-
Notifications
You must be signed in to change notification settings - Fork 0
Implement 2D LIDAR in RTABMap #57
Description
RTABMap (Real Time Appearance Based Mapping) is the SLAM software we are using; check below for documentation and an introduction to the software:
https://github.com/introlab/rtabmap
https://github.com/introlab/rtabmap/wiki
Also look through these demos:
https://docs.ros.org/en/humble/p/rtabmap_demos/
Although RTABMap is usually referred to (and is currently configured) as a purely visual SLAM software using a depth camera to generate a 3D RGBD pointcloud, 2D LIDAR can be used to supplement this and quickly fill in the costmap from all angles to prevent collisions in blindspots.
Your task has two major parts:
- Integrate the 2D LIDAR into our current RTABMap configuration in
igvc_slam/sim_rtabmap.launch.pyby altering/adding parameters (there are some examples). Don't worry about getting the LIDAR to publish a topic - that's up to sim/hardware (sim implementation pending merge of Added LIDAR functionality #54) - Investigate any implications of using 2D LIDAR with RTABMap. For example, can we still use VIO (Visual-inertial odometry) if LIDAR is being added? Should we also be using some LIDAR-based odometry along with visual approaches? If so, how do we even get started? Do we need to add anything external to RTABMap?
Produce a document of your research, and upload it to our GitHub wiki or send it to me. Try to accommodate for any implications as much as possible in your PR - feel free to add more packages or code beyond RTAB parameters as necessary
Also feel free to discuss what you find along the way with your subsystem team/gavin/me
This wiki page I created contains all possible RTABMap arguments. It will likely be helpful
https://github.com/RPI-IGVC-2025/2026RobotCode/wiki/RTABMap-Arguments