Research on 2D Mapping for Mobile Transport Robots Based on SLAM

Many existing studies predominantly focus on the performance of Simultaneous Localization and Mapping (SLAM) algorithms within single environments. This research provides a comparative analysis of two SLAM algorithms, Gmapping and Cartographer, evaluating their mapping performance across diverse environments from multiple dimensions. To facilitate this comparison, a mobile handling robot equipped with various sensors was designed and constructed. Mapping experiments were conducted in both a simplified laboratory setting and a complex warehouse environment using the 3D simulation tool Gazebo and the visualization tool Rviz on the ROS platform. The performance of the two algorithms was assessed based on several parameters, including map jaggedness, boundary omissions, mapping duration, and CPU usage. The results demonstrate that the Cartographer algorithm produces smoother and more accurate maps in complex environments compared to the Gmapping algorithm, with jaggedness accounting for 19% of the overall map and boundary omissions at approximately 14%. However, it requires slightly more computational resources and time than the Gmapping algorithm.