Magnetometer Bias Insensitive Magnetic Field Matching Based on Pedestrian Dead Reckoning for Smartphone Indoor Positioning

Magnetic field signals are ubiquitous, stable, and have little effect on the human body. Due to these qualities magnetic field matching (MFM) has become a popular indoor positioning method and has attracted considerable attention within the smartphone geolocation community. However, the need to frequently calibrate the magnetometer bias seriously undermines the applicability and stability of MFM. In this paper we present a magnetometer bias insensitive MFM based on pedestrian dead reckoning (PDR) for smartphones in the indoor environment. An inertial navigation system based PDR is designed for generating the relative trajectory and attitude. The relative trajectory is used to correlate the magnetic field feature time sequence to improve the distinguishability and the attitude is used for projecting the reference magnetic field feature from ${n}$ -frame to ${b}$ -frame, to perform the MFM in the ${b}$ -frame to eliminate the influence of the magnetometer bias. The results of eight experiments in four smartphones showed that the proposed method could effectively eliminate the influence of the magnetometer’s bias in the MFM, and finally reached mean positioning accuracy of 0.77 m (RMS).