How to improve the localization of the FUNMAP

Hi, i open this topic to discuss about the localization with the FUNMAP. As you know there some problem in the localization with the FUNMAP in comparison with the navigation stack that use the AMCL.
I was thinking to use the a map created with navigation stack to send the localization estimated with the AMCL to the FUNMAP, which have the localization deactivated. With this method the FUNMAP wouldn’t have to do a local localization each time that move to a goal.
What do you think this will help or will have problems with the FUNMAP?

Hi @Humberto_Andres_Hida, this is good idea. All that is required from FUNMAP’s perspective is a localization estimate in the form of a transform from “odom” to “base_link”. In the default launch file, this is provided by the laser_scan_matcher node. By substituting laser scan matching with local localization through AMCL particle filtering, the maps generated through FUNMAP would benefit from AMCL’s better state estimation.

1 Like