A Data Fusion Algorithm of GNSS/INS/ Odometer Integrated System In Consideration of Total Odometer Errors

Abstract

The low cost GNSS/INS integrated system applied in the land vehicle can benefit from the information fusion of odometer and motion aided constraints. However, the performance of the fusion system is affected by four key issues. Firstly, odometer scale factor is not a constant value and differs with the change of temperature, tire pressure and vehicle load. Secondly, the odometer measures velocity in the vehicle body frame (VBF) rather than the inertial sensor frame, whereas misalignment between the inertial measurement unit (IMU) and the VBF generally exists. Thirdly, the IMU origin doesn't coincide with the odometer origin, lever arm influences the position accuracy. Fourthly, poor road condition may make the wheel spin or slide, and eventually leads to odometer failure. To solve aforementioned problems, this paper attempts to provide a scale factor and misalignment estimation method, a lever arm compensation (LAC) approach, and an odometer fault detection and isolation (FDI) indicator according to the odometer position error propagation equation. In addition, a two-cascaded Kalman filter is designed based on the context awareness to fuse all data information. Simulation experiment demonstrates the effectiveness of the scale factor and misalignment calibration algorithm, lever arm compensation approach and fault detection indicator. Moreover, the processed fusion system can significantly improve the positioning accuracy in GNSS-hostile environment.