Solution to geomagnetic systems in areas with smooth magnetic fields

To solve the divergence problem of traditional Kalman filter-based geomagnetic navigation systems caused by their unobservability in the areas with smooth magnetic fields, a new algorithm was proposed. The algorithm uses linearized planes to fit a smooth magnetic field, the vector properties of the magnetic field to construct linear equations, and multi-level submaps to realize fast and accurate position estimation. The observation equation of the Kalman filter was constructed based upon the position estimation. Theoretical analysis and simulations proved the effectiveness of the algorithm.