Week Lecture 12

The Laplace Transform

From continuous time Fourier transform To Laplace transform

Recap: For Continuous Time Fourier Transform,


X ( j )   x(t ) exp
 jt
dt ANALYSIS



1
 X ( j ) exp d SYNTHESIS
jt
x(t ) 
2 

The signal exp jt is not the only complex exponential signal which we can use as input into an
LTI system. We can also instead use a more general eigenfunction which is of the form signal
exp st , where s    j

If the input x(t) into an LTI system is exp st , then the output
  

 h( ) x(t   )d   h( )exp  h( )exp d

s ( t  )  s
y(t )  x(t )* h(t )  h(t )* x(t )  dt  exp st

  

The output is also of the form exp st but the amplitude is modified by the term H ( s)


 h( )exp d . H ( s) is an eigenvalue (complex constant) which depends on

 s
where H ( s) 


the complex variable s only.

Definition of Laplace Transform



 h(t )exp
 st
The term H ( s)  dt is the definition of the Laplace transform of h(t), the impulse


response of the LTI system. However we do not restrict ourselves in applying the Laplace
transform to impulse response of LTI systems only. We can in fact also use and apply Laplace
transform for any continuous time signals i.e x  t   X  s 

The Laplace transform of a general continuous time signal x(t) is defined as



 x(t )exp dt where s    j

 st
X ( s) 


1
Relationship Between Laplace transform and Fourier transform

i) If we replace j in the continuous time Fourier transform with s    j , we obtain the

the Laplace transform. Therefore the Laplace transform is a generalization of the
continuous time Fourier transform i.e if   0 , the Laplace transform becomes the Fourier
transform.

ii) Another relationship between the Laplace and the Fourier transform is that the Laplace
transform can be interpreted as the Fourier transform of a modified x(t) i.e x(t )exp t .

Given x(t ) , the Fourier Transform is X ( j )   x(t )exp
 jt
dt


Given x(t )exp t , the Fourier Transform is

 
  j  t
X ( j )  

x(t )exp  t exp  jt dt   x(t )exp

dt


If we let s    j , then X ( s)   x(t )exp
 st
dt


Therefore the Laplace transform X (s) is actually the Fourier transform of x(t )exp t

Notice that the integration is from  to  , where in this case X ( s) is called the bilateral
Laplace transform of x(t ) . There is another type of Laplace transform called the Unilateral

Laplace transform defined as X ( s)   x(t ) exp  st dt , where the integration limit is from 0 to  ,
0

Example 1

This example illustrates a signal whose Fourier transform do not exist but has Laplace
transform

Consider the signal 𝑥(𝑡) = 𝑒 2𝑡 𝑢(𝑡) which the Fourier transform do not exist.
+∞ +∞ 2𝑡 −𝑗𝜔𝑡 +∞ (2−𝑗𝜔)𝑡
𝑋(𝑗𝑤) = ∫−∞ 𝑒 2𝑡 𝑢(𝑡)𝑒 −𝑗𝜔𝑡 𝑑𝑡 = ∫0 𝑒 𝑒 𝑑𝑡 = ∫0 𝑒 𝑑𝑡
∞
𝑒 (2−𝑗𝜔)𝑡
⌈ ⌉
(2 − 𝑗𝜔) 0
∞ ∞
𝑒 (2−𝑗𝜔)𝑡 𝑒 (2−𝑗𝜔)𝑡
This term ⌈ (2−𝑗𝜔) ⌉ not converge because when t ⇾ ∞, term ⌈ (2−𝑗𝜔) ⌉ ⇾∞
0 0

2
So the Fourier transform of 𝑥(𝑡) = 𝑒 2𝑡 𝑢(𝑡) do not actually exist in the normal sense..

Now take the Laplace transform of 𝑥(𝑡) = 𝑒 2𝑡 𝑢(𝑡)

+∞ +∞ 2𝑡 −𝑠𝑡 +∞ −(𝑠−2)𝑡
𝑋(𝑠) = ∫−∞ 𝑒 2𝑡 𝑢(𝑡)𝑒 −𝑠𝑡 𝑑𝑡 = ∫0 𝑒 𝑒 𝑑𝑡 = ∫0 𝑒 𝑑𝑡
∞
𝑒 −(𝑠−2)𝑡 1
⌈ ⌉ =
−(𝑠 − 2) 0 𝑠−2

1
So now we can say that the Laplace transform of 𝑒 2𝑡 𝑢(𝑡) is 𝑆−2 PROVIDED that Re(s-2) >0

Or (𝜎 − 2) > 0 𝑜𝑟 𝝈 > 𝟐

Example 2

Determine the bilateral Laplace transform and its region of convergence of the signal
x(t )  exp at u(t )

  
exp ( sa )t 1
 exp dt   exp
 at  st  ( sa )t
X (s)  u (t )exp dt  
 0
( s  a ) 0 s  a
1
Note that the Laplace transform for x(t )  exp at u(t ) is PROVIDED that Re(s+a) > 0
sa

or Re(s) > -a

Example 3

3
Determine the bilateral Laplace transform and its region of convergence of the signal
x(t )   exp at u(t )

 0
exp ( sa )t
0
1
X ( s)    exp  at
u (t )exp  st
dt    exp  ( s a )t
dt  
 
( s  a)  s  a
1
The Laplace transform for x(t )   exp at u(t ) is PROVIDED that Re(s+a) < 0
sa

or Re(s) < -a

Summary: In example 2 and example 3, even though the 2 signals are different, the algebraic
expressions for the respective Laplace transforms are the SAME. The only difference lies in their
respective regions of convergence (ROC). Therefore when specifying the Laplace transform of a
signal, both the algebraic expression and the ROC i.e the range of values of s for which the
expression of X(s) is valid are required.

The S-Plane

A convenient way to display the region of convergence (ROC) is to plot it on the s-plane. The s-
plane consists of the horizontal Re(s) axis (also known as  axis) and the vertical Im(s) axis
(also known as j axis). The shaded region represents the set of points in the s-plane
corresponding to the region of convergence of the Laplace transform of a particular signal..

4
Example 4

Find the Laplace transform and the region of convergence of x(t )  3exp2t u(t )  2expt u(t )

We first perform the Laplace transform of each term individually and note its ROC.
 

 3exp dt   3exp  (2s )t dt

2t 2t  st
For the Laplace transform of x(t )  3exp u (t ), X(s) = u (t )exp
 0

3exp  2s t

3
X ( s)   3exp  (2s )t dt   . The region of convergence is Re (2+s) > 0 or Re(s)
0
(2  s) 0
2s
>-2. Since Re(s) is  , we can also say that the ROC is   2
 

 2exp dt   2exp  (2 s )t dt

t t  st
For the Laplace transform of x(t )  2exp u (t ), X(s) = u(t )exp
 0

3exp 1s t

2
X ( s)   2exp  (1 s ) t
dt   . The region of convergence is Re (1+s) > 0 or
0
(1  s) 0
1 s
Re(s) >-1 or   1

Thus combining the Laplace transforms of the 2 signals we obtain

3 2
3exp2t u (t )  2exp t u (t ) 
 . The region of convergence of the OVERALL signal is
s  2 s 1
the COMBINED region of convergence. The combined region of convergence must be such that
BOTH terms must converge. In this case the ROC is s > -1

5
Zero –Pole Plot

In example 4 above, we noted that for the continuous time signal x(t),
3 2
x(t )  3exp2t u (t )  2exp t u(t )  X ( s)   . By combining the 2 terms on the right
s  2 s 1
s 1 s 1
hand side, we obtain 3exp2t u (t )  2exp t u(t )  2  ROC is Re (s) > - 1
s  3s  2  s  1 ( s  2) ,

We note that the Laplace transform can be expressed as a ratio of polynomial in the complex
N (s)
variable s i.e X ( s)  . X ( s) is called rational whenever X ( s) can be expressed as a ratio of
D( s )
N (s)
.
D( s)

In the expression for X(s), the roots of the denominator are referred to as the poles of X(s). The
poles are the values of s such that X(s) becomes infinite. The roots of the numerator in X(s) are
referred to as the zeros of X(s). The zeros are values of s such that X(s) becomes zero.

The representation of X(s) through its poles and zeros in the s-plane is referred to as the pole-
zero plot of X(s)

Thus, a convenient pictorial way to describe a rational Laplace transform is to perform the pole-
zero plot in the s-plane and also shade the ROC.

s 1
As an example, if X ( s)  , X(s) has poles are s= -1,s = -2 and zero at s = 1
 s  1 (s  2)

6
The Region of Convergence (ROC)

We have noted that 2 different signals can have identical algebraic expression for their Laplace
transform and they are only distinguishable by their respective ROCs. Here we explore some
important properties of ROC and how these properties can provide insights to the characteristic
of the continuous time signal x(t) in the time domain.

Property 1: The ROC of X(s) consists of strips parallel to the j axis in the s-plane

i) This property follows from the condition for the convergence of the Laplace transform


 x  t  exp dt   , That the region of convergence of X(s) depends only of  , the


 t
i.e


real part of s. This means that if X(s) converge for a particular value of  , then it should
also converge for   j where  is for all frequencies, which is now a line. If we have
a range of values of  , we now have a stripe parallel to the j axis in the s-plane

ii) For a signal x(t) which has a Fourier transform, its Laplace transform X(s) must include the
j axis,. This is because S    j , the Laplace transform X(s) is simply the Fourier
transform when   0 , therefore the ROC of X(s) must include the j axis,.

iii) The ROC of X(s) MUST be a connected region. Disconnected regions or disjointed stripes
are not allowed.

Property 2: For rational Laplace transform, the ROC contains no poles.

Since X(s) is infinite at a pole, by definition the Laplace transform of x(t) i.e X(s) will not
converge and thus do not exist if it contain a pole.

Property 3: If x(t) is of finite duration and is absolutely integrable, then the ROC is the entire s-
plane.

If the continuous time signal x(t) is of finite duration i.e  x  t  dt   ,

no matter what values

of  you apply in the weighting function exp t , the integration always converge. Thus the
ROC is not affected by  . In such a case, the ROC includes the entire s-plane

7
Property 4: If the continuous time signal x(t) is right sided, and if the line Re(s) =  0 is in the
ROC, then all values of s for which Re(s) >  0 will also be in the ROC

Property 5: If the continuous time signal x(t) is left sided, and if the line Re(s) =  0 is in the
ROC, then all values of s for which Re(s) <  0 will also be in the ROC

Property 6: If the continuous time signal x(t) is two sided, and if the line Re(s) =  0 is in the
ROC, then the ROC will consists of a strip in the s-plane that include Re(s) =  0

8
Property 7: If the Laplace transform X(s) of x(t) is rational, then its ROC is bounded by poles or
extends to infinity. In addition, no poles of X(s) are contained in the ROC

Property 8: If the Laplace transform of X(s) of x(t) is rational, then if x(t) is right-sided, the ROC
is the region in the s-plane to the right of the rightmost pole.

If x(t) is left-sided, the ROC is in the region in the s-plane to the left of the leftmost pole.

The Inverse Laplace Transform

  j
1
X  s  exp
2 j   
The inverse Laplace transform is defined as x(t )  st
ds
j

For finding the inverse of the Laplace transform, we usually do not use the above equation. We
will usually use the partial fraction expansion and look-up table for Laplace Transform Pair to
find the inverse Laplace transform instead.

9
Example

1
Let X ( s)  Re(s) >-1
( s  1)( s  2)

To obtain the inverse Laplace transform, we first perform a partial fraction expansion to obtain

1 A B
X ( s)   
( s  1)( s  2) s  1 s  2

By the cover up rule method, we can easily obtain the coefficients A = 1 and B = -1

1 1
Thus, X ( s)  
s 1 s  2

We know that 2 different continuous time signals x(t) can have the same Laplace transform.

1
Remember that the inverse Laplace transform for X ( s)  can be expt u(t ) if the ROC for
s 1
t
X(s) is Re(s) >-1 or it can also be  exp u(t ) if the ROC for X(s) is Re(s) < -1

In this case how can we obtain the correct inverse Laplace transform for the combined time
functions?

The trick is to look at the properties for the ROC

In this example, the ROC for X(s) is given as Re(s) > -1 . So the first thing to note is that the
ROC for the individual terms in the partial expansion must includes Re(s) > -1 Since the given
ROC is to the right of both poles, the same must be true also for the individual terms, By
property 8, the ROC in the s-plane which corresponds to the right of the rightmost pole, the
signal x(t) is right sided. Based on the conclusion that x(t) is a right sided signal, the correct
inverse Laplace transform is

x(t )  expt  exp2t  u(t ) Re(s) > -1

Example

The Laplace Transform Xs  of a continuous-time signal x t  is given as below:

𝑠−2
𝑋(𝑠) = 𝑠2 +2𝑠−15 − 5 < 𝑅𝑒{𝑠} < 3

Also sketch the pole-zero pattern of X s  and its region of convergence and find the signal x(t).

10
 𝑠−2 𝑠−2
Since 𝑋(𝑠) = 𝑠2 +2𝑠−15 = (𝑠+5)(𝑠−3) , there is one zero at s=2, one pole at s = -5 and one pole at s

=3

The signal x(t) can be contained by finding the inverse Laplace transform. But usually it is done
by using partial fraction.

𝑠−2 𝐴 𝐵
𝑋(𝑠) = = +
(𝑠 + 5)(𝑠 − 3) (𝑠 + 5) (𝑠 − 3)
−5−2 7
By using the cover up rule A = −5−3 = 8

3−2 1
By using the cover up rule B = 3+5 = 8

7 1
𝑋(𝑠) = 8 + 8
(𝑠 + 5) (𝑠 − 3)
7 −5𝑡 1
Therefore 𝑥(𝑡) = 𝑒 𝑢(𝑡) − 8 𝑒 3𝑡 𝑢(−𝑡)
8

11
Analysis and Characterization of LTI Systems Using Laplace Transform

One of the important applications of the Laplace transform is in the analysis and characterization
of LTI systems. Its role stems from the convolution property which says that convolution of 2
signals in the time domain has the effect of multiplication in the frequency domain.

Y ( s)  H ( s) X ( s)

where X(s) is the Laplace transform of the input, H(s) is the Laplace transform of the impulse
response, and Y(s) is the Laplace transform of the output.

The response of LTI systems to complex exponentials is such that if the input x(t )  exp st , with s
in the ROC of H(s), then the output will be H ( s)expst . The input exp st is the eigenfunction of
the LTI system, and H(s) is the eigenvalue equal to the Laplace transform of the impulse
response

If the ROC of H(s) includes the imaginary axis and when s  j only, H(s) is called the
frequency response. In the broader context of the Laplace transform, H(s) is called the transfer
function or system function.

The transfer function characteristics are closely related to the LTI system properties. We now
explore 2 important properties here.

i) Causality

For a causal LTI system, the impulse response is zero for t < 0. This means that the
impulse response of an causal LTI system is right sided

The ROC associated with the system function for a causal system is a right half plane.
However the converse is not true i.e given a ROC is to the right of the rightmost pole
does not guarantee that a system is causal, it only guarantee that the impulse response
is right sided.

However, for a system with a RATIONAL system function, causality of the system is
equivalent to the ROC being on the right half plane to the right of the rightmost pole.

Example

Consider a system with impulse response h(t )  expt u(t ) .

We know that since h(t) = 0 for t < 0, this system is causal. The Laplace Transform
 
1
 exp dt   exp  ( s1)t dt 
t  st
for this signal is H ( s)  u (t )exp Re(s) > -1
 0
s 1

12
 1
We observe that H ( s)  is rational, has a pole at s = -1 and its ROC is greater
s 1
than -1 i.e it is to the right of the rightmost pole. This observation is consistent with
the fact that a system with a rational function and the ROC is in the right half plane to
the right of the rightmost pole, the system is causal.

Example

exp s
Consider the system function H ( s)  Re(s) > -1
s 1

Now this system function is not rational because of the exp s term in the numerator.
(Rational function must have a polynomial of s in the numerator and a polynomial of
s in the denominator)

We observe that this system has a pole = -1, the ROC is to the right of the rightmost
pole. Based on this observation, we know that the impulse response is right-sided,
that’s all we can say here. As H(s) is not rational, we cannot say that the system is
causal. Let’s further investigate to see if the system is causal. We can do this by
finding the actual time function of x(t) i.e the inverse Laplace transform.

1
We know from the earlier example, expt u (t )  Re(s) > -1
s 1

By the time shifting property the factor exps has the effect of a time shift in the time
domain. x(t  t0 )  exp s X (s) or x(t  t0 )  exps X (s)

exp s
Thus exp (t 1) u (t  1)  ,
s 1

Thus the impulse response is h(t )  exp(t 1) u(t  1) .This response is non zero for
t < 0 i.e it is non zero for 1  t  0 , thus the system is not-causal.

ii) Stability

The ROC of H(s) is also related to the stability of a LTI system. We know that the
stability of a system is equivalent to its impulse response being absolutely integrable
i.e when its Fourier Transform converges. Since the Fourier Transform is equal to the
Laplace Transform when   0 i.e Re(s) =0, we have this conclusion.

An LTI system is stable if and only if the ROC of its transfer function H(s) includes
the entire j axis i.e Re(s) = 0

13
 Example

s 1
Consider and LTI system with system function H ( s) 
( s  1)( s  2)

Notice here we did not specify the ROC. Different ROCs (and therefore different impulse
responses of time functions) can be associated with the given H(s). Knowing additional
information such as causality or the stability, the appropriate ROC can be specified.

Given the extra information that the system is causal, then the impulse response can be
2 1 
immediately deduced as h(t )   exp t  exp 2t  u (t ) . However, the pole is located at
3 3 
s =- 1 and s = 2. Being causal means the ROC must be located to the right of the right
most pole i.e s > 2. With s > 2, the ROC do not include the j axis, therefore the system
is unstable

On the other hand, if the system is known to be stable and the ROC is as shown below,

the corresponding impulse response corresponding to the same H(s) will be deduced as
2 1 1
h(t )  exp t u (t )  exp 2t u (t ) . (N.B the negative sign in front of is required for the
3 3 3
u(-t) term. Refer to Laplace transform of elementary functions table). Now this time function is
now absolutely integrable which is consistent with the fact that the system is stable.

14
Finally, if the ROC for H(s) shown below,

2 1 
The system is anti-causal and unstable with h(t )    exp t  exp 2t  u (t )
3 3 

One last observation is that if a causal LTI system has a rational function, then for this system to
be stable, the ROC must not contain a pole and it must include the j axis. This is as good as
saying that all the poles must lie on the left half of the s-plane i.e all the poles must have negative
real parts.

Example

Consider the system with impulse response h(t )  expt u(t ) , the Laplace transform is
  1
H ( s)   expt u (t ) exp  st dt   exp  ( s 1)t dt  , Re(s) >-1 .
 0 s 1

We can see that there is a pole at s = -1. Since the ROC is at Re(s) >-1, we can conclude that the
system is stable. This is also confirmed by the fact that h(t) is absolutely integrable i.e




x(t ) dt  

Now if we have another system where the impulse response h(t )  exp2t u(t ) , the Laplace
  1
transform is, H ( s)   exp2t u (t ) exp st dt   exp  ( s 2)t dt 
, Re(s) >2 , we have now a
 0 s2
pole at s = 2. Since the pole does not have a negative real part, the system is unstable. This is also

confirmed by the fact that h(t) is not absolutely integrable i.e 

x(t ) dt  

15
Example

If the transfer function for a system is as shown, determine if this system is stable

𝟏
𝑿(𝒔) =
𝒔𝟐 −𝒔−𝟐
𝟏 𝟏
Since 𝑿(𝒔) = 𝒔𝟐 −𝒔−𝟐 = (𝒔+𝟏)(𝒔−𝟐)

The pole is at s = 2 and s =- 1

Based on the rule for R.O.C, the ROC must be between s = -1 and s = . Since this ROC include
the jω axis, the system is STABLE

LTI Systems Characterized by Linear Constant Coefficient Differential

Equations.(LCCDE)

The point to note about LCCDE is that its system function H(s) does not include the specification
of the region of convergence. You will need additional information such as knowledge about the
causality and stability of the system before you can make any conclusion about the region of
convergence. For example if we say that the given LCDDE is causal, then the ROC will be to the
right of the rightmost pole.

Example

Consider an LTI system for which the input and output satisfy the LCCDE

dy (t )
 3 y(t )  x(t )
dt

Applying the Laplace transform to each term and also using the differentiation and linearity
property, we obtain sY (s)  3Y (s)  X (s)

Y ( s)
Since the system function is defined as H ( s)  , we obtain the system function as
X ( s)
1
H (s) 
s3

The LCCDE in itself does not provide the complete specification of the LTI system. Additional
information is required. Eg if we know that the system is causal (condition of initial rest), then
the ROC will be to the right of the rightmost pole which in this case Re(s) > -3. The
corresponding impulse response in this causal case is h(t )  exp3t u(t )

16
Likewise if the system is known to be anti-causal, the ROC would be Re(s) < -3 and the impulse
response for the same H(s) will be h(t )   exp3t u(t ) .

Properties of Laplace transform

Many of the properties of the Laplace Transform are similar to the Fourier transform.

i) Linearity

Because the Laplace transform is linear, we can write:

ax1 (t )  bx2 (t )  aX1 (s)  bX1 (s) with ROC containing the intersection of R1 and R2

where X1(s) is the Laplace transform of x1(t) and X2(s) is the Laplace transform of x2(t).

ii) Time Shifting

x(t )  X (s) , with ROC = R

then x(t  t0 )  X (s) exp st0 with ROC = R

iii) Convolution Property

x1 (t )  X1 (s) with ROC = R1

x2 (t )  X 2 (s) with ROC = R2

Then x1 (t )* x2 (t )  X1 (s) X 2 (s)

The ROC of X1 (s) X 2 ( s) includes the intersection of the ROC of X1(s) and X2(s) and
maybe larger if pole –zero cancellation occurs.

iv) Differentiation in the Time Domain

x(t )  X (s) with ROC = R

dx(t )
 sX ( s ) with ROC containing R
dt

Differentiation of a signal x(t) in the time domain has the effect of multiplication by s in
the s-domain.

17
v) Differentiation in the s-domain


 x(t )exp
 st
X ( s)  dt



dX ( s)
Differentiating both sides w.r.t. s, we obtain    t  x(t )exp  st dt
ds 

Therefore, x(t )  X (s) with ROC = R

dX ( s)
tx(t )  with ROC = R
ds

vi) Integration in the Time Domain

Integration is the reverse of differentiation, we would expect that integration has the
effect of division by s in the s-Domain

x(t )  X (s) with ROC = R

 x  d  s X (s) with ROC containing R  Res  0

1
then


Laplace Transform of Elementary Functions

Signal Transform ROC

 (t ) 1 All s

1
u (t ) Re(s)  0
s

1
u (t ) Re(s)  0
s

t n1 1
u (t ) Re(s)  0
 n  1! sn

18
 t n1 1
 u (t ) Re(s)  0
 n  1! sn

1
expt u(t ) Re(s)  
s 

1
 expt u(t ) Re(s)  
s 

t n1 1
exp  t u (t ) Re(s)  
 n  1! s  
n

t n1 1
 exp  t u (t ) Re(s)  
 n  1! s  
n

 (t  T ) exp sT All s

cos 0t u(t ) s

Re(s)  0
s  02
2

0
sin 0t u(t ) Re(s)  0
s  02
2

19