What is the clear difference between /reset_simulation and /reset_world. I initially was using /reset_simulation to reset the state of my environment periodically (in the context of RL).
This worked well until I added the functionality of deleting and spawning a new goal model when. From my observation the physical location of the model seems to "glitch". I understand that it has something to do with my condition to when I respawn (/delete_entity , /spawn_entity) the model. But it is clear that I am misunderstanding the process of /reset_simulation.
Often the model is just is not present in the gazebo world when it should be. I guess this could be /reset_simulation resetting the state of gazebo to when the model was not spawned.
I then looked into /reset_world which I believe is the better use case. But it doesn't work well with the slam_toolbox which is a crucial component of my package. The slam_toolbox complains about out of data TF frames (or something to that effect).
So to summarise I'm looking for clarification on the two services. The /reset ... service that I'm looking for is more to reset the robots to their original position in the world.
Currently using ROS2 Humble