Analysis of a practical application of geodetic methods of robust estimation in a vessel positioning based on radar observations

Abstract

This paper analyses whether the application of robust estimation methods can prevent incorrect vessel positioning, thereby improving navigation safety. Empirical analyses were carried out assuming that the observational vector y comprises observations acquired from coastal radar stations operating within the “VTS Gdańsk Bay” system. Particular attention was paid to a situation in which one of observations has a gross error. Numerical tests were carried out considering several variants of the number of redundant observations. Since the significance of the subject matter in this study arises from the dynamism of position changes of a measurement item, measurements cannot be repeated at the same position. For the same reason, it is very important to choose the optimum method of working out the observations. The results of the robust adjustment of radar observations were compared with the results of the conventional least squares method.