Quaternion-Based Attitude Estimation Kalman Filter Using Global Optimization

Abstract

This paper presents a modified Kalman filter for estimating quaternion using inertial and magnetic sensors. When the initial estimation error is large, the convergence rate of the multiplicative extended Kalman filter tends to be slow due to the assumption of small estimation errors. In this paper, a new measurement equation is proposed, in which a quaternion is directly estimated instead of estimating multiplicative estimation errors. Through simulation and experiment data, we demonstrate that the proposed algorithm is robust to large initial estimation errors.