- a robot command/programming language, equivalent to what you have for industrial robots - that’s only necessary if you want to skip the vendor’s controller
If we have our own controller, that is ROS based, it almost eliminates the need for a robot command language. However, I do think there is value in providing a programming interface that is more familiar to industrial robot programmers. We are actively developing something called moveit_simple, that attempts to simplify programming, while still benefiting from power of ROS.
- Trajectory generation, planning, and smoothing at the joint level is also missing. This would have to be added to the low level controller in order to achieve industrial robot level motion.
- Collaborative functionality. Might as well jump into this growing market, rather than focus on legacy applications.
How to make money with this business?
This is by far the harder question. An open source controller still requires a robot. Not everyone is going to build a custom robot (or not enough to make it a worth while business). If the controller worked with off the shelf robots, that would help, but I don’t see getting the cooperation of existing robot vendors as very likely.