Is there a way to detect self-collisions in parallell within an articulated robot without simulating the collisions? For RL purposes, this would allow us to reset the scene with a large negative reward when self-collisions occur, without the need to simulate the collisions, which can be very time-consuming. In other words is there a parallell version of the overlap test( https://docs.omniverse.nvidia.com/isaacsim/latest/reference_python_snippets.html?highlight=ccd#do-overlap-test)
which can be used on an Articulation view when simulating many robots in parallell.
Best,
Lars