An Adaptive Tracking Estimator for Robust Vehicular Localization in Shadowing Areas

The vehicular localization is a crucial technique in autonomous vehicle systems. This paper addresses an adaptive tracking method for robust vehicular localization even in shadowing conditions. The proposed tracking estimator consists of attitude estimation and position estimation. For the attitude estimation, the tracking estimator relies on inertial measurement unit (IMU) and a global positioning system (GPS). According to the shadowing condition, the presented estimator adaptively selects the IMU data and the GPS information in order to continuously estimate the vehicle attitude. Then, the tracking estimator utilizes the attitude estimates and the on-board diagnostics II (OBD-II) information in order to accurately estimate the vehicle position. The adaptive tracking approach allows robust vehicular localization since the vehicle attitude can reliably be achieved in shadowing and non-shadowing areas. Furthermore, the adaptive tracking method effectively exploits one extended Kalman filter for the estimations of vehicle attitude and position. This significantly alleviates the complexity of the tracking estimator. The simulation results verify that the adaptive tracking estimator accurately tracks the vehicle trajectory even in shadowing areas. The results also exhibit that the proposed tracking estimator outperforms the conventional tracking estimators in the aspects of localization performance.