An Efficient and Robust Indoor Magnetic Field Matching Positioning Solution Based on Consumer-Grade IMUs for Smartphones

Abstract

This paper proposes an indoor magnetic field matching positioning (MFMP) scheme based on consumer-grade inertial measurement units (IMUs). The proposed MFMP can efficiently generate a magnetic field map and achieve robust matching positioning without calibrating the magnetometer bias. In the magnetic field map generation stage, a pedestrian positioning and orientation system (P-POS) is employed to provide the precise position and attitude of the smartphone during the data collection, and rasterization and bilinear interpolation methods are utilized to generate a three-dimensional grid magnetic map. In the real-time positioning stage, the position and attitude generated by pedestrian dead reckoning (PDR) are used to improve the position distinguishability of the magnetic field features and obtain the transformation relationship (from the navigation frame to the sensor frame). And the differential magnetic field features in the sensor frame are used to achieve matching positioning independent of the magnetometer bias. The slight differences in the magnetic field maps based on different smartphones show that the proposed scheme can efficiently generate a high precision magnetic map. Additionally, the positioning results of multiple tests using multiple smartphones show that the proposed scheme is less affected by the magnetometer bias. Furthermore, it has similar positioning performance in different smartphones, achieving continuous and robust meter-level positioning accuracy.