GEOMAGNETIC FIELD MODELING ERRORS AND OPTIMIZATION OF RECURSIVE ESTIMATION

THE CONCEPT OF USING RECURSIVE ESTIMATION METHODS FOR GEOMAGNETIC FIELD MODELING WAS DEMONSTRATED (BY BTS) IN A PREVIOUS LIMITED INVESTIGATION. THE RESEARCH PROPOSED HERE ATTEMPTS TO SOLVE SOME OF THE PROBLEMS ENCOUNTERED IN THAT STUDY SO THAT THE FULL POTENTIAL OF THE RECURSIVE ESTIMATION APPROACH MAY BE DEMONSTRATED. SPECIFICALLY, WE PROPOSE TO ESTABLISH EFFICIENT ALGORITHMS FOR COMPUTING THE OPTIMAL STATE NOISE COVARIANCE MATRIX USING MAXIMUM LIKELIHOOD ESTIMATION, TO FIND A MORE OPTIMAL METHOD OF TREATING OBSERVATORY "BIASES" IN FILTERING, AND TO ANALYZE AND IDENTIFY THE SOURCES OF GEOMAGNETIC MODELING ERROR WHICH IS CURRENTLY PRESENT IN THE LEAST SQUARES INVERSE PROBLEM FOR THE MAIN GEOMAGNETIC FIELD.