Newsgroups: comp.robotics
Path: brunix!uunet!elroy.jpl.nasa.gov!swrinde!cs.utexas.edu!utnut!nott!cunews!bws-pc.carleton.ca!Brian_Sullivan
From: Brian_Sullivan@Carleton.CA (Brian_Sullivan)
Subject: Re: Fuzzy Logic ?
Message-ID: <Brian_Sullivan.12@Carleton.CA>
Sender: news@cunews.carleton.ca (News Administrator)
Organization: Carleton University
References: <C5JAtA.CvM@cbnewsk.cb.att.com> <1993Apr16.125034.21792@amoco.com>
Date: Mon, 19 Apr 1993 15:02:23 GMT
Lines: 99

In article <1993Apr16.125034.21792@amoco.com> zjoc01@hou.amoco.com (Jack O. Coats) writes:

>In article CvM@cbnewsk.cb.att.com, nastasi@cbnewsk.cb.att.com (joseph.l.nastasi) writes:
>>>In article <1993Apr15.100344.1@uwovax.uwo.ca>, lgardi@uwovax.uwo.ca writes:
>>>> Hi,
>>>>   I was wondering if anyone out there uses Fuzzy Logic in their robot
>>>> designs. 
>>>
>>>Okay, I'll bite: what is fuzzy logic and how can it make my life happier,
>>>my hair shinier, my robot more confused that it already is, etc...
>>>
>>>> <<<RED FISHY WINS>>>
>>>> Lori Gardi                         (519) 661-2111 ext 8695
>>>> Dept. of Astronomy,                        lgardi@uwovax.uwo.ca
>>>> University of Western Ontario
>>>> London, ON, CANADA, N6A 3K7
>>>
>>>Thanks in advance,
>>>Joe Nastasi
>>>nastasi@mtuxo.att.com

>Fuzzy logic is some of what you have been doing all your programming carreer.

>Example of an fuzzy intelegent thermostat:

>  Client says:                          Client means:
>    I want my house to stay                I want my house to stay in the 
>    at a comfortable temperature.          temperature range of 68F to 74F
>                                           I feel HOT above 74 and COLD below 68.

>   One implementation of this would be to:
>     1. have heater come on when temp reaches 68, raise temp to 72 and shut off.
>     2. have a/c come on when temp reaches 74, and cool to 70.

>   Results are that client stays in 'comfortable' range without having to readjust 
>   their thermostat.

>The thoughts behind fuzzy logic is users/clients state things in fuzzy or relative
>terms, i.e. cold/hot bright/dim little/much loud/quiet etc.  And what is meant sometimes
>overlaps.  Fuzzy logic takes these 'feelings' and 'perceptions' and (behind the
>scenes) quantifys them to allow responses to be put into real world control situations.

>Is this clear now?
>Let me know if I can help further... Jack jocoats@amoco.com
>---
>                         Ideas: mine              LawyerSpeak: theirs

Comp.ai.fuzzy has a nice FAQ that deals with the subject. 

Suffice to say that the about example simplifies things behond the point that
fuzzy logic applies.

With fuzzy logic somthing is not False(0) or True(1), but has a degree of 
truth between 0 and 1. A degree of truth means how true is it. This should 
not be confused with probability, which is the likelyhood of some thing 
being absolutely true or false.

That said, the real value of fuzzy logic is not that it allow things to be 
expressed in real world terms but that it has the mathmatics to work with 
these concepts.

Fuzzy logic could let the following:
	::temp 	= indoor temp
	::air	= outdoor temp
	::cycle = minutes heat/minutes cooling/ minutes nothing

	if temp<40 and air<40 
		then cycle=25/0/5
	if temp<40 and air>=40 and air <50
		then cycle=20/0/10
	if temp<40 and air>=50 and air <60
		then cycle=5/0/5
	.
	.
	.
	if temp>=40 and temp < 50 and air <40
		then cycle=20/0/10

	.
	.
	.

be replace by:
	if temp low and air low 
		then heat alot
	if temp high and air low
		don't do anything(it should cool down on it's own)
	if temp low and air high
		don't do anything(it should heat up on it's own)
	if temp high and air high
		then cool alot.

Basically fuzzy logic can allow many complex if-tehn rulles with simpler
fuzzy statmemts.

At least thats my understanding ...
and air low 		cool(
