附有相对权比的总体最小二乘平差

Total Least Squares Adjustment with Weight Scaling Factor

  • 摘要: 推导了加权情况下附有相对权比的总体最小二乘平差方法,提出了确定相对权比的验前单位权方差法和目标函数最小化法。模拟算例表明,当观测值和系数矩阵的验前单位权方差已知且比较准确时,验前单位权方差法得到的结果与参数真值的差值范数最小;目标函数最小化法的目标函数估值最小,与参数真值的差别比验前单位权方差法的结果稍大。

     

    Abstract: The weighted total least squares adjustment with weight scaling factor is deduced.The prior unit weight variance method and minimum objective function method are proposed to determine weight scaling factor.Though the simulated example,some conclusions are obtained.When the prior unit weight variances of observation and coefficient matrix are known and accurate,the difference norm between the true value and estimation from the prior unit weight variance method is smallest of all;the estimation of the objective function by minimum objective function method is smallest,while the difference between true value and parameter estimation is larger than the result using the prior unit weight variance method.

     

/

返回文章
返回