object manipulation or tabletop detection with non-PR2 [closed]
I want my Scitos robot with Schunk arm to detect and grab objects, likely from a table. I looked at all those packages (object_manipulation, tabletop_object_perception, pr2_object_manipulation_launch, pr2_tabletop_manipulation_apps and the like) but the documentation is too small or too PR2-concentrated to understand all those complex packages and their dependecies. And whether one is lacking a deprecated note.
A top-level graph would be great, or a robot independent step by step guide.
I have a octomap from kinect for collision, a follow_trajectory_action server and the arm_navigation node running. Which packages or nodes do I need to do tabletop manipulations, and which are needless as they only help when using a PR2?
Right now I have a robot, and Im looking to do some object detection, did you finally manage to accomplish it with a robot different than PR2?
Actually I went for other topics. The only thing I got to see working was the planar segmentation from kinect data. [launchfile](http://wiki.ros.org/pcl_ros/Tutorials/SACSegmentationFromNormals%20planar%20segmentation) and [my code](https://gist.github.com/felix-kolbe/6579354) I cannot find online.