COMS3100/7100
Introduction to Communications
Lecture 4: LTI Systems
This lecture: 1. The Impulse Response & Convolution 2. Interconnections of LTI Systems 3. System Properties & the Impulse Response 4. The Step Response Ref: HvV pp. 97141, CCR ch. 3, Couch ch. 2
COMS3100/COMS7100 Intro to Communications
Lecture 4 - LTI Systems
1 / 21
The Impulse Response & LTI
One of the most important properties of an LTI system is that the system is characterised by its impulse response. Given the response of the system to an impulse, the response to any other signal can be computed in a straightforward manner.
COMS3100/COMS7100 Intro to Communications
Lecture 4 - LTI Systems
2 / 21
The Impulse Response & not-LTI
As the name suggests the impulse response is the response of a system given an impulse. All systems have this but only in LTI systems does this allow us to characterise the response to other input signals using this.
COMS3100/COMS7100 Intro to Communications
Lecture 4 - LTI Systems
3 / 21
The Discrete-Time Case
Given an input impulse, let the output of the system the discrete-time impulse response be denoted h[n].
[n ]
h [n ]
COMS3100/COMS7100 Intro to Communications
Lecture 4 - LTI Systems
4 / 21
Discrete-Time Case - Convolution
The output of a discrete-time LTI system is the discrete-time linear convolution of the input with the impulse response. We write
y [n ] = x [n ] h [n ] =
k=
x[k]h[n k].
(1)
COMS3100/COMS7100 Intro to Communications
Lecture 4 - LTI Systems
5 / 21
COMS3100/COMS7100 Intro to Communications
Lecture 4 - LTI Systems
6 / 21
Convolution Example
Find y[n] = h[n] x[n] where, h[n] = (t) + 3(t + 2) and x[n] = n2 .
y[0] = x[0] y[1] = x[1] y[2] = x[2] + 3x[0] y[3] = x[3] + 3x[1]
COMS3100/COMS7100 Intro to Communications
Lecture 4 - LTI Systems
7 / 21
Convolution Example
Find y[n] = h[n] x[n] where, h[n] = (t) + 3(t + 2) and x[n] = n2 .
y[0] = x[0] y[1] = x[1] y[2] = x[2] + 3x[0] y[3] = x[3] + 3x[1]
COMS3100/COMS7100 Intro to Communications
Lecture 4 - LTI Systems
7 / 21
Convolution Example
Find y[n] = h[n] x[n] where, h[n] = (t) + 3(t + 2) and x[n] = n2 .
y[0] = x[0] y[1] = x[1] y[2] = x[2] + 3x[0] y[3] = x[3] + 3x[1]
COMS3100/COMS7100 Intro to Communications
Lecture 4 - LTI Systems
7 / 21
Convolution Example
Find y[n] = h[n] x[n] where, h[n] = (t) + 3(t + 2) and x[n] = n2 .
y[0] = x[0] y[1] = x[1] y[2] = x[2] + 3x[0] y[3] = x[3] + 3x[1]
COMS3100/COMS7100 Intro to Communications
Lecture 4 - LTI Systems
7 / 21
Convolution Example
Find y[n] = h[n] x[n] where, h[n] = (t) + 3(t + 2) and x[n] = n2 .
y[0] = x[0] y[1] = x[1] y[2] = x[2] + 3x[0] y[3] = x[3] + 3x[1]
COMS3100/COMS7100 Intro to Communications
Lecture 4 - LTI Systems
7 / 21
The Continuous-Time Case
We can now develop the idea of an impulse response for a continuous-time LTI system in a nearly analogous fashion to discrete time. The following identity is the continuous-time analogue of (1): x (t) = x ( ) (t ) d .
COMS3100/COMS7100 Intro to Communications
Lecture 4 - LTI Systems
8 / 21
Continuous-Time Case Properties
If the input to a continuous-time LTI system is (t), label the output h (t) the continuous-time impulse response.
(t )
/
h (t)
COMS3100/COMS7100 Intro to Communications
Lecture 4 - LTI Systems
9 / 21
Continous-Time Case - Convolution
The output is the continuous-time linear convolution of x (t) and h (t). We write x ( )h (t ) d . y (t) = x (t) h (t) =
COMS3100/COMS7100 Intro to Communications
Lecture 4 - LTI Systems
10 / 21
Interconnected LTI Systems
Cascaded Connection of LTI Systems Suppose we cascade LTI systems with impulse responses h1 [n], h2 [n].
x [n ]
h 1 [n ]
x [n ] h 1 [n ]
/
h 2 [n ]
y [n ]
COMS3100/COMS7100 Intro to Communications
Lecture 4 - LTI Systems
11 / 21
LTI Systems - Associativity
We can eectively combine the two systems because convolution is associative, i.e., (x[n] h1 [n]) h2 [n] = x[n] (h1 [n] h2 [n]).
x [n ]
/
h 1 [n ] h 2 [n ]
/
y [n ]
COMS3100/COMS7100 Intro to Communications
Lecture 4 - LTI Systems
12 / 21
LTI Systems - Commutative
We can also swap the order of cascaded LTI systems because convolution is commutative, i.e., h 1 [n ] h 2 [n ] = h 2 [n ] h 1 [n ].
x [n ]
/
h 2 [n ]
x [n ] h 2 [n ]
/
h 1 [n ]
y [n ]
COMS3100/COMS7100 Intro to Communications
Lecture 4 - LTI Systems
13 / 21
Parallel Connection of LTI Systems
Suppose we connect two LTI systems in parallel.
/ h1 [n]
x[n]
/ / h2 [n]
H OI N M JK L + O
/ y[n]
Convolution is distributive over addition, i.e.,
x[n] (h1 [n] + h2 [n]) = x[n] h1 [n] + x[n] h2 [n].
COMS3100/COMS7100 Intro to Communications
Lecture 4 - LTI Systems
14 / 21
LTI Systems - Combing
Therefore, we can combine the two systems:
x[n]
/
h1 [n] + h2 [n]
/ y[n]
Each of these results can be carried over into continuous time.
COMS3100/COMS7100 Intro to Communications
Lecture 4 - LTI Systems
15 / 21
System Properties - Memory
Three of the properties previously examined for systems namely memory, causality & stability are reected in the form of the impulse response in LTI systems. If an LTI system is memoryless then the impulse response must satisfy
h[n] = 0
when n = 0
or
h (t) = 0
when t = 0.
COMS3100/COMS7100 Intro to Communications
Lecture 4 - LTI Systems
16 / 21
System Properties - Casual
If an LTI system is causal then
h [n ] = 0
when n < 0
or
h (t) = 0
when t < 0.
Some authors, e.g., CCR, also call any signal with the above properties causal. I prefer the term right-sided for these signals.
COMS3100/COMS7100 Intro to Communications
Lecture 4 - LTI Systems
17 / 21
System Properties - Stability + Impulse Response
If an LTI system is stable then the impulse response must be absolutely summable in the discrete-time case or absolutely integrable in the continuous-time case, i.e., | h [n ]| < |h (t)| dt < . or
n =
COMS3100/COMS7100 Intro to Communications
Lecture 4 - LTI Systems
18 / 21
Step Response
Since impulses are dicult to generate and work with, real LTI systems (like RLC networks) are often experimentally characterised by measuring their step response instead.
COMS3100/COMS7100 Intro to Communications
Lecture 4 - LTI Systems
19 / 21
Step Response (2)
The step response is simply a systems output given a step input.
u (t)
/ s (t)
u [n ]
/ s [n ]
COMS3100/COMS7100 Intro to Communications
Lecture 4 - LTI Systems
20 / 21
Step Response (2)
For the discrete-time step response, we have
s[n] = u[n] h[n] =
k=
h[k]u[n k] =
n k=
h[k].
Similarly, in continuous time, s (t) = u (t) h (t) = h ( )u (t ) d =
h ( ) d .
Hence, the step response yields the impulse response since
h[n] = s[n] s[n 1]
and
h (t) =
d s (t). dt
COMS3100/COMS7100 Intro to Communications
Lecture 4 - LTI Systems
21 / 21
