Using cubic splines to mitigate systematic errors in GPS relative positioning

Abstract

Systematic errors can have a significant effect on GPS observable. In medium and long baselines the major systematic error source are the ionosphere and troposphere refraction and the GPS satellites orbit errors. But, in short baselines, the multipath is more relevant. These errors degrade the accuracy of the positioning accomplished by GPS. So, this is a critical problem for high precision GPS positioning applications. Recently, a method has been suggested to mitigate these errors: the semiparametric model and the penalised least squares technique. It uses a natural cubic spline to model the errors as a function which varies smoothly in time. The systematic errors functions, ambiguities and station coordinates, are estimated simultaneously. As a result, the ambiguities and the station coordinates are estimated with better reliability and accuracy than the conventional least square method.