Lidar-Based 2D SLAM for Mobile Robot in an Indoor Environment: A Review

Abstract

This paper presents a review and comparison of the common 2D SLAM (Simultaneous Localization and Mapping) systems in an indoor static environment, by utilizing the ROS-based SLAM libraries on a experimental mobile robot equipped with a 2D LIDAR module, IMU and wheel encoders. The three common algorithms (GMapping, Hector-SLAM, Google Cartographer) are the metrical map generating approaches of SLAM, which are categorized as filter-based or graph-based SLAM. The experimental results are acquired from the similar robot trajectory in both the simulated and real-world environment for further analysis under different circumstances. Overall, this paper describes the strength and weaknesses of the algorithms and visualizes the differences in terms of constructed maps, as it is mandatory to select the most appropriate system according to the intended application, as well as to identify the potential direction of optimization in the future.