2008 · 72 citations · 4 references
EngineeringMechatronics EngineeringRobotics EngineeringGeomagnetismRobot KinematicsSensor FusionAttitude ControlAttitude DeterminationCalibrationKalman FilteringMagnetic SensorsSensor CalibrationAttitude Determination AlgorithmsAttitude Determination SystemsAttitude EstimationAttitude Reference SystemsExtended Kalman FilterHeading Reference SystemsMagnetic Sensor CalibrationPositional Measurement Errors
The calibration method of the soft iron and hard iron distortion based on attitude and heading reference system (AHRS) can boil down to the estimation of 12 parameters of magnetic deviation, normally using 12-state Kalman filter (KF) algorithm. The performance of compensation is limited by the accuracy of local inclination angle of magnetic field and initial heading. A 14-state extended Kalman filter (EKF) algorithm is developed to calibrate magnetic deviation, local magnetic inclination angle error and initial heading error all together. The calibration procedure is to change the attitude of AHRS and rotate it two cycles. As the strapdown matrix can hold high precision after initial alignment of AHRS in short time for the gyropsilas short-term precision, the magnetic field vector can be projected onto the body frame of AHRS. The experiment results demonstrate that 14-state EKF outperforms 12-state KF, with measurement errors exist in the initial heading and local inclination angle. The heading accuracy (variance) after compensation is 0.4 degree for tilt angle ranging between 0 and 60 degree.
4