TRANSFORMATION OF COORDINATES WITH ROBUST M-ESTIMATION AND MODIFIED HAUSBRANDT CORRECTION

Transformation of coordinates allows to convert coordinates from one geodetic system to another. Determination of transformation parameters is performed by the least-squares method. Unfortunately, the least squares method isn't immune to outliers. It means that if from any reason one or more of the reference point's coordinate is not correct the transformation parameters will be estimated with this errors. Therefore there is a necessity to develop a method that will be immune to outliers. In this paper robust estimation method for coordinate transformation is proposed to complete this task. To avoid influence of blunders in coordinates of reference points, three types of robust estimation were analyzed. But to assure that reference point's coordinates will remain unchanged after transformation, one needs to apply the Hausbrandt correction. The Hausbrandt correction algorithm also isn't immune to outliers. So it has to be modified to ensure it's robustness to a blunders. The results of coordinate transformation with robust estimation and modified Hausbrandt correction were compared with Helmert transformation with Hausbrandt correction.