Robust Kalman filtering for attitude estimation using low-cost MEMS-based sensors

This paper presents a robust and practical attitude estimation algorithm based on low-cost MEMS inertial and magnetic sensors. Since accelerometers measure motion other than gravity and the geomagnetic field is sensitive to the surrounding environment, a robust measurement update method is proposed to attenuate the influences of motion disturbances and geomagnetic disturbances. The advantage of the proposed approach is that the gain matrix can be switched according to the standardized residual. Therefore, the disturbances existing in the observations are controlled. Experimental results using a homemade quadrotor platform and low-cost IMU show good real-time performance with low computational cost in a microcontroller.