Linear Time-Invariant (LTI) Systems

Dr. Ray Kwok

SJSU
Fall 2013
The unit impulse function

The unit impulse function AKA Dirac delta function is given as:

Very large, t = 0 (t )
(t ) =
0, otherwise
1

K K t
2 1 0 1 2 3

The value at t = 0 is very large, and ( t ) = 0 for t =/ 0

The duration is very short.

The area is one.

The unit impulse function is a mathematical model to represent signals that are
highly localized in time.

Dr. Ray Kwok LTI Systems 2

Systems, Networks, and Circuits

Network theory is mainly concerned with network topology (interconnections of
component)

System theory is mainly concerned with input - output relationship.

Electrical system is divided to:
 Linear system described by set of linear equations.
 nonlinear system described by set of nonlinear equations.

And into:
 Time-invariant
 time-varying system

And into:
 Passive system
 Active system

Also
 Lumped system
 Distributed system

And

Dr. Ray Kwok LTI Systems 3

Mathematical model

To aid analyzing and designing systems, mathematical models are formulated.

A mathematical model describes the behavior of physical system or device in
terms of a set of mathematical equations, with schematic diagram of the device
connection and the symbols of its component and numerical values.

Several models for continuous systems and techniques for system design and
analysis by analytical methods are proposed.

A system or transform maps one signal that is called input signal x(t) into another
signal which is called the output signal or response y(t) :
y ( t ) = T { x ( t )}

System block diagram: (the operator)

Dr. Ray Kwok LTI Systems 4

Properties and characteristics of continuous systems

Properties and characteristics of continuous systems:

 Linearity

Homogeneity

Additivity
 Time-invariance
 Causality
 stability

x ( t ) = input y ( t ) = output
( excitation ) ( response )

Dr. Ray Kwok LTI Systems 5

Linear system

A system is linear if and only if it satisfies the principle of homogeneity and the
principle of additvity.

The principle of homogeneity.
If an input x1(t) applied to a linear system produces the output y1(t), when a
scaled input signal by constant C is applied to the linear system, x2 ( t ) = Cx1 ( t )
the output y2 ( t ) = Cy1 ( t ) results.

x1 ( t )
LINEAR
y1 ( t )
SYSTEM

Cx1 ( t ) LINEAR
SYSTEM Cy1 ( t )

Dr. Ray Kwok LTI Systems 6

Linear system

The principle of additivity.

If an input x1(t) applied to a linear system produces the output y1(t), and x2 (t)
produces y2(t) , when a new input x1(t) + x2(t) is applied to the linear system, the
output y1(t) + y2(t) results.

x1 ( t ) y1 ( t )

x2 ( t ) y2 ( t )

x1 ( t ) + x2 ( t ) y1 ( t ) + y2 ( t )

Dr. Ray Kwok LTI Systems 7

Linear system

The principle of superposition.

If an input x1(t) applied to a linear system produces the output y1(t), and x2 (t)
produces y2(t) , when a new input Ax1(t) +B x2(t) is applied to the linear system,
the output Ay1(t) + By2(t) results.

x1 ( t ) y1 ( t )

x2 ( t ) y2 ( t )

Ax1 ( t ) + Bx2 ( t ) Ay1 ( t ) + By2 ( t )

Dr. Ray Kwok LTI Systems 8

Time invariance

If an input x1(t) applied to a system produces the output y1(t), when a time-
shifted version of input x2(t) = x1(t - to) is applied to the linear system, the
output y2(t) = y1(t - to) results for arbitrary x1(t) and to and for all t , then the
system is said to be a time invariant system.

Loosely speaking, the system parameter do not change with time.

The same input applied at different times will produce outputs that are identical
in shape and size but shifted in time

x1 ( t ) y1 ( t )

x2 ( t ) = x1 ( t t0 ) y2 ( t ) = y1 ( t t0 )

Dr. Ray Kwok LTI Systems 9

Causality

A causal system is a non-anticipative system in that the output does not precede,
or anticipate, the input.

i.e. The systems output depends only on the past and current input, but not on
the future inputs.

All nature, physics operate under this principal called causality (cause and
effect).

The impulse response of a causal system must be 0 for all t < 0. [ i.e. the
impulse input is (t) ].

Dr. Ray Kwok LTI Systems 10

Memoryless system and system with memory

A system has memory if its output signal depends on past or/and future values of
its input signal. System with a memory is also called dynamic system.

A system is memoryless if the output signal depends only on the present value of
the input signal. A memoryless system is also called instantaneous system.

The resistor is an example of a memoryless system:

v ( t ) = Ri ( t )

While the capacitor has memory:
t
v ( t ) = v ( t0 ) + i ( ) d
t0

Dr. Ray Kwok LTI Systems 11

Stability

System stability can be defined from several points of view.

The bounded input bounded output (BIBO) criterion

A system is bounded-input bounded-output (BIBO) stable if for any bounded input
defined by x
x k1

The corresponding output y is also bounded defined by y k2

where k2 and k2 are finite real numbers.

Dr. Ray Kwok LTI Systems 12

System invertibility and inverse systems

A system is said to be invertible if distinct inputs lead to distinct output. An inverse

system is a system that when is cascaded with duplicate of itself yields an output
equal to the input of cascaded system

x (t ) y (t ) x (t )

Dr. Ray Kwok LTI Systems 13

Thevenins Theorem

A linear two-terminal circuit can be replaced by an equivalent circuit consisting of

a voltage source in series with a resistor where Vth is the open-circuit voltage at
the terminals and is the input or equivalent resistance at the terminals when the
independent sources are turned off.

Nortons Theorem

A linear two-terminal circuit can be replaced by an equivalent circuit consisting of
a current source in parallel with a resistor, where Is is the short-circuit current
through the terminals and R is the input or equivalent resistance at the terminals
when the independent sources are turned off.

VTh

Dr. Ray Kwok LTI Systems 14

Operation for linear systems

Operation for linear systems

Continuous systems
 Laplace transform, (s)
 Convolution integral (time)
 Correlation integral, (time)
 Fourier series and transform
 DFT approximation
 Convolution (frequency, f )
 Correlation (frequency, f )

Discrete systems
 z- transform,
 Convolution sum (time)
 Correlation sum, (time)
 Discrete time Fourier transform (DTFT) implemented by FFT
 Discrete Fourier transform (DFT) implemented by Fast Fourier transform (FFT)
 Convolution (frequency, )
 Correlation (frequency, )

Dr. Ray Kwok LTI Systems 15

Models for continuous and discrete systems

Models for contiguous and discrete systems

Continuous systems
 Differential equation, DEs
 Transfer function, H(s)
 Frequency response, H(j)
 State differential equations
 Unit impulse response, h(t)
 Signal flowgraph or block diagram

Discrete systems
 Difference equation, DEs
 Transfer function, H(z)
 Frequency response, H ( e j )
 State difference equations
 Unit sample response, h[n]
 Signal flowgraph or block diagram

Dr. Ray Kwok LTI Systems 16

Linear differential equation

The output y(t) and input x(t) of a LTIC system are related by a linear
differential equation with constant-coefficient of the form
d n y (t ) d n1 y ( t ) dx d mx
an + an1 + L + a0 y = b0 x ( t ) + b1 ( t ) + L + bm m
dx n dx n1 dt dt

The right hand side terms are often lumped together and called forcing function
as: dx d mx
f ( t ) = b0 x ( t ) + b1 ( t ) + L + bm
dt dt m
with initial condition y ( 0 ) ,L, y .( 0 )
n 1

The complete response is of the form y t = yh 0 t + y f 0 t () () ()

()
where yh 0 t the homogenous response is the solution to the differential
equation with f ( t ) = 0 and contain n arbitrary constant.

and y f 0 ( t ) the forced response is that one particular solution to the differential
equation that contains no part of the yh 0 ( t ) .
The n arbitrary constant may be found by applying the values of y ( 0 ) ,L, y ( 0 )
n 1

to it.

Dr. Ray Kwok LTI Systems 17

Linear differential equation

The homogenous response yh 0 ( t ) also called natural response, free response,
or complementary response.

And the transient response is y f 0 ( t )

The terms forced, particular integral, final and steady state are used
interchangeably.

A more concise representation using the finite summation
k =n
k y ( t ) k =m k x ( t )

k =0
ak
t k
= bk
k =0 t k

Note: the notation is restricted to the practical situation where the number of the
derivatives of the output is greater than or equal to the number of input derivative,
that is n m .

The order of a differential equation is the order the highest derivative of the output
function that appears in the equation.

Dr. Ray Kwok LTI Systems 18

The characteristic equation

The characteristic equation (CE) of the system is found by substituting a trial
solution y ( t ) = Ce st into the homogenous differential equation. k =n k y ( t )
a
k =0
k
t k
=0

Obtained by setting to zero terms involving the input and its derivative, with the

( ) + a s ( Ce ) + L + a s ( Ce ) = 0
result: 0 st 1 st n st
a0 s Ce 1 n

Since Ce st can not be zero (this corresponds to the trivial solution y (t ) = 0

it can be factored out. The remaining terms must satisfy the algebraic
equation a s 0 + a s1 + L + a s n = 0
0 1 n

This result is known as the characteristic equation and can also be written as
k =n

a s
k =0
k
k
=0

Dr. Ray Kwok LTI Systems 19

The characteristic equation
k =n

The n roots of the equation
a s
k =0
k
k
=0

s1 = r1 , s2 = r2 L sn = rn

are called characteristic roots. The characteristic equation can be written in factored
form: k =n

a s
k =0
k
k = an ( s r1 )( s r2 )L( s rn ) = 0

where r1 , r2 ,L , rn , may be real or complex. If the DE has real coefficient, complex

roots must appear in conjugate pair. The solution of the homogenous DE
k =n
k y (t )

k =0
ak
t k
=0
for the given set of initial condition

y (t ) , y ( t ) / t , n1 y ( t ) / t n1
t =0 t =0 t =0

Dr. Ray Kwok LTI Systems 20

The characteristic equation

The solution of the homogenous DE k =n

a s
k =0
k
k
=0

For the given set of initial condition

y (t ) , y ( t ) / t , n1 y ( t ) / t n1
t =0 t =0 t =0
is called the initial condition (IC) solution and for simple (non-repeating) roots is of
the form
yIC ( t ) = C1e + C2e + L + Cne
r1t r2t rnt

n
or in compact form
yIC ( t ) = Ck e rk t

k =1

where the Ck for k = 1, 2,L n are coefficient that must be determined in order
to satisfy the given set of initial condition and rk for k = 1, 2,L n , are the
characteristic roots.

Dr. Ray Kwok LTI Systems 21

The CE with multiple roots

( s sk )
q

If the CE contain multiple roots indicated by the factor terms of the
form q 1 sk t q 2 sk t
C1t e + C2t e + L + Cq e sk t

will appear in the initial condition

Dr. Ray Kwok LTI Systems 22

Linear differential equation

Example:
Given the first order differential equation dy


+ py = q p and q = C
dx
Suggested solution
y = yh 0 + y f 0

d ( yh 0 + y f 0 )

Substituting to original differential equation
+ p ( yh 0 + y f 0 ) = q
dx
To find the homogenous solution, set forced function to zero:


dyh 0
+ pyh 0 = 0
dx

The yh 0 ( t ) is

yh 0 = Ae px

Dr. Ray Kwok LTI Systems 23

Distortionless transmission

What are the conditions for distortionless transmission?

By distortionless transmission, we mean that output signal of system is an exact
replica of the input signal except for two minor modifications.
1 A possible scaling of amplitude
2 A constant time delay

A signal x(t) is transmitted through a system without distortion if the output signal
is defined by y(t)

x (t ) y (t )

y ( t ) = k x ( t t0 )

Dr. Ray Kwok LTI Systems 24

Distortionless transmission

Here the constant k accounts for a change in amplitude and t0 constant
accounts for a delay in transmission:
Y ( j ) = k X ( j ) e j t0

The frequency response of a distortionless system is:

Y ( j )
H ( j ) = = k e j t0
X ( j )

Correspondingly, the impulse response of the system is

h(t ) = k (t to )

Dr. Ray Kwok LTI Systems 25

Condition for CT LTI distortionless transmission

Conditions for distortionless transmission of CT LTI system

with transfer function H ( j ) is

1) The magnitude response H(j ) must be constant for all frequencies of interest.
H(j ) = k

2) For the same frequencies of interest H ( j ) is linear in frequency with

slope t0 and intercept of zero.
H ( j ) =- t0
where t0 accounts for a delay in transmission through the system.

Dr. Ray Kwok LTI Systems 26

Condition for CT LTI distortionless transmission

For DT LTI system with transfer function

1 The magnitude response H(e j ) is constant
for all frequencies of interest
H(e j ) = k

2 For the same frequencies of interest, the phase

response H (e j ) is linear in frequency
H (e j ) = n0
where n0 accounts for an integer delay in transmission
through the system.

Dr. Ray Kwok LTI Systems 27

Frequency Response

Frequency response for distortionless transmission through a linear time-invariant

system
H ( j )
K

arg H ( j )

Slope = -t0

a) Magnitude response
b) Phase response

Dr. Ray Kwok LTI Systems 28