[Discourse.ros.org] [ROS Projects] Master thesis subject to help ROS
The most significant problems (aka _opportunities_) off the top of my mind are:
1) The common GUI tools are decent but they could be better. For example, after loading files with the Moveit Setup Assistant, you can't reload a new set of files. You have to close the program and begin again.
2) Aligning components in a URDF is incredibly tedious. The ROS-I CAD-TO-ROS project started to address this issue, but it has pretty much stalled. It would be great to have some Solidworks-type "mating" functionality. Huge opportunity here.
3) How to easily and accurately locate a camera in a tf frame? (Maybe there's a package to help with that. I will do some more searching.)