EKF based multiple-mode attitude estimator for quadrotor using inertial measurement unit

A novel switchable multiple-mode sensor fusion algorithm for attitude estimation of quadrotor is proposed in this paper. It is a robust algorithm which shows good performance in flight conditions where sustained translational acceleration and wind disturbance exist. This algorithm is based on quaternion-based extended Kalman filter (EKF) in direct configuration and uses measurements from a strap-down MEMS IMU equipped with 3-axis gyroscope and 3-axis accelerometer. Based on a high-fidelity dynamic model of quadrotor, a new measurement model is proposed for flight conditions with translational acceleration. In order to use sensor data appropriately in different flight conditions, a decision tree is designed as a switching strategy to divide the fusion process into three modes and switch the estimator between them according to the outputs of the IMU and the innovation vector. It is demonstrated by simulation and experiment that the proposed algorithm can maintain accurate estimation of the quadrotor's attitude even under the effect of sustained translational acceleration and wind disturbance, which lead to poor performance of traditional algorithms which assume that 3-axis accelerometer measures the gravity vector.