How to incorporate "logic" into a component-based robot controller...

I would like to have your opinion about the following...

Every more or less advanced robot control system needs support for
"logical operations". (CNC circles talk about a "PLC".)

Since we want a _very strict_ separation between "policy" and
"mechanism" (or call it "infrastructure" and "application" if you
like), it seems to me that a possible solution could be this:

- The Orocos framework provides "Logic Component(s) (such as "FSM
  component" (FInite State Machine), and a "Petri Net component" and a
  "Discrete Event Systems component" and an "Event handling component"
  etc).  These components provide application-neutral support for various
  logic processing in the application. 
  Eventually, the different logic components can/should be integrated.

- These Logic Components have a standardized interface, with which all
  other components can perform their "logic" needs.

One of the big advantages of this approach is that there will be a
much larger group of people interested in these general purpose logic
libraries than in just a "PLC".

What would be the disadvantages?


Herman

-- 
  K.U.Leuven, Mechanical Engineering, Robotics Research Group
<http://people.mech.kuleuven.ac.be/~bruyninc> Tel: +32 16 322480