Quaternion-based robust extended Kalman filter for attitude estimation of micro quadrotors using low-cost MEMS

摘要

In this paper, a quaternion-based robust extended Kalman filter (EKF) is developed for the attitude estimation of micro quadrotors. Since the raw data of low-cost MEMS is sensitive to the environmental conditions, the attitude estimation method is developed to minimize the influence of outlier and to estimate the bias effectively. In order to overcome the singularity of certain orientation, the quaternion is involved in the filter parameter. A cortex-M4 based micro quadrotor is developed to implement the proposed quaternion-based robust extended Kalman filter, and simulation results show good estimation accuracy and satisfactory real-time performance.