Real-time estimation of magnetometer calibration parameters in unmanned autonomous vehicle

One of the most commonly used sensors in navaid systems are magnetometers which sense the magnetic field where they have been. Three axis magnetometers with the ability to measure the Earth's magnetic field are used in UAVs navigations due to provide reference for INS as a navaid system. This reference value that is called magnetic variation defined as a angle of between Geographic(true) North and Magnetic North. The deviation(variation) angle can calculated by using horizontal components of the magnetometer results and is used by INS in order to eliminate cumulative growing errors in long term operation. Main subject of this thesis is estimation of magnetometer calibration parameters in real time via finding the local(real) magnetic filed components in the position where the UAV during flight. Thanks to the method being developed inaccuracy magnetometer's measurements are corrected using the model based on probabilistic approach. To compare success of the developed method an offline method is built based on LS algorithm. In the literature accepted as one of the offline calibration algorithms is called ellipsoid fitting which also has different types depends on various mathematical backgrounds is demonstrated based on eigendecomposition centered algorithm in this thesis. This method is not always suitable for use because the data obtained according to the very different orientations needed for the operation of the algorithm are not sufficiently provided from the flight maneuvers. So, results of developed ellipsoid fitting is not use for comparison with results of real time calibration method.