摘要
本文对观测值误差的配赋情况进行了研究,提出了一种统计量,以此统计量进行选权迭代,具有很强的剔除粗差能力。本文还对动态网的抗差问题做了研究;利用Masreliez滤波和极大验后MAP滤波,推出了一套适于动态大地网和工程变形网的抗差滤波模型。算例充分体现了上述几种抗差方法的可行性和有效性。
This paper deals with two problems. Firstly, the adjustment of observation errors to the residual vector V was re serehed, a new robust estimation method that based on the W--estimation was Porposed. The experiment shown that the method is efficient. Secondly, Non--Gaussian filtering with the linear state and observation relation in the dynamic networks was reserched. Used the Marseliez filtering and MAP filtering, two robust models that are suitable for dynamic control networks and deformation observation networks were ascertained. The feasiblity and the advantages of these methods were affirmed by the examples.
出处
《测绘科学技术学报》
1992年第4期17-23,共7页
Journal of Geomatics Science and Technology
关键词
抗差估计
抗差滤波
平差
Robust estimation, Robust filtering, Adjustment
