A Parallel Kalman Filter for Estimation of Magnetic Disturbance and Orientation Based on Nine-axis Inertial/Magnetic Sensor Signals

Magnetic disturbance is one of the main factors that deteriorate the accuracy of orientation estimation methods based on inertial/magnetic sensor signals. This paper proposes a parallel Kalman filter( KF) that explicitly detects magnetic disturbances and thus can accurately estimate 3D orientation in magnetically disturbed environments. Due to the parallel nature of the proposed KF, even severe magnetic disturbances only affect yaw estimation, while roll and pitch values remain accurate. Consequently, the proposed KF can be effectively used in various applications that involve magnetically inhomogeneous environments, such as robots, ships, and planes.