2D LIDAR SLAM Localization System for a Mobile Robotic Platform in GPS Denied Environment

Abstract

Agricultural mobile robots using Global Navigation Satellite System (GNSS)–based signal for navigation can be easily occluded and the position and attitude errors calculated by inertial navigation sensor (INS) will accumulate over time, which will seriously affect functions like localization and navigation. Therefore, based on the above shortcomings, this research uses 2D LIDAR (two-dimensional, Light Detection and Ranging) SLAM (Simultaneous Localization and Mapping) as the scheme for the outdoor positioning of the mobile robot. The Cartographer SLAM algorithm was selected in this study and was operated under the ROS (Robot Operating System) platform. After a series of comparative experiments, it can be concluded that the positioning accuracy at the normal speed of 1.5 km/h is about 0.2–0.3 m considering the error of human manipulation and the accuracy of attitude around 3–4 degrees. This 2D LIDAR-based localization framework helps for developing an autonomous navigation system for the static agricultural environment that can be operated under GNSS denied environment.