Research on 3D LiDAR Point Cloud Matching Localization Method

In 3D laser SLAM, the laser point cloud matching accuracy plays a decisive role in map building and localization. In order to better analyze the point cloud matching accuracy of LiDAR.In this paper, an experimental platform is built for two different point cloud matching algorithms commonly used, the nearest point iterative algorithm ICP and the normal distribution transform algorithm NDT. And the experiments are conducted in the experimental field to obtain the relative and absolute positional errors. The experimental results show that the localization accuracy of NDT is higher than that of ICP, and the robustness of NDT is significantly higher than that of ICP especially at the corner.