Magnetic Field-based Indoor Localization of a Tracked Robot with Simultaneous Calibration

Magnetic-field based indoor localization uses distortions of the Earth magnetic field caused by magnetic material in the surrounding of the platform that is to be localized. Existing magnetic localization methods compare measurements from a magnetometer with a magnetic field map to estimate the position of the platform, on which the sensor is mounted. For the comparison it is typically assumed that the magnetometer is calibrated. Unfortunately, for some platforms calibration is not straightforward, particularly when they are large or heavy. In this paper we therefore propose a simultaneous localization and calibration approach. The approach is based on a Rao-Blackwellized particle filter that makes use of the conditional linearity of magnetometer measurements and calibration parameters. The feasibility of the proposed particle filter is evaluated with three data sets recorded with a differential drive robot in an indoor environment.