Schmidt ST-EKF for Autonomous Land Vehicle SINS/ODO/LDV Integrated Navigation

Abstract

Autonomous land navigation has been studied extensively in recent years to improve vehicle location accuracy in the satellite-denial environment. However, many previous studies of strapdown inertial navigation system/odometer (SINS/ODO) integrated navigation mainly focus on the horizontal location and neglect the height positioning since the troublesome problem of vertical channel divergency without extra height information assistance. In contrast, this article considered both horizontal and vertical positioning accuracy by SINS/ODO/laser doppler velocimeter (LDV)-based integration framework and proposed a novel algorithm named Schmidt ST-EKF which takes advantage of both ST-EKF’s consistency by nonlinear error definition and Schmidt filter’s improvement on system observability. Detailed system equations and observation equations of Schmidt ST-EKF-based SINS/ODO/LDV integrated navigation are derived and followed by concrete experiment results, which showed that the SINS/ODO/LDV integration method had better horizontal location accuracy than SINS/ODO integration by 35% and SINS/LDV integration by 26% and that applying Schmidt ST-EKF could efficiently reduce height positioning error. A further experiment about the correlating time of pitch mounting angle’s influence is conducted and revealed that the Schmidt ST-EKF is more robust than ST-EKF. Consequently, the algorithm proposed in this article for SINS/ODO/LDV has strong engineering applicability in the vast class of land vehicle applications.