[G05-1] Indoor and outdoor navigation

[G05-1-04] The Estimation of Error Models of MEMS-IMU and its application to develop the GNSS/MEMS-IMU/On-board Vehicle sensor based positioning System

^○Yong Lee, Jay Hyoun Kwon (University of Seoul)

The determination of the precise position and attitude of a vehicle is an essential issue in ensuring the safety of the driver and pedestrian. Recently, vehicles equipped with various sensors have been introduced to support the safety of drivers. In addition, GNSS/INS integration technology is broadly applied in the field of the mobile mapping system (MMS) and unmanned aerial vehicle (UAV) due to the development of low-cost micro-electro-mechanical system (MEMS). MEMS-based inertial measurement unit (IMU) has an advantage in terms of the small size as well as the low cost, but its error rapidly increases in a short time. Especially, the stability of the position could not be guaranteed in the case of the GNSS signal outage. Therefore, it is necessary to model the behavior of the MEMS-based IMU to maintain the stability of positioning. Also, the way to compensate the error of MEMS-based IMU using supplement sensors is required to improve the precision of the position in the GNSS/INS integration.
In this study, three types of error models, random constant, 1st order Gauss-Markov and 3rd order Autoregressive, were applied to estimate the MEMS-based IMU error. The estimated models were applied as a form of closed-loop in the extended Kalman filter (EKF) based navigation algorithm to compensate the error, and the effect of the models were evaluated based on the static and kinematic experiments. Also, the position, velocity and attitude estimated from various on-board sensors are applied to improve the performance of the navigation when GNSS signal is blocked