Error Analysis and Stochastic Modeling of MEMS based Inertial Sensors for Land Vehicle Navigation Applications

Although GPS measurements are the essential information for currently developed land vehicle navigation systems (LVNS), the situation when GPS signals are unavailable or unreliable due to signal blockages must be compensated to provide continuous navigation solutions. In order to overcome the unavailability or unreliability problem in satellite based navigation systems and also to be cost effective, Micro Electro Mechanical Systems (MEMS) based inertial sensor technology has pushed the development of lowcost integrated navigation systems for land vehicle navigation and guidance applications. In spite of low inherent cost, small size, low power consumption, and solid reliability of MEMS based inertial sensors, the errors in the observations from the MEMS-based sensors must be appropriately treated in order to turn the observations into useful data for vehicle position determination. The error analysis would be conducted in the time domain specifying the stochastic variation as well as error sources of systematic nature. This thesis will address the above issues and present algorithms to identify and model the error sources in MEMS-based inertial sensors. A Kalman filter will be described and applied to analyze the performance of a minimum configured GPS/IMU system for vehicle navigation applications. The performance of the testing system has been assessed via a comparison to Precise Point Position (PPP) reference data. The testing results indicate the effectiveness of the discussed error analysis and modeling method. iii