| ROS Resources: Documentation | Support | Discussion Forum | Index | Service Status | ros @ Robotics Stack Exchange |
Can we use the same .STL files for both visual representation and collision representation while building URDF?
Yes, you can.
But you probably don't want to.
The collision geometry is typically used by motion planners (such as MoveIt) to perform collision checks, and in most cases, performance of that step is inversely proportional to the fidelity of the collision geometry you've supplied.
So in general I would recommend to only use the same meshes for both if you are OK with reduced performance or the visual mesh is low fidelity already.
Note: whether planning performance really is slower with high fidelity meshes for collision of course depends on the planning software used. Some will allow you to approximate collision geometry using convex hulls/shapes, or perhaps even capsules.
As MoveIt is often used with URDFs and you're asking this on ROS Answers, I've provided the main rationale for why we have separate visual and collision elements.
Asked: 2020-12-28 14:34:12 -0500
Seen: 782 times
Last updated: Dec 29 '20
ROS Answers is licensed under Creative Commons Attribution 3.0 Content on this site is licensed under a Creative Commons Attribution Share Alike 3.0 license.
