184

4/THE ONESIDED LAPLACE TRANS'OIIi

may be written as:
jet) fr(t) * L 5(1 - fiT)
"-.
(e) Using the convolution theorem, verify the Laplace transform of the
periodic functionj(t) is:
(d) Find the inverse transform of:
Fr(s)
.r.
+-
I - e-'
G(s) sO _ e' 4')'
Re(s) >
and plot g(t). .. (I) b
(e) Plot the pole-zero diagram of G (s) and venfy your result findmg g y
the residue the9ry.
4-10. Given:
g(t) Lf(t - 2f1)
.-.
where
f(t) I,
0, otherwise
is the input to a system with system function H(s) 2/(s + 3).
zero-state output y (I).
4-11. (a) Find 8(s) - Y(s)/X(s) for the system shown.
(b) For what values of K is the system stable?
x(t) ' ___
---'=-{ + r

5 - 2
"-------"

y(t)
Chapter 5*
The Two-sided Laplace
Transform
INTRODUCTION
Chapter 5 treats the two-sided or bilateral Laplace transform whose main
application is the solving of L TIC continuous systems with random or signal plus
random inputs. The chapter considers the transform analysis of the continuous
material of Chapter 3.
The material is traversed by means of what is now our standard treatment
of any transform. The different stages are:
I. The transform is defined and a number of transforms are evaluated. We
utilize to the fullest our knowledge of one-sided transforms to evaluate
two-sided ones.
2. The properties and theorems are given and attention is focused on the
transform of convolution and correlation integrals.
3. The inverse transform is treated either by using previously mastered
partial fraction techniques referring to tables, or using the residue
theory from complex variables.
4. L TIC systems are solved for the cases of'random and signal plus noise
inputs.
5-1 DEFINITION AND EVALUATION
OF SOME TRANSFORMS
The two-sided or bilateral Laplace transform of a real function!(t) is defined
as:
(5-1)
186 5/THE TWO-SIDED LAPLACE TF''''8FORI
for complex s = tJ + jw. Fo(s) if it exists will do so for a region of
plane, 07, < Re(s) < 0'2' called the region of convergence. Normally, t
"8" is excluded and it will be clear from the context whether the
two-sided Laplace transform is being used. Again, alternate notations,
f(l) andf(t) -- F(:.') will be used.
A number of transforms will now be evaluated and the concept of
region of convergence will be expanded.
EXAMPLE 5-1
Find the two-sided Laplace transforms of the following functions and
the region of convergence:
(a) J; (I) = 3e
2
'u(t)
(b) 12(/) = 3e
21
u(
pc) h(t) = 3e-
2t
u(t) + 4e'u( -t)
L (d) 1,(1) - 3e-"u(l) - 4e'u(l)
(el I,(t) = _3e-
2
'u(_t) + 4e'u(-t)
(f) f6(t) = Ale"ru(t) + A
2
eO'u( -t) in generalforall appropriate real

Solution
Therefore
PI (s) [.,0 Oe-
JI
dt + [., 3r'e _sr dt
3
s 2'
Re(s) > 2
!J(t), FI(s), and the region of convergence are shown in Figure 5-1
(b) 1,(1) - 3e"u( -I)
F2(S) = J 0 3e
2r
e-sr dl
= J 0 3e(2-,)' dt
3
--- -0,
2 -,
since then = 0
Therefore
writing s = u + jw
ifu < 2
for2-u>O
Re(s) < 2
fJt), and the region of convergence are shown in Figure 5-1 (b). '
[(I)
3
3
,0) :H[""
'"
3e 11
'"
I "I
A, eO/
A I ,,0/
3
s -
-3
_, - 2
" (; - I)(s +
F(,.) = _3 ___ 4_
s-I
-,. - II

F(,) = _3_ _ 4_
,+J s-I
-s - 11
(s I)(s + 21
(AI - All, -A,tJ + Ale.
(, - "'1(, --
Region of Convergence
jw
Figure 5-1 The two-sided Laplace transforms and their regions of convergence for
Example 5-\,
188
\
5/THE TWO-SIDED LAPLAce TRANSFORM
(e) f3(t) = 3e-
2
'u(t) + 4e
H
u( -I)
F3(S) = 4e'e-" dt + dt
__ e(-l-s)/ + e(-l-S)I
4 I' 3 I"
I - s -$ - 2 0
--+0 if+O if +--
-4 1 1 3
s - 1 rr>-2 S + 2
-$ - 11
-2 < Re(s) < +1
For this function which exists for both positive and negative time we note
that the behavior for negative time puts an upper bound on an allowable (J
and the function behavior for positive time puts a lower hound on an '
allowable u. Therefore we obtain the strip of convergence Ul < (J' < U2.f3(t),
F3(s), and the region of convergence are plotted in Figure 5-1 (e).
(d) 14(/) = (3e-
21
- 4e+l)u(t)
3 1 4 1 F4(S) -- + 0 - -- + 0
s + 2 S - 1 ifa>+l
-$ - 11
Re(s) > 1
We notice the bilateral transform expression is the same as in part (e)
hut the region of convergence is different.f4(t). F4(s). and the region of '
convergence are shown in Figure 5-1 (d).
(e) fs(t) -3e-
2
,u( -t) + 4e'u( -t)
3 1 4 1 F,(s) --+ 0 - -- + 0
s+2 s-1
Therefore
-s - 11
F,(s) (s _ 1)(s + 2)'
Re(s) < -2
Since e-
2
,u( - t) is an increasing exponential of negative time its
convergence factor fT < - 2 determines the region of convergence.
Again, the Laplace transform expression is identical to parts (c) and
(d) but the region of convergence is different. f,(1), F5($), and the
region of convergence are pld'lted in Figure 5-1 (e).
(fl f6(t) = A1e"'u(t) + -t)
It can quite easily be shown or by now be clear that:
F6(S) + 0 1_ - + 0 I.
S - a S f3
(A, - A
2
)s - A
1
f3 + A
2
a
(s - a)(s - (3)
ifu->a n u-<f3
5-1 DEFINITION AND EVALUATION OF SOME TRANSFORMS 189
The bilateral Laplace transform will exist for all a and fl such that a <
f3 and then the region of convergence is a < Re(s) < fl. The different
possible situations for are shown in Figure 5-1 (C).
Two-sided Transforms Using One-sided
Transforms
The evaluation of two-sided Laplace transforms involves the same amount of
work as doing two one-sided Laplace transforms and indeed a table of one-sided
Laplace transforms may be used to find two-sided ones. We will now develop this
technique. Given:
f(l) f,(t)U(I) + f,(t)u( -I)
F(s) = FI(s) + J 0 f2(t)e-" dt
-,.
Letting t -p, we obtain:
and
dt = -dp, -00 < t < o gives 00 > p > 0
F(s) = Fl(S) + 1f2(-p)e'" (-dp)
F(s) F,(s) + Llf,( -I)U(I)] 1,--,
(5-2)
Iffl( -t)u(t) has a region of convergence fT > fT2. thenf2(t)u( -t) has a region of
convergence u < - u-;.
EXAMPLE 5-2
Using Equation 5-2. find the two-sided Laplace transform of:
(a) f(t) e"u( -I)
(b) f(l) 3e-
2
'u(t) + le-'u( -I)
Solution
(a) F(s) L[,-"u(l)] 1 , __ ,
=s!2Is __
(b) f(t) 3e-
21
u(t) + te-'u(-t)
Using Equation 5-2, we obtain:
F(s) - s! 2 + L[ -le'u(t)] 1 $ __ '
3 -1 1
---+ , '
s+2 (s-1) , __ s
u> -2 n u-<-1
190
5ITHE TWO-SIDED LAPLACE TRANSFt)R',
3
3$2+5s+1
(5 + 2)(s _ 1)2'
-2<0-< -1
Example 5-1 should now be repeated using Equation 5-1.
There are a few very important functions for which a two-sided L'pl"ce
transform does not exist. The functionsj;(t) = 1,f2{t) = Cos (wof + 4,
1(t) = 2:= __ ., g(t - nT) a periodic function, do not possess a Laplace '".nSlfo'm
as each function for t < 0 requires (Y < 0, whereas the function behavior for t >
requires u > O. Therefore, no value of s exists for which the transform conv,,,,,.,,,.
These functions are not to be confused with the causal functions:
f,(1) - u(1), [,(t) - Coo (wot + <I)u(1), ,nd f,(t) - g(1 - nn]u(1)
Finally, we are rarely interested in the two-sided Laplace transforms
functions whose transforms are not the ratio of two polynomials in s. As we
see, the most common functions for which we find two-sided Laplace 'n,",,fotcm.
are correlation functions for finite energy waveforms and autocorrelation
tions for ergodic noise waveforms. Both these functions are even and
positive or negative time may often be coIP".ent<,d by 'he p,oduc' of pol)m"miah
and exponentials.
5-2 IMPORTANT THEOREMS OF BILATERAL
LAPLACE TRANSFORMS
Table 4-2 of Chapter 4 listed an extensive set of theorems and properties
one-sided Laplace transform. Some important theorems for the bilatera.li
transform are given in Table 5-1. Since the convolution and correlation t:
are of the utmost importance when applying this material to systems
random inputs we will prove and demonstrate these. Discussion of the cmmpiex
convolution theorem is deferred until after the inverse transform is considered.
EXAMPLE 5-3
Prove the convolution theorem and comment on the region of conver-
gence.
Solution.
and
then
The convolution theorem states that if:
f(t) - F('),
g(t) - G(,),
f(t).g(t) - F(,)G(,)
52 IMPORTANT THEOREMS OF BILATERAL LAPLACE TRANSFORMS
TABLE 5-'
Theorem
Linearity
Time-scaling
Shifting
Convolution
Cros:;-correlation
Autocorrelation
Time function
ox(l) + by(l)
x(ot)
X(I - a)
x(t)*y(t)
X(p)y(1 - p) dp
x(l) ffi y(l)

X(I) x(r)
Two-sided Laplace tranaform
oX(s) + byes)
a> max (a." (1",,), a < min (a." a,,)
I:I
X
(;)
oO",<a<oO"" 0>0
00", < 0" < 00"" 0 < 0
e-'" Xes). 0", <0"<0",
Xes) yes),
Y(s)X( -s),
191
X(s)X( os), max(a" -O",)<a<min(O"" -0",)'
_ x(p)x(p _ I) dp
For causal functions all one-sided Laplace transform theorems carryover.
'X(I) _ XIs),
y(l) y(s),
O",<O"<fJ, or ".,<lI<<r.,
""Yo < "" < ""
By definition:
L[f(t).g(t)] - L"[l: f(p)g(t - p) d+-" dt
Interchanging the order of integration, we obtain:
L [f(t).g(t)] - L-f(P)[ L" g(1 - p ),-" dt] dp
Letting t - P = u:
dt = du, -co<t<co gives -co < u < co
L [f(t).g(t)] - 1: f(p )[1: g(u),-,>,t dU] dp
-J: f(p ),-"G(,) dp
- G(,)F(,) (5-3)
If we write R(s) = G(s)F(s), then R(s) will converge for all s for which
both G(s) and F(s) converge. For example, if F(s) converges -3 < (J < 1
and G(s) converges -2 < (J" < 4, then R(s) would converge - 2 < (J < I. In
general, R(.s) converges over the intersection of the points
(5-4)
192
5/THE TWO-SIDED LAPLACE
EXAMPLE 5-4 .
Find the transforms of the following functions and denote the regIOns
convergence if the transform exists:
(a) !J(t) = e-1'u(t)*e'u( -I)
(b) 12(/) = e
2
'u(t)*e-'u( -t)
(el e-
2I
,I*u(t)
Solution
(a) Let x(t) = e-
2t
u(t)
and then
Lety(t) = e'u(-t)
and then
1
X(.<)
s+
1
Y(s) - - ,
s - 1
By the convolution theorem:
rJ> -2
-1
L[x(t).y(t)[ (s + 2)(s _ I)'
-1
(1)-2(1(1<+1
(s + 2)(s I)'
-2<0"<+1
(b) Let X(/) = e1'u(t)
and then
Lety(t) = e-'u(-t)
and then
(J < - 1
By the convolution theorem:
-1
L[x(t).y(t)[ t< 2)(s + I)'
Since (J > 2 n (]' < - 1 = IjJ the Laplace transform does not exist.
(el Let X(I) = e-
21tl
= e-Uu(t) + e1'u( -I) and this has a Laplace trans-
form:
1 I -4
Xes) = s + 2 - s _ 2 (5 + 2)(s 2)'
Lety(t) = u(t)
5-2 IMPORTANT THEOREMS OF BILATERAL LAPLACE TRANSFORMS
and then
By the convolution theorem:
-4
L[x(t)'Y(I)[ s(.< + 2)(s _ 2)'
-4
+ 2){s 2)'
193
Our anticipation of evaluating inverse two-sided transforms should be
mounting since we now have an alternative to convolution,
EXAMPLE 5-5
Prove the correlation theorems:
(a) L[x(t) Ell Y(I)) Y(s)X( -,)
Ib) L[x(l) Ell x(t)) X(,)XI -s)
Solution
(a) By definition:
L[xlt) Ell y(t)) 1:[1: y(p)xlp - t) drje-" dl
Interchanging the order of integration, we obtain:
L[x(t) Ell Y(I)) L' y(p{L" x(p - tj,''' d}P
Substitute p t = I
L[xlt) Ell y(I)) ,L' Yip) [';:.-- x(lV'('-O (-dl)l dp
= yep )e-'P dp x(l)e'l dl = Y(05)X( (5-5)
where C(s) = Y(s)X( -s) converges for all s for which both Yes) and
X( -s) converge, To interprete the region of convergence for X( -s), we
note that if Xes) contains a pole at s = -51' then X( contains a pole
at = 51. With a little more thought we conclude that if the region of
convergence for Xes) iSO"xl < 0" < 0".. " then the region of convergence for
X( -05) is O"x, < -0" < O"Xj or -O"x, < 0" < and the region of
convergence for C(s) is the intersection of the points:
(5-6)
(b) In the casey{t) = x(t) we find by the croSs-correlation theorem that:
L[x(t) EEl x(r)J =
15-7)
194
5/THE TWO-SIDED LAPLACE TR,'N',"OR",
If the region of convergence for X(s) is 0"] < 0" < 0"2, then X(s)X( -s)
a region of convergence max(u] , -0"2) < 0" < min(O"l' -0"1) if it exists. ,
EXAMPLE 5-6
Find the Laplace transforms of the following correlation functions:
(a) e-
2
'u(t) E.B e-
2I
u(t) = C'"Jt)
(b) e--'u(t) E.B e
2t
u( -t) = (.-it)
(e) e
2t
u( -I) E.B e-'u(t) = Cy..-(/)
Solution
1
(a) e -2'U(t) ...... __
,+ 2
L[e-
2
'u(t) E.B e-
2t
u(t)] = I 1
(, + 2) (-, + 2)
1
(b) x(t) = e-'u(t) - - --,
.\' + 1
(
21 - 1
Y t) = e u( -t) ...... --,
,-2
-1
(s + 2)(5 - 2)'
-1
-2<0"<+2
0" > -I,
0" < -2
Therefore L[e-tu(t) E.B e
2I
u(_t)]
-I
~ - - - - - ' - ~
s - 2 -5 +
(, - 2)(, - 1)'
1
(, - 2)(, - I)'
(e) L[e2'u(-1)@e-'u(1)]='_I_(--=-'-)1
5 + I s - 2 s __ ,
1
(s + I)(s + 2)'
0">-1 nO">-2
(s + l)(s + 2)'
5-3 THE INVERSE TWO-SIDED LAPLACE TRANSFORM
195
The inverse Laplace transform will provide an alternative approach for
correlation. For example, when finding xU) EB yet) we have the choice of
evaluating:
[x(t) EB Y(I)] ~ 1"" y(p)x(p - I) dp or for example
0" > -I
Close examination of Example 5-6(a), (b), and (c) would lead to general
relations concerning symmetry of the poles for the Laplace transform of
correlation functions, which will be developed later in the chapter.
5-3 THE INVERSE TWO-SIDED LAPLACE
TRANSFORM
The uniqueness theorem for the inverse two-sided Laplace transform states:
,C'[.L(f(I))] ~ f ( l ) (5-8)
and in this section we discuss two techniques for finding inverses_ These are:
1. the use of partial fraction expansions plus table reference.
2. the classical evaluation from the formal definition of the inverse using
the residue theory from complex variables.
5-3-1 Inverse Transforms Using Partial Fraction
We confine our discussion to Laplace transforms which are the ratio of two
polynomials in s:
F(s) = N(s) = b",s'" + b",_I
Srn
-
1
+ ... + b
o
D(s) a/ + a
n
_
1
S
n
1 + ... + ao
Where the order of D(s) is at least one order higher than N(s). There is associated
with F(s) a region of convergence, 0"1 < 0" < 0"2, where D(s) contains two
consecutive poles at s = 0"1 + jWI and s = 0"2 + jW2 and we assume any
cancellation of poles by zeros has been carried out. From Section 5-1 we know
that a pole of F(s) at s = -S1 where Re( -Sl) < 0"1 contributes Ae-S"u(t) if 51 is
real and a pole of F(s) at s = -S2 contributes Be -"I u( - t) if ~ l is real and -S2 >
0"2 We now use previously learned partial fraction theory to find inverses_ A short
general table of two-sided transforms is given in Table 5-2, which we will utilize
when finding inverses.
5/THE TWO-SlOED LAPLACE TRAN"FC'A1i
5-2 SOME FUNDAMENTAL TWO-SIDED LAPLACE TRANSFORMS
F(s)
A
A
(s + a)
A
(s + a)'
A
(5 + a)'
A,s + A,
(s + a)' + fJ'
EXAMPLE 5-7
Region 01 convergence I{t)
(1) -DC
(1< -a
u> -a Ate-a'u(t)
(J< -(J(
<1> -a
a real, f1 real and positive
(1< -a
Find the inverse Laplace transforms of the following functions using
fractions:
35 + 2
(a) FJ(s) ~ (s + 1)2(s _ 2)'
-1<0"<2
3s + 2
(b) F
2
(s) = (s + \)2(S 2)'
0- < -1
3s + 2
(e) F3(S) = (s + 1)2(s _ 2)'
25 + 3
(d) F4(S) = (5 + 4)(S2 + 2s + 3)'
-4<0"<-1
Solution
35 + 2
(a) F1(s)-(s+ 1)2(s 2)' -1<u<2
A B C
---+ +--
s+1 (s+I)2 s-2
where C = 8/(3)2 = 0.89 and B = -1/-3 = 0.33.
find:
2 0.89
- ~ A + 0.33 + -2
-2 -
5-3 THE INVERSE TWO-SIDED LAPLACE TRANSFORM
Therefore
A = - I - 0.33 + 0.44
= -0.89
F (,) _ -0.89 0.33 0.89
1
5
+
1
+(5+1)2+
s
_
2
'
and from Table 5-2:
!J(t) = (-O.8ge-' + O.33te ')u(t) - O.8ge
2l
u(_t)
FI (s) and!, (t) are shown in Figure 5-2(a).
3,. + 2
(-0.89 + 0.33r)e
(a)
3s + 2
(J + 1 Ills -1) ,
(b)
je
(-0.89 + 0.331)-1 + 0.8ge11
k)
-,-II y'20_44 em (1.4/- 45) I
/
20 + 3
If
Is + 4)(,2 + 20' + 3)
-045 -4,
<0,
Figure 5-2 The transforms and their inverses for Example 5-6.
197
198
5/THE TWO-SIDED LAPLACE TAAN"FC)A'
(b)
F
2
(s) = 35+2
($ + Ircs - 2)'
0- < -1
-0.89 0.33 0.89
= + 7('-'-+-'-'--1 +
Therefore h{l) = [O.8ge-
t
- O.33Ir' _ a.8ge2'1 u( -I)
F2(S) and..r;(t) arc shown in Figure 5-2(b).
(e)
FJ(s) = 3$ + 2
(s + Ij2(s _ 2)'
.>2
Therefore !J{t) = [-O.8ge-
t
+ 0.33te-t + a.8ge2llu{f)
FJ(s) and!J{t) are shown in Figure 5-2(c).
(d) F,(,) _ 2s + 3
Where
($ + 4)(S2 + 25 + 3)'
25 + 3
(, + 4)[(, + I)' + 2)
25 + 3
(s + 4)(s + I + )1.4)(s + I _ )1.4)
-0.45 AI A*
=-+ + I
5+4 s+ 1+)1.4 s+ 1-)1.4'
A
J
=2(-1-jl.4)+3 .
(3 j1.4)( j2.83) - 0.22 + ,0.22
'nd
Ill) = -O.45e-
4
'u(t) - e-,(0,44Cos IAt + O.44Sin I At)u( -t)
F4(S) arc plotted in Figure S-2(d).
. From 5-7 it is seen that all the work required for
two-sIded transforms by partia) fractions was already master',"d",u
o
''', inS
ed Now the right of (1 contribute to the function for n'g,,';"
time and pIck up a mmus sIgn.
5-3-2
Inverse Two-sided Laplace Transforms Using
Residues
,
The on complex. variables lists some of the more important
pertammg to system analysIs and they will be utilized in this section Th
Laplace transform of F(s) = N(s)jD(s) Where the order ofD(): '
N() 'h h . s Ismore
s ,wit t e regIOn of convergence (11 <: (1 <: (12 is defined as;
5-3 THE INVERSE TWOSIDED LAPLACE TRANSFORM
199
where C is the straight line s(8) = (1 + j8, -0') < 8 < 00. The contour Cis shown in
Figure 5-3(a). We now discuss the evaluation of Equation 5-9 for the cases of t
positive and negative.
Fort>O
Consider dosing the contour C to the left with contour C
1
as shown in Figure
5-3(b). For t > 0, F(s)e+$! -- 0 at all points on C
1
and by Jordan's lemma
j., F(s)e+s
'
ds = 0.
Therefore
/(/) = F(s)e" ds + -2
1
.1 F(s)e-" ds
27rJ 7rJ C,
C
--1-.1i F(s)e-" ds
27rJ c+c,
= L [residues of the poles of F(s)e-" to the left of iT] (5-10)
Forl<O
Consider dosing the contour C to the right with contour C
2
as shown in Figure
5-3(c). For 1 < 0, F(s)e" -- at all points on C
2
and by Jordan's lemma
!c F(s)e" ds = 0.
,
Therefore
1(/) = F(s)e" ds + -2
1
.1 F(s)e" ds
27rJ 7rJ C,
C
--I-.i F(s)e" ds
27rJ c+c,
= -L [residues of the poles of F(s)e" to the right of iTJ (5-11)
The minus sign Occurs because the closed contour C + C
2
is traversed in a
c,
c
c
c
",
",
",
(a) (b)Casel>O (c)Caser<o
Figure 5-3 (a) The contour C for finding the inverse Laplace transform; (b) closing C
. With C, for t > 0; (c) closing Cwith C
2
for t <: O.
I
200
5/THE TWO-SIDED LAPLAce TA"NSIFOI""
clockwise manner. Summarizing. we conclude:
If F(,) - N(,)/D(,),
N(s), then
0"1 < (] < (J2 where the order of D(s) is higher than
For t > 0
f(t) = 2: [residues of the poles of F(s)e" to the left of 11]
For t < 0
J(t) = -L [residues of the poles of F(s)e"" to the right of u]
In addition, the uniqueness theorem for the Laplace transform is:
f(t) - _I .j"'i-[f"f(p)e-"'dpLd'
21fJ -0> r
c
which says that a time function and its Laplace transform constitute a
pair.
We evaluate some previously considered inverses using residue theory.
EXAMPLE 5..8
Using residues, find the inverse Laplace transforms of the follo,,;nl
functions:
-1
(a) FI(s) --,'
,+
I
(b) F,(,) - --,'
, +
u < -2
u> -2
3s + 2
(c) Fl(5) (s + 1/(5 _ 2)'
3s + 2
(d) Fis) (s + 1 )l(S _ 2)'
3s + 2
-1<u<2
u <-I
(e) FS(5) = 1)2(5 _ 2)'
2s + 3
(f) F6(S) (s + 4)(52 + 25 + 3)'
-4<u< -I
Solution
(a) As shown in Figure 5-4(a) C + C
1
enclose no poles of and
thereforeJ(t) 0, if I> O.
For t < 0, C + C
2
encloses the pole at s = -2
3s + 2
(., + 1)1(, 2)
], +
I< + 11
1
(., - 2)
J, + 2
1\+ I )l(s -
::s + 3

2
(>0
(-0.89 + 0.331)e-lt
Ie)
(089 - 0,33I)e' ,- (j,8,!"l,
(d)
(-0,89+0331)1' '+089,2,
-O.45e-'
-e-'l 0.44 Co. L4t + 0.44 Sin I At J
(f)
Figure 5-4 The inverse transforms for Example 5-7.
202
and thus
Therefore
S/THE TWO-SIDED LAPLACE TR,'"SFQ,
1,(/) = - [reSidue of pole at s = - 2 for If"j
'+2
[(' + 2) --=-'-- ," I J
s + 2 ,--2
(b) 12(/) = 0, t <: 0 as C + (; encloses no poles in Figure 5-4(b).
0,
h(/) = [reSidue of pole at s = -2 for _1_ '''j
'+2
Therefore h(/) = e-
2
'u(t)

)F(,)- _ 3s+2
, -('+1)'('-2)'
as shown in Figure 5-4(c).
For t > Q
h(l) = [residue of the second-order pole at s = _ I for F(sk"J
- + 2 e"l
ds s - 2 --1
For / <: 0
Therefore
=.=-! te--' + est ($ - 2)(3) - (3s + 2)1 I
-3 (s 2)2 , __ I
= D.33!e-' - O.8ge--'
h(t) = - [residue of pole = 2 for F(.r)e"]
3(2)+2
21
= - (2 + 1)2 e
= _O.8ge
2t
h(t) = (-0.&.9 + O.33t)e-'u(t) _ O.8ge2tu( -I)
This result agrees with Example 5-7(a) where the inverse was found
partial fractions.
(d)F(,)- 3s+2
4 - (s + l)\s _ 2)'
0" <: -I
as shown in Figure 5-4(d).
f 4 (t) = 0 for t > 0
since there are no poles to the left of u.
53 THE INVERSE TWOSIDED lAPLACE TRANSFORM
For t < 0
/4(t) = - [sum of residues of poles of F(s)e"']
= (-0.33te-' + 0.8ge-' - 0.8ge
2t
)u( -t)
which agrees with Example 5-7(b).
35 + 2
(e) F
5
(5) = (5 + 1)2(5 2)' 0>2
as shown in Figure 5-4(e).
/5(1)=0,1<0
since there are no poles to the right of u.
For t > 0
Therefore
!set) = [sum of the residues of the poles of F(5)e"']
!s(t) = (0.33te-' - 0.8ge-' + O.8ge
2
')u(t)
25 + 3
(f) F6(5) = (5 + 4)(52 + 25 + 3)' -4<u< -1
as shown in Figure 5-4(f).
For I> 0
For t < 0
25 + 3
(s + 4}(5 + 1 + j1.4)(s + I j1.4)
.h,(t) = [residue of the pole at s = - 4 for F(s)e$l]
5 -4'
= --e
II
= -0,45e
4
,
203
h(l) = - [sum of the residues of the poles at s = -I - j1.4 and _ I + j1.4]
_ [2( -1 - j1.4) + 3 e(-l-1L4), + r*e(-I +11.4)']
(3 j1.4)( j2.8) ,
where
r l = 0.22 + jO.22
16(t) = - [(0.22 + jO.22)e( '-11.4)t + (0.22 _ jO.22)eH
-[e-'(OA4Cos IAt + OA4Sin 1.4t)]
Finally, as in Example 5-7(d) we obtain:
1;,(1) - -0.45e-
4
'u(t) + [-O.44e-'(Cos IAt + Sin 1 At)Ju( -t)