Comparative analysis of different manners of execution of Huber’s robust estimation
AbstractObservations made in geodetic network must be processed in order to find the parameters that we search for, such as coordinates of points or their displacement. In practice, the most popular method used for this purpose is the least squares adjustment. However, according to research, it is susceptible to gross errors in observations. In recent years, the so-called robust estimation methods have gained great popularity. In these methods, the impact of gross errors is eliminated by iterative modification of observation weights. One of these methods is Huber’s robust estimation, in which the objective function is a spline. For observations with residuals within the assumed range, the component of the objective function is identical to objective function in the method of least squares. However, for the remaining observations the component of the objective function is similar to that in the least absolute deviations method. In literature, various ways of implementation of Huber’s method have been described. The study presents results of numerical analyses of effectiveness of Huber’s method. Various approaches to weight modification, described in literature, have been applied. Conclusions on research conducted have been presented.
|Journal series||Reports on Geodesy, ISSN 0867-3179, , (0 pkt)|
* presented citation count is obtained through Internet information analysis and it is close to the number calculated by the Publish or Perish system.