GNSS Differential Positioning by Robust Estimation

Abstract

Robust estimation yields more accurate results than least-squares estimation in the presence of undesirable errors. It has been recently applied to the process of GNSS differential positioning in the standard fashion of iteratively reweighted least squares. We propose an improvement of the technique, based on a previous result by which robust estimation is truly robust only if considered as a global optimization problem. Robust estimation by global optimization techniques will prove to be an even more powerful method, capable of solving problematic baselines that are nonsolvable by classic methods.