Grid_localization is an elevation grid map based localization solution for urban environments using LiDAR sensor.
Matching step in this project is excuted between 3D point cloud (current measurement) and elevation grid map (reference map), using ICP based on MRPT.
MRPT-1.5 is required to build the project. See how to get MRPT on Ubuntu.
See DEMO here.
1.Point cloud generation.
Generate a global point cloud with the trajectory formed form EKF using odometer, imu, and rtk-gps.
2.Elevation grid map generation.
Generate the elevation grid map with the global point cloud. Each grid map covers a 80*80m^2 area and is generated every 20m to ensure there is enough shared area between two grid maps (to ensure the vehicle is located in the central area of each grid map, better for localization).
3.Localization against elevation grid map.