> Please show we a PIC using fuzzy logic that does a control application better
> than a classical control circuit.  Classical control systems have been in
> existence since the 1800's.  They are not inherently unstable and they do not
> operate poorly.  The cruise control in your car is a good example.  My cruise
> control works great, operates smoothly, is safe, stable and uses classical
> control systems.

I am sure you are aware of its use in current electronics devices,
e.g., camcorders, mainly either it is cheaper or there is no better
alternatives. To my knowledge it is replacing conventional
controllers in automobile industry, including speed control. In a
workshop, a scientist from Germany presented throttle control for
Volkswagen factory. This is not very difficult task. Real life
performance results were showing much better than VW's existing
controllers, especially when you turn on/off a\c, change the load of
the motor, and he said that he got the funding for that. He seemed to
be a happy man.

Then a Japanese engineer presented world's first autonomous
helicopter flight control, highly complicated problem, accomplished
by a fuzzy control scheme.

More importantly, there systems which can't be modeled as a linear or
nonlinear they are simply fuzzy systems. You can't employ any
conventional techniques. Here is the tool for us to play those class
of problems. There is the excitement.

> If you are interested in a critical view,  view the URL below.  Bob Pease, the
> author of Pease Porridge, tackles some Fuzzy control systems.  The fuzzy
 system
> used a 486 computer to control a ball on balance beam system.  Bob did the
 same
> thing with better results using four op-amps.  Now maybe the 486 computer was
> vast overkill, so redo the code into a PIC using fuzzy control and get the
 same
> level of results as the four op-amp classical method.  If fuzzy control cannot
> compete with four op-amps on either price or performance then it is not
 useable
> to me as an engineer.
>

I would not compete over ball and beam problem. Use the opamps and
tuned parameters of Bob. But what if your client asks you what you
expect PID performance if there is a sudden obstacle on the beam, or
increased friction, changed mass some facts of real life? I would
expect to see saturated signals driving motors crazy or bad tuned
controller performance which can easily be avoided in a fuzzy scheme.



Finally, there is no "the perfect controller" for all systems.
However, fuzzy logic control is a promising alternative for some
class of problems with its advantages and disadvantages and I really
would have liked to hear from people using Microchip Fuzzy development
tools.


__Aydin