Aw too bad.  Not fuzzy logic, just fuzzy thinking.  +ADw-G+AD4-

+ACI-Active Stochastic noise cancellation system+ACI-  sounds really cool, though.
Maybe I'll use that.


-----Original Message-----
From: Stephan Kotze +ADw-stephank+AEA-iafrica.com+AD4-
To: Lawrence Lile +ADw-lilel+AEA-toastmaster.com+AD4-
Date: Sunday, January 02, 2000 11:52 AM
Subject: Re: Is This Fuzzy Logic?


+AD4-Apologies for taking so long to reply, still recovering from celebrations.
+AD4-what you actually have there is an active stochastic noise cancellation
+AD4-system. very close to a Fast Fourier transform algorithm (your basically
+AD4-extracting a fundemental frequency). Systems like this use a learning
+AD4-algoritm (high end DSP's) to actively adapt. Your system has it hard coded
+AD4-by you.
+AD4-Sorry, its not fuzzy logic
+AD4-
+AD4-Stephan
+AD4------ Original Message -----
+AD4-From: Lawrence Lile +ADw-lilel+AEA-TOASTMASTER.COM+AD4-
+AD4-To: +ADw-PICLIST+AEA-MITVMA.MIT.EDU+AD4-
+AD4-Sent: Thursday, December 30, 1999 4:51 PM
+AD4-Subject: Is This Fuzzy Logic?
+AD4-
+AD4-
+AD4APg- Stephan,
+AD4APg-
+AD4APg-
+AD4APg- I wrote a very successful program that I claimed used Fuzzy Logic, but I
+AD4APg- really have no idea if this fits the definition.  As you wrote your
+AD4APg- dissertation on the subject, you can probably answer with authority:  Is
+AD4APg- This Fuzzy?  Here's a synopsis:
+AD4APg-
+AD4APg- I am measuring a very noisy signal, represented at my PIC by a
+AD4-capacitance.
+AD4APg- Using AN512 techniques, I turn this into an RC time delay measured at the
+AD4APg- TMR0 input.  My program waits around looking for a TMR0 increment, like
+AD4-this
+AD4APg- pseudocode:
+AD4APg-
+AD4APg-
+AD4APg- +ADsAOwA7-  Measurement - wait for TMR0 to go to 1
+AD4APg- TMR0 +AD0- 0+ADs-
+AD4APg- if TMR0 +AD0- 1
+AD4APg-               goto +ACM-1
+AD4APg- if TRM0 +AD0- 1
+AD4APg-               goto +ACM-2
+AD4APg- ....... etc. etc.
+AD4APg- if TRM0 +AD0- 18
+AD4APg-               Goto +ACM-18
+AD4APg-
+AD4APg- +ADsAOwA7- Assign wieghts to each measurement
+AD4APg-
+AD4APg- +ACM-1  Weight +AD0- 1
+AD4APg-             goto analyze+AF8-weight
+AD4APg- +ACM-2  Weight +AD0- 2
+AD4APg-             goto analyze+AF8-weight
+AD4APg- +ACM-3  Weight +AD0- 3
+AD4APg-             goto analyze+AF8-weight
+AD4APg- ..... etc. etc.
+AD4APg- +ACM-18 Weight +AD0- 18
+AD4APg-             goto analyze+AF8-weight
+AD4APg-
+AD4APg- I wrote it in straight line assembler (which I will spare you from)  for
+AD4APg- speed.  Looped code was way too slow for my purposes.
+AD4APg-
+AD4APg- The WEIGHT variable is used to analyze my data.  If the weight was less
+AD4-than
+AD4APg- 2 or more than 16, I threw it out as a noise pulse.  I then add the
weight
+AD4APg- of 15 successive measurements.  If the resulting sum was greater than
135,
+AD4APg- my machine turns on.  If the resulting sum is between 90 and 135, my
+AD4-machine
+AD4APg- stays in the state it was already in (hysteresis).  If the sum is less
+AD4-than
+AD4APg- 90, my machine turns off.  Practically, the device was really successful.
+AD4APg- The setpoints were picked by making dozens of measurements (with a PIC
+AD4APg- hooked up to some LED's for an output) in the actual machine (learning).
+AD4APg-
+AD4APg- The routine was developed because the signal is so noisy that even when
+AD4-the
+AD4APg- machine was supposed to be ON, sometimes a signal was recieved that was
+AD4APg- thoroughly in the OFF range.  But never were 15 signals in a row this far
+AD4APg- out of range.
+AD4APg-
+AD4APg- Is this an example of Fuzzy logic or +AFs-sic+AF0- fuzzy thinking (fuzzy meaning
+AD4APg- sloppy,  inaccurate, imprecise and nonscientific in the second usage+ADw-G+AD4-)
+AD4APg- ????
+AD4-