DT LTI Systems:

Unit Impulse (Sample)

Response and Convolution Sum

1
DT- LTI Systems:
Unit Impulse Response
Characterization

 LTI system: Linear and Time Invariant system

 Some times it is called LSI (Linear and Shift Invariant) systems.
 LTI systems have significant signal processing applications.
 Identity system, Delay and Advance systems, Accumulator, Forward difference
and backward difference systems, Moving average systems are examples of LTI
systems.
 The input x(n) and output y(n) of LTI system are related by its unit impulse
response h(n).
 LTI system is completely characterized by its unit impulse response.
 For LTI systems: output y(n) is h(n) if input x(n) is unit impulse δ(n).
 We can say convolution sum of x(n) with h(n) provides y(n).
 Symbolically it is written as y(n)=x(n)*h(n); where * denotes convolution
operation.
2
Convolution Sum of x(n) and h(n):

Here k is dummy variable

 Proof:

3
Computation of Convolution Sum
 Two methods used :
 Direct computation from equation: Analytical method
 Graphical method
 Involves:
1. Folding
2. Shifting
3. Multiplication
4. Summation

4
Finite length convolution
 Find convolution of x (n)={1,2,2,1} and h(n)={1,-1,2}.
↑ ↑
Ans.:y(n)={1,1,2,3,3,2}
↑
From the finite length convolution examples we can deduce
that:
 Lowest index of y(n) = lowest index of x(n) + lowest index of
h(n)
 Highest index of y(n) = highest index of x(n) + highest index of
h(n)
 If length of x(n) is L and length of h(n) is M then length of y(n) is
N=L+M-1 5
Find y(n)=x(n)*h(n) for

6
Infinite length convolution:
 Mostly use combination of graphical method and
analytical method
 Find y(n)=x(n)*h(n) for:
1. x(n)=u(n), h(n)=u(n)
2. 𝑥 𝑛 = 𝑎𝑛 𝑢 𝑛 , ℎ 𝑛 = 𝑢 𝑛
3. x(n)=2nu(-n), h(n)=u(n)
Ans.
1. y (n)  (n  1)u (n)
1  a n 1
2. y (n)  u ( n)
1 a
3. y (n)  2 n 1
7
DT LTI Systems:
Properties

8
Properties of LTI system in terms
of properties of convolution
 Identity property
 Shifting property
 Commutative property
 Distributive property
 Associative property

9
Identity property
 x(n)*δ(n) = x(n)

10
Shifting property
 x(n)*δ(n-n0) = x(n-n0); where no is +ve integer
 x(n)*δ(n+n0) = x(n+n0)

 Shifting of convoluted signal:

 y(n)=x(n)*h(n) => y(n-n0)= x(n-n0)*h(n) = x(n)*h(n-n0)
 Scaling of convoluted signal:
 y(n)=x(n)*h(n) => Ay(n)= Ax(n)*h(n) = x(n)*Ah(n); where A is
constant
11
Commutative property
 x(n) * h(n) = h(n) * x(n)

12
Distributive property
 x(n) * [h1(n) + h2(n)] = x(n) * h1(n) + x(n) * h2(n)

13
Associative property
 [x(n) * h1(n)] * h2(n) = x(n) * [h1(n) * h2(n)]

14
Combination of Associative and
Commutative Property:

 In cascaded LTI systems, the order of connection doesn’t matter.

 The overall response remains same for connection (a) and (d). 15
Ex. Consider the interconnection of LTI systems in figure below

Find the impulse response of overall system

Ans. h(n)=δ(n)
16
 DT LTI Systems:
Information from Unit
Impulse Response

17
What information can be obtained
from Unit Impulse response?
 Unit step response
 Causality
 Stability
 Invertibility
 Static or Dynamic
 Finite Impulse Response (FIR) or Infinite Impulse
Response (IIR)

18
The Unit Step Response (s(n)) of an
LTI System:
 It describes output y(n) when input x(n) is unit step
signal u(n).

 s(n)=u(n)*h(n)=h(n)*u(n)
 n
0; k  n
  h(k )u (n  k )   h( k ) as u (n  k )  
k   k   1; k  n

 Accumulation (Running sum) of unit impulse

response gives unit step response.
 Also from accumulator description we can write:
 h(n)=s(n)-s(n-1)
19
Causality of LTI system
 For the causal system,
 h(n) = 0, for n < 0
 Hence an LTI system is causal if and only if its impulse response
is zero for negative values of n
 So, the unit impulse response of a causal LTI system must be
zero before unit impulse occurs.
 More generally, the causality for a linear system is equivalent to
the condition of initial rest (relaxed system).
 If the input to a causal system is zero up to some point in time,
then the output must be zero up to that time.
 The equivalence of causality and initial rest applies only to linear
systems.
 For e.g. The system y(n)=ax(n)+b; a and b constants, is causal
but non-linear and not satisfying initial rest condition.
 For a causal LTI system the convolution sum is given by:

20
Stability of LTI system
 System is stable if bounded input x(n) produces bounded output
y(n).
 If x(n) is bounded then | x(n) | < Mx < ∞ for all n.
y(n) = ∑ h(k) x(n – k)
| y(n) | = | ∑ h(k) x(n – k) |
| y(n) | < ∑ | h(k) | | x(n – k) |
| y(n) | < Mx ∑ | h(k) |
| y(n) |<∞ if ∑ | h(k) |< ∞
 This specifies that a LTI system is stable if its impulse response is
absolutely summable.

 This is the sufficient condition to check the stability of the LTI

system.
 In fact, this condition is also a necessary condition, since it can
be shown that, if this condition is not satisfied, there are
bounded inputs that result in unbounded outputs. 21
FIR and IIR systems
 FIR is acronym for Finite Impulse Response and IIR for Infinite
Impulse Response.
 If the unit impulse response of a linear and time invariant (LTI)
system is of finite length then system is called Finite Impulse
Response (FIR).
 If the unit impulse response of a LTI system is of infinite length
then system is called Infinite Impulse Response (IIR).
 For FIR system, h(n) is finite length sequence, thus Sh is always
summable. Therefore, FIR systems are always stable.
 But for IIR system, if h(n) tends to zero as n tends to infinite,
then summation in equation converges to a finite value and
system becomes stable. Otherwise, IIR system is unstable.

22
 Determine causality and stability of following LTI systems,
Also state FIR/IIR?
 (i) h(n)={1,2,2,1}
 |
 (ii) h(n)={1,2,2,1}
 |
 (iii) h(n)=(½)nu(n)
 (iv) h(n)=(2)nu(n)
 Also classify each in FIR/IIR.
 Ans. (i) Non causal, stable, FIR (ii) Causal, stable, FIR (iii)
Causal, stable, IIR (iv) Causal, Unstable, IIR
23
Unit impulse response of a static
and dynamic LTI systems:
 Static: Memoryless: Output depends on present input
only.
 So for static system h(n)=0; n≠0.
 Hence in this case the unit impulse response: h(n)=Kδ(n);
where K=h(0) is a constant.
 If this condition is not satisfied, system is dynamic: with
memory.

24
Unit Impulse response of an
inverse system:
 If an LTI system is invertible, its inverse is also an LTI
system.

 If h(n) is unit impulse response of a given LTI system and

h1(n) is unit impulse response of its inverse system, the
cascade of h(n) and h1(n) gives identity system.
 So, h(n)*h1(n)=δ(n), if h1(n) is unit impulse response of
inverse system of the system with unit impulse response
h(n).
25
 DT LTI Systems:
Linear Constant Coefficient
Difference Equations

26
Linear Constant Coefficient
Difference Equation: LCCDE
 In general in a causal LTI system the output depends
on present input, M past inputs and N past outputs.
 A causal LTI system is described by following general
LCCDE:
a0 y ( n)  a1 y ( n  1)  a2 y ( n  2)  ........  a N y ( n  N )
 b0 x( n)  b1 x( n  1)  b2 x( n  2)  .......  bM x( n  M ); M N
N M
  ak y ( n  k )   bk x( n  k )
k 0 k 0

1  N M

OR y ( n)    k
 a y ( n  k )   bk x ( n  k ) 
a0  k 1 k 0 
 Here {ak};k=0,1,____,N and {bk};k=0,1,___M are
constant co-efficients.
 This LCCDE is called Nth order equation and the
system is called Nth order system. 27
LCCDE: Recursive and Nonrecursive
systems:
 A system whose output y(n) at a time n depends on any number
(N≥1) of previous output values [y(n-1), y(n-2),…,y(n-N)] is
called recursive system.
 If output y(n) of system is depends only on present and/or past
input signal samples, system is nonrecursive. Here N=0. It is
zeroth order system.
 In case of non recursive system: {ak}=0; k=1,2,..,N.
M
b 
 y (n)     x(n  k )
k

k 0 a  0 
 bn
 ;0  n  M
 h ( n )   a0
0; otherwise

 So, non recursive system has finite impulse response, while
recursive system has infinite impulse response.
 However, it should be noted that FIR systems can be
implemented nonrecursively or recursively, but IIR systems can
only be implemented recursively.
28
Block diagram representation of DT
LTI systems:
 Building blocks of DT-LTI systems:

 Adder

 Constant
Multiplier

 Delay Element

29
 Consider first order causal LTI system described by
LCCDE: y(n)-a y(n-1)=bx(n); a,b are constant.
 y(n)=a y(n-1)+bx(n)
 It s a recursive system and having feedback (closed
loop).

30
Examples:
 Represent in block diagram form.
1. y(n) – 0.75 y(n – 1) + 0.125 y(n – 2)= x(n)
2. y(n) – 0.7 y(n – 1) + 0.1 y(n – 2) = 2 x(n) – x(n – 2)
3. y(n) – 0.7 y(n – 1) + 0.12 y(n – 2) = x(n – 1) + x(n – 2)
4. y(n) – 0.9 y(n – 1) + 0.81 y(n – 2) = x(n)
5. y(n) - ½ y(n – 1) = x(n-1)
6. y(n) – 3 y(n – 1) – 4 y(n – 2) = x(n) + 2 x(n – 1)
7. y(n) – 4 y(n – 1) – 4 y(n-2) = x(n) + 3 x(n – 1)

31
Solution of Linear Constant
Coefficient Difference Equation
 The difference equation does not completely specify the
output in terms of input.
 To obtain it we have to solve LDDCE; for that we need
auxiliary conditions called initial conditions.
 We focus on initial rest conditions (causal and LTI
systems): i.e. x(n)=0 for n<no gives y(n)=0 for n<no.
 With initial rest condition, the system described by
difference equation is LTI and causal.
 For non recursive systems, we do not need any auxiliary
conditions to obtain output y(n).
 The simple way to solve LCCDE is using iterative method.
But it will not provide solution in closed form. 32
Causal LTI system analysis using
iterative (recursive) method:
Examples
 1. Obtain unit impulse response of the LTI system described by
following LCCDE. Assume initial rest (relaxed) conditions.
y(n)-a y(n-1)=x(n); a is constant.
Ans. h(n)=(a)nu(n)
 2. Obtain unit step response of the LTI system described by following
LCCDE. Assume initial rest conditions.
y(n)-a y(n-1)=x(n); a is constant.
Ans. s(n)=((1-an+1)/(1-a))u(n)
 3. Verify the relationship between unit impulse and unit step response
for example 1 and 2.
 4. Obtain unit impulse response of the LTI system described by
following LCCDE. Assume initial rest (relaxed) conditions.
y(n)-0.5y(n-1)=x(n-1)
Ans. h(n)=(0.5)n-1u(n-1)

33
Closed form solution of LCCDE:
 To obtain closed form solution, two methods can be used
to solve the linear constant coefficient difference equations:
direct (classical) method and indirect method (Using Z-
transform described in subsequent chapter).
 The direct solution method assumes that the total solution
is the sum of two parts:
 y(n) = yh(n) + yp(n)
 The part yh(n) is known as the homogeneous or
complementary solution, whereas yp(n) is called the
particular solution.
 The solution to the homogenous equation is often referred
to as the natural or transient response of the system.
 The particular solution is often called forced or steady state
response of the system as it depends on force (input)
applied.
34
Homogeneous solution:
 Assume that the input x(n) and all x(n-k) = 0
 ∑ ak yh(n – k) = 0
 Assume that the solution is in the form of an exponential: yh(n) = λn
 Substitute solution in characteristics equation
 ∑ ak λn-k = 0
 Solve it for λ: λ has N solutions λ1, λ2, …, λN: characteristic roots or
eigenvalues.
 If all characteristics roots are distinct:
yh(n) = c1λ1n + c2λ2n + … + cN λnN
 c1, c2, … , cN are arbitrary constants determined from N auxiliary
conditions.
 If two or more repeated roots: e.g. if a root λ1 repeats k times
yh(n) = (c1 + c2 n + c3 n2 + … + ck nk – 1)λ1n + ck+1λ2n + … + cN λnN
 It may be happened that, the complex roots occur in characteristics
equation in complex conjugate pairs; can be treated exactly as we
would treat real roots
35
Particular solution:
 For particular solution, it is necessary to find the sequence
yp(n) that satisfies the difference equation for given x(n) for
n > 0.
 To solve given difference equation, we assume for yp(n) will
have the same form as the input as shown in table.
 The constants Ki are found by substituting the solution
yp(n) into the difference equation and solving it for values
of n such that none of the term vanishes on both sides.
Input x(n); n>0 Particular solution yp(n) ; n>0
A K

An K1n + K2

A an (a ≠ λi ) K an

A an (a = λi ) K njan where j is the smallest integer that will make Knjan

different form characteristic mode term
A cos(nω0) K1 cos(nω0) + K2 sin(nω0)

A sin(nω0) K1 cos(nω0) + K2 sin(nω0)

A an sin(nω0) K1 an cos(nω0) + K2 an sin(nω0)

A cos (βn + θ ) K1 cos (βn + φ )

36
Examples:
 Determine the response of the following LTI systems
described by their constant coefficient difference equation
to the specified input. Assume all systems are initially
relaxed.
1. y(n) - a y(n – 1) = x(n) with x(n) = u(n)
2. y(n) – 0.75 y(n – 1) + 0.125 y(n – 2)= x(n) with x(n) = u(n)
3. y(n) – 0.7 y(n – 1) + 0.1 y(n – 2) = 2 x(n) – x(n – 2) with x(n) =
( ½ )n u(n)
4. y(n) – 0.7 y(n – 1) + 0.12 y(n – 2) = x(n – 1) + x(n – 2) with
x(n) = n u(n)
5. y(n) – 0.9 y(n – 1) + 0.81 y(n – 2) = x(n) with x(n) = u(n)
6. y(n) - ½ y(n – 1) = x(n) with x(n) = 10 cos (π/4 n) u(n)
7. y(n) – 3 y(n – 1) – 4 y(n – 2) = x(n) + 2 x(n – 1) with x(n) =
(½)n u(n)
8. y(n) – 4 y(n – 1) + 4 y(n-2) = x(n) + 3 x(n – 1) with x(n) =
n(⅓)n u(n)
37
Answers:
1−𝑎𝑛+1
1. 𝑢 𝑛
1−𝑎
1 1 𝑛 1 𝑛 8
2. 𝑢 𝑛 −2 𝑢 𝑛 + 𝑢 𝑛
3 4 2 3
1 𝑛 1 𝑛 10 1 𝑛
3. 12.2 𝑢 𝑛 − 10.2 𝑢 𝑛 − 𝑛 𝑢 𝑛
2 5 3 2
4. 48.8828 0.4 𝑛 𝑢 𝑛 − 36.528 0.3 𝑛 𝑢 𝑛 + 4.7619𝑛𝑢 𝑛 − 12.358𝑢 𝑛
𝑛 𝜋
5. −1.088 0.9 cos 𝑛 + 84.72° 𝑢 𝑛 + 1.09𝑢(𝑛)
3
1 𝑛 𝜋
6. −1.91 𝑢 𝑛 + 13.58 cos 𝑛 − 28.7° 𝑢(𝑛)
2 4
48 2 5 1 𝑛
7. 4 𝑛𝑢 𝑛 − −1 𝑛 𝑢 𝑛 − 𝑢 𝑛
35 15 21 2
𝑛 𝑛 1 𝑛 1 𝑛
8. 0.456 2 𝑢 𝑛 − 0.221𝑛 2 𝑢 𝑛 − 0.456 𝑢 𝑛 + 0.04𝑛 𝑢 𝑛
3 3

38
 Unit Impulse Response from
LCCDE:
 Case: M<N
 Obtain yh(n)
 Take yp(n)=0
 y(n)=yh(n)+yp(n)=yh(n)
 Find constants in y(n) by considering x(n)=δ(n) and
zero initial conditions.
 The obtained solution is unit impulse response h(n).

39
Examples:
 Determine the unit impulse response of the
following LTI systems described by their constant
coefficient difference equation. Assume all
systems are initially relaxed.
1. y(n) - a y(n – 1) = x(n)
2. y(n) – 0.75 y(n – 1) + 0.125 y(n – 2)= x(n)
3. y(n) – 0.9 y(n – 1) + 0.81 y(n – 2) = x(n)
4. y(n) – 3 y(n – 1) – 4 y(n – 2) = x(n) + 2 x(n – 1)
5. y(n) – 4 y(n – 1) + 4 y(n-2) = x(n) + 3 x(n – 1)

40
Answers:
1. h(n)=anu(n)
2. h(n)=[2(1/2)n-(1/4)n]u(n)
𝜋
3. h(n)=1.154 cos 𝑛 − 30° 𝑢 𝑛
3
6 1 𝑛
4. h(n)= (−1)𝑛 𝑢 𝑛 − 4 𝑢(𝑛)
5 5
5
5. h(n)= 2𝑛 + 𝑛2𝑛 𝑢(𝑛)
2

41
 DT LTI Systems:
Initially relaxed and Non
relaxed systems

42
Initially Non-relaxed System: Non-
zero initial conditions
 System has nonzero initial state
 At n=0 the input is applied to the system
 Two sources of energy: Initial state (existing n<0-) and Applied input
(exists for n>0+)
 n=0-: just before application of input
 n=0+: just after application of input
 Total Response= Response due to initial state only + Response due to
input only
 Due to linear system:
 Total Response=Zero input response + Zero state response
 y(n)=yzi(n)+yzs(n)
 Evaluation of yzi(n) involves only homogeneous solution but with
non-zero initial conditions
 Evaluation of yzs(n) involves both homogeneous solution and
particular solution with zero initial conditions.
 The zero state response can be also obtained using convolution integral
if we know the unit impulse response h(t) of the system.
 So far we have evaluated only zero state response in all previous
discussions. 43
Example:
 Determine zero input, zero state and total response of the LTI system
described by following LCCDE and initial conditions.
 (i) y(n) – 0.75 y(n – 1) + 0.125 y(n – 2)= x(n) with x(n) = u(n) and y(-1)
= y(-2) = 1
Solution:
−1 1 𝑛 3 1 𝑛
yzi(n)= 𝑢 𝑛 + 𝑢 𝑛
8 4 4 2
1 1 𝑛 1 𝑛 8
yzs(n)= 𝑢 𝑛 −2 𝑢 𝑛 + 𝑢 𝑛
3 4 2 3
5 1 𝑛 5 1 𝑛 8
y(n)= yzi(n) + yzs(n) = 𝑢 𝑛 − 𝑢 𝑛 + 𝑢 𝑛
24 4 4 2 3
(ii) y(n)-0.6y(n-1)-0.16y(n-2)=5x(n) with x(n)=4-nu(n) and y(-1)=0, y(-2) =
25/4
 Solution:
 yzi(n)=((1/5)(-0.2)n+(4/5)(0.8)n)u(n)
 yzs(n)=(0.444(-0.2)n+5.81(0.8)n-1.26(4)-n)u(n)
 y(n)=(0.644(-0.2)n+6.61(0.8)n-1.26(4)-n)u(n)
44
45