1

ECE 2221
SIGNALS AND SYSTEMS

ECE2221SignalsandSystems,Sem32010/2011,Dr.
SigitJarot
 CourseObjectives

Toprovideananalysisskillofthecontinuoustimesignals
andsystemsasreflectedtotheirrolesinengineering
practice.

Toexposestudentstoboththetimedomainand
frequencydomainmethodsofanalyzingsignalsand
systems.

Toillustratethepotentialapplicationsofthiscourseasa
Prerequisitecoursetocommunicationengineeringand
principles,digitalsignalprocessingandcontrolsystem.

ECE2221SignalsandSystems Dr.SigitPWJarot 2
 OBE(OutcomeBasedEducation)
LearningOutcomes
Aftercompletionofthiscoursethestudentswillbeableto:

Classify,characterizeandconductbasicofsignalsandsystems.

Analyzecontinuoustimesignalsandsystemsintimedomainusingconvolution.

AnalyzecontinuoustimesignalsandsystemsinfrequencydomainusingLaplacetransform.

AnalyzecontinuoustimesignalsandsystemsinfrequencydomainusingFourierseriesand
Fouriertransform.

Acquireintroductorylevelknowledgeofdiscretetimesignalsandsystems,andsampling
theory.

Workingrouptoperformbasicsimulationofsignalsandsystemsanalysis.

ECE2221SignalsandSystems Dr.SigitPWJarot 3
 CourseSynopsis

IntroductiontoSignals

IntroductiontoSystems

TimeDomainAnalysisofContinuousTimeSystems

FrequencyDomainSystemAnalysis:theLaplaceTransform

MIDTERMExamination

SignalsAnalysisusingtheFourierSeries

SignalsAnalysisusingtheFourierTransform

IntroductiontoDiscreteTimeSignalsandSystemsAnalysis

FINALExamination

ECE2221SignalsandSystems Dr.SigitPWJarot 4
 Outline
5

Conceptofsystemandsystemclassification
Lineartimeinvariantcontinuoustimesystems
Linearityandtimeinvariance

Causalityandstability

SystemRepresentation
Systemsrepresentedbydifferentialequations

Convolutionintegral

Systeminterconnection

ECE2221SignalsandSystems,Sem32010/2011,Dr.
SigitJarot
 SystemConcept
6

Systemisamathematicaltransformationofaninput
signal(orsignals)intoanoutputsignal(orsignals)

Input Output
Signal System Signal

SIGNAL
Asetofdataorinformation.
Afunctionorsequenceofvaluesthatrepresentsinformation.
Afunctionofoneormorevariables(e.g.time,frequency,space,..)that
conveysinformationonthenatureofaphysicalphenomenon.

SYSTEM
Asystemisanentitythatprocesses asetofsignals(inputs)toyield
anothersetsofsignals(outputs)

ECE2221SignalsandSystems,Sem32010/2011,Dr.
SigitJarot
 ResponseofaLinearSystem

Asystemsoutputfort 0 istheresultof2independentcauses:
1. Theinitialconditionatt = 0 zeroinputresponse
2. Theinputx(t) fort 0 zerostateresponse
Decompositionproperty:

Totalresponse=zeroinputresponse+zerostateresponse

x(t) y(t) = y0(t) + x(t) ys(t)

1 t
y (t ) vC (0) Rx(t ) x( )d t0
C 0
zeroinputresponsezerostateresponse
 TheimportanceofImpulseResponseh(t)

Zerostateresponseassumesthatthesystemisinrest
state,i.e.allinternalsystemvariablesarezero.
Derivingandunderstandingzerostateresponsedependson
knowingtheimpulseresponseh(t) toasystem.
Anyinputx(t) canbebrokenintomanynarrowrectangular
pulses.Eachpulseproducesasystemresponse.
Sincethesystemislinearand
timeinvariant,thesystem
responsetox(t) isthesumofits
responsestoalltheimpulse
components.
h(t)isthesystemresponsetothe
rectangularpulseatt=0 asthe
pulsewidthapproacheszero.
 ZerostateResponse(1)

Wenowconsiderhowtodeterminethesystem
response y(t) toaninputx(t) whensystemisin
zerostate.
Defineapulsep(t) ofunitheightandwidth at
t=0 .
Inputx(t) canberepresentedassumofnarrow
rectangularpulses.
Thepulseatt=n hasaheightx(t)=x(n) .
Thiscanbeexpressedasx(n)p (t-n).
Thereforex(t) isthesumofall[x(n)/ ] .such
pulses,i.e,
x(t ) lim x(n ) p (t n )
0

x(n )
lim p (t n )
0

 ZerostateResponse(2)

Theterm[x(n)/ ]p(t-n) representsapulsep(t-n) with

heightx(n).
As 0, heightofstrip ,butarearemainx(n) and

x(n )
p (t n ) x(n ) (t n )

x(t ) lim x(n ) (t n )

0

ZerostateResponse(3)
 ZerostateResponse(4)

x(t) y(t)
 ZerostateResponse(5)

Therefore, y (t ) lim x(n )h(t n )

0


y (t ) x( )h(t )d

Knowingh(t),wecandeterminetheresponsey(t) toanyinputx(t).
Observetheallpervasivenatureofthesystemscharacteristicmodes,which
determinestheimpulseresponseofthesystem.

LTISystem
x(t) y(t)
h(t)
y(t) = x(t) * h(t)

y (t ) x( )h(t )d

 Howtodeterminetheunitimpulseresponseh(t)(1)

Giventhatasystemisspecifiedbythe
followingdifferentialequation,determine
itsunitimpulseresponseh(t).
d2y dy dx
3 2 y (t )
dt 2 dt dt

Rememberthegeneralsystemequation:

Q( D) y (t ) P( D) x(t )
Itcanbeshownthattheimpulseresponseh(t) isgivenby:

h(t ) [ P( D) yn (t )]u (t )
whereu(t) istheunitstepfunction,andyn(t) isalinearcombinationofthe
characteristicmodesofthesystem.

yn (t ) c1e 1t c2 e 2t ...... c N e N t
 Howtodeterminetheunitimpulseresponseh(t)?(2)

Theconstantsci aredeterminedbythefollowinginitialconditions:

( N 2)
yn (0) y n (0) y n (0) .... y n (0) 0
( N 1)
yn (0) 1
Noteyn(k)(0) isthekth derivativeofyn(t) att=0.

TheaboveistrueifM,theorderofP(D) islesstheN,theorderofQ(D)
(whichisgenerallythecaseformoststablesystems).
 Example

Findtheunitimpulseresponseofasystemsspecifiedbythe
equation:
( D 2 3D 2) y (t ) Dx(t )
 Exercise1

Findtheunitimpulseresponseofasystemsspecifiedbythe
equation:
( D 4 D 3) y (t ) ( D 5) x(t )
2
 Exercise

Findtheunitimpulseresponseofasystemsspecifiedbythe
equation: ( D 2 4 D 3) y (t ) ( D 5) x(t )
 Exercise2

Findtheunitimpulseresponseofasystemsspecifiedbythe
equation:
( D 5 D 6) y (t ) ( D 7 D 11) x(t )
2 2
 Exercise
Findtheunitimpulseresponseofasystemsspecifiedbythe
equation:
( D 5 D 6) y (t ) ( D 7 D 11) x(t )
2 2
 TheConvolution Integral

Thederivedintegralequationoccursfrequentlyinphysical
sciences,engineering,andmathematics.
Itisgiventhename:theconvolutionintegral.
Theconvolutionintegraloftwofunctionx1(t) andx2(t) is
denotedsymbolicallyasx1(t)* x2(t)
Andisdefinedas
x1 (t ) x2 (t ) x ( )x (t )d

1 2

Notethattheconvolutionoperatorislinear,i.e.itobeysthe
principleofsuperposition.
 Properties ofConvolution(1)

CommutativeProperty (orderofoperandsdoesnotmatter):

AssociativeProperty(orderofoperatordoesnotmatter):

DistributiveProperty:
 Properties ofConvolution(2)

ShiftProperty:
if

then

also

ImpulseProperty
Convolutionofafunctionx(t) withaunitimpulseresultsinthe
functionx(t).
 Properties ofConvolution(3)

WidthProperty:
Durationofx1(t)=T1 ,anddurationofx2(t)=T2 ,

thendurationofx1(t)*x2(t)=T1 + T2

CausalityProperty:
 ConvolutionTable(1)

Usetabletofindconvolutionresultseasily:
ConvolutionTable(2)
ConvolutionTable(3)
 Whentheinputiscomplex

Whathappensifinputx(t) isnotreal,butiscomplex?
Ifx(t) = xr(t)+ jxi(t) ,wherexr(t) andxi(t) aretherealand
imaginarypartofx(t) ,then

y (t ) h(t ) xr (t ) jxi (t )
h(t ) xr (t ) jh(t ) xi (t )
yr (t ) jyi (t )

Thatis,wecanconsidertheconvolutionontherealand
imaginarypartseparately.
 Convolutionusing
GraphicalMethod

ECE2221SIGNALSANDSYSTEMS
 TheConvolution Integral

Thederivedintegralequationoccursfrequentlyinphysical
sciences,engineering,andmathematics.
Itisgiventhename:theconvolutionintegral.
Theconvolutionintegraloftwofunctionx1(t) andx2(t) is
denotedsymbolicallyasx1(t)* x2(t)
Andisdefinedas
x1 (t ) x2 (t ) x ( )x (t )d

1 2

Notethattheconvolutionoperatorislinear,i.e.itobeysthe
principleofsuperposition.
 ConvolutionIntegral

Theconvolutionintegraloftwofunctionx1(t) andx2(t) is
denotedsymbolicallyasx1(t)* x2(t)

Andisdefinedas
x1 (t ) x2 (t ) x ( )x (t )d

1 2

Systemoutput(i.e.zerostateresponse)isfoundbyconvolving
input x(t) withsystemimpulseresponseh(t) .

LTISystem
x(t) y(t)
h(t)
y(t) = x(t) * h(t)

y (t ) x( )h(t )d

 Convolution Integral

ThedefinitionoftheconvolutionintegralinLTIsystems:

Theoutput isthe sum/integrationof scaled and shifted impulseresponses

y (t ) x(t ) h(t ) x( )h(t )d

 Demonstrationofconvolutionsum
input function,x
6

0
-4 -2 0 2 4 6 8 10 12
point number n
system function,h
4

0
-4 -2 0 2 4 6 8 10 12
point number n
 convolutionsumforn=0
input-function,x[m] 5
Thevalueoftheconvolutionsumisthe
sumoftheproducttermsoverm:
0
0 5 10
y[n ] x[m] h[n m]
point number m m
h[n-m] for n = 0

4
Forn=0:
2
y[0] x[m] h[0 m]
0
0 5 10 m
point number m =0
Convolution Sum
1 50
product

0.5

0 0
0 5 10 0 5 10
point number m point number n
 convolutionsumforn=1

input-function,x[m]
5
Thevalueoftheconvolutionsumisthe
sumoftheproducttermsoverm:
0
0 5 10 y[n ] x[m] h[n m]
point number m m
h[n-m] for n = 1

4
Forn=1:
2
y[1] x[m] h[1 m]
0
0 5 10 m
point number m =0
Convolution Sum
1 50
product

0.5

0 0
0 5 10 0 5 10
point number m point number n
 convolutionsumforn=2
input-function,x[m] 5
Thevalueoftheconvolutionsumisthe
sumoftheproducttermsoverm:
0
0 5 10 y[n ] x[m] h[n m]
point number m m
h[n-m] for n = 2

4
Forn=2:
2
y[2] x[m] h[2 m]
0
0 5 10 m
point number m =1
Convolution Sum
1 50
product

0.5

0 0
0 5 10 0 5 10
point number m point number n
 convolutionsumforn=3

input-function,x[m] 5
Thevalueoftheconvolutionsumisthe
sumoftheproducttermsoverm:

0 y[n ] x[m] h[n m]

0 5 10
point number m m
h[n-m] for n = 3

4
Forn=3:
2
y[3] x[m] h[3 m]
0
0 5 10 m
point number m =4
Convolution Sum
2 50
product

0 0
0 5 10 0 5 10
point number m point number n
 convolutionsumforn=4

input-function,x[m] 5 Thevalueoftheconvolutionsumisthe
sumoftheproducttermsoverm:

0 y[n ] x[m] h[n m]

0 5 10 m
point number m
h[n-m] for n = 4

4
Forn=4:
2
y[4] x[m] h[4 m]
0
0 5 10
m
point number m = 10
Convolution Sum
4 50
product

0 0
0 5 10 0 5 10
point number m point number n
 convolutionsumforn=5

input-function,x[m] 5 Thevalueoftheconvolutionsumisthe
sumoftheproducttermsoverm:

0 y[n ] x[m] h[n m]

0 5 10
point number m
m
h[n-m] for n = 5

4
Forn=5:
2
y[5] x[m] h[5 m]
0 m
0 5 10
point number m = 20
Convolution Sum
50
5
product

0 0
0 5 10 0 5 10
point number m point number n
 convolutionsumforn=6

input-function,x[m] 5 Thevalueoftheconvolutionsumisthe
sumoftheproducttermsoverm:

0
y[n ] x[m] h[n m]
0 5 10 m
point number m
h[n-m] for n = 6

4
Forn=6:

y[6] x[m] h[6 m]

2

0 m
0 5 10
point number m
= 30
Convolution Sum
50
product

0 0
0 5 10 0 5 10
point number m point number n
 convolutionsumforn=7

input-function,x[m] 5
Thevalueoftheconvolutionsumisthe
sumoftheproducttermsoverm:

0 y[n ] x[m] h[n m]

0 5 10
point number m
m
h[n-m] for n = 7

4
Forn=7:
2
y[7] x[m] h[7 m]
0
0 5 10
m
point number m = 40
Convolution Sum
50
10
product

0 0
0 5 10 0 5 10
point number m point number n
 convolutionsumforn=8
input-function,x[m] 5
Thevalueoftheconvolutionsumisthe
sumoftheproducttermsoverm:
0
0 5 10 y[n ] x[m] h[n m]
point number m m
h[n-m] for n = 8

4
Forn=8:
2
y[8] x[m] h[8 m]
0 m
0 5 10
point number m = 43
Convolution Sum
50
product

10

0 0
0 5 10 0 5 10
point number m point number n
 convolutionsumforn=9

input-function,x[m] 5 Thevalueoftheconvolutionsumisthe
sumoftheproducttermsoverm:

0 y[n ] x[m] h[n m]

0 5 10
m
point number m
h[n-m] for n = 9

4
Forn=9:
2
y[9] x[m] h[9 m]
0 m
0 5 10
point number m = 38
Convolution Sum
20 50
product

10

0 0
0 5 10 0 5 10
point number m point number n
 convolutionsumforn=10

input-function,x[m]
5
Thevalueoftheconvolutionsumisthe
sumoftheproducttermsoverm:

0 y[n ] x[m] h[n m]

0 5 10
point number m
m
h[n-m] for n = 10

4
Forn=10:
2
y[10] x[m] h[10 m]
0 m
0 5 10
point number m = 24
Convolution Sum
50
20
product

10

0 0
0 5 10 0 5 10
point number m point number n
 convolutionsumforn=11
input-function,x[m]
5
Thevalueoftheconvolutionsumisthe
sumoftheproducttermsoverm:
0 y[n ] x[m] h[n m]
0 5 10
point number m m
h[n-m] for n = 11

4
Forn=11:
2
y[11] x[m] h[11 m]
0 m
0 5 10
point number m =0
Convolution Sum
1 50
product

0.5

0 0
0 5 10 0 5 10
point number m point number n
 AnimatedIllustrationsofConvolutionIntegral
46

http://www.jhu.edu/signals/convolve/
 GraphicalMethod
Stepbystep Procedure

1.Chooseonesignaltobex(t),theothertobeh(t) ;drawthembothin axis

2.ROTATEoneofthesignalh() aboutthe axis
3.SHIFTtherotatedsignalh(-) totherightbyt

1
3 2

y (t ) x(t ) h(t )

x ( )h (t
4
) d
5
4.MULTIPLYx() withtheflipped/shiftedversionofsignalh()
5.CalculatetheareaunderthisproductusingcorrectlimitofINTEGRATION
6.Step45producestheequationx(t)*h(t) overspecifiedregion
7.Repeatstep45forallregionsofinterest
 Example
ConvolutionusingGraphicalMethod

Convolvethefollowingtwofunctions:
f(t) g(t)
2 3

*
t t
2 2 2

Replacet withinf(t) andg(t)

Choosetoflipandslideg() sinceitissimplerandsymmetric
3 g(t)
Functionsoverlaplikethis:
2
f()

2
2+t 2+t
 Example
ConvolutionusingGraphicalMethod

Convolutioncanbedividedinto5parts
I. t<2
3 g(t)
Twofunctionsdonotoverlap
2
Areaundertheproductofthe f()
functionsiszero

2
II. 2 t <0 2+t 2+t

Partofg(t)overlapspartoff(t) 3 g(t)
Areaundertheproductofthe 2
f()
functionsis

2
2+t 2+t
2t
32 t
2t
2 2
3t 2
0 3( 2)d 3 2 2 2 62 t 2 6
0
 Example
ConvolutionusingGraphicalMethod

III. 0 t <2 3 g(t)

Here, g(t)completelyoverlapsf(t)
2
Areaundertheproductisjust f()
2
2

2

0 3 2 d 3
2 2 6 2+t
2
2+t
0

IV. 2 t <4 3 g(t)

Partofg(t)andf(t)overlap 2
f()
Calculatedsimilarlyto2 t <0

V. t4 2
2+t 2+t
g(t)andf(t)donotoverlap
Areaundertheirproductiszero
 Example
ConvolutionusingGraphicalMethod

Resultofconvolution(5intervalsofinterest):
0 for t 2 No overlap
3
t 2 6 for 2 t 0 Partial overlap
2
y (t ) f (t ) * g (t ) 6 for 0 t 2 Complete overlap
3 2
t 12 t 24 for 2 t 4 Partial overlap
2
0 for t 4 No overlap

y(t)
6

t
2 0 2 4
 Convolutionusinggraphicalmethod(1)

Determinegraphicallyy(t)=x(t)* h(t) forx(t)=e-t u(t) andh(t)=e-2t u(t) .

Remember:variableof
integrationis,nott .
Convolutionusinggraphicalmethod(2)
 TimeDomainRepresentationsofLTISystems

2.4TheConvolutionIntegral
1.Acontinuoustimesignalcanbeexpressedasaweightedsuperpositionof
timeshiftedimpulses.
Thesiftingpropertyoftheimpulse!


x(t) x( ) (t - )d (2.10)
-
Inputx(t) Outputy(t)
2.ImpulseresponseofLTIsystemH: LTIsystem
Output: H

y t H x t H

x t d Linearityproperty

y(t) x( )H{ (t - )}d (2.10)
-
3. h(t)=H{ (t)} impulseresponseoftheLTIsystemH
Ifthesystemisalsotimeinvariant,then
Atimeshiftedimpulsegeneratesa
H{ (t - )} h(t - ) (2.11) timeshiftedimpulseresponse
output
y(t) x( )h(t )d (2.12)
-
Fig.2.9.

SignalsandSystems_SimonHaykin&BarryVanVeen 56
 TimeDomainRepresentationsofLTISystems
Convolutionintegral:

x t h t x h t d

2.5ConvolutionIntegralEvaluationProcedure
1.Convolutionintegral:

y(t) x( )h(t )d (2.13)
-
2.Definetheintermediatesignal: =independentvariable,
wt x h t t=constant

h(t ) =h ( ( t))isareflectedandshifted(by t)versionofh().

3.Output:
Thetimeshiftt determinesthetime
y(t) w t ( )d (2.14) atwhichweevaluatetheoutputof
-
thesystem.

SignalsandSystems_SimonHaykin&BarryVanVeen 57
 TimeDomainRepresentationsofLTISystems
Procedure2.2: ReflectandShiftConvolutionIntegralEvaluation
1.Graphbothx()andh(t )asafunctionoftheindependentvariable .To
obtainh(t ),reflecth() about =0toobtainh( )andthenh( ) shiftby
t.
2.Beginwiththeshiftt largeandnegative.Thatis,shifth( ) tothefarlefton
thetimeaxis.
3.Writethemathematicalrepresentationfortheintermediatesignalwt ().
4.Increasetheshiftt (i.e.,moveh(t ) towardtheright)untilthemathematical
representationforwt ()changes.Thevalueoft atwhichthechangeoccurs
definestheendofthecurrentsetandthebeginningofanewset.
5.Lett beinthenewset.Repeatstep3and4untilallsetsofshiftst andthe
correspondingmathematicalrepresentationsforwt ()areidentified.This
usuallyimpliesincreasingt toaverylargepositivenumber.
6.Foreachsetsofshiftst,integratewt ()from = to = toobtainy(t).

SignalsandSystems_SimonHaykin&BarryVanVeen 58
 TimeDomainRepresentationsofLTISystems

Example2.6 ReflectandshiftConvolutionEvaluation
Given x t u t 1 u t 3 and h t u t u t 2 asdepictedinFig.210,
Evaluatetheconvolutionintegraly(t)=x(t) h(t).

Figure2.10(p.117)
InputsignalandLTIsystemimpulseresponseforExample2.6.

SignalsandSystems_SimonHaykin&BarryVanVeen 59
 TimeDomainRepresentationsofLTISystems

Figure2.10(p.117)
InputsignalandLTIsystemimpulseresponseforExample2.6.
<Sol.>
1.Graphofx()andh(t ):Fig.2.11(a).
2.Intervalsoftimeshifts:Fourintervals
1stinterval:t <1
2ndinterval:1t <3
3rdinterval:3 t <5
4thinterval:5t
3.Firstintervaloftimeshifts:t <1 wt()=0
4.Secondintervaloftimeshifts:1t <3
1, 1 t
wt Fig.2.11(b).
0, otherwise
SignalsandSystems_SimonHaykin&BarryVanVeen 60
 TimeDomainRepresentationsofLTISystems

Figure2.11 t
(p.118)
Evaluationofthe
convolutionintegral
forExample2.6.(a)
Theinputx()
depictedabovethe
reflectedandtime
shiftedimpulse
response.(b)The
productsignalwt()
for1 t <3.(c)The
productsignalwt()
for3 t <5.(d)The
systemoutputy(t).

SignalsandSystems_SimonHaykin&BarryVanVeen 61
 TimeDomainRepresentationsofLTISystems
5.Thirdinterval:3 t <5
1, t 2 3
wt Fig.2.11(c).
0, otherwise
6.Fourthinterval:5t wt()=0
7.Convolutionintegral:
1)Fort <1andt 5:y(t)=0
2)Forsecondinterval1t <3,y(t)=t 1
3)Forthirdinterval3t <5,y(t)=3 (t 2)
0, t 1
t 1, 1 t 3

y t
5 t , 3 t 5
0, t 5

SignalsandSystems_SimonHaykin&BarryVanVeen 62
Example2.7 RC CircuitOutput
FortheRC circuitinFig.2.12,assumethatthecircuitstimeconstantisRC =1sec.Ex.1.21
showsthattheimpulseresponseofthiscircuitish(t)=e t u(t).
Useconvolutiontodeterminethecapacitorvoltage,y(t),resultingfromaninputvoltagex(t)
=u(t) u(t 2).

Figure2.12(p.119)
RCcircuitsystemwiththevoltage
sourcex(t)asinputandthevoltage
measuredacrossthecapacitory(t),
asoutput.
 TimeDomainRepresentationsofLTISystems
<Sol.> RCcircuitisLTIsystem,soy(t)=x(t) h(t).
1.Graphofx()andh(t ):Fig.2.13(a).
1, 0 2 e t , t
h t e u t
t
x and
0, otherwise 0, otherwise
2.Intervalsoftimeshifts:Threeintervals
1stinterval:t <0
2ndinterval:0t <2
3rdinterval:2 t
3.Firstintervaloftimeshifts:t <0 wt()=0
4.Secondintervaloftimeshifts:0t <2
e t , 0 t
Fort >0, wt Fig.2.13(b).
0, otherwise
5.Thirdinterval:2 t
e t , 0 2
wt Fig.2.13(c).
0, otherwise

SignalsandSystems_SimonHaykin&BarryVanVeen 64
 TimeDomainRepresentationsofLTISystems
6.Convolutionintegral:
1)Fort <0:y(t)=0
2)Forsecondinterval0t <2:

0
t
y t e t d e t e 1 e
t

0
t

3)Forthirdinterval2t:
y t e
2

0
t
d e t
e e 1 e
2
0
2 t

0, t0

y t 1 e t , 0t 2 Fig.2.13(d).
2
e 1 e ,
t
t2

SignalsandSystems_SimonHaykin&BarryVanVeen 65
 TimeDomainRepresentationsofLTISystems
Figure2.13(p.120)
EvaluationoftheconvolutionintegralforExample2.7.(a)The
inputx()superimposedoverthereflectedandtimeshifted
impulseresponseh(t ),depictedasafunctionof.(b)The
productsignalwt()for0 t <2.(c)Theproductsignalwt()fort
2.(d)Thesystemoutputy(t).

SignalsandSystems_SimonHaykin&BarryVanVeen 66
 TimeDomainRepresentationsofLTISystems

Example2.8 AnotherReflectandShiftConvolutionEvaluation
Supposethattheinputx(t)andimpulseresponseh(t)ofanLTI systemare,respectively,
givenby
x t t 1 u t 1 u t 3 and h t u t 1 2u t 2
Findtheoutputofthesystem.

SignalsandSystems_SimonHaykin&BarryVanVeen 67
 TimeDomainRepresentationsofLTISystems
<Sol.>
1.Graphofx()andh(t ):Fig.2.14(a).
2.Intervalsoftimeshifts:Fiveintervals
1stinterval:t <0
2ndinterval:0t <2
3rdinterval:2 t<3
4thinterval:3t<5
5thinterval:t 5
3.Firstintervaloftimeshifts:t <0 wt()=0
4.Secondintervaloftimeshifts:0t <2
1, 1 t 1
wt Fig.2.14(b).
0, otherwise
5.Thirdintervaloftimeshifts:2t <3
1, 1 3
wt Fig.2.14(c).
0, otherwise
6.Fourthintervaloftimeshifts:3t <5

SignalsandSystems_SimonHaykin&BarryVanVeen 68
 TimeDomainRepresentationsofLTISystems
Figure2.14(p.121)EvaluationoftheconvolutionintegralforExample2.8.
(a)Theinputx()superimposedonthereflectedandtimeshiftedimpulse
responseh(t ),depictedasafunctionof.(b)Theproductsignalwt()for
0 t <2.(c)Theproductsignalwt()for2 t <3.(d)Theproductsignalwt()
for3 t <5.(e)Theproductsignalwt()fort 5.Thesystemoutputy(t).

SignalsandSystems_SimonHaykin&BarryVanVeen 69
 Interconnected Systems(1)

Parallelconnectedsystem

y (t ) h1 (t ) x(t ) h2 (t ) x(t )

Cascadesystem

y (t ) [h1 (t ) h2 (t )] x(t )
Andcommutativeproperty
 Interconnected Systems(2)

Integration

Alsotruefordifferentiation

Letx(t)=(t) andy(t)=h(t)
(x(t)isanimpulse,andh(t) istheimpulseresponseofthesystem),
Theng(t),thestepresponseis: t
g (t )

h( )d
 Total Response

Asystemsoutputfort 0 istheresultof2independentcauses:
1. Theinitialconditionatt = 0 zeroinputresponse
2. Theinputx(t) fort 0 zerostateresponse
Decompositionproperty:

Totalresponse=zeroinputresponse+zerostateresponse

x(t) y(t) = y0(t) + x(t) ys(t)

N
total response c e
k 1
k
k t
x (t ) h (t )

zeroinputcomponent zerostatecomponent