非线性最小二乘参数平差

在分析线性化的非线性参数平差的近似直接解法与近代解法基础上,利用非线性最小二乘的精确正交性条件方程,推导出顾及到二次项的非线性参数平差的一种新的计算公式。为了公式表达的简洁,精度评定的需要以及非线性平差本质的分析,在公式推导过程及最终的平差计算公式中,引入了矩阵分解及非线性度量曲率矩阵。最终算例分析说明该方法的有效性及实用性。     

非线性普通秩亏自由网平差

在文献〔1〕的非线性普通秩亏自由网平差模型强度的曲率度量基础上,利用非线性最小二乘的精确正交性条件方程,提出两种顾及二阶偏导数的非线性普通秩亏自由网平差的计算方法。由于非线性度量曲率立体阵的引入,使得平差计算公式表达简洁,且便于非线性平差本质的分析。最终算例分析说明该方法的有效性及实用性     

非线性条件平差的迭代解法

本文通过具体的算例,探讨了按附有限制条件的间接平差法求解非线性条件平差问题,详细地论述了非线性条件平差的迭代解法。     

非线性条件平差的迭代解法

本文通过具体的算例,探讨了按附有限制条件的间拉平产左法求解非线性条件平差问题,详细地论述了非一条件平差的迭代解法。     

非线性与线性平差偏差的分布特征

测量中的平差模型大都是非线性型的,实际处理常使其线性化,进行线性模型平差,平差值与其理论值必然存在差异。文中对非线性函数作任何逼近时所产生的参数偏差的概率分布特征进行讨论,导出了函数偏差与平差参数偏差的期望和方差公式,它们是进一步研究非线性平差统计分析的基础。     

测角网坐标平差模型的Bootstrap参数估计方法

针对线性最小二乘法处理非线性模型产生模型误差的问题,文章将高斯牛顿迭代法引入测角网坐标平差模型中,给出测角网坐标平差模型的高斯牛顿迭代法计算过程.考虑到非线性平差模型的参数估计值是有偏估计,结合Bootstrap重采样方法对参数估值进行改善,提出测角网坐标平差模型的Bootstrap参数估计方法,并给出详细的迭代流程图.针对等精度与不等精度角度观测数据,设计两个测角网案例.实验结果表明,测角网坐标平差模型的高斯牛顿迭代解法能够削弱线性近似带来的模型误差影响,其参数估值优于传统的线性近似方法;而测角网坐标平差模型的Bootstrap参数估计方法比高斯牛顿迭代解法在提高测角网坐标平差参数估值质量方面更加有效.实验证明将高斯牛顿迭代解法应用于测角网坐标平差模型的必要性与实用性,也证明将Bootstrap重采样参数估计方法与高斯牛顿迭代解法结合并用于测角网坐标平差的可行性与有效性.     

广义逆矩阵与测量平差

在介绍带权广义逆基础上,分析和雄导了间接平差、条件平差、附有条件的间接平差、附有参数的条件平差、秩亏自由网平差和最小二乘配置等计算公式。     

实数编码自适应遗传算法辅助实现变形监测网平差

变形监测网的平差问题实际上是一种带约束条件的函数优化问题。本文从附加带权基准方程的变形监测网平差统一模型入手,探讨了实数编码遗传算法在变形监测网非线性参数平差应用中的有效性问题:针对平差基准的等式约束条件,通过等式变换对平差参数进行降维处理,达到了约束优化问题向无约束优化问题的转变,实现了不同基准条件下变形监测网的非线性遗传算法参数平差。     

非线性整体最小平差迭代算法

整体最小二乘法不仅考虑观测向量的误差而且还考虑系数矩阵的误差,平差理论相对更为严密。在研究经典整体最小二乘法的基础之上,对系数矩阵元素是表达式或函数情况的非线性整体最小二乘模型进行了描述,用拉格朗日极值条件式推导了基于牛顿型解法的非线性整体最小二乘平差计算公式,并设计了一种对应的迭代算法。最后设计了两组模拟试验分析在观测向量和系数矩阵的输入向量等精度观测和非等精度观测两种情况下参数和验后方差的估计特点。试验结果表明,非线性整体最小二乘平差法获得的参数估计值比最小二乘平差法获得的估计结果更接近参数的实际值,方差分量(或中误差)估计结果也更接近先验值,本文给出的迭代算法是有效的。     

经典平差模型的通用模型

在对测量数据进行处理时,常用的严密平差模型有条件平差、间接平差,针对不同的测量网型和不同的需求会选用不同的平差模型,因此解法不同,同时增加了一定的计算量。为了以后方便计算并且能够使用同一种模型,本文提出这几种模型共用的一种模型,附有参数的条件平差模型,此种模型在不同的测量网下都可使用,解法简单单一,只需根据实际情况列出条件方程进行计算。最后再用具体实例计算验证,表明这一模型可用。     

广义非线性最小二乘测量参数平差的快速差分迭代解算

针对数字化建设中涉及的多类型、多精度、多源、动态非线性最小二乘测量参数平差问题,提出了一种新的解算模型。该模型通过构造适当的差分点列,避免了一阶或二阶导数计算,对于模型复杂或者导数不存在的函数模型而言,具有适应面广、算法简单等特点,为广义非线性数据处理中的参数估计问题的解决开拓了又一新的思路。     

A closed-form of Newton method for solving over-determined pseudo-distance equations

The Newton method has been widely used for solving nonlinear least-squares problem. In geodetic adjustment, one would prefer to use the Gauss–Newton method because of the parallel with linear least-squares problem. However, it is proved in theory as well as in practice that the Gauss–Newton method has slow convergence rate and low success rate. In this paper, the over-determined pseudo-distance equations are solved by nonlinear methods. At first, the convergence of decent methods is discussed after introducing the conditional equation of nonlinear least squares. Then, a compacted form of the Hessian matrix from the second partial derivates of the pseudo-distance equations is given, and a closed-form of Newton method is presented using the compacted Hessian matrix to save the computation and storage required by Newton method. At last, some numerical examples to investigate the convergence and success rate of the proposed method are designed and performed. The performance of the closed-form of Newton method is compared with the Gauss–Newton method as well as the regularization method. The results show that the closed-form of Newton method has good performances even for dealing with ill-posed problems while a great amount of computation is saved.     

Reference frame stability and nonlinear distortion in minimum-constrained network adjustment

The aim of this paper is to investigate the influence of the minimum constraints (MCs) on the reference frame parameters in a free-net solution. The non-estimable part of these parameters (which is not reduced by the available data) is analysed in terms of its stability under a numerical perturbation of the constrained datum functionals. In practice, such a perturbation can be ascribed either to hidden errors in the known coordinates/velocities that participate in the MCs or to a simple change of their a priori values due to a datum switch on a different fiducial dataset. In addition, a perturbation of this type may cause a nonlinear variation to the estimable part of the reference frame parameters, since it theoretically affects the adjusted observations that are implied by the network’s nonlinear observational model. The aforementioned effects have an impact on the quality of a terrestrial reference frame (TRF) that is established via a minimum-constrained adjustment, and our study shows that they are both controlled through a characteristic matrix which is inherently linked to the MC system (the so-called TRF stability matrix). The structure of this matrix depends on the network’s spatial configuration and the ‘geometry’ of the datum constraints, while its main role is the filtering of any MC-related errors into the least-squares adjustment results. A number of examples with different types of geodetic networks are also presented to demonstrate the theoretical findings of our study.     

基于非线性平差模型的坐标转换公式

用线性近似法以及非线性参数估计法讨论了直角坐标系七参数转换模型，指出改进的Gauss-Newton方法具有理论严密、计算简洁、易于编程、精度较高等持点。对于处理类似坐标转换的非线性模型具有重要的理论和实践意义。     

伪距单点定位两域联合平差解的等价性分析

针线性观测模型观测值域联合平差解和平差值域联合平差解完全等价,但实际中联合平差的观测模型多为非线性模型,参数估计时常常需要进行线性化处理,存在线性化引起的模型误差,该误差对两域平差解的等价性影响值得探讨. 以伪距单点定位(SPP)为例,论文推导了其平差值域联合平差求解公式,分析了两域联合平差不能完全等价的原因,并用实例进行了数值对比. 研究表明:在配置相同的情况下,不同全球卫星导航系统(GNSS)组合SPP的两域平差位置解在毫米量级上数值相同,就SPP应用而言,两域平差解可认为相同. 论文内容对于一些GNSS其他应用也具有一定的参考意义.      

基于线性模型的二维直角坐标变换

二维直角坐标变换模型在工程领域中应用非常广泛,该模型是一个非线性模型,对于多公共点的二维直角坐标变换需要进行模型的线性化,计算待求参数的近似值,再利用间接平差原理计算各待求参数最或然值。本文在非线性模型基础上推导出一个线性模型,利用该模型进行二维直角坐标变换,无需计算待求参数近似值,对于涉及到二维直角坐标变换的问题,可以简化其解算过程。     

非线性最小二乘参数平差迭代算法

在非线性最小二乘问题现有的3类主要算法－－高斯－牛顿法、阻尼最小二乘法和最小二乘的拟牛顿法的基础上，引入了综合性能更优的非线性规划的SQPM（序列二次规划法）算法，并且为进一步提高SQPM算法迭代的收敛性，对其步长策略进行了改进。改进的SQPM算法成为无需精确计算参数概略值的非线性最小二乘参数平差的实用和有效算法。     

An iterative solution of weighted total least-squares adjustment

Total least-squares (TLS) adjustment is used to estimate the parameters in the errors-in-variables (EIV) model. However, its exact solution is rather complicated, and the accuracies of estimated parameters are too difficult to analytically compute. Since the EIV model is essentially a non-linear model, it can be solved according to the theory of non-linear least-squares adjustment. In this contribution, we will propose an iterative method of weighted TLS (WTLS) adjustment to solve EIV model based on Newton–Gauss approach of non-linear weighted least-squares (WLS) adjustment. Then the WLS solution to linearly approximated EIV model is derived and its discrepancy is investigated by comparing with WTLS solution. In addition, a numerical method is developed to compute the unbiased variance component estimate and the covariance matrix of the WTLS estimates. Finally, the real and simulation experiments are implemented to demonstrate the performance and efficiency of the presented iterative method and its linearly approximated version as well as the numerical method. The results show that the proposed iterative method can obtain such good solution as WTLS solution of Schaffrin and Wieser (J Geod 82:415–421, 2008) and the presented numerical method can be reasonably applied to evaluate the accuracy of WTLS solution.     

Generalyzed variance-covariance propagation law formulae and application to explicit least-squares adjustments

Standard formulae overlook the contribution of a number of terms in the derivation of variance-covariance matrices for parameters in nonlinear least squares adjustment. In a large class of nonlinear mathematical models, these terms can contribute to an important error in the estimation of parameter variances. Improved formulae are derived. A numerical example is given and the use of our improved formula in the case of least-squares adjustment in the explicit case (L=F(X)) is fully documented.     

半参数回归与模型精化

就一般情况给出了半参数平差的算法，并结合一种特定的情况，讨论了正规化矩阵半正定时的计算方法，给出了相应的公式，最后构造了一个模拟的平差问题，对半参数法和最小二乘法的计算结果进行了比较，计算表明，半参数法能够发现并识别模型误差或观测值中的系统误差。