Low-Cost Vehicle Navigation System Based on IMU/GNSS/Odometer/Barometer Fusion in Diverse Environments

Huan Che,Hai Zhang,Mingo Tsai,Eric Lee,Kelly Zhang,Liang Li

DOI: https://doi.org/10.1007/978-981-15-3715-8_41

2020-01-01

Abstract:A vehicle GNSS navigation device may provide inaccurate position results or even fail to work in urban environments when the GNSS signals are blocked or reflected by tree canopies or buildings. To achieve more reliable and continuous vehicle navigation solution, this paper proposes a multi-sensor navigation system based on a consumer-grade MEMS IMU, fused with a GNSS receiver, an odometer and a barometer. An IMU error model is constructed in this integrated system to estimate and compensate for the errors of the gyros and accelerometers in motion because without any aided sensors, the accuracy of a consumer-grade MEMS-IMU would degrade fast due to its large bias and noise. Meanwhile, the vehicle motion state is detected to further aid the navigation system to ensure the system can continue to work even in the absence of GNSS or odometer. In addition, a plug-and-play multi-sensor fusion filter derived from Kalman filter framework is designed to ensure that the system operates normally when the available sensors switch over due to the vehicle’s moving through different scenarios. The filter employs UD decomposition to update the measurement so as to achieve real time positioning even on a low cost MCU processor. The proposed multi-sensor integrated system is implemented based on navigation chip of Allystar Technology Co. The fields result shows that, under challenging urban canyon environments, the developed integrated system can achieve continuous, stable and reliable positioning, and in particular during GNSS outage, the positioning accuracy can maintain an error within 1% of the traveled distance without fusing odometer while within 0.5% when odometer is deployed. With the aiding of a barometer, the elevation positioning accuracy can be maintained at about 2 m.