES数据格式存储语义信息,阐述语义信息的封装、解析方法以及基于语义的数据预处理、模型重建过程。实验研究表明,基于语义的测量信息传递方法能够有效提高CAD模型重建效率。     

几何缺席推理研究

提出了实体几何缺席推理理论，研究了相应的三维实体重构算法，并进行了实例验证。     

基于未知输入观测器的非线性广义系统执行器故障检测和重构

针对一类含有未知扰动广义非线性系统的执行器故障,本文提出一种重构算法。首先设计未知输入观测器对干扰鲁棒,作为故障检测观测器。检测到发生故障后,通过提出含有误差比例项和积分项的故障估计算法,形成自适应观测器,实现准确快速地估计故障,同时估计状态变量。根据李雅普诺夫稳定理论给出估计误差一致最终有界的充分条件。最后仿真验证该类观测器和重构算法的有效性。     

基于多项式重构的拒绝服务反向追踪方案

提出了利用代数方法反向追踪信息的方案,该方案把追踪重构问题当作多项式重构问题,使用代数编码理论技术提供鲁棒的传送和重构方法,是DoS攻击过程中的反向追踪问题的一种新的解决方法。     

基于洛伦兹范数的医学图像超分辨率重建研究

在超分辨率图像重建(SR)模型中,为了达到良好的重建效果,选择一个合适的代价函数是研究的重点。采用SR重建模型中的差错项选择了洛伦兹范数,正则化项选择了吉洪诺夫正则化,重建过程采用了迭代方法。提出的算法可以有效地解决医学图像SR重建过程中的去异值点和图像边缘保持的两大关键问题,达到良好的重建效果。为了验证上述算法的有效性,就一系列添加了运动模糊和不同噪声的低分辨率MRI医学图像进行了SR重建,并且与基于L2范数的重建算法的重建效果进行了比较分析。实验结果显示,所提算法具有良好的实用性和有效性。     

融合诊断中信息容错性的证据重构方法

在基于证据理论的故障融合诊断过程中,错误的待融合信息会引起故障的漏诊.针对此类容错性问题,提出了一种证据重构方法,在已有的故障诊断识别框架的基础上,利用多传感器信号的方差计算信号的相对可靠程度,构造新的信度函数,对待融合证据进行了重新分配,以减少出错信息对融合结果的影响,同时对该方法中可调参数的选取原则进行了理论分析.融合诊断实验在隔振器硬件实验平台上进行,诊断对象为振动发散故障,针对的信息容错性问题为基础加速度信号断路情况.对比性实验及结果分析表明,采用该方法能够在单路信号传输出错的情况下及时诊断振动发散故障,增加了融合诊断系统的容错性.     

相空间重构优化研究

分析了传统相空间重构方法的不足,提出了相空间重构优化的标准和重构质量的评价方法.由于相空间重构过程中,重构参数的改变直接导致了吸引子的伸展和折叠行为.并形成相应的空间结构.通过无序度可以有效地刻画重构吸引子的不同性态,当无序度最小时可保证吸引子具有最规则的结构.重构质量的评价主要是通过检测吸引子拓扑结构的稳定性来实现,这比其他检验方法更具有客观性.仿真实验结果表明该方法的可行性.     

基于压缩感知的遥感成像稀疏重构性能分析

压缩感知是一种新型的信息论，打破了传统的Shannon-Nyquist采样定理，能够以少量数据完成信号采样。稀疏重构是压缩感知由理论到实际的关键环节，为了将压缩感知有效地应用于遥感成像领域，研究了稀疏重构对遥感成像过程的影响。针对稀疏重构理论模型，分析了重构误差的成因；同时，针对典型的凸优化类算法和贪婪类算法，利用峰值信噪比指标对遥感图像重构误差进行评价。在仿真实验中，定量考察遥感图像在不同压缩采样率、不同重构算法下的稀疏重构性能。结果表明，稀疏重构算法能够成功重构遥感图像，各算法在不同压缩采样率下均表现出了较好的重构质量，整体上能够满足遥感成像应用，验证了压缩感知稀疏重构方法在遥感成像中应用的可行性。     

A GREEDY ALGORITHM FOR A GENERALIZATION OF THE RECONSTRUCTION PROBLEM

The aim of the reconstruction problem is to determine from among all relevant structure systems those which allow us to reconstruct a given overall system to an acceptable degree of approximation. This paper considers a more general form of the reconstruction problem. Instead of restricting the reconstruction to structure systems, substates in general are allowed in the reconstruction. Furthermore, the overall system need not be known in its entirety to formulate a reconstruction. A very effective and efficient algorithm for this generalized reconstruction problem is presented.     

直接三维重建研究

本文提出了一个基于多结点插值样条和有限级数展开法的直接三维图像重建方法。该重建算法可以充分利用被重建物体的先验三维信息。对于减少稀疏系数矩阵的维数问题，本文也作了研究，实验结果证明了本方法是可行的。     

Innovation approach to reduced-order estimation of complementary states.

A reduced-order filler algorithm is developed for estimation of the complementary states of a linear system driven by white noise. The estimator gives an unbiased estimate of the states, and the error in the estimate is orthogonal to the subspace of the observations of those states that are common to the observations and the estimated states, The filter performance is suboptimal relative to the full-order optimal linear filter, but benefits are reaped from computational savings in the filtering and associated Riccati equations. The algorithms are developed using a concept of the ‘reduced-order innovation process’. The design uses only a priori information for z, the states to be estimated. The estimator is shown to be unique. The requirements of the stability of the filter are also presented, and its determination can precede its implementation. The basic limitation of this approach is that one is constrained by the minimum order of the filter, because of the assumption in (2.11) that the states to be estimated are complementary to the observed states.     

多尺度Retinex算法在简牍图像增强中的应用

简牍图像增强是简牍文字修复和文字信息提取最为关键的一个步骤.以长沙简牍博物馆收藏的简牍文物为实物模型,以该馆提供的简牍数字图像为研究对象,提出了基于多尺度Retinex(MSR)算法的简牍图像增强方法,论述了本算法的原理及实现方法.采用MSR算法来消除光照对简牍图像的退化影响,以达到增强图像暗区细节信息的目的;同时针对MSR输出图像偏暗,提出了直方图拉伸算法.试验结果表明:本算法能够提高简牍图像对比度的同时,能显著地增强图像暗区的细节信息并且使简牍图像具有较好、较高的色彩保真度.本方法克服了目前常规的图像增强方法的不足,能够满足简牍考古人员的要求.     

线性逼近获取深度图象算法研究

三维表面重建是计算机视觉领域中的一个重要课题，也是计算机视觉中最困难的问题之一。该文提出了一种用线性逼近方法、基于灰度信息从图像获取物体深度信息，从而实现物体三维表面重建的算法。该算法原理简单，容易实现，速度快，具有较高的实用价值。