Attitude-Independent Magnetometer Calibration Based on Adaptive Filtering
Wenshi Zeng, Qiang Bian, Junji Gao, Lubin Chang, Yude Tong
Abstract
In order to calibrate the tri-axis magnetometer errors, an attitude-independent calibration method based on adaptive filtering is proposed. This method can estimate all the error parameters precisely without the need of external equipment to provide stable magnetic field and reference attitude information. In this paper, incremental meta-learning Incremental Delta-Bar-Delta (IDBD) algorithm is applied to adaptive filtering to achieve magnetometer calibration instead of the traditional least squares (LS) algorithm. The results of this calibration method will be compared with the classical calibration method in the earth’s magnetic field as well. The experiments show that the sensor error after calibration by the proposed method is significantly suppressed and reduced by more than 93 % compared with that before calibration and the calibration effect of IDBD algorithm is much better than that of LS algorithm. The proposed method in this paper can not only effectively improves the measurement accuracy of the tri-axis fluxgate sensor, but also is potentially applicable to various kinds of three-axis sensors.