MIT-SPARK
Integration Issues with Kimera-PGMO and Kimera-Semantics in Multi-Robot Context
Hi, I am integrating Kimera-PGMO and Kimera-Semantics with a setup that successfully runs the front-end, back-end, and single-robot VIO, as shown in the attached image. However, I've encountered issues during integration, particularly with frame transformations for Kimera-Semantics in the multi-robot context.
I have followed the instructions in the provided launch file and modified the topic names as needed. My goal is to display kimera_semantics_node/mesh and kimera_vio_ros/odometry in the map of a single robot or to show the mesh in the world coordinates, as depicted in the paper.
Are any specific frame transformations required for Kimera-Semantics in a multi-robot setup? Did I need to include any critical steps for achieving the desired visualization per the paper? Additionally, my PGMO module shows the following error when receiving dpgo_ros/path messages. Could this be affecting the mesh optimization process?
I would appreciate any insights or guidance on resolving these integration issues. Thank you!
Do I need to have a specific world frame set up to display my mesh correctly?
Add a transformation between {robot_name}/map to world may help you, like this https://github.com/MIT-SPARK/Kimera-Multi/issues/3#issuecomment-1793603701
