A critical step in the deployment of Wyca vehicles is the creation of a map from scratch, with no environmental modification and by using the vehicle only.
In other words: driving the vehicle around to make it learn about the place it is going to be used.
Technically, on an indoor mobile robot, this capability comes from a field of algorithm called SLAM, for simultaneous localization and mapping.
One of the most commonly used open source SLAM implementations in the mobile robot community is cartographer, which was open sourced and fully integrated in ROS by google in 2017. Wyca uses some part of it in its map creation workflow.
Unfortunately, Google is not pursuing the development of the project and the porting of cartographer in ROS2 was incomplete and from an old version with obsolete APIs. As it was a blocking point for Wyca for the transition to ROS2, our team ported it completely to ROS2 early 2021. Then this porting was merged in april 2022 to the official ROS2 cartographer release maintained by the Open Source Robotics Foundation. It will be part of the ROS2 Humble release.
En savoir plus : https://discourse.ros.org/t/rolling-and-soon-humble-release-of-both-cartographer-and-cartographer-ros-v2-and-call-for-testing/25137
Guillaume Doisy, Chief Robotics Officer at Wyca Robotics