Velodyne rays don't detect actor animations

Hi all,

I followed the steps in the actor tutorial on Gazebo8 http://gazebosim.org/tutorials?tut=actor&branch=actor

and I can spawn animated characters that move in front of a simulated Kinect sensor and a Velodyne (VLP16, block laser sensor). The Kinect depth image can clearly detect the animated character, however the Velodyne rays can't. Looking back at the tutorial, it says that "They [animated characters] will, however, have a 3D visualization which can be seen by RGB cameras, and 3D meshes which can be detected by GPU based depth sensors."

Does that mean that non-GPU based laser sensors will not be able to see the animated character? If not, is there a reason why the block laser would not see it?

Thanks!

Asked by florian7 on 2016-07-20 16:51:01 UTC

Comments

Have you found any solution on that?

Asked by klapetos on 2017-12-06 13:34:34 UTC

Answers

Yes, at the moment non-GPU sensors cannot see actor skeleton animations.

The reason is that non-GPU sensors locate collision shapes, which are managed by the physics engine. Animated skeletons in actors don't have collisions (for now). But other links animated within an actor may have collisions which should be seen by non-GPU sensors.

GPU based sensors, on the other hand, use visual shapes, which are managed by the render engine, and actor skeletons have those.

Asked by chapulina on 2016-07-20 17:03:55 UTC

Comments

Thanks for the clarification! Do you think writing a new, GPU-enabled, block laser plugin for gazebo is the way to go, then? Is there a technical reason OSRF hasn't implemented this already? Or is it due to something else?

Asked by florian7 on 2016-07-20 17:48:53 UTC

@ chapulina I am using Gazebo 11 and I see the same behavior where gazebo actors are being detected by non-GPU laser. Is there a plan to add this capability in the near future to Gazebo?

Asked by PKGosal on 2020-08-19 22:46:31 UTC