A Performance Comparison of Geomagnetic Field-Based Vector Field SLAM Approaches

Abstract

This paper presents performance comparison results of geomagnetic field-based vector field SLAM (simultaneous localization and mapping) approaches for mobile robot localization. Generally, building structures distort the Earth’s magnetic field inside the building, and the variation of the magnetic field over time is extremely small. Therefore, the geomagnetic field can be applied to mobile robot localization. The geomagnetic field-based vector field SLAM uses only the Earth’s magnetic field signals and odometry data to estimate the robot state. Recently, there have been a lot of studies about the possibility of applying the geomagnetic field for mobile robot localization in indoor environments. To estimate the geomagnetic field, the geomagnetic field-based SLAM approaches use bilinear interpolation or bicubic interpolation. However, the approaches have a limitation that its performance highly depends on interpolation method. In this paper, we analysis the performance of the geomagnetic field-based SLAM with respect to interpolation methods. We provide performance comparison results by performing simulations.