chapter 11 fourier series, integrals, and transforms 11.1 fourier...

101
Chapter 11 Fourier Series, Integrals, and Transforms 11.1 Fourier Series A function f (x) is called periodic if there is some positive number p such that R x x f p x f ), ( ) ( (1) p is called a period of f (x)

Upload: others

Post on 26-Jan-2021

101 views

Category:

Documents


13 download

TRANSCRIPT

  • Chapter 11 Fourier Series, Integrals,

    and Transforms

    11.1 Fourier Series

    A function f (x) is called periodic if there is

    some positive number p such that

    Rxxfpxf ),()((1)

    p is called a period of f (x)

  • constcxfxx )(,cos,sin

    are periodic functions.

    xexx ,, 2 are not periodic. Properties of periodic function:

    1. For any integer n,

    )()( xfnpxf (2)

    2. If f(x) and g(x) have period p, then

    )()()( xbgxafxh

    also has the period p.

  • 3. If a periodic function f (x) has a smallest

    period p(>0), then it is often called the

    fundamental period of f (x) .

    The fundamental period of xx cos,sin is 2

    The fundamental period of xx 2cos,2sin is

    constxf )( without fundamental period.

  • 2 is the common period of

    ,sin,cos,,2sin,2cos,sin,cos,1 nxnxxxxx

    A trigonometric series has the form

    1

    0

    22110

    )sincos(

    2sin2cossincos

    n

    nn nxbnxaa

    xbxaxbxaa (4)

    (4) can be used to represent any periodic

    function with the period 2

    We call (4) as Fourier series.

  • Assume the f(x) is a periodic function of period

    that can be represented by a trigonometric

    series,

    2

    1

    0 )sincos()(n

    nn nxbnxaaxf(5)

    We want to determine the coefficients nn ba and

    of the corresponding series (5).

  • dxxfa

    dxnxbdxnxadxadxxfn

    nn

    )(

    2

    1

    sincos)(

    0

    1

    0

    Similarly, we have

    1

    0

    cossincoscos

    coscos)(

    n

    nn mxdxnxbdxmxnxa

    mxdxamxdxxf

  • 0)sin(2

    1)sin(

    2

    1

    cossin

    0

    )cos(2

    1)cos(

    2

    1

    coscos

    ,0cos

    dxxmndxxmn

    dxmxnx

    mn

    mn

    dxxmndxxmn

    dxmxnx

    dxmx

  • Thus we obtain the result

    ,2,1 ,cos)(1

    mdxmxxfam

    Finally, we got

    1

    0

    sinsinsincos

    sinsin)(

    n

    nn mxdxnxbdxmxnxa

    mxdxamxdxxf

  • 0)sin(2

    1)sin(

    2

    1

    sincos

    0

    )cos(2

    1)cos(

    2

    1

    sinsin

    ,0sin

    dxxmndxxmn

    dxmxnx

    mn

    mn

    dxxmndxxmn

    dxmxnx

    dxmx

  • Thus we obtain the result

    ,2,1 ,sin)(1

    mdxmxxfbm

    Euler formulas:

    ,2,1sin)(1

    )(

    ,2,1cos)(1

    )(

    )(2

    1)( 0

    ndxnxxfbc

    ndxnxxfab

    dxxfaa

    n

    n

    (6)

  • Fourier series of f (x) with (6)

    1

    0 )sincos(n

    nn nxbnxaa

    Ex. 1 Find the Fourier coefficients of f (x)

    xk

    xkxf

    0

    0)( and

    )()2( xfxf

    (7)

  • Solution. f (x) is an odd function, so we have

    n

    nn

    k

    nn

    knx

    n

    k

    dxnxk

    dxnxxfb

    dxnxxfa

    dxxfa

    n

    n

    even 0

    odd4

    )cos1(2

    cos2

    sin2

    sin)(1

    0cos)(1

    0)(2

    1

    0

    0

    0

  • Thus the Fourier series of f(x) is

    xxx

    k5sin

    5

    13sin

    3

    1sin

    4

    (8)

    2/xFurthermore, we set

    7

    1

    5

    1

    3

    11

    4

    25sin

    5

    1

    23sin

    3

    1

    2sin

    4

    2

    k

    kkf

    47

    1

    5

    1

    3

    11

    (by Leibniz, 1673).

  • The trigonometric system

    In formulas, for any integers m and n,

    ,sin,cos,,2sin,2cos,sin,cos,1 nxnxxxxx

    is orthogonal on the interval x

    ) (including 0sincos

    )( 0coscos

    )( 0sinsin

    nmdxmxnx

    nmdxmxnx

    nmdxmxnx

  • Given a piecewise continuous function f (x)

    of period , we can get the coefficients

    according to (6) and then obtain a Fourier series

    of f(x), denoted as

    2

    1

    0 )sincos(~)(n

    nn nxbnxaaxf

    Under what condition we get to know

    ?)sincos()(1

    0

    n

    nn nxbnxaaxf

  • Theorem 1 If a periodic function f (x) is

    piecewise continuous in the interval

    and has a left-hand derivative and right-hand

    derivative at each point of the interval, then

    x

    1

    0 )sincos(2

    )0()0(

    n

    nn nxbnxaaxfxf

    Particularly, if f (x) is continuous at 0x

    1

    0000 )sincos()(n

    nn nxbnxaaxf

  • Ex. 2 In example 1, f (x) has a jump at 0x

    kfkf )00(,)00( . So we obtain

    2

    )00()00(

    0

    0*5sin5

    10*3sin

    3

    10sin

    4

    ff

    k

  • 11.2 Functions of Any Period p=2L

    We want to find a trigonometric series to

    represent a function f (x) of period p=2L, where

    . Let , then g (t) has

    period . Thus, we have

    0L )/()( Ltftg

    2

    1

    0 )sincos(~)(n

    nn ntbntaatg

    where

  • ,2,1sin)(1

    )(

    ,2,1cos)(1

    )(

    )(2

    1)( 0

    ndtnttgbc

    ndtnttgab

    dttgaa

    n

    n

    We replace t by , then Lx /

    1

    0 )sincos(~)()(n

    nnL

    xnb

    L

    xnaa

    L

    xgxf

    (5)

  • with Fourier coefficients

    ,2,1sin)(1

    )(

    ,2,1cos)(1

    )(

    )(2

    1)( 0

    ndxL

    xnxf

    Lbc

    ndxL

    xnxf

    Lab

    dxxfL

    aa

    L

    Ln

    L

    Ln

    L

    L

    (6)

  • here we only verify (b)

    dxLxnxfL

    LxdLxnLxg

    Lxtdtnttga

    L

    L

    L

    L

    n

    )/cos()(1

    )/()/cos()/(1

    )/ setting(by cos)(1

  • Ex. 1 Find the Fourier series of

    2,42

    210

    11

    120

    )(

    LLp

    x

    xk

    x

    xf

    Solution. It is an even function. 0nb

  • ,2,1,12

    )12(

    2)1(20

    2sin

    2

    2cos

    2cos)(

    2

    124

    1)(

    4

    1

    1

    1

    0

    2

    2

    1

    1

    2

    20

    mmn

    m

    kmn

    n

    n

    k

    dxxn

    kdxxn

    xfa

    kdxkdxxfa

    m

    n

    Hence the result is

    ,5,3,1

    2

    5cos

    5

    1

    2

    3cos

    3

    1

    2cos

    2

    2)(

    x

    xxxkk

    xf

  • Ex. 2 Find the Fourier series of f (x)

    2,42 20

    02)(

    LLp

    xk

    xkxf

    Solution. f (x) is an odd function, so we have

    0/cos)(1

    0)(2

    10

    dxLxnxfL

    a

    dxxfL

    a

    L

    Ln

    L

    L

  • n

    nn

    k

    nn

    knx

    n

    k

    dxnxk

    dxLxnxfL

    bL

    Ln

    even 0

    odd4

    )cos1(2

    cos2

    sin2

    /sin)(1

    0

    0

    Thus the Fourier series of f (x) is

    2/5sin

    5

    12/3sin

    3

    12/sin

    4xxx

    k

  • Ex. 3 Find the Fourier series of

    LLp

    LttE

    tLtu ,

    22

    0sin

    ,00)(

    Solution. By formulas (6) we have

    Edxx

    EdttEa 0

    /

    00 sin

    2sin

    2

  • 11

    )1cos(

    1

    )1cos(

    2

    12sin2

    ])1sin()1[sin(2

    )]1sin()1[sin(2

    cossin

    0

    0

    0

    /

    0

    /

    0

    nn

    xn

    n

    xnE

    ndxxE

    dxxnxnE

    dttntnE

    dttntEan

  • mnnn

    Emn

    mnnn

    Emn

    nn

    n

    n

    nEn

    2)1)(1(

    2120

    21

    1

    1

    1120

    11

    1)1cos(

    1

    1)1cos(

    2

    10

  • 10

    12/

    11

    )1sin(

    1

    )1sin(

    2

    1]12[cos2

    ])1cos()1[cos(2

    )]1cos()1[cos(2

    sinsin

    0

    0

    0

    /

    0

    /

    0

    n

    nE

    nn

    xn

    n

    xnE

    ndxxE

    dxxnxnE

    dttntnE

    dttntEbn

  • Thus

    tt

    Et

    EEtu

    4cos

    53

    12cos

    31

    12sin

    2)(

  • 11.3 Even and Odd Functions

    Half-Range Expansions

    Properties of the even or odd functions:

    1.If g (x) is an even function, then

    dxxgdxxgLL

    L 0 )(2)(

    2.If h (x) is an odd function, then

    0)( dxxhL

    L

  • Theorem 1 The Fourier series of an even

    function of period 2L is a cosine series

    1

    0 cos)(n

    nL

    xnaaxf

    with Fourier coefficients

    ,2,1,cos)(2

    ,)(1

    0

    00

    ndxL

    xnxf

    La

    dxxfL

    a

    L

    n

    L

    (1)

    (2)

  • The Fourier series of an odd function of

    period 2L is a sine series

    1

    sin)(n

    nL

    xnbxf

    with Fourier coefficients

    ,2,1,sin)(2

    0 ndxL

    xnxf

    Lb

    L

    n

    (3)

    (4)

  • In the case of period , we give for an even

    function simply 2

    1

    0 cos)(n

    n xnaaxf

    with coefficients

    ,2,1,cos)(2

    ,)(1

    0

    00

    ndxxnxfa

    dxxfa

    n

    (1*)

    (2*)

  • In the case of period , we give for an odd

    function simply 2

    1

    0 sin)(n

    n xnbbxf

    with coefficients

    (3*)

    ,2,1,sin)(2

    0 ndxxnxfbn

    (4*)

  • Theorem 2 The Fourier coefficients of a sum

    are the sums of the corresponding

    Fourier coefficients of The Fourier

    coefficients of cf are c times the corresponding

    Fourier coefficients of f

    21 ff

    21 and ff

  • Ex. 1. Find the Fourier series of f (x)

    xk

    xxf

    02

    00)(

    )()2( xfxf

    Solution. )()( xgkxf where

    xk

    xkxg

    0

    0)(

    is an odd function.

  • For g (x), we have

    n

    nn

    k

    nn

    knx

    n

    k

    dxnxk

    dxnxxgbn

    even 0

    odd4

    )cos1(2

    cos2

    sin2

    sin)(1

    0

    0

    xxx

    kxg 5sin

    5

    13sin

    3

    1sin

    4~)(

    ,0,5sin

    5

    13sin

    3

    1sin

    4)( xxxx

    kkxf

  • Ex. 2 Find the Fourier series of f (x)

    xxf )( and )()2( xfxf

    Solution. xxf )( x is odd.

    nn

    nxn

    x

    dxnxxdxnxxbn

    cos2

    cos2

    sin2

    sin1

    0

    0

    xxxxxf ,3sin

    3

    12sin

    2

    1sin2)(

  • Half-Range Expansions

    Given a function Lxxf 0 ),(

    we can extent to

    1) the even extension; 0),()( xLxfxf

    2) the odd extension; 0),()( xLxfxf

    We say that both 1) and 2) are half-range

    expansions.

  • Ex. 3 Find the two half-range expansions of

    LxLxLL

    k

    LxxL

    k

    xf

    2/),(2

    2/0,2

    )(

    Solution. (a) Even periodic extension.

    288

    2

    )(221

    22

    2

    2/

    2/

    00

    kLL

    L

    k

    dxxLL

    kdxx

    L

    k

    La

    L

    L

    L

  • nydyyknydyynydyy

    k

    nydyynydyyk

    dxL

    xnxL

    L

    kdx

    L

    xnx

    L

    k

    La

    n

    n

    L

    L

    L

    n

    cos)1(14

    cos)1(cos4

    cos)(cos4

    cos)(2

    cos22

    2/

    02

    2/

    0

    2/

    02

    2/

    2/

    02

    2/

    2/

    0

    Thus 0na for the odd n

  • oddmm

    kevenm

    mm

    k

    mym

    mm

    k

    dymym

    myym

    k

    mydyyk

    a m

    22

    22

    2/

    022

    2/

    0

    2/

    02

    2/

    022

    40

    cos12

    2cos4

    1sin

    4

    8

    2sin2

    12sin

    2

    18

    2cos8

  • the even half-range expansion is

    x

    Lx

    Lx

    L

    kkxf

    10cos

    5

    16cos

    3

    12cos

    4

    2)(

    222

    (b) Odd periodic extension.

  • nydyyknydyynydyy

    k

    dxL

    xnxL

    L

    kdx

    L

    xnx

    L

    k

    Lb

    n

    n

    L

    L

    L

    n

    sin)1(14

    sin)1(sin4

    sin)(2

    sin22

    2/

    0

    1

    2

    2/

    0

    12/

    02

    2/

    2/

    0

    Thus 0nb for the even n

  • For the odd n

    x

    Lx

    Lx

    L

    kxf

    5sin

    5

    13sin

    3

    1sin

    8)(

    222

    2sin

    8sin

    18

    coscos18

    sin8

    22

    2/

    02

    2/

    0

    2/

    02

    2/

    02

    n

    n

    kny

    nn

    k

    dynynyyn

    k

    nydyyk

    bn

    Hence the odd half-range expansion is

  • 11.4 Complex Fourier Series

    The Fourier series

    1

    0 )sincos()(n

    nn nxbnxaaxf(1)

    can be written in complex form by the Euler’s

    formula

    nxinxe inx sincos (2)

  • inxinxinxinx eei

    nxeenx 2

    1sin ,

    2

    1cos

    Thus we obtain

    inx

    nn

    inx

    nn

    inxinx

    n

    inxinx

    n

    nn

    eibaeiba

    eebi

    eea

    nxbnxa

    )(2

    1)(

    2

    12

    1

    2

    1

    sincos

  • We write

    nnnnnnn ccibacibaca 2/)(,2/)(,00

    Then the Fourier series (1) becomes

    n

    inx

    n

    n

    inx

    n

    inx

    n

    n

    nn

    ecececc

    nxbnxaaxf

    1

    0

    1

    0

    )(

    )sincos()((4)

    ,2,1,0 ,)(2

    1

    ndxexfc inxn

    (5)

  • For the function of period 2L, the complex

    Fourier series gives as

    n

    Lxin

    necxf/)( (6)

    where

    ,2,1,0 ,)(2

    1 / ndxexf

    Lc

    L

    L

    Lxin

    n

    (7)

  • Ex. 1 Find the complex Fourier series of a 2

    periodic function xexfx ,)(

    Solution. We note that nine )1(

    sinh)1(

    )1()1(

    )()1)(1(2

    )1()1(

    )1(2

    1

    2

    1

    2n

    in

    eeinin

    in

    ein

    dxeec

    n

    n

    inxxxinx

    n

  • Hence the complex Fourier series is

    xen

    ine inx

    n

    nx ,1

    1)1(

    sinh2

    From this let us derive the real Fourier series

    )sincos()sin(cos )sin)(cos1()1(

    nxnxninxnnxnxinxinein inx

    Omitting the imaginary part, we have

  • x

    nxnnxn

    nxnnxn

    e

    n

    n

    n

    nx

    )sin(cos1

    1)1(21

    sinh

    )sin(cos1

    1)1(

    sinh

    20

    2

  • 11.6 Approximation by Trigonometric

    Polynomials

    N

    n

    nn nxbnxaaxf1

    0 )sincos()((1)

    Is (1) a good approximation of f (x)?

    Let a trigonometric polynomial of degree N

    N

    n

    nnN nxBnxAAxF1

    0 )sincos()((2)

    is a best approximation of f (x), What is ? )(xFN

  • Or what is in (2), such that nn BAA ,,0

    dxFfE N

    2)((3)

    achieves the minimum for all trigonometric

    polynomials of degree N.

    (3) is called the total square error of relative

    to the function f (x) on the interval

    )(xFN

    x

    dxFdxfFdxf

    dxFfE

    NN

    N

    22

    2

    2

    )(

  • where

    N

    n

    nn

    N

    n

    nn

    N

    n

    nnN

    BAA

    dxnxBnxAA

    dxnxBnxAAdxF

    1

    222

    0

    1

    22222

    0

    2

    1

    0

    2

    )(2

    sincos

    sincos

    N

    n

    nnnn

    N

    n

    nnN

    bBaAaA

    dxnxfBnxfAfAdxfF

    1

    00

    1

    0

    )(2

    sincos

  • Thus we have

    N

    n

    nnnn

    N

    n

    nn

    N

    n

    nn

    N

    n

    nnnn

    bBaAaA

    baadxf

    BAA

    bBaAaAdxfE

    1

    222

    00

    1

    222

    0

    2

    1

    222

    0

    1

    00

    2

    ])()[()(2

    )(2

    )(2

    )(22

  • It is clear that E reaches the minimum

    0)(2min*1

    222

    0

    2

    N

    n

    nn baadxfEE

    (6)

    if and only if NN bBbBaAaA ,,,, 111100

    Hence we prove the theorem 1 below.

  • Theorem 1 The total square error of in

    (2) relative to the function f (x) on the interval

    is minimum if and only if the

    coefficients of in (2) are the Fourier

    coefficients of f (x). The minimum value is

    given by (6).

    x

    )(xFN

    )(xFN

  • The Bessel inequality

    .any for ,1

    )(2 2

    1

    222

    0 NdxfbaaN

    n

    nn

    (7)

    The Parseval’s identity

    dxfbaan

    nn

    2

    1

    222

    0

    1)(2(8)

  • Ex. 1 Compute the total square error of

    with N=3 relative to

    )( )( xxxf

    )(xFN

    Solution. xxf )( We have

    nn

    nxn

    x

    dxnxxdxnxxbn

    cos2

    cos2

    sin2

    sin1

    0

    0

  • xxxxF 3sin

    3

    12sin

    2

    1sin2)(3

    567.3]2[

    ])3/2(122[)(*

    94923

    38

    22222

    dxxE

  • Ex. 2 Compute the value of

    222

    1

    3

    1

    2

    11

    n

    Solution. xxxf ,)( We have

    nn

    nxn

    x

    dxnxxdxnxxbn

    cos2

    cos2

    sin2

    sin1

    0

    0

    xxxxf 3sin

    3

    12sin

    2

    1sin2~)(

  • According to the Parseval’s identity

    3

    2

    3

    2114

    2

    0

    32

    12

    xdxxnn

    So we have

    6

    1

    3

    1

    2

    11

    2

    222

    n

  • 11.7 Fourier Integrals

    Consider a periodic function of period

    2L that can be represented by Fourier series

    )(xfL

    L

    nwxwbxwaaxf n

    n

    nnnn

    ,)sincos()(1

    0

    According to the Euler’s formula, we have

    L

    Lnn

    n

    L

    LnLn

    L

    LLL

    vdvwvfxw

    vdvwvfxwL

    dxxfL

    xf

    sin)(sin

    cos)(cos1

    )(2

    1)(

    1

  • We now set LL

    n

    L

    nwww nn

    )1(1

    dwwvdvvfwx

    vdvwvfwxxf

    vdvwvfwxw

    vdvwvfwxw

    dxxfL

    xf

    L

    Lnn

    n

    L

    LnLn

    L

    LLL

    sin)(sin

    cos)(cos1

    )(

    sin)()sin(

    cos)()(cos1

    )(2

    1)(

    0

    1

    (1)

    (3)

  • We denote

    vdvwvfwB

    vdvwvfwA

    sin)(1

    )(

    ,cos)(1

    )(

    (4)

    Then we can rewrite this in the form

    dwwxwBwxwAxf sin)( cos)()(0

    (5)

    (5) is called a representation of f (x) by a

    Fourier integral.

  • Theorem 1 If f (x) is piecewise continuous in

    very finite interval and has a right-hand

    derivative and a left-hand derivative at every

    point and if the (4) is well defined, then f (x)

    can be represented by a Fourier integral (5).

    At a point where f (x) is discontinuous the value

    of the Fourier integral equals the average of the

    left- and right-hand limits of f (x) at the point.

  • Ex. 1 Find the Fourier integral representation

    of the function

    1||0

    111)(

    x

    xxf

    Solution. From (4) 0)( wB

    w

    wvdvwvdvwvfwA

    sin2cos

    2cos)(

    1)(

    1

    0

    and (5) gives the answer

    dww

    wwxxf

    0

    sincos2)(

  • 10

    14/

    102/sincos

    0

    x

    x

    x

    dww

    wwx

    We set 0x then

    2/sin

    0

    dww

    w

  • If f (x) is even, then B(w)=0, and then

    vdvwvfwA

    0

    cos)(2

    )(

    (5) then reduces to the Fourier cosine integral.

    dwwxwAxf

    0

    cos)()((11)

    (10)

    If f (x) is odd, then A(w)=0, and then

    vdvwvfwB

    0

    sin)(2

    )(

    dwwxwBxf

    0

    sin)()((13)

    (12)

    (5) then reduces to the Fourier sine integral.

  • If f (x) is defined on a half range, we can find

    the even or odd representation, respectively.

    Evaluation of integrals

    Ex. 2 Find the Fourier cosine and sine integrals

    of )0,0(,)( kxexf kx

    Solution. (a) from (10) we have

  • )(

    2

    cossin)(

    2

    cos2

    )(

    22

    0

    22

    0

    wk

    k

    wvwvk

    we

    wk

    k

    vdvwewA

    kv

    kv

    The Fourier cosine integral representation

    )0,0( ,cos2

    )(0 22

    kxdwwk

    wxkexf kx

    )0,0( ,2

    cos

    0 22

    kxekdw

    wk

    wx kx

  • (b) from (12) we have

    )(

    2

    cossin)(

    2

    sin2

    )(

    22

    0

    22

    0

    wk

    w

    wvwvk

    we

    wk

    w

    vdvwewB

    kv

    kv

    The Fourier sine integral representation

    )0,0( ,sin2

    )(0 22

    kxdwwk

    wxwexf kx

    )0,0( ,2

    sin

    0 22

    kxedwwkwxw kx

  • 0

    02/

    00

    1

    sincos

    0 2

    xe

    x

    x

    dww

    wxwwx

    x

    Ex. 3 Evaluation of Integrals for the follows

    Solution. From (4) we have for the rhs function

    20 1

    1coscos)(

    1)(

    wvdvwevdvwvfwA x

    20 1sinsin)(

    1)(

    w

    wvdvwevdvwvfwB x

  • So we have

    0

    02/

    00

    1

    sincos

    0 2

    xe

    x

    x

    dww

    wxwwx

    x

  • 11.8 Fourier Cosine and Sine Transforms

    For an even function f (x), we have

    dwwxwAxf

    0

    cos)()((1)

    with vdvwvfwA

    0

    cos)(2

    )(

    Then

    xdxwxfwfc

    0

    cos)(2

    )(ˆ

    and

    (3)

    (2)

    dwwxwfxf c

    0

    cos)(ˆ2

    )(

  • )(ˆ wfc is called the Fourier cosine transform of

    f (x), and f (x) is called the inverse Fourier

    cosine transform of )(ˆ wfc

    For an odd function f (x), we have

    dwwxwBxf

    0

    sin)()((4)

    with Then vdvwvfwB

    0

    sin)(2

    )(

  • xdxwxfwfs

    0

    sin)(2

    )(ˆ

    and

    (6)

    (5)

    dwwxwfxf s

    0

    sin)(ˆ2

    )(

    )(ˆ wfs is called the Fourier sine transform of

    f (x), and f (x) is called the inverse Fourier

    sine transform of )(ˆ wfs

    We also denote

    ffFffFffFffF scsscc sc }ˆ{ ,}ˆ{ ,ˆ}{ ,ˆ}{ 11

  • Ex. 1 Find the Fourier cosine and Fourier sine

    transforms of the function

    ax

    axkxf

    0

    0)(

    Solution. From (2) and (5) we have

    w

    awkxdxwkwf

    a

    c

    sin2cos

    2)(ˆ

    0

    w

    awkxdxwkwf

    a

    s

    cos12sin

    2)(ˆ

    0

  • Ex. 2 Find }{x

    c eF

    Solution. By (2) we get

    2

    02

    0

    1

    /2

    )sincos(1

    2

    cos2

    }{

    w

    wxwwxw

    e

    xdxweeF

    x

    xx

    c

  • It is easy to show that the Fourier cosine and

    sine transforms are linear operations,

    }{}{}{}{}{}{

    gbFfaFbgafFgbFfaFbgafF

    sss

    ccc

    (7)

    Theorem 1 Let f (x) be continuous and

    absolutely integrable on the x-axis, let f ’(x) be

    piecewise continuous on each finite interval,

    and let , then xasxf ,0)(

    )}({)}({

    )0(2

    )}({)}({

    xfwFxfF

    fxfwFxfF

    cs

    sc

    (8)

  • Proof. This follows from the definitions

    )};({)0(2

    sin)(cos)(2

    cos)(2

    )}({

    00

    0

    xfwFf

    wxdxxfwxwxf

    wxdxxfxfF

    s

    c

    )}({0

    cos)(sin)(2

    sin)(2

    )}({

    00

    0

    xfwF

    wxdxxfwxwxf

    wxdxxfxfF

    c

    s

  • It is easy to show by theorem 1 that

    )0(2

    )}({

    )}('{)}({

    )0(2

    )}({

    )0(2

    )}({)}({

    2

    2

    wfxfFw

    xfwFxfF

    fxfFw

    fxfwFxfF

    s

    cs

    c

    sc

    (9)

  • Ex. 3. Find the Fourier cosine transform of axexf )( where a>0

    Solution. By differentiation, )()(2 xfaxf

    22

    22

    2)}({

    )},({2

    )}({ )}({

    wa

    axfF

    xfFaxfFwxfF

    c

    ccc

  • 11.9 Fourier Transform

    We consider the Fourier integral

    dwwxwBwxwAxf sin)( cos)()(0

    vdvwvfwBvdvwvfwA

    sin)(

    1)( ,cos)(

    1)(

    Combining together, we have

    dvdwxvwvf

    dvdwwxwvwxwvvfxf

    )](cos[)(1

    sinsin coscos)(1

    )(

    0

    0

  • And then

    dvdwxvwvfxf )](cos[)(2

    1)(

    (1)

    We also get that

    0 )](sin[)(2

    1

    dvdwxvwvf

    (2)

    (1)-i(2) and resulting

    dvdwevf

    dvdwxvwixvwvfxf

    xviw

    )()(2

    1

    )](sin[)](cos[)(2

    1)(

    (4)

  • Here (4) can be rewritten as

    dwedvevfxf iwxiwv

    )(

    2

    1

    2

    1)(

    (5)

    Furthermore

    dxexfwf iwx)(2

    1)(ˆ

    (6)

    is called the Fourier transform of f (x), and

    dwewfxf iwx)(ˆ

    2

    1)(

    (7)

    is called the inverse Fourier transform of )(ˆ wf

    )(ˆ)}({ wfxfF )()}(ˆ{1 xfwfF

  • Sufficient for the existence of the Fourier

    transform (6) are the following two conditions:

    1. f (x) is piecewise continuous.

    2. f (x) is absolutely integrable on the x-axis.

    Ex. 1 Find the Fourier transform of f (x)=k if

    0

  • Ex. 2 Find the Fourier transform of 2

    )( axexf , where a>0. Solution.

    dxa

    iw

    a

    iwxa

    dxiwxaxeF ax

    22

    2

    22exp

    2

    1

    )exp(2

    1}{

    2

  • a

    w

    aaa

    w

    dvvaa

    w

    dxa

    iwxa

    a

    iw

    4exp

    2

    11

    4exp

    2

    1

    )exp(1

    4exp

    2

    1

    2exp

    2exp

    2

    1

    22

    22

    22

  • Theorem 1 The Fourier transform is a linear

    operation. That means

    )}({)}({)}()({ xgbFxfaFxbgxafF

    Proof. It is easy to get according to the

    definition of Fourier transform.

    Theorem 2 Let f (x) be continuous on the x-asis

    and Furthermore, let

    f ’(x) be absolutely integrable on the x-asis. Then

    || 0)( xasxf

    )}({)}({ xfiwFxfF (9)

  • Proof. From the definition, we have

    )}({)(2

    1

    )()(2

    1

    )(2

    1)}({

    xfiwFdxexfiw

    dexfexf

    dxexfxfF

    iwx

    iwxiwx

    iwx

    )}({)}({)}({ 2 xfFwxfiwFxfF

  • Ex. 3. Find the Fourier transform of 2

    )( xxexf

    Solution. Let 2

    )( xexg by Ex. 2, we have

    4/22

    2

    1}{)}({ wx eeFxgF

    Notice that )(22)(2

    xfxexg x

    4/2

    22)}({

    2)}({

    2

    1)}({ we

    iwxgF

    iwxgFxfF

  • The convolution f*g of f (x) and g (x) is defined

    by

    dvvgvxfdvvxgvfxgf )()()()())(*(

    (11)

    Theorem 4. Suppose that f (x) and g (x) are

    piecewise continuous, bounded, and absolutely

    integrable on the x-axis. Then

    }{}{2}*{ gFfFgfF (12)

  • Proof. By the definition

    }{){2

    )()(2

    1

    )()(2

    1

    )()(2

    1

    )()(2

    1}*{

    )(

    gFfF

    dpepgdvevf

    dvdpepgvf

    dvdxevxgvf

    dxdvevxgvfgfF

    iwpiwv

    pviw

    iwx

    iwx

  • By taking the inverse Fourier transform,

    )(ˆ)(ˆ))(*(

    dwewgwfxgf iwx(13)