gsr fuzzylogic
Post on 19-Feb-2018
226 Views
Preview:
TRANSCRIPT
7/23/2019 GSR Fuzzylogic
http://slidepdf.com/reader/full/gsr-fuzzylogic 1/19
Artifcial Intelligence - Fuzzy Logic
-Gopal Singh Rawat
7/23/2019 GSR Fuzzylogic
http://slidepdf.com/reader/full/gsr-fuzzylogic 2/19
What is Fuzzy Logic?
• Fuzzy Logic (FL) is a etho! o" reasoning thatrese#les huan reasoning$ %he approach o" FLiitates the way o" !ecision a&ing in huans thatin'ol'es all intere!iate possi#ilities #etween !igital'alues S an! *+$
• %he con'entional logic #loc& that a coputer canun!erstan! ta&es precise input an! pro!uces a !efniteoutput as %R, or FALS which is e.ui'alent tohuan/s S or *+$
• Fuzzy Logic Systes (FLS) pro!uce accepta#le #ut!efnite output in response to incoplete a#iguous!istorte! or inaccurate ("uzzy) input$
7/23/2019 GSR Fuzzylogic
http://slidepdf.com/reader/full/gsr-fuzzylogic 3/19
• %he in'entor o" "uzzy logic Lotf A. Zadeho#ser'e! that unli&e coputers the huan
!ecision a&ing inclu!es a range o" possi#ilities#etween S an! *+ such as 0
• 2R%AI*L S
• 3+SSI4L S• 2A**+% SA
• 3+SSI4L *+
• 2R%AI*L *+
• %he "uzzy logic wor&s on the le'els o" possi#ilitieso" input to achie'e the !efnite output$
7/23/2019 GSR Fuzzylogic
http://slidepdf.com/reader/full/gsr-fuzzylogic 4/19
Ipleentation
• It can #e ipleente! in systeswith 'arious sizes an! capa#ilitiesranging "ro sall icro-controllersto large networ&e! wor&station-#ase! control systes$
• It can #e ipleente! in har!wareso"tware or a co#ination o" #oth
7/23/2019 GSR Fuzzylogic
http://slidepdf.com/reader/full/gsr-fuzzylogic 5/19
Why Fuzzy Logic?
• Fuzzy logic is use"ul "or coercialan! practical purposes$
• It can control achines an!consuer pro!ucts$
• It ay not gi'e accurate reasoning#ut accepta#le reasoning$
• Fuzzy logic helps to !eal with theuncertainty in engineering$
7/23/2019 GSR Fuzzylogic
http://slidepdf.com/reader/full/gsr-fuzzylogic 6/19
Fuzzy Logic Systems Architecture
•
It has "our ain parts as shown
7/23/2019 GSR Fuzzylogic
http://slidepdf.com/reader/full/gsr-fuzzylogic 7/19
• Fuzzifcation Module 0 It trans"ors thesyste inputs which are crisp nu#ers into
"uzzy sets$ It splits the input signal into f'esteps such as 0
• Knowledge Base 0 It stores IF-%5* rulespro'i!e! #y e6perts$
L3 6 is Large 3ositi'e
73 6 is 7e!iu 3ositi'e
S 6 is Sall
7* 6 is 7e!iu *egati'e
L* 6 is Large *egati'e
7/23/2019 GSR Fuzzylogic
http://slidepdf.com/reader/full/gsr-fuzzylogic 8/19
• Inerence !ngine 0 It siulates thehuan reasoning process #y a&ing "uzzy
in"erence on the inputs an! IF-%5* rules$
• "euzzifcation Module 0 It
trans"ors the "uzzy set o#taine! #y thein"erence engine into a crisp 'alue$
7/23/2019 GSR Fuzzylogic
http://slidepdf.com/reader/full/gsr-fuzzylogic 9/19
Mem#ershi$ Function
• %he e#ership "unctions wor& on "uzzy sets o"'aria#les$
• 7e#ership "unctions allow you to .uanti"y linguistic
ter an! represent a "uzzy set graphically$ Ae#ership "unction "or a "uzzy set A on theuni'erse o" !iscourse 8 is !efne! as 9A:8 ; <=>$
•5ere each eleent o" 8 is appe! to a 'alue#etween = an! >$ It is calle! e#ership 'alue or!egree o" e#ership$ It .uantifes the !egree o"e#ership o" the eleent in 8 to the "uzzy set A$
7/23/2019 GSR Fuzzylogic
http://slidepdf.com/reader/full/gsr-fuzzylogic 10/19
• 6 a6is represents the uni'erse o" !iscourse$
• y a6is represents the !egrees o" e#ership in the <= >inter'al$
• %here can #e ultiple e#ership "unctions applica#le to"uzzi"y a nuerical 'alue$ Siple e#ership "unctions areuse! as use o" cople6 "unctions !oes not a!! oreprecision in the output$
•
All e#ership "unctions "or L3 73 S 7* an! L* areshown as #elow 0
7/23/2019 GSR Fuzzylogic
http://slidepdf.com/reader/full/gsr-fuzzylogic 11/19
• %he triangular e#ership "unction shapes are ostcoon aong 'arious other e#ership "unction shapessuch as trapezoi!al singleton an! Gaussian$
•
5ere the input to @-le'el "uzzifer 'aries "ro ->= 'olts to>= 'olts$ 5ence the correspon!ing output also changes$
!%am$le o a Fuzzy Logic System
• Let us consi!er an air con!itioning syste with @-le'el "uzzylogic syste$ %his syste a!Busts the teperature o" aircon!itioner #y coparing the roo teperature an! thetarget teperature 'alue$
7/23/2019 GSR Fuzzylogic
http://slidepdf.com/reader/full/gsr-fuzzylogic 12/19
Algorithm• Cefne linguistic'aria#les an! ters$
• 2onstruct e#ership
"unctions "or the$• 2onstruct &nowle!ge#ase o" rules$
• 2on'ert crisp !ata into"uzzy !ata sets using
e#ership "unctions$("uzzifcation)
• 'aluate rules in therule #ase$ (in"erenceengine)
•
2o#ine results "roeach rule$ (in"erenceengine)
• 2on'ert output !ata intonon-"uzzy 'alues$
(!e"uzzifcation)
7/23/2019 GSR Fuzzylogic
http://slidepdf.com/reader/full/gsr-fuzzylogic 13/19
Logic Ce'elopent
Ste$ &' "efne linguistic (aria#les and terms
• Linguistic 'aria#les are input an! output 'aria#les in the "oro" siple wor!s or sentences$ For roo teperature col!war hot etc$ are linguistic ters$
%eperature (t) D E'ery-col! col! war 'ery-war hot
• 'ery e#er o" this set is a linguistic ter an! it can co'ersoe portion o" o'erall teperature 'alues$
Ste$ )' *onstruct mem#ershi$ unctions orthem
• %he e#ership "unctions o" teperature 'aria#le are as
shown 0
7/23/2019 GSR Fuzzylogic
http://slidepdf.com/reader/full/gsr-fuzzylogic 14/19
Ste$+' *onstruct ,nowledge #ase rules
•
2reate a atri6 o" roo teperature 'alues 'ersus targetteperature 'alues that an air con!itioning syste ise6pecte! to pro'i!e$
7/23/2019 GSR Fuzzylogic
http://slidepdf.com/reader/full/gsr-fuzzylogic 15/19
oomem$ /arget
0ery1*old
*old 2arm 3ot 0ery13ot
eryH2ol! *oH2hange 5eat 5eat 5eat 5eat
2ol! 2ool *oH2hange
5eat 5eat 5eat
War 2ool 2ool *oH2hange
5eat 5eat
5ot 2ool 2ool 2ool *oH2hange
5eat
eryH5ot 2ool 2ool 2ool 2ool *oH2hange
• 4uil! a set o" rules into the &nowle!ge #ase in the "or o" IF- %5*-LS structures$
Sr .4o.
*ondition Action
> IF teperatureD(2ol! +R eryH2ol!) A*CtargetDWar %5*
5eat
IF teperatureD(5ot +R eryH5ot) A*CtargetDWar %5*
2ool
J IF (teperatureDWar) A*C (targetDWar) %5* *oH2hange
7/23/2019 GSR Fuzzylogic
http://slidepdf.com/reader/full/gsr-fuzzylogic 16/19
Ste$ 5' 6#tain uzzy (alue
• Fuzzy set operations per"or e'aluation o" rules$ %heoperations use! "or +R an! A*C are 7a6 an! 7in
respecti'ely$ 2o#ine all results o" e'aluation to "or afnal result$ %his result is a "uzzy 'alue$
Ste$ 7' 8erorm deuzzifcation• Ce"uzzifcation is then per"ore! accor!ing to e#ership
"unction "or output 'aria#le$
7/23/2019 GSR Fuzzylogic
http://slidepdf.com/reader/full/gsr-fuzzylogic 17/19
A$$lication Areas o Fuzzy Logic
• Automoti(e Systems Autoatic Gear#o6es
Four-Wheel Steering
ehicle en'ironent control
• *onsumer !lectronic9oods
5i-Fi Systes 3hotocopiers
Still an! i!eo 2aeras
%ele'ision
• "omestic 9oods 7icrowa'e +'ens
Re"rigerators %oasters
acuu 2leaners
Washing 7achines
• !n(ironment *ontrol Air
2on!itionersKCryersK5eaters
5ui!ifers
%he &ey application areas o" "uzzy logic are as gi'en 0
Ad t "i d t
7/23/2019 GSR Fuzzylogic
http://slidepdf.com/reader/full/gsr-fuzzylogic 18/19
Ad(antages oFLSs
• 7atheatical concepts
within "uzzy reasoning are'ery siple$
• ou can o!i"y a FLS #y Busta!!ing or !eleting rules !ueto e6i#ility o" "uzzy logic$
• Fuzzy logic Systes can ta&eiprecise !istorte! noisyinput in"oration$
• FLSs are easy to constructan! un!erstan!$
•
Fuzzy logic is a solution tocople6 pro#les in allfel!s o" li"e inclu!inge!icine as it rese#leshuan reasoning an!!ecision a&ing$
"isad(antages oFLSs
• %here is no systeaticapproach to "uzzysyste !esigning$
• %hey areun!erstan!a#le onlywhen siple$
• %hey are suita#le "or
the pro#les which !onot nee! highaccuracy$
7/23/2019 GSR Fuzzylogic
http://slidepdf.com/reader/full/gsr-fuzzylogic 19/19
4i#liography K Sources :
• http:KKwww$seattlero#otics$orgKenco!erKarMNK"uzKin!e6$htl
• wi&ipe!ia$org
•
tutorialspoint$co
top related