How to obtain the ray detection of a large number of agents by GPU

When Isaac gym is used, the ArticulationView in Isaac core can process the motions of all agents at once, which is very good.
However, if my agents need ray detection, depth map or collision information, how can I obtain the information of all agents in the form of GPU at once.

Hi @P90-RushB ,

We do not currently have GPU APIs for ray detection or collision information. Depth sensors should be available in the coming release, but the number of sensors we can have in the scene is very limited at the moment. We’ll be introducing GPU APIs for collision detection in later releases as well.

1 Like

Hi @P90-RushB - I just want to make sure if my problem is similar to yours.

Does this mean that you tried to move multiple agents at once. The moving works but if there are lidars on each agent, they wont detect each other?

You can use ray to detect the agents, but not in the form of GPU, so the efficiency will be low.

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.