ABSTRACT

To resolve the problem due to drift and noise in attitude estimation process of MEMS inertial measuring unit on quadrotor, an improved algorithm based on complementary filtering is presented. A test platform of quadrotor carried out with MPU9250 as attitude measuring unit is established. Both under static and dynamic conditions, attitude data obtained from different methods are collected and compared, including Kalman filtering algorithm for information integration, data fusion with traditional complementary filtering, and our improved filtering fusion algorithm. Experimental results show that the present improved attitude fusion algorithm has advantages in terms of higher estimation precision and lesser drift and noise errors of the final attitude angle in different conditions, and is easy to implement in low cost aircraft control system.