Quality Evaluation Method of 2D SLAM

Abstract

A quantitative evaluation of a 2D map is important and becomes a challenging task for indoor survey mapping. The paper presents the method for comparison of built 2D maps generated from Google’s Cartographer. We focus on the comparison of map quality without ground truth data between a reference map from a set of LiDAR with IMU and wheel-encoder and compared maps from a set of LiDAR with IMU, wheel-encoder, and/or UWB. Three metrics for comparison the map quality are presented in this paper including RMSE, blur detection, and drifts in translation and rotation. The finding shows that a set of LiDAR with IMU and UWB can be comparable to a set of LiDAR with IMU and wheel-encoder in terms of map quality. Moreover, the proposed method can be further used for the relative evaluation of a feasibility study.