MEMS Magnetometer Calibration Using Modified Ellipsoid Fitting Method

Low-cost sensors based on Micro-Electro-Mechanical-Systems (MEMS) are typically used for attitude determination in navigation systems. Measured value of magnetometer is subjected to different kind of errors such as random noise, constant bias, nonorthogonality, scale factor deviation and more importantly hard iron and soft iron effects. Therefore, it is needed to calibrate these sensors for accurate measurement. One of the most common methods in calibrating magnetic sensors is least squares ellipsoid fitting. But, the common least squares ellipsoid fitting can be inefficient for real time applications. In this paper, a modified ellipsoid fitting method is proposed. In this method, nonlinear optimization is developed to minimize a novel cost function. The efficacy of the new algorithms is demonstrated through experimental results.