Dynamic magnetic field compensation for micro UAV attitude estimation

This paper presents a method to calibrate and compensate dynamically the magnetic field created by wires of different devices in the context of micro robotic localization. The compensation step consists in mapping available information on the vehicle state to the dynamic magnetic perturbation created by each device and compensate this perturbation. The model parameters used for this mapping are estimated in a prior step thanks to calibration data. The compensation is independent of the calibration of the static errors and does not require current sensors. The proposed method improves the quality of magnetic measurements and relaxes the distance restriction between sensors and power electronics which is primordial to develop compact systems. We illustrate and validate experimentally our approach on a micro UAV with four rotors.