Sensitivity-based adaptive particle filter for geomagnetic indoor localization

Geomagnetic field is almost stable in the indoor environment. The interference sources which are distributed randomly throughout the whole area can form an identical pattern. In this paper, we propose a sensitive based adaptive particle filter scheme to locate the target with geomagnetic field. Firstly, we construct the feature map using interpolation method after gathering intensity and three-component of geo-magnetic field by a smartphone. Then we divide the space into some subspaces and utilize the measured values to find best match map. Finally, we integrate perceptual information into particle filter for adaptive processing. Our scheme can adjust the filter parameters to match different feature points automatically, which enhance the relationship between the target positions and the feature map using particle filter. The experimental results indicate that comparing to the traditional particle filter, our scheme is more stable. The positioning error achieves an average loss of 0.51 m through numerical experiments.