Attitude estimation of an accelerated rigid body with sensor fusion based-on switching extended Kalman filter

In this paper a new method is proposed for attitude estimation of an accelerated rigid body. The sensors are used consist of Inertial Measurement Unit (a 3- axis gyroscope with uncompensated bias and a 3-axis accelerometer) and a 3-axis magnetometer. To improve the accuracy of attitude estimation an extended Kalman filter (EKF) with a switching rule is used to fuse the sensory information. In practical applications, under the external acceleration, the output of accelerometer is not reliable due to its sensivity to vibration and non-separability the dynamic acceleration and gravity. To overcome this problem a switching rule is designed based on acceleration detection. If external acceleration is detected in rigid body movement, only magnetometer outputs will be used in EKF equations; otherwise, both of the accelerometer and magnetometer outputs will be employed. Through numerical simulations, the efficiency of the proposed method is illustrated under external acceleration.