4.

Approximation of continuous time systems

with discrete time systems
A. Theory (it helps if you read the course too)
The continuous-time systems are replaced by discrete-time systems even for the
processing of continuous-time signals.

1) Impulse Invariance Method:

hd [ n ] = T ha ( nT )

The relation between the Z transform and the Laplace transform:

2 k

H d ( z ) z =esT = H a s j

k =
The frequency response of the digital system is the same with the frequency response of
the analog system of limited band for frequency less than half of sampling frequency

H a ( ) = H d ( )

=T

and M

The relation between the s and z planes

r = eT
.
s = + j ;z = re j = x + jy ; z = e sT re j = eT e jT
2

T
k

2) Step invariance method:

sd [ n ] = sa ( nT )

3) The method of the differential equation approximation

Differential eq:
N
d k y (t ) M
d k x (t )
a
b
=

k
k
dt k
dt k
k =0
k =0

Use approximation:
dya ( t )
dt

t = nT

ya ( nT ) ya ( nT T ) yd [ n ] yd [ n 1]
=
T
T

Finite difference eq:

N

ak
k =0

1
Tk

( 1)
p =0

Ckp y [ n p ] = bk
k =0

1
Tk

( 1)
p =0

Ckp x [ n p ]

The relation between the Z transform and the Laplace transform:

Hd ( z ) = Ha ( s)
The relation between the s and z planes
s=

1 z 1
T

s=

1 z 1
T

4) Bilinear Transformation Method

Hd ( z ) =

T 1 + z 1
= Ha ( s)
2 1 z 1

s = 2 1 z 1
T 1+ z

B. Problems
Problem 1. We have a linear time-invariant system with the transfer function:
s+2
.
Ha ( s) =
2
( s + 1) ( s + 3)

a) Determine its impulse response, ha ( t ) ;

b) Find the impulse response of the digital system equivalent to the analog one,
using the impulse response invariance method hd [ n ] ;
c) Find H d ( z ) , the Z transform of hd [ n ] ;

Solution.
a) To determine the impulse response from the transfer function, we decompose it into
simple fractions:
Ha (s) =

A
B
C
+
+
2
s + 1 ( s + 1) s + 3

s+2
1
1
= B=
s
=
1
s+3
2
2
s+2
1
C = ( s + 3) H a ( s )
=
=
2 s =3
s =3
4
( s + 1)
B = ( s + 1) H a ( s )
2

A=

s =1

d
2
( s + 1) H a ( s )

ds

s =1

d s + 2
ds s + 3

s =1

s +3 s 2

( s + 3)

s =1

1
4

We identify each term using the tables of Laplace transforms:

Ha (s) =

( s + 1)

1 1
1
1
1 1

+

2
s
+
1 2 ( s + 1)
s +
3
4
4 


7

1 e t t
()
4

1 e3t ( t )
4

d 1

ds s + 1

x (t ) X ( s )

( s + 1)

dX ( s )
ds

= x ( t )( t ) e st dt = tx ( t ) e st = L {tx ( t )} ( s )

= L te t ( t ) ( s )

ha ( t ) =

( s + 1)

te t ( t )

1 t
1
1
e ( t ) + te t ( t ) e 3t ( t )
4
2
4

T nT
T 2 nT
T
b) hd [ n ] = T ha ( nT ) = e [ n ] +
ne [ n ] e 3nT [ n ]
4
2
4
c) We identify each term using the tables of Z transforms:
T nT
T T n
T /4
e [ n] =
e
[ n]
4
4
1 e T z 1
n
T
T
T /4
e 3nT [ n ] = e 3T [ n ]
4
4
1 e 3T z 1
n
T 2 nT
T2
T2
e T z 1
ne [ n ] =
n ( e T ) [ n ]
2
2
2 (1 e T z 1 ) 2

( )
(

Hd ( z) =

T /4
T 2 e T
z 1
T /4
+

T 1
T 1 2
1 e z
2
(1 e z ) 1 e 3T z 1

Problem 2. We have H a ( s ) the transfer function of an analog system, linear and timeinvariant,
1
.
Ha ( s) =
2
s
+
s
+
1
2
( )(
)

a) Determine its impulse response, ha ( t ) ;

b) Compute the impulse response for the digital system in the case of the impulse
invariance method.
c) The step response is sa ( t ) . For the digital system, its impulse response is hd [ n ]
and its step response is sd [ n ] . Assume we used the impulse response invariance
method to approximate the analog system with a digital one:
hd [ n ] = T ha ( nT ) ,

Verify that the relation:

sd [ n ] =

k =

ha ( kT )

is true.
d) Compute the step response of the digital system, using the step response
invariance method:
sd [ n ] = sa ( nT ) .
Solution.
a) To determine the impulse response from the transfer function, we decompose it into
simple fractions and we identify each term using the tables of Laplace transforms:
Ha (s) =

( s + 1)( s + 2 )

A = ( s + 1) H a ( s )

s =1

A
B
C
+
+
2
s + 1 ( s + 2)
s+2

( s + 2)
=

1
s +1

d
2
( s + 2 ) H a ( s )

ds

B = ( s + 2) H a ( s )
2

C=

s =2

=>

2 s =1

s =2

=1

= 1

d 1

ds s + 1

= 1

( s + 1) s =2
ha ( t ) = e t ( t ) te 2t ( t ) e 2t ( t )
s =2

s =2

b) Impulse invariance method

The impulse response of the equivalent digital system is:

hd [ n ] = T ha ( nt ) = Te nT [ n ] nT 2 e 2 nT [ n ] Te 2 nT [ n ]

c) Impulse invariance method

The step response in discrete-time is the response to the unit step:

sd [ n ] = [ n ] hd [ n ] =

k =

hd [ k ] [ n k ] =

Which means that: sd [ n ] = T

k =

k =

hd [ k ] = T

k =

ha ( kT )

ha ( kT ) . The relation is true in the general case.

d) Step invariance method

Let S a ( s ) be the Laplace transform of the step response for the analog system:

Sa ( s ) =

Ha (s)
s

A
B
C
D
1
= +
+
+
2
s ( s + 1)( s + 2 ) s s + 1 ( s + 2 ) s + 2

We compute the step response of the analog system using the decomposition in simple
fractions and tables of Laplace transforms:
A = sS a ( s )

s =0

B = ( s + 1) Sa ( s )

( s + 1)( s + 2 )
s =1

C = ( s + 2 ) Sa ( s )

D=

s =2

s =0

1
4

1
s ( s + 2)
=

d
2
( s + 2 ) Sa ( s )

ds

2 s =1

1
s ( s + 1)
s =2

s =2

= 1
=

1
2

d 1
ds s 2 + s

s =2

2s + 1
( s2 + s )

s =2

3
= 3/ 4
4

1
1
3
=> sa ( t ) = ( t ) e t ( t ) + e2t ( t ) + e 2t ( t )
4
2
4
1
1
3
sd [ n ] = sa ( nT ) = [ n ] e nT [ n ] + e 2 nT [ n ] + e 2 nT [ n ]
4
2
4

Problem 3. We have an analog, linear time-invariant system with the impulse response
ha ( t ) and a digital linear time-invariant system hd [ n ] equivalent to the analog system

with the impulse response invariance method.

a) If ha ( t ) = e t ( t ) find the transfer function of the analog system;
b) Find the transfer function of the digital system;
c) Sketch the amplitude-frequency characteristic of the analog system and the
discrete-time system.
Solution.

a) H a ( s ) = L et ( t ) =

1
s +1

T
b) hd [ n ] = T ha ( nT ) = Te nT [ n ] H d ( z ) =
T 1
1 e z

c) H d ( ) =

T
1 e

e j

1
1
H a ( ) =
1 + j
1+ 2
T
Hd () =
=
T
1 e cos + jeT sin

d) H a ( ) =

(1 eT cos )

+ e2T sin 2

1 + e2T 2eT cos

For T=0.1 the amplitude-characteristic are almost the same for frequency less than half of
sampling frequency 2/T=10=31.14.

Ha()

Hd( )| = T
0.8
T=0.1s
0.6

0.4

0.2

-30

-20

-10

10

20

30

Plot it using Matlab. What do you notice when the frequency is higher than half of the
sampling frequency?

>>
>>
>>
>>

w=-200*pi:0.01*pi:200*pi;
Ha=1./sqrt(1+w.^2);
Hd=0.1./sqrt(1+exp(-0.2)-2*exp(-0.1).*cos(0.1*w));
plot(w,Ha,w,Hd); grid on

Problem 4 Consider the analog, linear time-invariant system with the differential
equation:
dy
+ 2 y (t ) = x (t )
dt
a) Find the impulse response and the transfer function of a discrete-time system that
approximates the analog system using the impulse invariance method;
b) Find again the impulse response and the transfer function if we use the differential
equation approximation;
c) The same for the bilinear transformation method.
d) Give implementation forms for the systems from a), b) and c) ;
e) For x ( t ) = sin t and a sampling period of T = 0,1 find the amplitude of the output

signal for the analog system and for the digital system, for the three cases,
respectively.
Solution.
a) impulse invariance method
Y (s)
dy
1
+ 2 y ( t ) = x ( t ) sY ( s ) + 2Y ( s ) = X ( s ) H a ( s ) =
=
dt
X (s) s + 2

T
ha ( t ) = e2t ( t ) hd [ n ] = Te 2nT [ n ] H d ( z ) =
2 nT 1
z
1 e
b) differential equation approximation
T
1
1
1 + 2T
H d ( z ) = H a ( s ) s =1 z 1 =
=
=
1
1
1
z

1
T
z
+

1
2
T
z 1
1
2+
1 + 2T
T
T 1
The impulse response of the digital system is: hd [ n ] =

1 + 2T 1 + 2T

[ n]

c) bilinear transformation method

H d ( z ) = H a ( s ) s = 21 z 1

T 1+ z 1

Hd ( z ) =

1
1
1 + z 1
=
=
s + 2 s = 21 z 1 2 1 z 1
2 1 z 1 + T (1 + z 1 )
2
+
T 1+ z 1
T 1 + z 1

1 + z 1
1
1 + z 1
Hd ( z) =
=
2 (1 + T ) 2 (1 T ) z 1 2 1 + T (1 T ) z 1
=

1
1
1
+

2 (1 T ) 1 T 1 + T (1 T ) z 1

1
1
1
+

2 (1 T ) (1 T )(1 + T ) 1 1 T z 1
1+ T

The impulse response of the digital system using the bilinear transformation:
n

1
1
1 T
[ n] +
hd [ n ] =

[n]
2 (1 T )
1 T 2 1+ T

d)

-2

y[n]

For the digital system from a)

b0 = 1; a0 = 1; a1 = e 2T
y [ n ] e 2T y [ n 1] = x [ n ]

e-2T

(1+2T)-1

For the analog system

y + 2 y = x

x[n]

x[n]

y[n]

For the digital system from b)

b0 = T ; a0 = 1 + 2T ; a1 = 1

(1 + 2T ) y [ n] y [ n 1] = Tx [ n ]

D
y[n-1]

For the digital system from c)

y[n] b0 = 1; b1 = 1;
a0 = 2 (1 + T ) ; a1 = 2 (1 T )

1
2 (1 + T )

x[n]

2 (1 + T ) y [ n ] + 2 (1 T ) y [ n 1] =

= x [ n ] + x [ n 1]

-2(1-T)

e)
For x ( t ) = sin t and a sampling period of T = 0,1 we have to find the amplitude of the

output signal, in the analog and in the digital case.

x ( t ) = sin t ; y ( t ) = H (1) sin t + arg { H (1)}

H ( ) = H ( s ) s = j =

1
1
H (1) =

2 + j
2+ j

1
1
; arg { H (1)} = arg {2 + j} = arctg
2
5
1
1
1

sin t arctg , amplitude

y (t ) =
2
5
5

H (1) =

For the digital systems we have:

n

n
1
1
x [ n ] = sin ( 0,1n ) = sin y [ n ] = H d sin + arg H d
10
10
10

10
For the digital system from a)
1
1
1
Hd ( z ) =
Hd () =
=
2T
2T 1
2T j
1 e
1 e
cos + j e 2T sin
z
1 e
z
1
1
Hd () =
=
2
1 + e4T 2e 2T cos
1 e 2T cos + e 4T sin 2

1
Hd =
10

1
1 + e0,4 2e0,2 cos

1
10

1 e 0,2

10

For the digital system from b)

T
T
T
Hd ( z ) =
Hd () =
=
j 1 + 2T cos j sin
1
1 + 2T z
1 + 2T e
T
T
Hd () =
=
2
2
2
1 + 4T + 4T cos
(1 + 2T cos ) + ( 2T sin )
0,1
0,1
1
Hd
=
= 0, 083
1, 44 1, 2
10

For the digital system from c)

1
1 + z 1
1
1 + e j
Hd ( z ) =

d( ) 2
2 1 + T (1 T ) z 1
1,1 0,9e j
1
Hd () =
2

(1 + cos ) + sin 2
2
2
(1,1 0,9 cos ) + ( 0,9sin )
2

1
2 + 2 cos
=
2 1, 21 + 0,81 1,98cos

Set the value of to 0,1 (homework).

Problem 5. Consider a continuous-time system with the impulse response ha ( t ) and the

transfer function H a ( s ) . The differential equation is:

N
dkx
=
b
k
dt k k =0 dt k
k =0
We approximate this system using a discrete-time system using a method similar to the
method of the differential equation approximation:
x [ n + 1] x [ n ]
dx

dt t =nT
T
N

ak

dk y

We define the difference of

( 0) x [ n ] = x [ n ]
{ }
x [ n + 1] x [ n ]
(1)
order 1: { x [ n ]} =
order 0:

order k:

( ) x [ n] = ( ) 1 x [ n]
{ }
{ { }}
k 1

a) Let H a ( s ) =

( s + 1)

. Find the transfer function of the digital equivalent system

Hd ( z ) ;

b) What is the connection between H a ( s ) and H d ( z ) ?

11

Solution

(1) x [ n] = x [ n + 1] x [ n]
{ }

T
(1) x [ n + 1] x [ n]
2 { x [ n ]} =

1 (1)
(1)
{ x [ n + 1]} { x [ n ]}

1 x [ n + 2] x [ n + 1] x [ n + 1] x [ n ]

T
T
T

1
= 2 x [ n + 2] 2 x [ n + 1] + x [ n ]
T
We notice that:
1 k
l
k { x [ n ]} = k Ckl x [ n + l ] ( 1)
T l =0
a) We have to find the transfer function of the digital equivalent system H d ( z ) using a
method similar to the differential equation approximation.

b1s
s
Ha ( s) = 2
=
s + 2s + 1 a0 + a1s + a2 s 2
b0 = 0; b1 = 1; a0 = 1; a1 = 2; a2 = 1
The linear differential equation is:

dy ( t ) d 2 y ( t ) dx ( t )
y (t ) + 2
+
=
dt
dt
dt 2

The finite difference equation is:

y [ n ] + 2

(1)

2
1
{ y [ n]} + ( ) { y [ n]} = ( ) { x [ n]}

meaning that:
y [ n] + 2

y [ n + 1] y [ n ]
T

x [ n + 1] x [ n ]
1
+ 2 y [ n + 2] 2 y [ n + 1] + y [ n ] =
T
T

or:
T 2 y [ n ] + 2T ( y [ n + 1] y [ n ]) +

y [ n + 2] 2 y [ n + 1] + y [ n] = Tx [ n + 1] Tx [ n ]
12

Applying in both sides the Z transform, we obtain:

T 2Y ( z ) + 2T ( z 1) Y ( z ) + z 2Y ( z ) 2 zY ( z ) + Y ( z ) = TzX ( z ) TX ( z )
or:
Y ( z ) T 2 + 2Tz 2T + z 2 2 z + 1 = X ( z )(Tz T )

Tz T
Hd ( z ) = 2
z + 2 (T 1) z + T 2 2T + 1

b) What is the connection between H a ( s ) and H d ( z ) ?

Hd ( z ) =

Tz T

( z + T 1)

=
2

T ( z 1)
z 1
T2
+ 1
T

z 1
T
meaning that: H d ( z ) =
2
z 1
+ 1

We can see that: H d ( z ) = H a ( s ) s = z 1

T

13

T ( z 1)

( z + T 1)

z 1
1
T
= T
T z 1 2
+ 1