What is a Communications System?

Actual

Real Life
Message

Real Life
Message

Input Signal

Replica

Output Signal

Input
Transducer

Output
Transducer
Electronic Signal

Electronic Signal
Transmitter

Transmitted
Signal

Channel

Received
Signal

Receiver

Noise and Distortion

Modulation/Coding
Redundancy

Signal to Noise Ratio (SNR)

Capacity

Demodulation
Error Correction

Department of Electrical and Computer Engineering

Classification of Signals
Continuous-time and Discrete-time Signals

Analog and Digital Signals

Deterministic and Random Signals
Periodic and Aperiodic Signals
Energy and Power Signals

Department of Electrical and Computer Engineering

Continuous vs. Discrete Time

Analog vs. Digital
Time

Value

Continuous

Continuous

Analog

Continuous

Discrete

Digital

Discrete

Continuous

Discrete

Discrete

Department of Electrical and Computer Engineering

Some Examples

Department of Electrical and Computer Engineering

Deterministic vs. Random Signals

Department of Electrical and Computer Engineering

Effect of SNR

Department of Electrical and Computer Engineering

Periodic vs. Aperiodic Signals

Signal is Periodic if :

g (t ) g (t nT )
Period of the Periodic Signal
Frequency of the Periodic Signal

=T
= f = 1/T

If the signal is not Periodic, it will be Aperiodic

Department of Electrical and Computer Engineering

Some Examples
g (t ) 3 sin( 400t )
g (t ) 2 t 2

g (t ) sin(12t ) sin( 6t )

Department of Electrical and Computer Engineering

Size of the Signal?

Energy of Signal

Eg

g (t )

dt

Power of Signal (Time Average Energy)

T /2

1
2
Pg lim
g (t ) dt

T T
T / 2
RMS value of Signal

Pg

Department of Electrical and Computer Engineering

Example of Energy and Power Signals

Neither

g (t ) e t / 2

Energy Signal

g (t ) u (t )e t / 2

Power Signal

g (t ) C cos(0t 0 )

Power Signal

g (t ) C1 cos(1t 1 ) C2 cos(2t 2 )

Department of Electrical and Computer Engineering

Example 1 neither energy nor power signal

g (t ) e t / 2

Eg

g (t )

dt

t / 2 2

dt

e dt e
t

Department of Electrical and Computer Engineering

Example 2 energy signal

g (t ) u (t )e t / 2

Eg

g (t )

dt

t / 2 2

dt

e dt e
t

t
0

0 1 1

Department of Electrical and Computer Engineering

Example 3 power signal

g (t ) C cos(0t 0 )
T /2

1
2
g
(
t
)
dt

T T
T / 2

Pg lim

T /2

1
2
2
lim
C
cos
(0t 0 )dt
T T
T / 2
T /2

1
C2
lim
[1 cos(20t 2 0 )]dt
T T
2
T / 2
C2
lim
T 2T

T /2

C2
dt lim

T 2T
T / 2

T /2

cos(2 t 2
0

T / 2

Department of Electrical and Computer Engineering

)]dt

2 T /2

C
T 2T

lim

C
T 2T

dt lim

T / 2

2 T /2

cos(2 t 2
0

)]dt

T / 2

C2
C2
lim
(T ) lim
(0)
T 2T
T 2T

C2

2
C2
Pg
2

For a sinosoidal signal regardless

of frequency and phase shift

RMS value C / 2
Department of Electrical and Computer Engineering

Example 4 power signal

g (t ) C1 cos(1t 1 ) C2 cos(2t 2 )
T /2

1
2
Pg lim
g
(
t
)
dt

T T
T / 2
T /2

T /2

1
1
2
2
2
2
lim
C
cos
(

)
dt

lim
C
cos
(2t 2 )dt
1
1
1
2
T T
T T
T / 2
T / 2
2C1C2
T
T

lim

T /2

cos( t ) cos( t
1

)dt

T / 2

C12 C22
2C1C2

lim
2
2 T T

T /2

cos( t ) cos( t
1

T / 2

Department of Electrical and Computer Engineering

)dt

T /2

C12 C22
2C1C2

lim
2
2 T T
Hint ->

cos( t ) cos( t
1

)dt

T / 2

cos(a b) cos(a b) 2 cos a cos b

C12 C22
C1C2
Pg

lim
2
2 T T

T /2

cos[(

2 )t 1 2 ]dt

T / 2

C1C2
lim
T T
C12 C22
Pg

2
2

T /2

cos[(

2 )t 1 2 ]dt

T / 2

Power of the sum of the two sinosoid signals with

distinct frequencies is equal to the sum of the
powers of the individual signals

Department of Electrical and Computer Engineering

Generalizing the Result

g (t ) Cn cos(n t n )
n 1

Cn2
Pg
n 1 2
It is called Parsevals Theorem

Department of Electrical and Computer Engineering

Some Important Operations on Signals

Time Shifting

x(t ) g (t a)

Time Scaling

t
x(t ) g ( )
a

Time Inversion

x(t ) g (t )

Department of Electrical and Computer Engineering

Unit Impulse Function

Definition:

(t ) 0

t0

(t ) 1

Sampling Property

g (t ) (t )dt g (0)

or

g (t ) (t a)dt g (a)

Department of Electrical and Computer Engineering

Unit Step Function

Definition:

u (t )
0

t0
t0

Can you represent u(t) in terms of unit impulse function?

t

u (t ) ( x)dx

Department of Electrical and Computer Engineering

Components of a Signal
Lets approximate a signal with another signal

g (t ) cx(t )

t1 t t2

Error of approximation is:

e(t ) g (t ) cx(t )
Energy (size) of error is

t1 t t2

t2

Ee e 2 (t )dt
t1
t2

[ g (t ) cx (t )]2 dt
t1

For best approximation, energy needs to be minimized

Department of Electrical and Computer Engineering

To minimize energy, a necessary condition is

t2

d
2
[
g
(
t
)

cx
(
t
)]
dt 0

dc t1
t2

or

dEe
0
dc

t2

t2

d
d
d
2
2 2
g
(
t
)
dt

c
x
(
t
)
dt

2cg (t ) x(t )dt 0

dc t1
dc t1
dc t1
t2

t2

t1

t1

2c x 2 (t )dt 2 g (t ) x(t )dt 0

t2

g (t ) x(t )dt
t1
t2

2
x
(t )dt

t2

g (t ) x(t )dt

E x t1

cx(t) is the projection

of g(t) on x(t)

t1

Department of Electrical and Computer Engineering

Example 1
g (t )
1

-1

0 t 2

g (t ) c sin t

c?

We know

1 2
c
g (t ) x(t )dt

E x t1
2

t sin 2t
E x sin 2 tdt
2
4
0

2
2
1

1
c g (t ) sin tdt sin tdt sin tdt
0

(1 1) (1 1) 4

Department of Electrical and Computer Engineering

Example 2
g (t )
1

-1

x(t )

0 t 2

g (t ) cx (t )

c?

We know

-1

1 2
c
g (t ) x(t )dt

E x t1

E x 1dt t 0 2
0

1
c
1dt 1dt 1
2 0

g(t) has a full projection on x(t) both are same

Department of Electrical and Computer Engineering

Example 3
g (t )
1

x(t )

0 t 2

g (t ) cx (t )

c?

We know

-1

1 2
c
g (t ) x(t )dt

E x t1

E x 1dt t 0 2
0

1
c
(1)dt (1)dt 1
2 0

g(t) has a full negative projection on x(t)

or g(t) and x(t) are opposite to each other
Department of Electrical and Computer Engineering

Example 4
g (t )
1

-1

x(t )

0 t 2

g (t ) cx (t )

c?

We know

1 2
c
g (t ) x(t )dt

E x t1

E x 1dt t 0 2
0

1
c
1dt (1)dt 0
2 0

g(t) does not have any projection on x(t)

or g(t) and x(t) are orthogonal to each other
Department of Electrical and Computer Engineering

Correlation between Signals

g (t ) cx(t )

We already know, if

t1 t t2

1 2
c
g (t ) x(t )dt

E x t1
Lets define a correlation co-efficient between two functions, g(t) and x(t), as

cn

1
Eg Ex

t2

g (t ) x(t )dt

t1 t t2

t1

Or to generalize

cn

1
Eg Ex

cn 1
Opposite

g (t ) x(t )dt

1 cn 1

Orthogonal

Itself

Department of Electrical and Computer Engineering

Energy of Sum of Orthogonal Signals

z (t ) g (t ) x(t )

t1 t t 2

If g(t) and x(t) are orthogonal, then

Ez Eg Ex
By definition:
t2

E z [ g (t ) x(t )]2 dt
t1
t2

t2

t2

t1

t1

t1

g 2 (t )dt x 2 (t )dt 2 g (t ) x(t )dt

Eg Ex

Department of Electrical and Computer Engineering

Orthogonal Signal Set

Lets define a signal set x1(t), x2(t), ., xN(t), such that

x
(
t
)
x
(
t
)
dt

t m n

t2

m n

En

m n

t1 t t2

Then, x1(t), x2(t), ., xN(t) are called orthogonal signal set

and if En is = 1, then the set is called orthonormal
Now lets assume

g (t ) c1 x1 (t ) c2 x2 (t ) c3 x3 (t ) ..... cN xN (t ) cn xn (t )
n 1

t2

Then for best

approximation i.e.,
minimized error energy

g (t ) x (t )dt
n

cn

t1
t2

2
x
n (t )dt

1 2

g (t ) xn (t )dt

En t1

t1

Department of Electrical and Computer Engineering

If the orthogonal set is complete, then error energy -> 0, i.e.,

approximation changes to equality
N

g (t ) c1 x1 (t ) c2 x2 (t ) c3 x3 (t ) ..... cN x N (t ) cn xn (t )
n 1

The above equation is called generalized Fourier Series

What about energy of g(t)? Parseals Theorem
N

E g c E1 c E2 c E3 ..... c E N cn2 En
2
1

2
2

2
3

2
N

n 1

Department of Electrical and Computer Engineering

An Example of Complete Orthogonal Signal Set

{1, cos0t , cos 20t , cos 30t.....

0 2f 0

sin 0t , sin 20t , sin 30t ,.....}

Follows by:

0
T cos n0t cos m0tdt T0 / 2

0
T sin n0t sin m0tdt T0 / 2

sin n t cos m tdt 0

0

m n
mn0

m n
mn0

for all m and n

T0

Department of Electrical and Computer Engineering

2
T0

Lets assume:

g (t ) a0 a1 cos0t a2 cos 20t a3 cos 30t .....

b1 sin 0t b2 sin 20t b3 sin 30t .....

t1 t t 2 T0

a0 [an cos n0t bn sin n0t ]

n 1

Remember:

1 2
cn
g (t ) xn (t )dt

En t1

Therefore,

2
an
T0
2
bn
T0

t1 T0

g (t ) cos n tdt
0

n 1,2,3...

t1
t1 T0

g (t ) sin n tdt
0

t1

n 1,2,3...

1
a0
T0

t1 T0

g (t )dt
t1

Trigonometric
Fourier Series

Department of Electrical and Computer Engineering

2
T0

Concept of
representing a
periodic signal
with a summation
of sinusoids

Department of Electrical and Computer Engineering

Compact Trigonometric Fourier Series

g (t ) a0 [an cos n0t bn sin n0t ]

n 1

C0 Cn cos(n0t n )
n 1

Compact Trigonometric
Fourier Series

Where

Cn an2 bn2

bn

an

n tan 1

Also called Fourier Spectrum

of periodic signal

C 0 a0

Department of Electrical and Computer Engineering

Example 1
T0 1sec f 0 1Hz 0 2

g (t )
1

0.25

0.75

1.25

-1

Without any mathematical calculations,

a0

bn

1
T0

2
T0

t 0 T0

g (t )dt 0

t0

t 0 T0

g (t ) sin( n t )dt 0
0

t0

Department of Electrical and Computer Engineering

Example 1 Cont.
T0 1sec f 0 1Hz 0 2

g (t )
1

0.25

0.75

1.25

-1

2
an
T0

t 0 T0

g (t ) cos(n t )dt
0

t0

0.5

0.5

0.5

an 2 g (t ) cos(2nt )dt 2 x 2 g (t ) cos(2nt )dt

0.25

0.5

0.25

4 cos(2nt )dt 4 (1) cos(2nt )dt

Department of Electrical and Computer Engineering

0.25

0.5

0.25

4 cos(2nt )dt 4 (1) cos(2nt )dt

0.25

0.5

0.25

4 cos(2nt )dt 4 cos(2nt )dt

4 sin( 2nt )
4 sin( 2nt )

2n
2n
0
0.25
0.25

0.5

2
n
2
n
4
n
sin( ) ( sin( )) sin( )
n
2
n
2
n
2
4
n
an sin( )
n
2
Department of Electrical and Computer Engineering

Example 1 Cont.
T0 1sec f 0 1Hz 0 2

g (t )
1

0.25

0.75

1.25

-1

4
n
an sin( )
n
2
a1

a4 0

a2 0

4
a5
5

4
a3
3

a6 0

Department of Electrical and Computer Engineering

Example 1 Cont.
T0 1sec f 0 1Hz 0 2

g (t )
1

0.25

0.75

1.25

-1

g (t ) a0 [an cos n0t bn sin n0t ]

n 1

g (t )

cos(0t )

4
4
4
cos(30t )
cos(50t )
cos(70t ) ...
3
5
7

4
4
4
cos((2 )t )
cos(3(2 )t )
cos(5(2 )t )
cos(7(2 )t ) ...

3
5
7
Department of Electrical and Computer Engineering

Example 2
T0 1sec f 0 1Hz 0 2

y(t )
2

0.25

0.75

1.25

y(t ) 1 g (t )
We know that
4
4
4
4
g (t ) cos((2 )t )
cos(3(2 )t )
cos(5(2 )t )
cos(7(2 )t ) ...

3
5
7
Therefore,
4
4
4
4
y(t ) 1 cos((2 )t )
cos(3(2 )t )
cos(5(2 )t )
cos(7(2 )t ) ...

3
5
7
Department of Electrical and Computer Engineering

Example 3
T0 f 0

y(t )

1
2
0
T0
T0

0.25T0

0.75T0

1.25T0

We know that with time period = 1 sec

4
4
4
4
y(t ) 1 cos((2 )t )
cos(3(2 )t )
cos(5(2 )t )
cos(7(2 )t ) ...

3
5
7

Therefore, with time period of T0

y(t ) 1

cos(0t )

4
4
cos(30t )
cos(50t ) ...
3
5

Department of Electrical and Computer Engineering

Example

1
2
T0 f 0 0
T0
T0

g (t ) T0 (t )

2T0

T0

T0

2T0

g (t ) C0 Cn cos(n0t n )
n 1

1
a0
T0

t 0 T0

1
t (t )dt T0
0

Remember definition of
impulse function

(t ) 1

Department of Electrical and Computer Engineering

Example cont.

1
2
T0 f 0 0
T0
T0

g (t ) T0 (t )

2T0

2
an
T0

2
bn
T0

T0

T0

2T0

t 0 T0

2
t (t ) cos(n0t )dt T0
0

Remember extended sampling

properties of impulse function

t 0 T0

(t ) sin( n t )dt 0
0

g (t ) (t ) g (0)

t0

Department of Electrical and Computer Engineering

Example cont.

1
2
T0 f 0 0
T0
T0

g (t ) T0 (t )

2T0

1
a0
T0
2
an
T0
bn 0

T0

T0

2T0

C 0 a0

1
T0

2
C n a b an
T0
2
n

2
n

bn
0
n tan
an
1

Department of Electrical and Computer Engineering

Example cont.

1
2
T0 f 0 0
T0
T0

g (t ) T0 (t )

2T0

T0

T0

2T0

g (t ) C0 Cn cos(n0t n )
n 1

1 2
cos(n0t )
T0 n1 T0

Department of Electrical and Computer Engineering

Example cont.
Time Domain Signal
T0 f 0

g (t ) T0 (t )

2T0

T0

T0

2T0

2 0

1
2
0
T0
T0

Fourier Spectrum

Cn
2
T0

3 0

4 0

Department of Electrical and Computer Engineering

Energy of Signal Revisited

Energy of Signal is defined

Eg

g (t )

dt

and not

Eg

2
g
(t )dt

To accommodate complex signals

Because

g (t ) g (t ) g (t ) g 2 (t )

for real signals

g (t ) g (t ) g * (t )

for complex signals

Similarly,

g (t ) x(t )

becomes g (t ) x* (t )

complex signals

Department of Electrical and Computer Engineering

Another Example of orthogonal Signal Set

(again periodic)

e
T0

jm0t

jn 0t

(n 0,1,2,...)

e dt e
jn 0t *

j ( m n )0 t

0 2f 0

dt

T0

cos(m n)0t j sin( m n)0t dt

T0

cos(m n)0tdt j sin( m n)0tdt

T0

T
0

T0

m n
mn 0

Department of Electrical and Computer Engineering

2
T0

Exponential Fourier Series

g (t )

jn 0 t
D
e
n

By orthogonality

1
jn 0t *
Dn g (t ) e
dt
T0 T0
or

1
Dn g (t )e jn0t dt
T0 T0

Department of Electrical and Computer Engineering

Compact to Exponential Fourier Series

g (t ) C0 Cn cos(n0t n )
n 1

Cn j ( n0t n ) j ( n0t n )
C0
e
e
n 1 2

Cn j n jn0t Cn j n jn0t
C0 e e
e e

2
2

n 1

D0 Dn e jn0t D n e jn0t

n 1

D0

D e

n
n0

jn 0t

jn 0 t
D
e
n

Department of Electrical and Computer Engineering

Lets look back on the Example of Impulse Train

1
2
T0 f 0 0
T0
T0

g (t ) T0 (t )

2T0

T0

T0

2T0

g (t ) C0 Cn cos(n0t n )
n 1

1 2
cos(n0t )
T0 n1 T0

1
D0 C0
T0
C n j n 1
Dn
e
2
T0

D n

C n j n 1
e

2
T0

Department of Electrical and Computer Engineering

Time Domain Signal

2T0

g (t ) T0 (t )

T0

Fourier Spectrum

1
2
0
T0
T0

T0

2T0

2 0

3 0

4 0

2 0

3 0

4 0

Cn

2
T0

Exponential Fourier
Spectrum

T0 f 0

Dn
1
T0

4 0 30 2 0

Department of Electrical and Computer Engineering