Attitude estimation method for small UAV under accelerative environment

Abstract An attitude estimation method which is robust for accelerative environment and applicable to small unmanned aerial vehicle (UAV) using low-cost gyroscope, magnetometer and global positioning system (GPS) receiver is developed. The accelerometer, which is essentially used for attitude heading reference system degrades the accuracy of estimated attitude when the UAV is under accelerative maneuver. For this reason, the accelerometer is excluded from the proposed method. On the other hand, pseudo-attitude, which is estimated from velocity measurement of the GPS receiver is easy to implement and provides stable attitude information for long term. However, it has low-rated output, time lag and deviation from true attitude. In order to complement the deviation of pseudo-attitude, magnetic-pitch angle determination method, which extracts the pitch angle information from magnetic vector in addition to yaw angle is developed in this study. Moreover, the time lag of pseudo-attitude is modeled using error of pseudo-attitude and time constant. Using this model, the time lag of pseudo-attitude is effectively compensated without large memory or computational load. Finally, the pseudo-attitude, magnetic-attitude and gyroscope measurement are fused based on Euler angle. As a result, an attitude which is high-rated and robust against acceleration is achieved. The proposed method is verified through both of the simulation and flight test. The results show that the proposed method maintains stable attitude accuracy even if the aircraft experiences sudden or continuous acceleration.