Evaluation of Several Double-differencing Strategies for Reduction of Unmodeled Correlated Errors in Carrier Phase GNSS Processing

K. O'Keefe, K. Johnson, and P. Alves

Abstract:	"In this paper, a result from early carrier phase GPS research, that all double differencing schemes lead to equivalent results, is investigated in the context of the GPS industry standard of using the highest elevation satellite as the base satellite. The paper begins with a discussion of double differencing. Five commonly stated reasons for using the highest elevation satellite are then discussed and discarded and a simplified version of an earlier proof of the equivalence of double differencing schemes is developed. Four strategies for double-differencing are tested in kinematic mode using a very short GPS baseline. Identical float position and velocity solutions are obtained confirming the theoretical result. Equivalent float ambiguity solutions are also obtained. The Lambda algorithm is used to obtain equivalent fixed solutions. The fact that the float ambiguities are equivalent, but not identical is discussed and it is then suggested that the optimal differencing strategy, though all provide identical results, is the one that leads to the least correlation between float ambiguities. If the Lambda algorithm is used, it finds the same final linear combination of ambiguities in all cases, but when it begins with less correlated ambiguities, fewer iterations of its permutation and decorrelation function are required."
Published in:	Proceedings of the 61st Annual Meeting of The Institute of Navigation (2005) June 27 - 29, 2005 Royal Sonesta Hotel Cambridge, MA
Pages:	917 - 927
Cite this article:	O'Keefe, K., Johnson, K., Alves, P., "Evaluation of Several Double-differencing Strategies for Reduction of Unmodeled Correlated Errors in Carrier Phase GNSS Processing," Proceedings of the 61st Annual Meeting of The Institute of Navigation (2005), Cambridge, MA, June 2005, pp. 917-927.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In