Magnetic field-aided 3D LiDAR localization in geometrically repetitive environments

Abstract

With rapid development of mobile robots, a robust localization algorithm that would enable the robot to acquire its accurate location in various environments, especially the geometrically repetitive and symmetric environment, is still a challenging task. Global Navigation Satellite System (GNSS) is a widely used and reliable method for localization in outdoor environments. However, when it comes to semi-indoor or indoor situation, GPS would become unstable or even unable to give out the location of robot. Consequently, a large number of solutions are proposed for localization in indoor scenarios over the past decades. Currently, there are two main kinds of localization methods: infrastructure-based and infrastructure-free. Infrastructure-based methods such as Wi-Fi, QR code and UWB can provide accurate location of robot. However, these kind of methods need to pre-install the infrastructures in the environment. In the contrary, infrastructure-free methods do not need to install and maintain the infrastructure. They only rely on the environment information collected by various kind of sensors. However, it is still challenging for infrastructure-free methods to localize in the geometrically repetitive environments. In such environment, it usually has similar structure and lack of distinctive geometric features. Since the 3D LiDAR can perceive the environments with lots of geometric features, this paper introduce the working principles of multiple 3D LiDAR-based localization methods in detail and analyzed their performances in geometrically repetitive environments. Moreover, since the geomagnetic field information is unique at the local environments, this paper will explore and propose a magnetic field-aided 3D LiDAR localization system for mobile robot localization in geometrically repetitive environments.