An Efficient UKF Based GPS/INS Augmented by Local Landmark Update

Rui Hirokawa and Kenji Nakakuki

Abstract:	In this paper, a highly precise and robust GPS/INS for ground vehicle is designed and evaluated. The proposed GPS/INS is based on a dual frequency GPS, a rate gyro sensor and an odometer. To overcome the performance degradation on GPS outage, the local landmark update based on a laser scanner measurement is also used. The conventional GPS/INS is based on the extended Kalman filter (EKF) using the first order linearization. The un-modeled higher order dynamics could cause the estimation error on the average and covariance of the EKF. In this paper, Unscented Kalman Filters (UKF) having more than second order accuracy is applied for the GPS/INS with local landmark update. As a preliminary study, a simplified analysis is conducted to clarify the dierence between the EKF and the UKF. In the analysis, long distance range measurements of GPS and short range and angle measurements of local landmarks are assumed. The eect of nonlinearity associated with the measurements in those filters are shown in the analysis. Then, the GPS/INS for a unmanned ground vehicle is designed and evaluated by the numerical simulations. The robustness of the system is evaluated in some partialsatellite- blockage cases having less than four satellites. It is shown that the UKF based navigation system has better performance than EKF based system when used with local landmark measurements.
Published in:	Proceedings of the 20th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS 2007) September 25 - 28, 2007 Fort Worth Convention Center Fort Worth, TX
Pages:	127 - 134
Cite this article:	Hirokawa, Rui, Nakakuki, Kenji, "An Efficient UKF Based GPS/INS Augmented by Local Landmark Update," Proceedings of the 20th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS 2007), Fort Worth, TX, September 2007, pp. 127-134.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In