Robust and Reliable Kalman Filtering of GPS Data

Ahmed H. Mohamed

Abstract:	This paper discusses the use of an Ll-Norm blunder detector within the Kalman filtering of GPS static single point positioning data, using a data snooping technique. The motivation for using this approach stems from the fact that conventional Least-Squares, though reliable, tends to smooth blunders into the whole set of observations, good and bad. On the other hand, blunders have little effect on robust estimation procedures based on the Ll-Norm, where the sum of the absolute errors is minimized for an optimal robust solution. Unfortunately, the Ll-Norm robust estimate is not as reliable as the Least-Squares estimate. To obtain both benefits, robustness and reliability, the Ll-Norm procedure is used within a Least- Squares Kalman filtering procedure as a blunder detector regime.
Published in:	Proceedings of the 9th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GPS 1996) September 17 - 20, 1996 Kansas City, MO
Pages:	1441 - 1450
Cite this article:	Mohamed, Ahmed H., "Robust and Reliable Kalman Filtering of GPS Data," Proceedings of the 9th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GPS 1996), Kansas City, MO, September 1996, pp. 1441-1450.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In