DGPS/INS/VMS Integration for High Accuracy Land-Vehicle Positioning

Hideo Kumagai, Tsuyoshi Kindo, Yukihiro Kubo and Sueo Sugimoto

Abstract:	In this paper, we investigate the land-vehicle positioning by DGPS/1NS/VMS integration, where DGPS means the transfer differential GPS. Our main purpose of the present investigation is to integrate these systems and to develop the high precision positioning system in all circumstances, especially in the poor PDOP and the lack of GPS satellite, using the Vehicle Motion Sensor (VMS) efficiently. Usually the DGPS/INS/VMS systems have the different hybrid Kalman filters and switch these filters suitably depending on the circumstance of the vehicle [I, 15]. In case of enough GPS satellites and good PDOP, the system adopts the DGPS/INS configuration and estimates the errors of INS. In other cases, the INS/VMS configuration is adopted, so that the VMS errors such as scale factor and boresight errors in addition to the INS errors are estimated. So in this existing method, there is the difficulty of initialization of the error covariance and the step response of the filter just after the filter mode is changed. Furthermore the initial value of the error covariance should be reset if the very large values of filter outputs come out comparing with the actual situation of the vehicle. In the present work, we propose a new algorithm, which consists of only one Kalman Filter without switching the filter configuration. In the proposed method, thus, the INS errors as well as the scale factor and the boresight errors of the VMS are estimated simultaneously. The proposed algorithm was examined by using the actual DGPS, VMS and INS data. The actual data were obtained by a running vehicle equipped with the stereo cameras for the GIS purpose. Throughout the experiments, the carrier-based kinematic GPS (KGPS) data were observed simultaneously to evaluate the system perlbrmance. The INS data were recorded at 51.2 Hz rate and others were 1.0 Hz respectively. And for the synchronization of the KGPS and estimated data, the 1 pulse per second (1PPS) signal of DGPS and time count of 1 MHz timer were recorded in the files. In the conclusion, the DGPS/INS/VMS integration system provides the continuous positioning data and high accuracy estimates of INS and VMS errors by combination of the each measurement without switching the filter algorithm. Our hybrid system enables to accurately survey the geographical information on the road in all circumstances; even in cases of the poor PDOP, the lack of GPS satellites in the city area, and the wheel slip (which is corrected by estimating the error of the scale factor).
Published in:	Proceedings of the 13th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GPS 2000) September 19 - 22, 2000 Salt Palace Convention Center Salt Lake City, UT
Pages:	1845 - 1853
Cite this article:	Kumagai, Hideo, Kindo, Tsuyoshi, Kubo, Yukihiro, Sugimoto, Sueo, "DGPS/INS/VMS Integration for High Accuracy Land-Vehicle Positioning," Proceedings of the 13th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GPS 2000), Salt Lake City, UT, September 2000, pp. 1845-1853.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In