Euler Angle Based Attitude Estimation Avoiding the Singularity Problem

摘要

Precise control of UAV (Unmanned Aerial Vehicle) or MAV (Micro Aerial Vehicle), requires the accurate determination of vehicle orientation. Attitude and heading reference systems (AHRS) are popularly used for vehicle orientation determination. In these systems, Euler angle estimation is more appropriate than quaternion estimation because the accelerometer measurement and magnetometer measurement are separated. Furthermore, the accelerometer and magnetometer measurements have independent effects on the tilt and heading angles. An Euler based attitude estimation system that uses two-stage extended Kalman filtering is proposed. To avoid singularity of Euler angle, a new heading estimation parameter is introduced and a filter mode switching algorithm is proposed. The experimental results indicated that the proposed system performed better than a quaternion algorithm based on magnetometer disturbance. Furthermore, the results indicated that the proposed algorithm was immune to the singularity problem.