🚀 go-pugleaf

From article <1992Sep29.144132.8473@csi.jpl.nasa.gov>, by mwette@csi.jpl.nasa.gov (Matt Wette):
> 
> I've had a very limited exposure to Fuzzy Logic.  However, from what
> I've seen Fuzzy Logic is memoryless.  This seems to be a severe
> limitation.  What is done to rectify the situation.  One possibility
> is to use state estimators.  Do others exist.  I could imagine that
> someone would want to use a dynamic logic function which maps the
> input sets to the output sets.  Is this done?
> 
> Matt
> -- 
> mwette@csi.jpl.nasa.gov

Fuzzy logic is not memoryless if the input quantities have memory!
A prime example of this is a simple control loop that uses error
and error rate of change as input variables.  The rate of change
variable has memory in that it needs to know the previous value
as well as the present value.  Fuzzy logic is a way of processing
data.  What that data is depends on what information you need
to control your system.  And for the typical control application where
the fuzzy output variable is a differential value, it gets inherently
integrated so viola -- you have memory again!

=======================
scotbri@rosemount.com
Scott Brigham
Rosemount, Inc.
Eden Prairie, MN
=======================