I am looking for some resources like publications and articles that compare Unity with Gazebo or other tools for robotic simulation and visualization. Ideally, with some supporting experiments and results.
But blog posts and videos or even your personal experiences are welcome as well, as I would like to know more about this topic.
I watched this podcast where Sarah Gibson mentioned various benefits of using Unity over Gazebo, but I am not sure which of these hold today and were tested or even published somewhere.
Iām working with the Unity Robotics package for round about a year now. My thoughts as short a possible:
after struggling with the ArticulationBodies we now use them just to interact with the physics on the lowest level. We helped ourself with a self made engine and controller script, just rotate the bodies (or just simply teleport the bodies).
in my point of viwe the URDF importer is not working well enough to integrate it
while having performance issues with the visualization of a lidar sensor we removed the visualization package after struggling here as well
the ROS bridge is the best of all changes and even if it is a bit complicated to run the bridge in parallel and e.g. handle costum messages, we use it quite often
Apart from my feedback and the discussion in the link, we do not use Gazebo in the most cases anymore. The visualization (mean the core module of Unity) is much easier to handle and is more robust than Gazebo.