Correction method of magnetic measurement error based on locus constraints

According to the general error model of magnetometer, a calibration method for geomagnetic measurement error is proposed, and its performance is analyzed. The proposed method adopts the least square algorithm under ellipse restriction, and estimates the 12 calibration parameters; then it uses these parameters to obtain the true geomagnetic field vector. Theory analyzes and simulation test shows that, in some instances, a locus-constraint calibration algorithm will be unobservable. And the method to examining whether it is unobservable is incorporated in the algorithm. The errors due to sort iron are different from the errors due to physical arrangement of the sensor axes. In the presence of soft-iron errors, locus-constraint algorithms can only provides 9 of the 12 calibration parameters. The remaining three parameters can be obtained by external information.