Chapter 2

Linear Time-Invariant Systems

(LTI Systems)
 Linear Time-Invariant Systems
• A system is said to be Linear Time-Invariant (LTI) if it possesses the basic system properties of linearity and
time-invariance.
• The input-output relationship for LTI systems is described in terms of a convolution operation.
DT Signal Decomposition in terms of shifted unit impulses 𝑥𝑛 𝑦𝑛
𝑥𝑛 LTI System

𝑥2 𝑥2
𝑥 −1 𝑥 1 𝑥 −1 𝑥1
𝑥 −2 𝑥0
𝑥 −2 𝑥0
∙∙∙ + ∙∙∙
+ + +
−2 −1 0 1 2
−2 −1 0 1 2 𝑛 𝑥 −2 𝛿[𝑛 + 2] 𝑥 −1 𝛿[𝑛 + 1] 𝑥 0 𝛿[𝑛] 𝑥 1 𝛿[𝑛 − 1] 𝑥 2 𝛿[𝑛 − 2]
Amplitude (constant)
∞ Impulse (signal) ෍𝛽 𝛿 𝑛 − 𝑘 ෍𝛽 ℎ 𝑛 − 𝑘

𝑥 𝑛 = ෍ 𝑥[𝑘]𝛿[𝑛 − 𝑘] 𝛿𝑛 ℎ𝑛 𝛿 𝑛−𝑘 ℎ 𝑛−𝑘

LTI System LTI System LTI System
𝑘=−∞
Sum of scaled impulses Impulse response Time invariance Linearity
∞ ∞
𝑥𝑛 𝑦 𝑛 = ෍ 𝑥 𝑘 ℎ[𝑛 − 𝑘] 𝑦 𝑛 = 𝑥[𝑛] ∗ ℎ[𝑛] = ෍ 𝑥 𝑘 ℎ[𝑛 − 𝑘]
LTI System 𝑘=−∞ 𝑘=−∞

Convolution sum
 𝑥𝑛 ℎ𝑛 𝑦𝑛
Convolution
∞
1

Example: Convolution 𝑦 𝑛 = ෍ 𝑥 𝑘 ℎ[𝑛 − 𝑘]

𝑘=−∞
Impulse response of 𝑦 𝑛 = ⋯ 𝑥 −1 ℎ 𝑛 + 1 + 𝑥 0 ℎ 𝑛 + 𝑥 1 ℎ 𝑛 − 1 + 𝑥 2 ℎ 𝑛 − 2 + ⋯
an LTI system
𝑥 0 ℎ 𝑛 − 0 = 0.5 ℎ[𝑛]
+
𝑥 1 ℎ 𝑛 − 1 = 2ℎ[𝑛 − 1]
input
𝑥 𝑛 has only 𝑥 0 and 𝑥 1

Convolution
Find output
∞ ∞ =
𝑦 𝑛 = 𝑥[𝑛] ∗ ℎ[𝑛] = ෍ 𝑥 𝑘 ℎ[𝑛 − 𝑘] = ෍ ℎ 𝑘 𝑥[𝑛 − 𝑘]
𝑘=−∞ 𝑘=−∞
There are only two non-zero values for the input

𝑦 𝑛 = 0 + 𝑥 0 ℎ 𝑛 − 0 + 𝑥 1 ℎ 𝑛 − 1 + 0 = 0.5ℎ 𝑛 + 2 ℎ 𝑛 − 1
 Convolution 2
Example:

3 𝑛=0 3 𝑛=0
ℎ𝑛 = 2 𝑛=1
𝑥𝑛 = 1 𝑛=1
1 𝑛=2 2 𝑛=2
0 𝑒𝑙𝑠𝑒𝑤ℎ𝑒𝑟𝑒 0 𝑒𝑙𝑠𝑒𝑤ℎ𝑒𝑟𝑒
Length =3 Length =3

∞
Solution:
𝑦 𝑛 = 𝑥[𝑛] ∗ ℎ[𝑛] = ෍ 𝑥 𝑘 ℎ[𝑛 − 𝑘]
𝑘=−∞

9 𝑛=0
9 𝑛=1
11 𝑛=2
𝑦𝑛 = 𝑛=3
5
2 𝑛=4
0 𝑒𝑙𝑠𝑒𝑤ℎ𝑒𝑟𝑒

Convolution Length = 𝑁1 + 𝑁2 − 1 = 3 + 3 − 1 = 5
 Convolution 3
Example: Find the output of an LTI system having a unit impulse response ℎ 𝑛 = 𝑢[𝑛], for the input 𝑥 𝑛 =∝𝑛 𝑢 𝑛
∞

𝑦 𝑛 = ෍ 𝑥 𝑘 ℎ[𝑛 − 𝑘]
𝑘=−∞

𝑛≥0
∝ 𝑘 0≤𝑘≤𝑛 𝑘 >𝑛 →ℎ 𝑛−𝑘 =0
𝑥 𝑘 ℎ[𝑛 − 𝑘] = ቊ 𝑘<0 →𝑥 𝑘 =0
0 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒

(Shifting right) 𝑛 𝑛+1 𝑦𝑛

1−∝
Thus, for 𝑛 ≥ 0, 𝑦 𝑛 = ෍ ∝𝑘 =
1 −∝
𝑘=0
output
𝑛

𝑠 = ෍ ∝𝑘 = 1 +∝ +∝2 + ⋯ +∝𝑛
𝑛<0 𝑘=0

∝ 𝑠 = ∝ +∝2 + ⋯ +∝𝑛+1
(Shifting left) 𝑠 −∝ 𝑠 = 1−∝𝑛+1 Thus, for all 𝑛,
output 1−∝𝑛+1
Thus, for 𝑛 < 0, 𝑦 𝑛 = 0 𝑦𝑛 = 𝑢[𝑛]
1 −∝
 The Representation of Continuous-Time Signals in Terms of Impulses
𝑥(𝑡) is approximated in
𝑥(𝑡) Defining:
terms of pulses or staircase

1ൗ 0≤𝑡≤Δ
𝛿Δ 𝑡 = ൝ Δ
0 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒

𝑡 1 0≤𝑡≤Δ
Δ𝛿Δ 𝑡 = ቊ
0 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
𝑥ො 𝑡 : the approximation of 𝑥(𝑡)
Δ𝛿Δ 𝑡 has a unit amplitude
𝑥(0) 0 ≤ 𝑡 ≤ Δ
At 𝑡 = 0 𝑥ො 0 = 𝑥(0)Δ𝛿Δ 𝑡 = ቊ
0 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
The complete pulse/staircase approximation of
𝑥(𝑡)is the sum
At 𝑡 = Δ 𝑥(Δ) Δ ≤ 𝑡 ≤ 2Δ
𝑥ො Δ = 𝑥(Δ)Δ𝛿Δ 𝑡 − Δ = ቊ
0 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
𝑥ො 𝑡 =∙∙∙ +𝑥ො −Δ + 𝑥ො 0 + 𝑥ො Δ +∙∙∙
∞
In general At 𝑡 = 𝑘Δ 𝑥ො 𝑡 = ෍ 𝑥(𝑘Δ)𝛿Δ 𝑡 − 𝑘Δ Δ
𝑘=−∞
𝑥(𝑘Δ) 𝑘Δ ≤ 𝑡 ≤ (𝑘 + 1)Δ
𝑥ො 𝑘Δ = 𝑥(𝑘Δ)Δ𝛿Δ 𝑡 − 𝑘Δ = ቊ
0 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
 The Representation of CT Signals in Terms of Impulses2
If we let Δ approach 0, 𝑥ො 𝑡 becomes closer and closer and equals 𝑥(𝑡) in the limit of 0

𝑥 𝑡 = lim 𝑥ො 𝑡 = lim ෍ 𝑥(𝑘Δ)𝛿Δ 𝑡 − 𝑘Δ Δ

Δ→0 Δ→0
𝑘=−∞

In the limiting case the sum approaches integral (area):

Δ → 0; 𝛿Δ 𝑡 → 𝛿(𝑡)
∞
∞ ∞
෍ ∙∙∙ Δ → න ∙∙∙ 𝑑𝜏 Consequently: 𝑥 𝑡 =න 𝑥 𝜏 𝛿 𝑡 − 𝜏 𝑑𝜏
𝜏=−∞ 𝜏=−∞
𝑘=−∞

Convolution integral
 The Convolution Integral1
• The impulse response ℎ(𝑡) of a continuous-time LTI system 𝑆 𝛿(𝑡) ℎ(𝑡)
CT LTI System
ℎ 𝑡 = 𝑆{𝛿 𝑡 } Sum (Integral) of weighted shifted impulses
∞ linearity ∞
𝑦 𝑡 =𝑆 𝑥 𝑡 =𝑆 න 𝑥 𝜏 𝛿 𝑡 − 𝜏 𝑑𝜏 = න 𝑥 𝜏 𝑆{𝛿 𝑡 − 𝜏 } 𝑑𝜏
For the input 𝑥(𝑡): 𝜏=−∞ 𝜏=−∞
weight impulse
• Since the system is time-invariant: ∞
Time-invariance
𝑆 𝛿 𝑡−𝜏 = ℎ(𝑡 − 𝜏) 𝑦 𝑡 =න 𝑥 𝜏 ℎ 𝑡 − 𝜏 𝑑𝜏 = 𝑥(𝑡) ∗ ℎ(𝑡)
𝜏=−∞
Example 1
Let, the input 𝑥 𝑡 = 𝑒 −𝑎𝑡 𝑢(𝑡) for 𝑎 > 0 to an LTI system
with unit impulse response ℎ 𝑡 = 𝑢(𝑡). Find the output
𝑦 𝑡 of the system.
∞
𝑦 𝑡 = 𝑥(𝑡) ∗ ℎ(𝑡) = න 𝑥 𝜏 ℎ 𝑡 − 𝜏 𝑑𝜏 𝑥 𝑡 ≠ 0 𝑓𝑜𝑟 𝑡 ≥ 0
𝜏=−∞ 1 𝜏≤𝑡
𝑓𝑜𝑟 𝑡 < 0 No overlap 𝑦 𝑡 = 0 ℎ 𝑡−𝜏 =𝑢 𝑡−𝜏 =ቊ
0 𝜏>𝑡
𝑓𝑜𝑟 𝑡 ≥ 0 ∞ 𝑥 𝑡 ℎ 𝑡−𝜏 =ቊ
𝑥 𝑡 0<𝜏≤𝑡
0 𝜏>𝑡
𝑦 𝑡 = න 𝑒 −𝑎𝜏 ℎ 𝑡 − 𝜏 𝑑𝜏 𝑓𝑜𝑟 𝑡 > 0
0 Thus, for all t, we can write
𝑡
1 −𝑎𝜏 1
𝑡 1
=න𝑒 −𝑎𝜏
𝑑𝜏 = 𝑒 ቤ = 1 − 𝑒 −𝑎𝑡 𝑦 𝑡 = 1 − 𝑒 −𝑎𝑡 𝑢(𝑡)
−𝑎 𝑎 𝑎
0 0
 The Convolution Integral2
𝑥 𝜏 = 𝑒 2𝑡 𝑢(−𝜏)
Example 2 ℎ 𝜏
𝑥 𝑡 = 𝑒 2𝑡 𝑢(−𝑡)
Find 𝑦 𝑡 = 𝑥(𝑡) ∗ ℎ(𝑡) , where
ℎ 𝑡 = 𝑢(𝑡 − 3)
∞
The system response is 𝑦 𝑡 = ‫=𝜏׬‬−∞ 𝑥 𝜏 ℎ 𝑡 − 𝜏 𝑑𝜏
𝜏𝜏
these two signals have regions of nonzero overlap
For 𝒕 − 𝟑 ≤ 𝟎: nonzero overlap for −∞ ≤ 𝜏 ≤ 𝑡 − 3 ℎ 𝑡−𝜏
Flip and shift by t

𝑡−3 𝑡−3
1 2𝜏 1 2(𝑡−3)
𝑦 𝑡 =න 𝑒 2𝜏 𝑑𝜏 = 𝑒 ቤ = 𝑒 For 𝑡 ≤ 3
−∞ 2 −∞
2

For 𝒕 − 𝟑 ≥ 𝟎: nonzero overlap for −∞ ≤ 𝜏 ≤ 0 𝜏

0 0
1 1 𝑦 𝑡
𝑦 𝑡 =න 𝑒 2𝜏 𝑑𝜏 = 𝑒 2𝜏 ቤ = For 𝑡 ≥ 3
−∞ 2 −∞
2

1 2(𝑡−3)
𝑒 For 𝑡 ≤ 3
𝑦 𝑡 =൞ 2
1ൗ For 𝑡 ≥ 3 𝑡
2
 𝑦 𝑡 =න
∞
𝑥 𝜏 ℎ 𝑡 − 𝜏 𝑑𝜏
The Convolution Integral3
𝜏=−∞

Convolution

Shift ℎ 𝑡 − 𝜏

Aggregate the overlapping

𝒕<𝟎 No overlapping
𝑦 𝑡 =0
𝑡
𝑦 𝑡 = න 2 ∙ 1 𝑑𝜏 0<𝜏<𝑡
0 𝟎<𝒕<𝟐
𝟎<𝒕<𝟐 𝑡
𝑦 𝑡 = 2 𝜏ቚ
0

𝑦 𝑡 = 2𝑡
 The Convolution Integral4

𝟐<𝒕<𝟒

𝟒<𝒕<𝟔

No-overlap 𝑦 𝑡 = 0
𝒕>𝟔
𝟎 𝟏 𝟐 𝟑 𝟒 𝟓 𝟔 𝟕
𝒕−𝟒 𝒕
 Properties of LTI Systems
• The characteristics of an LTI system are completely determined by its impulse response ℎ 𝑛 . This property holds
in general only for LTI systems only.

• The unit impulse response of a nonlinear system does not completely characterize the behavior of the system.

Consider a discrete-time system with unit impulse response:

If the system is LTI, we get the system output (by convolution): 𝑦 𝑛 = ෍ ℎ 𝑘 𝑥[𝑛 − 𝑘]
𝑘=−∞
There is only one such LTI system for the given h[n].

However, there are many nonlinear systems with the same response, h[n]. Two different Non-Linear systems
with same impulse response

2 1, 𝑛 = 0,1
ℎ 𝑛 = 𝛿 𝑛 +𝛿 𝑛−1 ℎ 𝑛 =ቊ
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
1, 𝑛 = 0,1
ℎ 𝑛 = 𝑀𝑎𝑥 𝛿 𝑛 , 𝛿 𝑛 − 1 ℎ 𝑛 =ቊ
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
 1 Commutative Property

Proof: (discrete domain)

Put 𝑟 = 𝑛 − 𝑘 ⇒ 𝑘 = 𝑛 − 𝑟

Similarly, we can prove it for continuous domain.

 2 Distributive Property 𝑥 𝑛 ∗ ℎ1 𝑛

Convolution is distributive over addition,

in discrete time
𝑥 𝑛 ∗ ℎ2 𝑛

𝑥 𝑛 ∗ ℎ1 𝑛 + 𝑥 𝑛 ∗ ℎ2 𝑛
in continuous time

Example:
𝑢 𝑛 = 1, 𝑛 ≥ 0
and
x[n] in nonzero for entire n, so direct convolution is difficult. Therefore, we will use commutative property.

1𝑓𝑜𝑟 𝑘 ≥ 0 1𝑓𝑜𝑟 𝑘 ≤ 𝑛
𝑛 𝑘
1
=෍
2
𝑘=0
𝑙 = −𝑘 𝑙 𝑚 =𝑙+𝑛 𝑚−𝑛 𝑚
𝑛 ∞ 1 ∞ 1 ∞ 1 1
𝑘 𝑛
෍ 2 =෍ =෍ =2 ෍ = 2𝑛 = 2𝑛+1
𝑘=−∞ 𝑙=−𝑛 2 𝑚=0 2 𝑚=0 2 1
1−
2

1𝑓𝑜𝑟 𝑘 ≤0 1𝑓𝑜𝑟 𝑘 ≤ 𝑛
 3 Associative Property

in continuous time

In discrete time

Proof:
 4 LTI Systems With and Without Memory
system
A system is memory-less if its output at any time depends only on the value 𝑥[𝑛] 𝑦[𝑛]
of the input at that same time. ℎ[𝑛]

𝑦[𝑛] depends on only 𝑥[𝑛] only if k = 𝑛,

System output: y n = 𝑥 𝑛 ℎ 0 = 𝑘 𝑥[𝑛]

A discrete-time LTI system can be memory-less if only:

Thus, the impulse response have the form: ℎ 𝑛 = 𝐾𝛿 𝑛 , with 𝐾 = ℎ 0 is a constant

the convolution sum reduces to

If k = 1, then the system is called identity system.

Similarly for continuous LTI systems.

a continuous-time LTI system is memory-less if
 5 Invertibility of LTI Systems
A system is invertible only if the input of the system appears at the output (if an inverse system exists, if distinct inputs
Leads to distinct outputs)
The system with impulse response ℎ1 (𝑡) is inverse of
the system with impulse response ℎ 𝑡 if:

∞
Proof If ℎ t = 𝛿 𝑡 𝑦 𝑡 = න 𝑥 𝜏 𝛿 𝑡 − 𝜏 𝑑𝜏 = 𝑥(𝑡)
−∞
Example1 CTS:
𝑡0 > 0 delay
Consider the LTI system consisting of a pure time shift 𝑦 𝑡 = 𝑥(𝑡 − 𝑡0 )
𝑡0 < 0 advance
The impulse response for the system (for 𝑥 𝑡 = 𝛿(𝑡)): ℎ 𝑡 = 𝛿(𝑡 − 𝑡0 ) impulse response 𝑥(𝑡) = 𝛿(t)

the system’s output (the convolution): 𝑦 𝑡 = 𝑥 𝑡 ∗ ℎ 𝑡 = 𝑥 𝑡 ∗ 𝛿 𝑡 − 𝑡0 = 𝑥(𝑡 − 𝑡0 )

To recover the input from the output (invert the system), all that is required is to shift the output back.
𝜏 = −1 𝜏 = 0 𝜏=5 𝜏=8
𝜏
The inverse system impulse response: ℎ1 𝑡 = 𝛿(𝑡 + 𝑡0 ) 𝛿 𝑡 0
∞ 0 0
then 𝛿 𝜏−5 𝛿 𝑡+5−𝜏 1 𝑡=0
𝑥 𝑡 ∗ ℎ 𝑡 = න 𝑥 𝜏 𝛿 𝑡 − 𝜏 𝑑𝜏 𝛿 𝑡 =ቊ
−∞ 0 𝑡≠0
1𝑓𝑜𝑟 𝜏 = 𝑡0
∞
ℎ 𝑡 ∗ ℎ1 𝑡 = 𝛿 𝑡 − 𝑡0 ∗ 𝛿 𝑡 + 𝑡0 = න 𝛿 𝜏 − 𝑡0 𝛿 𝑡 + 𝑡0 − 𝜏 𝑑𝜏 = 𝛿(𝑡) identity system (𝑦 𝑡 = 𝑥 𝑡 ∗ 𝛿 𝑡 = 𝑥(𝑡))
−∞
 Invertibility of LTI Systems: Example2 DTS
Consider an LTI system with impulse response: ℎ 𝑛 = 𝑢[𝑛] 𝑢 𝑛 − 𝑘 = 0 𝑓𝑜𝑟 𝑘 > 𝑛
summer or accumulator
∞ 𝑛
Response of this system (convolution sum): 𝑦 𝑛 =෍ 𝑥 𝑘 𝑢[𝑛 − 𝑘] 𝑦 𝑛 =෍ 𝑥𝑘
𝑘=−∞ 𝑘=−∞
first difference equation
𝑛−1
Inverse system
𝑦 𝑛 =𝑥 𝑛 +෍ 𝑥𝑘 𝑦 𝑛 = 𝑥 𝑛 + 𝑦[𝑛 − 1] 𝑥 𝑛 = 𝑦 𝑛 − 𝑦[𝑛 − 1] 𝑦 𝑛 = 𝑥 𝑛 − 𝑥[𝑛 − 1]
𝑘=−∞

Impulse response (𝑥 𝑛 = 𝛿[𝑛]): ℎ1 𝑛 = 𝛿 𝑛 − 𝛿[𝑛 − 1]

𝑢𝑛
1

Verification: ℎ 𝑛 ∗ ℎ1 𝑛 = 𝛿 𝑛
−2 −1 0 1 2 3 4 𝑛
𝑢 𝑛−1 1
ℎ 𝑛 ∗ ℎ1 𝑛 = 𝑢[𝑛] ∗ 𝛿 𝑛 − 𝛿[𝑛 − 1]
= 𝑢[𝑛] ∗ 𝛿 𝑛 − 𝑢[𝑛] ∗ 𝛿[𝑛 − 1] 𝑛
1 2 3 4
−2 −1 0

= 𝑢[𝑛] − 𝑢[𝑛 − 1] Identity system ℎ 𝑛 = 𝛿 𝑛 𝑢 𝑛 −𝑢 𝑛−1

1
=𝛿 𝑛
−2 −1 1 1 2 3 4 𝑛
ℎ 𝑛 ∗ ℎ1 𝑛 = 𝛿 𝑛
the impulse responses are inverses of each other
 6 Causality of LTI Systems
• The output of a causal system depends only on the present and past values of the input to the system.
∞

𝑦 𝑛 = ෍ 𝑥 𝑘 ℎ[𝑛 − 𝑘] 𝑦 𝑛 must not depend on 𝑥 𝑘 for 𝑘 > 𝑛 to be causal

𝑘=−∞

Therefore, for a discrete-time LTI system to be causal: 𝑥 𝑘 ℎ 𝑛 − 𝑘 = 0 for 𝑘 > 𝑛

𝛽 =𝑛−𝑘
ℎ 𝑛 − 𝑘 = 0 for 𝑛 − 𝑘 < 0 ℎ 𝛽 = 0 for 𝛽 < 0 ℎ 𝑛 =0 for 𝑛 < 0

Causality for LTI system is equivalent to the condition of initial rest (output must be 0 before applying the input)
𝑓𝑜𝑟 𝑘 > 𝑛 ∶ ℎ 𝑛 − 𝑘 = 0 𝑓𝑜𝑟 𝑘 < 0 ∶ ℎ 𝑘 = 0

Both the accumulator (ℎ 𝑛 = 𝑢[𝑛]) and its

inverse (ℎ 𝑛 = 𝛿 𝑛 − 𝛿 𝑛 − 1 ) are causal.

• Similarly, for a continuous-time LTI system to be causal:

Past and present values of 𝑥(𝑡 − 𝜏) are for 𝜏 ≥ 𝑡

⇒ the output 𝑦(𝑡 )depends only on past and present values of 𝑥(𝑡) if ℎ(𝑡) = 0 for 𝑡 < 0
 7 Stability of LTI Systems
• A system is stable if every bounded input produces a bounded output (BIBO).

Consider, an input x[n] to an LTI system that is bounded in magnitude:

Suppose that we apply this to the LTI system with impulse response h[n].
We take max(𝑥 𝑛 ) = 𝐵 𝑎+𝑏 ≤ 𝑎 + 𝑏

Example:
Therefore, if
An LTI system with pure time shift is stable.

The system is stable if the impulse response ℎ 𝑛 is

absolutely summable.
• Similar case in continuous-time LTI system. An accumulator (DT domain) system is unstable.

the system is stable if the impulse response is

absolutely integrable. Similarly, an integrator (CT domain) system is unstable.
 8 Unit Step Response of An LTI System
• the unit step response 𝑠[𝑛] or 𝑠(𝑡), the output corresponds to the input 𝑥 𝑛 = 𝑢[𝑛] or 𝑥 𝑡 = 𝑢 𝑡 .
• it is worthwhile relating the unit step response to the impulse response
Response to the input h[n] of a
commutative property
LTI system with unit impulse
∞ 1 𝑓𝑜𝑟 𝑘 ≤ 𝑛 𝑛
response u[n].
Convolution 𝑠 𝑛 =෍ ℎ 𝑘 𝑢[𝑛 − 𝑘] = ෍ ℎ𝑘 an accumulator
𝑘=−∞ 𝑘=−∞
𝑛−1 impulse response of an accumulator
𝑛
𝑠 𝑛 =ℎ 𝑛 +෍ ℎ 𝑘 = ℎ 𝑛 + 𝑠[𝑛 − 1] ℎ 𝑛 = 𝑠 𝑛 − 𝑠[𝑛 − 1] ℎ𝑛 =෍ 𝛿[𝑘] = ቄ
1 𝑛≥0
= 𝑢[𝑛]
𝑘=−∞ 𝑘=−∞ 0 𝑛<0
𝑢[𝑛] is the unit impulse response of the accumulator. Example: Find the step response 𝑠 𝑛 of
a system with impulse response ℎ 𝑛 :
𝑛 𝑛
1
ℎ𝑛 = 𝑢[𝑛] 𝑠 𝑛 = ෍ ℎ[𝑘]
2
𝑘=−∞
Relation between step response s[n] and impulse response h[n] 𝑛 𝑘 𝑛 𝑘
1 1
𝑠𝑛 = ෍ 𝑢[𝑘] = ෍
2 2
𝑘=−∞ 𝑘=0
𝑛+1
1 𝑛+1
1− 2 1
= =2 1− 𝑢[𝑛]
1 2
1−2
 LTI Systems Described by Differential Equation
(Linear Constant-Coefficient Differential Equation)

A general Nth-order linear constant-coefficient differential equation 𝑁 𝑘 𝑀 𝑘

that relates the input 𝑥(𝑡) to the output 𝑦(𝑡) is given by 𝑑 𝑦(𝑡) 𝑑 𝑥(𝑡)
෍ 𝑎𝑘 = ෍ 𝑏𝑘
𝑑𝑡𝑘 𝑑𝑡𝑘
Example1:
𝑘=0 𝑘=0

consider a first-order differential equation 𝑎1 = 1, 𝑎0 = 2 𝑎𝑛𝑑 𝑏0 = 1

where the input to the system is:

𝑦𝑝 (𝑡) the particular solution

The complete solution is
𝑦ℎ (𝑡) The homogeneous solution,
 Finding the particular solution 𝑦𝑝 (𝑡) (signal of the
 Finding the homogeneous solution (x(t)=0, first order system
same form as the input: forced response)
response ) 𝑦ℎ 𝑡 = 𝐴𝑒 𝑠𝑡
Same form as input 𝑦𝑝 𝑡 = 𝑌𝑒 3𝑡 Determine 𝑌 Determine 𝑠 and A

From differential equation: From differential equation:

𝐾 𝑠𝐴𝑒 𝑠𝑡 + 2𝐴𝑒 𝑠𝑡 = 0 ⇒ 𝐴 𝑠 + 2 𝑒 𝑠𝑡 = 0 ⇒ 𝑠 = −2
3𝑌𝑒 3𝑡 + 2𝑌𝑒 3𝑡 = 𝐾𝑒 3𝑡 ⇒ 3𝑌 + 2𝑌 = 𝐾 ⇒ 𝑌 =
5 𝑦ℎ 𝑡 = 𝐴𝑒 −2𝑡
𝐾
𝑦𝑝 𝑡 = 5 𝑒 3𝑡 , 𝐾 𝑟𝑒𝑎𝑙 𝑎𝑛𝑑 𝑡 > 0
Complete solution:
 Example_contd
• To find A suppose that the auxiliary condition is 𝑦 0 = 0, i.e. , at t = 0, 𝑦 𝑡 = 0
Using this condition into the complete solution, we get:
𝐾 3𝑡
𝑦 𝑡 = 𝑒 + 𝐴𝑒 −2𝑡 With 𝑦 0 =0
5

Example2: (Discrete System)

1
Find 𝑦 𝑛 of the system with the difference equation 𝑦 𝑛 − 𝑦 𝑛 − 1 = 𝑥[𝑛]
2
1
We have the output 𝑦 𝑛 = 𝑥 𝑛 + 𝑦 𝑛−1
2
Consider the input 𝑥 𝑛 = 𝑘 𝛿[𝑛] and initial condition 𝑦 −1 = 0 (rest)
1
𝑦 0 = 𝑥 0 + 𝑦 −1 = 𝑘
2
1 1 Setting 𝑘 = 1 we obtain the impulse response for the system
𝑦 1 =𝑥 1 + 𝑦 0 = 𝑘 𝑛
2 2 2 1
1 1 ℎ𝑛 = 𝑢[𝑛]
𝑦 2 =𝑥 2 + 𝑦 1 = 𝑘 2
2 2
⋮ 𝑛 impulse response with infinite duration
1 1
𝑦 𝑛 =𝑥 𝑛 + 𝑦 𝑛−1 = 𝑘 infinite impulse response ( IIR) systems.
2 2
 Block Diagram Representations of Systems
systems described by linear constant -coefficient difference and differential equations can be represented in terms
of block diagram interconnections of elementary operations (adder, scaler, unit delay).

adder scaler
unit delay
Example:
Consider the causal system described Consider the causal continuous−time system
by the first-order difference equation described by a first−order differential equation
𝑦 𝑛 + 𝑎 𝑦 𝑛 − 1 = 𝑏 𝑥[𝑛] 𝑑𝑦 𝑡
+ 𝑎𝑦 𝑡 = 𝑏 𝑥(𝑡)
𝑦 𝑛 = 𝑏 𝑥[𝑛] − 𝑎 𝑦 𝑛 − 1 𝑑𝑡
𝑏 1 𝑑𝑦 𝑡
𝑦 𝑡 = 𝑥 𝑡 −
𝑎 𝑎 𝑑𝑡