-
|
I am writing our own motion planning for a robotic arm with collision avoidance. I am thinking the most direct way To find the xyz position of obstacles would be to use the octomap output and check if a position is occupied (ignoring positions around the goal object, which I have). The Octomap library has: a leaf iterator, a getCoordinate and a isNodeOccupied methods. Is this a sensible approach, or is there another solution already implemented in rtabmap? Thanks! |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 1 reply
-
|
If your robotic arm is not on a mobile base, you may be able to just use octomap_server directly. For 3D output format in RTAB-Map, we have raw point cloud, mesh (with standalone library we can generate a mesh) or OctoMap. For your case, I guess the OctoMap makes more sense with its OcTree structure to know space that is occupied, unknown or empty. |
Beta Was this translation helpful? Give feedback.
-
|
Thank you @matlabbe I will look into it. |
Beta Was this translation helpful? Give feedback.
If your robotic arm is not on a mobile base, you may be able to just use octomap_server directly. For 3D output format in RTAB-Map, we have raw point cloud, mesh (with standalone library we can generate a mesh) or OctoMap. For your case, I guess the OctoMap makes more sense with its OcTree structure to know space that is occupied, unknown or empty.