GROUND VEHICLE ATTITUDE ESTIMATION THROUGH MAGNETIC-INERTIAL SENSOR FUSION

Abstract

The main goal of this paper is to design accurate and model-independent vehicle attitudes estimator. The proposed estimator uses low cost 3 axis magnetometer and vehicle yaw rate and lateral acceleration sensors. The estimation skim can be divided into 2 parts: Euler angle estimation and sideslip estimation. An extended Kalman filter combined with a deterministic Euler angle estimator is utilized for robust and fast converging Euler angle estimation. Using the result of Euler angle estimator, a sideslip observer based on compensated planar kinematic model is proposed. The feasibility of proposed algorithm is verified with the commercial simulation tool, Carsim. Simulation includes possible noises of sensor based on manufacturer datasheets.