Closed froj closed 10 years ago
If it's just a question of motor cabling it should be done in the hardware abstraction. For more complex systems (like the holonomic base) there should be a transformation function applied after the PID mapping control outputs to motors.
Scaling factors for output are no longer necessary, because we have float
P, I and D parameters.
I agree that this should be done in abstraction layer because it is not always trivial to decide where to switch the sign if you do it in the controller and you can easily end up with a positive feedback system...
OK, I think we all agree.
Sometimes we need to flip the sign of the filter output because of how our hardware is constructed. Should this be a property of our filter or some other abstraction layer?