Estimation of reliable vehicle dynamic model using IMU/GNSS data fusion for stability controller design

Abstract

This study deals with the enhancement of reliability of vehicle dynamic model for the stability controller design by estimating the effect of uncertainties in both tire forces and vehicle dynamics. In this respect, any selected initial kinetic model of vehicle is updated at each instant by adding some complementary terms using a novel prediction approach. The proposed estimation method uses the information fusion between the inertial measurement unit (IMU) and the aided systems including the global navigation satellite system (GNSS) and the calibrated compass system. The bias of acceleration sensors and the drift of gyroscope are compensated using a novel adaptive algorithm by which the estimator weights are tuned automatically. Therefore, the vehicle longitudinal and lateral velocities together with the yaw rate are accurately estimated with a high frequency. The proposed estimation algorithm is mathematically analyzed for the stochastic stability and experimentally evaluated through real-world vehicular tests. Accordingly, a nonlinear controller for vehicle directional stability is designed based on the updated dynamic model within the CarSim software environment as a virtual vehicle experiment platform. The obtained results reveal that, the designed data fusion algorithm remarkably improves the estimation performance in the presence of system and measurement uncertainties. Also, the controller designed by the estimated vehicle dynamic model will be reliable and robust in the presence of uncertainties.