基于随机模型基准的Helmert方差分量估计

在多类观测数据联合平差中,存在某类观测值的验前协方差阵正确而其他类不正确的情况,传统Helmert方差分量估计在求解此类问题时没有对正确的验前协方差阵加以区别,从而造成正确的随机模型经估计后同样也被调整。针对上述情况,首先分析了Helmert方差分量估计迭代收敛结果的实质,然后提出了随机模型基准的概念,并推导了基于随机模型基准的Helmert方差分量估计公式。经计算表明,新公式完全可行,可以用于解决实际问题。

Helmert Type for Estimating Variance Components Based on Datum of Stochastic Model

Concerning the combined adjustment of heterogeneous types of observations,one case maybe occur: prior covariance matrix of some type of observations is correct,while that of others are wrong.But in the traditional formulae of Helmert type for estimating variance components,these prior covariance matixes have not been differentiated,so the correct stochastic models have been adjusted.In order to solve this case,the concept of datum of stochastic model is proposed,and furthermore,new formulae for estimating variance component of Helmert type are derived.One numerical example is reported in order to demonstrate the validity of new formulae when being applied to actual combined adjustment.