Separating Vertical Deflection from Attitude Errors by Proper Gravity Modeling in a Tightly Integrated GPS/INS System

Jin Wang

Abstract:	It is commonly known that attitude errors, the vertical deflection (DOV), as well as accelerometer errors, act on an inertial system concurrently and it is possible to distinguish their effects only in a certain spectral range with the aid of GPS positioning. Therefore, the separation is largely dependent on how well we model the gravity disturbance, assuming the model and the parameters of inertial sensors given by the manufacturer are accurate. In a high accuracy GPS/INS system, such as in Airborne Integrated Mapping System (AIMS), effectively separating the DOV from orientation errors will definitely be of great benefit. From a theoretical point of view, a higher order Gauss-Markov (GM) model would better approximate the gravity field. The objective of this paper is to compare the system performance with different GM models in practical settings. The DOV data obtained from a three-dimensional gridded (2’´2’) database was used as a gravity reference, as well as inputs to the system, for the purpose of comparison. The test results have shown that the DOV can be extracted to an accuracy of less than 1.5 arcsec with both 1 st and 3 rd GM gravity models, while the orientation difference is at the level of a few arcsec as compared to the situation with known DOV data inputs. The 1 st GM model works somewhat better than the 3 rd GM model. The reason for this is explained in this paper.
Published in:	Proceedings of the 1999 National Technical Meeting of The Institute of Navigation January 25 - 27, 1999 Catamaran Resort Hotel San Diego, CA
Pages:	709 - 716
Cite this article:	Wang, Jin, "Separating Vertical Deflection from Attitude Errors by Proper Gravity Modeling in a Tightly Integrated GPS/INS System," Proceedings of the 1999 National Technical Meeting of The Institute of Navigation, San Diego, CA, January 1999, pp. 709-716.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In