Performance Analysis of a Low Cost MEMS IMU and GPS Integration

Yudan Yi, Dorota A. Grejner-Brzezinska and Charles K. Toth

Abstract:	"Due to the complementary error characteristics of GPS and inertial sensors, INS/GPS integration has been broadly used for the high accuracy positioning, timing and navigation tasks. The capabilities to obtain high accuracy 3-D position, velocity and attitude are mandatory for direct geo-referencing of imaging sensors without ground control in photogrammetric applications and other similar precise positioning applications. The accuracies of the integrated systems for such applications are mainly dominated by the error characteristics of the inertial sensors during the loss of GPS lock, where the free inertial navigation errors grow with time. Consequently, the relatively slow-degrading quality of medium- to high-accuracy IMUs can allow longer GPS gaps, ranging from minutes to hours, depending on the application requirements, while low accuracy IMUs will only bridge a few seconds of a GPS outrage. In mobile mapping, the majority of the existing INS/GPS implementations are predominantly based on the navigation-grade IMU sensors. However, with the development of the low-cost MEMS IMU sensors, it is now necessary to investigate the possibility of GPS integration with these sensors as a primary georeferencing sensor for land-based mobile mapping systems. In this paper, a low cost MEMS IMU (Crossbow MEMS IMU400CC-100) and GPS integration is investigated and tested. The INS/GPS Airborne Integrated Mapping System (AIMSª), which was originally developed with a navigation-grade Litton LN-100, was modified to accommodate the low-cost MEMS IMU sensor error characteristics in the Extended Kalman filter. The sensor errors, together with their characteristics and corresponding empirically identified noise error models, based on the static datasets are summarized and analyzed first. Then, the augmented error state vector is discussed. In order to achieve better navigation performance of the relatively low-cost MEMS IMU sensors with a rather high measurement noise, a signal de-noising technique based on the wavelet decomposition is introduced. This technique can lower the measurement noise by smoothing the raw inertial accelerations and angular rates. Some practical considerations for the low-cost MEMS IMU, such as static initial coarse/fine alignment, is also addressed. Finally, the performance analysis of the INS/GPS integration based on the static and kinematic tests will be presented for the low-cost MEMS IMU, with a special consideration on the free inertial navigation performance. The preliminary results indicate that the MEMS/GPS integration with a careful signal de-noising and sensor-customized noise modeling may satisfy the requirements of the lower accuracy mobile mapping applications."
Published in:	Proceedings of the 61st Annual Meeting of The Institute of Navigation (2005) June 27 - 29, 2005 Royal Sonesta Hotel Cambridge, MA
Pages:	1026 - 1036
Cite this article:	Yi, Yudan, Grejner-Brzezinska, Dorota A., Toth, Charles K., "Performance Analysis of a Low Cost MEMS IMU and GPS Integration," Proceedings of the 61st Annual Meeting of The Institute of Navigation (2005), Cambridge, MA, June 2005, pp. 1026-1036.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In