Chapter 10

Discrete-Time Linear Time-Invariant Systems

Sections 10.1-10-3
 Representation of Discrete-Time Signals
We assume Discrete-Time LTI systems

The signal X[n] can be represented using

unit sample function or unit impulse
function: [n] x[n] x[k ] [n k ]
k

Remember:
x[n ] [n k ]
x [ k ],n k
0,else

Notations:
x0 [n] x[n] [n 0] x[0] [n 0] x[0], n 0
x1[n] x[n] [n 1] x[1] [n 1] x[1], n 1 notes
Representation of Discrete-Time Signals - Example

x[n] x1[n] x0 [n] x1[n]

x[1] [n (1)] x[0] [n] x[1] [n (1)]
 Convolution for Discrete-Time Systems
LTI system response can be described using: [n] h[n]
System
For time-invariant: [n-k]h[n-k]
Impulse Response of a System

For a linear system: x[k][n-k]x[k]h[n-k]

Remember:
x[n] x[k ] [n k ]
k
Thus, for LTI:

x[n] x[k ] [n k ] y[n] x[k ]h[n k ] x[n] * h[n]
k k

We call this the convolution sum y[n] x[k ]h[n k ] x[n] * h[n]
k

y[n] h[k ]x[n k ] h[n] * x[n]
k
Remember:
h[ n] * [n n0 ] h[n n0 ] * [n] h[n n0 ]
 Convolution for Discrete-Important Properties
By definition

y[n ] h[n ] * [n ] h[n ]

Remember (due to time-invariance property):

h[n] * [n n0 ] h[n n0 ] * [n] h[n n0 ]

Multiplication

[n ]g [n n0 ] [n ]g[ n0 ]
 Properties of Convolution
Commutative
Associative
Distributive
 Example
Given the following block diagram
Find the difference equation
Find the impulse response: h[n]; plot h[n]
Figure 10.3
Is this an FIR (finite impulse response) or IIR system?
Given x[1]=3, x[2]=4.5, x[3]=6, Plot y[n] vs. n
Plot y[n] vs. n using Matlab FIR system contains finite
number of nonzero terms

1
Difference equation y[n] ( x[n] x[n 1] x[n 2])
3

To find h[n] we assume x[n]=[n], thus y[n]=h[n]

1
y[n] h[n] ( [n] [n 1] [n 2])
Thus: h[0]=h[1]=h[2]=1/3 3
Since h[n] is finite, the system is FIR
In terms of inputs:
y[n] x[k ]h[n k ] x[n] * h[n]
k

y[n] h[k ]x[n k ] h[n] * x[n]
k

... x[n 3]h[3] x[n 2]h[2] x[n 1]h[1] x[n 0]h[0] x[n 1]h[1] ....
x[n 2]h[2] x[n 1]h[1] x[n 0]h[0]
 Example cont.
Given the following block diagram
Find the difference equation
Find the impulse response: h[n]; plot h[n]
Is this an FIR (finite impulse response) or IIR system?
Given x[1]=3, x[2]=4.5, x[3]=6, Plot y[n] vs. n Figure 10.3
Plot y[n] vs. n using Matlab

In terms of inputs:

Calculate for n=0, n=1, n=2, n=3, n=4, n=5, n=6

...y[0]=0
y[n]n=0; x[n 3]h[3] x[n 2]h[2] x[n 1]h[1] x[n 0]h[0] x[n 1]h[1] ....
n=1; y[n]=1
x[nn=2; y[2]=2.5
2]h[2] x[n 1]h[1] x[n 0]h[0]
n=3; y[2\3]=4.5 Try for different values of n
n=4; y[4]=3.5
n=5; y[5]=2
n=6; y[6]=0

Example cont. (Graphical Representation)

X[m] X[n-k]

h[0]=h[1]=h[2]=1/3 x[1]=3, x[2]=4.5, x[3]=6

y[ n] ... x[n 3]h[3] x[ n 2]h[ 2] x[ n 1]h[1] x[n 0]h[0] x[n 1]h[1] ....
x[n 2]h[2] x[n 1]h[1] x[n 0]h[0]
 Example
Consider the following difference equation:y[n]=ay[n-1]+x[n]
Draw the block diagram of this system
Find the impulse response: h[n]
Is it a causal system?
Is this an IIR or FIR system?
 Example
Consider the following difference equation:y[n]=ay[n-1]+x[n];
Draw the block diagram of this system
Find the impulse response: h[n]
Is this an IIR or FIR system?

We assume x[n]=[n]
y[n]=h[n]=ah[n-1]+[n];
y[0]=h[0]=1
y[1]=h[1]=a
y[2]=h[2]=a^2
y[3]=h[3]= a^3 Causal system (current and past)

h[n]=a^n ; n>=0

It is IIR (unbounded)
 Example
Assume h[n]=0.6^n*u[n] and x[n]=u[n] x[n] y[n]
h[n]
Find the expression for y[n]
Plot y[n]
Plot y[n] using Matlab

y[n] h[k ]x[n k ] h[n] * x[n]
k y[0]=1
n y[1]=1.6
0.6 u[n k ]u[k ] 0.6
k
k

k 0
k
2.5[1 0.6 n 1 ]; n 0 ..
y(100)=2.5
Steady State Value is 2.5
Remember These Geometric Series:
 Properties of Discrete-Time LTI Systems

Memory: y[n] x[k ]h[n k ] x[n] * h[n]
k
A memoryless system is a pure gain

system: iff h[n]=K[n];
K=h[0] = constant & h[n]=0 otherwise
y[n] h[k ]x[n k ] h[n] * x[n]
k

y[n] ... x[n 3]h[3] x[n 2]h[2] x[n 1]h[1] x[n 0]h[0] x[n 1]h[1] ....
x[n 2]h[2] x[n 1]h[1] x[n 0]h[0]

Causality Note that if k<0depending on

future; Thus h[k] should be zero
y[n] has no dependency on future values to remove dependency on the
future.
of x[n]; thus h[n]=0 for n<0 (note h[n] is
non-zero only for [n=0].
y[n] h[k ]x[n k ] x[n]h[0] ... x[0]h[n] x[1]h[n 1] ...
k 0
n
y[n] x[k ]k[n k ] x[n]h[0] ... x[0]h[n] x[1]h[n 1] ...
k
 Properties of Discrete-Time LTI Systems

Stability
BIBO: |x[n]|< M
Absolutely summable: | h[k ] |
k

Invertibility:
If the input can be determined from output
It has an inverse impulse response
Invertible if there exists: hi[n]*h[n]=[n]
 Example 1
Assume h[n]= u[n] (1/2)^n x[n] y[n]
h[n]
Memoryless?
Casual system?
Stable?

Has memory (dynamic): h[n] is not

K[n] (not pure gain); h[n] is non-zero
Is causal: h[n]=0 for n<0
Stable:
1 1

k
| h[k ] | | |
k

k 0 2 1 0.5
2
 Example 2
Assume h[n]= u[n+1] (1/2)^n x[n] y[n]
h[n]
Memoryless?
Casual system?
Stable?

Has memory (dynamic): h[n] is

not K[n] (not pure gain)
Is NOT causal: h[n] not 0 for
n<0; h[-1]=2

1 k 1
Stable:
k
| h[k ] | | | 2
k 1 2 1 0 .5
4
 Example 3
Assume h[n]= u[n] (2)^n x[n] y[n]
h[n]
Memoryless?
Casual system?
Stable?

Has memory (dynamic): h[n] is

not K[n] (not pure gain)
Is causal: h[n]=0 for n<0
Not Stable:

| h[
k
k ] |
| 2
k 0
| k