Failure Scenarios to be Considered with Kinematic High Precision Relative GNSS Positioning

Andreas Wieser, Mark G. Petovello and Gérard Lachapelle

Abstract:	Failure detection, identification, and model adaptation (DIA) are essential elements of quality control. They rely directly on the formulation of failure scenarios and can only be successful if all likely failures are included. This is an extremely challenging task with kinematic positioning, because DIA requires high redundancy and the number of failures which can possibly be detected is limited. Standard procedures try to eliminate potential biases by an iterative rejection of individual observations. They often fail with real GNSS data, because failures may mask each other and may not be detectable by such tests. In view of computational burden it is hardly possible to test for arbitrary multiple blunders strictly, even if the redundancy were sufficient. We show a way to mitigate this problem by careful selection and mathematical formulation of failure scenarios. In a Kalman Filter, failures appear as individual or multiple outliers among measurements and predicted state vector. Starting with a formulation of the Kalman Filter measurement update as an extended Gauss-Markoff model, we derive specific tests for the likely failures and discuss possible limitations of common tests like the local overall model test (LOM) and the individual local slippage test. The discussion primarily aims at real-time processing of double differenced carrier phase observations in order to obtain relative positioning accuracies of 1–2 cm. However, the results are perfectly applicable also when processing other observation types e.g., pseudo-ranges or undifferenced observations, or when integrating other sensors. Data obtained from an experiment involving a computer controlled translation stage are used to demonstrate the application and performance of the previously derived analysis tools. The investigation is essential for all kinematic high precision positioning tasks, where testing for individual outliers on the epoch level is not strong enough to detect multiple blunders and realistic contamination of the data.
Published in:	Proceedings of the 17th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS 2004) September 21 - 24, 2004 Long Beach Convention Center Long Beach, CA
Pages:	1448 - 1459
Cite this article:	Wieser, Andreas, Petovello, Mark G., Lachapelle, Gérard, "Failure Scenarios to be Considered with Kinematic High Precision Relative GNSS Positioning," Proceedings of the 17th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS 2004), Long Beach, CA, September 2004, pp. 1448-1459.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In