Is it good idea, first: 3D scan environment (mapping), and Next: Just localizations ? (not Simultaneous )

[damavand1]

Hi

I know SLAM, means Simultaneous localization and mapping
but, what if I first make a good 3D scan of environment by human using Kinect like camera and save the file

and in operation time, just localize robot using Kinect+rtabmap+saved map ?

[matlabbe]

Yes, it is often the most common workflow for robot navigation:

1. Teleop (or send goals if navigation stack is up) the robot with rtabmap in SLAM mode (default) to create the map
2. After the map is created, restart rtabmap in LOCALIZATION mode (set parameter Mem/IncrementalMemory to false and remove --delete_db_on_start argument if previously set). Rtabmap will localize in previous map without adding new data.