Searching \ for 'Re[5]: FuzzyTech MP' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/index.htm?key=fuzzytech+RE%5D
Search entire site for: 'FuzzyTech MP'.

Truncated match.
PICList Thread
'Re[5]: FuzzyTech MP'
1996\04\08@110305 by Pete Brink

picon face
     It might help if everyone were aware of what actually happens in
     a fuzzy system.  The fuzzy-Tech system actually creates a series
     of PID or PI control algorithms depending upon the input variables
     to the control.  A different control for each set of inputs.

     This is the main advantage of a fuzzy system.  It can model
     multivariate control outputs (and inputs) much more easily than a
     traditional PID.  To perform an engine control based upon numerous
     inputs and controlling two outputs (fuel flow and torque) using a
     PID required several hundred hours of development time using many
     different software iterations (It was a turbo-prop engine.)
     Furthermore, the two output system required a complex matrix
     system called H-Infinity to perform the calculations with the
     other output in mind.  A fuzzy system would probably be able to
     come up with a system to at least test the engine in a matter a
     days.

     You are all correct, you cannot verify a fuzzy system.  And for
     that reason alone I would not use it on an aircraft, not to
     mention that I prefer an optimized PID controller for both
     performance and economy.  However, in non-critical systems, the
     fuzzy system does create a method for creating a control system
     quickly and efficiently.  The only way an engineer will ever trust
     a fuzzy logic is to create a control using something like
     fuzzy-Tech MP and look at the assembly code it generates.  Much
     like using a C Compiler there will always be ways to optimize.

     Pete Brink
     Senior Software Engineer
     Microchip Technology, Inc.
     spam_OUTPete.BrinkTakeThisOuTspamMicrochip.com

1996\04\09@101819 by terogers

flavicon
face
Pete Brink wrote:

>  .... You are all correct, you cannot verify a fuzzy system.....

Dear Pete:

This might be the crux of a lot of the controversy surrounding fuzzy
logic.

Please define verify in this context.

Thanks - Tom Rogers  VP - R&D  Time Tech, Inc.

More... (looser matching)
- Last day of these posts
- In 1996 , 1997 only
- Today
- New search...