Comparative Study of popular LiDAR-based SLAM Algorithms

Abstract

Autonomous navigation relies on the robot’s ability to perceive the environment, plan routes, avoid obstacles, and localize itself in space. Simultaneous Localization and Mapping (SLAM) is crucial for this task, allowing the robot to build a map of the environment while localizing itself within it. This paper presents a comparative study among four widely used SLAM algorithms (Gmapping, Hector SLAM, Cartographer, and Karto SLAM), implemented on a real mobile robot equipped with LiDAR and encoders, aiming to evaluate the quality of the maps generated by each algorithm. The results reveal a good overall performance, with Cartographer standing out for producing maps with higher structural similarity to the reference map.