The design of an extended Kalman filter for magnetometer and MEMS gyroscope/accelerometer integrated attitude determination

MEMS (Micro Zlectromechanical System) gyroscope is used widely in many applications, but it can't provide a steady attitude alone for a long time and must be integrated with other sensors. A quaternion based extended Kalman filtering algorithm for magnetometer and MEMS gyroscope/accelerometer integrated attitude determination is presented. This algorithm judges whether or not use the Kalman filter to calculate attitude according to the output of accelerometer and compensates the drift of gyroscope in the Kalman filter using the attitude computed from the magnetometer and accelerometer data. The effectiveness of this algorithm is demonstrated by experiment and simulation.