chapter 2

64
2/9/2014 Communication Systems 1 Communication Systems Instructor: Engr. Dr. Sarmad Ullah Khan Assistant Professor Assistant Professor Electrical Engineering Department CECOS University of IT and Emerging Sciences [email protected] Chapter 2 Dr. Sarmad Ullah Khan Signals and Signal Space 2

Upload: fawad-masood-khattak

Post on 18-Nov-2014

125 views

Category:

Engineering


1 download

DESCRIPTION

By Fawad Masood Khattak Visit For More: https://www.facebook.com/FawadMasoodkhankhattak?fref=ts OR https://www.facebook.com/fawadmasood.kttk

TRANSCRIPT

Page 1: Chapter 2

2/9/2014

Communication Systems 1

Communication Systems

Instructor: Engr. Dr. Sarmad Ullah Khan

Assistant ProfessorAssistant ProfessorElectrical Engineering Department

CECOS University of IT and Emerging [email protected]

Chapter 2

Dr. Sarmad Ullah Khan

Signals and Signal Space

2

Page 2: Chapter 2

2/9/2014

Communication Systems 2

Outlines

• Signals and systems• Size of signal

Dr. Sarmad Ullah Khan

Size of signal• Classification of signals• Signal operations• The unit impulse function• Signals versus Vectors• Correlation

O th l i l• Orthogonal signals• Trigonometric Fourier Series• Exponential Fourier Series

3

Outlines

• Signals and systems• Size of signal

Dr. Sarmad Ullah Khan

Size of signal• Classification of signals• Signal operations• The unit impulse function• Signals versus Vectors• Correlation

O th l i l• Orthogonal signals• Trigonometric Fourier Series• Exponential Fourier Series

4

Page 3: Chapter 2

2/9/2014

Communication Systems 3

Signal and system

• SignalSignal is a set of Information / Data

Dr. Sarmad Ullah Khan

Signal is a set of Information / Data

For example, Telephone, Telegraph, Stock Market

Price

Signals are function of independent variable time

In electric charge distribution over surface signalIn electric charge distribution over surface, signal

is function of space rather than time

5

Signal and system

• SystemSystem process the received signal

Dr. Sarmad Ullah Khan

System process the received signal

System might modify or extract information from

signal

For example, Anti aircraft missile launcher may

want to know the future location of targetwant to know the future location of target

Anti aircraft missile launcher gets information

from radar (INPUT)

Radar provides target past location and velocity 6

Page 4: Chapter 2

2/9/2014

Communication Systems 4

Signal and system

• SystemAnti aircraft missile launcher calculates the future

Dr. Sarmad Ullah Khan

Anti aircraft missile launcher calculates the future

location (OUTPUT) using the received

information

• Definition

System gets a set of signals as INPUT and yield aSystem gets a set of signals as INPUT and yield a

set of signals as OUTPUT after proper processing

System might be a physical device or it might be

an algorithm 7

Signal and system

Dr. Sarmad Ullah Khan

8

Page 5: Chapter 2

2/9/2014

Communication Systems 5

Outlines

• Signals and systems• Size of signal

Dr. Sarmad Ullah Khan

Size of signal• Classification of signals• Signal operations• The unit impulse function• Signals versus Vectors• Correlation

O th l i l• Orthogonal signals• Trigonometric Fourier Series• Exponential Fourier Series

9

Size of Signal • Size of an entity is a quantity that shows

largeness or strength of entity

Dr. Sarmad Ullah Khan

g g y

• It is a single value/number measure

• How signal (amplitude and duration) can berepresented by a single number measure?

• For example, person’s width ‘r’ and height ‘h’

• To be more precise, single value measure of aperson is its volume

10

Page 6: Chapter 2

2/9/2014

Communication Systems 6

Size of Signal • Signal Energy Area under a signal g(t) is its SIZE

Dr. Sarmad Ullah Khan

g g( )

Signal Size takes two values “Amplitude” and“Duration”

This measuring approach is defective for largesignals having positive and negative portions

11

Size of Signal • Signal Energy Positive portion is cancelled by negative portion

Dr. Sarmad Ullah Khan

p y g presulting in small signal

This can be solve by calculating area under g2(t)

12

Page 7: Chapter 2

2/9/2014

Communication Systems 7

Size of Signal • Signal Energy If signal is complex signal, then

Dr. Sarmad Ullah Khan

g p g ,

h ibl h i d | ( )| Other possible approach is area under |g(t)|

Energy measure is more desirable

13

Size of Signal • Signal Power Signal energy must be finite for it to be

Dr. Sarmad Ullah Khan

g gymeaningful

Necessary conditions to make it finite Amplitude goes to zero as time approaches infinity

Signal must converge

IfIf Amplitude does not go to zero

Then Energy is infinite

14

Page 8: Chapter 2

2/9/2014

Communication Systems 8

Size of Signal

Dr. Sarmad Ullah Khan

Does this mean that a 60 hertz sine wave feeding intoyour headphones is as strong as the 60 hertz sine wavecoming out of your outlet? Obviously not. This iswhat leads us to the idea of signal power. 15

Size of Signal • Signal Power To make energy finite and meaningful, time

Dr. Sarmad Ullah Khan

gy g ,average of signal energy is taken into consideration

For comple signals For complex signals

Signal power is time average of signal amplitude 16

Page 9: Chapter 2

2/9/2014

Communication Systems 9

Size of Signal

Dr. Sarmad Ullah Khan

• Square root of signal power is the Root MeanSquare (RMS) value of signal 17

Size of Signal • Are all energy signals also power signals?• No. In fact, any signal with finite energy will have

Dr. Sarmad Ullah Khan

zero power.

• Are all power signals also energy signals?• No, any signal with non-zero power will have infinite

energy.

• Are all signals either energy or power signals?• No. Any infinite-duration, increasing-magnitude

function will not be either. (e.g. f(t)=t is neither)

18

Page 10: Chapter 2

2/9/2014

Communication Systems 10

Size of Signal

• Remark:

Dr. Sarmad Ullah Khan

• The terms energy and power are not used intheir conventional sense as electrical energyor power, but only as a measure for the signalsize.

19

Example 2.1• Determine the suitable measures of the signals given

below:

Dr. Sarmad Ullah Khan

• The signal (a) amp 0 as t infinity Therefore• The signal (a) amp. 0 as t infinity .Therefore, the suitable measure for this signal is its energy, given by

gE 8444)2()(= 0

0

1

22 =+=+= ∞

∞−

dtedtdttg t

20

Page 11: Chapter 2

2/9/2014

Communication Systems 11

Example 2.1The signal in the fig. Below does not --- 0 as t . However it is periodic, therefore its power exits. ∞

Dr. Sarmad Ullah Khan

21

Example 2.2

Dr. Sarmad Ullah Khan

(a)

22

Page 12: Chapter 2

2/9/2014

Communication Systems 12

Example 2.2

Dr. Sarmad Ullah Khan

Remarks:A sinusoid of amplitude C has power of regardless of its frequency and phase .23

Example 2.2

Dr. Sarmad Ullah Khan

24

Page 13: Chapter 2

2/9/2014

Communication Systems 13

Example 2.2

Dr. Sarmad Ullah Khan

We can extent this result to a sum of any number of sinusoids with distinct frequencies.

25

Example 2.2

Dr. Sarmad Ullah Khan

Recall that ThereforeThereforeThe rms value is

26

Page 14: Chapter 2

2/9/2014

Communication Systems 14

Outlines

• Signals and systems• Size of signal

Dr. Sarmad Ullah Khan

Size of signal• Classification of signals• Signal operations• The unit impulse function• Signals versus Vectors• Correlation

O th l i l• Orthogonal signals• Trigonometric Fourier Series• Exponential Fourier Series

27

Classification of signals

• Continuous-time and discrete-time signalsA l d di it l i l

Dr. Sarmad Ullah Khan

• Analog and digital signals• Periodic and Aperiodic signals• Energy and power signals• Deterministic and random signals• Causal vs Non-causal signals• Causal vs. Non-causal signals• Right Sided and Left Sided Signals• Even and Odd Signals

28

Page 15: Chapter 2

2/9/2014

Communication Systems 15

Classification of signals

• Continuous-time Signal

Dr. Sarmad Ullah Khan

A signal that is specified for every value of time t e.g. Audio and video signals

29

Classification of signals

• Discrete-time signal A i l h i ifi d f di l f

Dr. Sarmad Ullah Khan

A signal that is specified for discrete value of time t = nT, e.g. Stock Market daily average

30

Page 16: Chapter 2

2/9/2014

Communication Systems 16

Classification of signals

• Analog signal A l i l d i i i l

Dr. Sarmad Ullah Khan

Analog signal and continuous-time signal are two different signals

Analog signal whose amplitude can have any value over a continuous range

Analog continuous time signal x(t) Analog discrete time signal x[n]31

Classification of signals

• Digital signal Di i l i l d di i i l

Dr. Sarmad Ullah Khan

Digital signal and discrete-time signal are two different signals

Discrete signal whose amplitude can have only a finite number of value

Digital continuous time signal Digital discrete time signal 32

Page 17: Chapter 2

2/9/2014

Communication Systems 17

Classification of signals

• Periodic signal A i l ( ) i id b i di if h i

Dr. Sarmad Ullah Khan

A signal g(t) is said to be periodic if there exist a positive constant T0, such that

g(t) = g(t+T0) for all t

Oth i i di i l Otherwise aperiodic signal

33

Classification of signals

• Properties of Periodic Signal P i di i l i

Dr. Sarmad Ullah Khan

Periodic signal must start at time t = - Periodic signal shifted by integral multiple of T0

remains unchanged A periodic signal g(t) can be generated by

periodic extension of any segment of g(t) withduration T0duration T0

34

Page 18: Chapter 2

2/9/2014

Communication Systems 18

Classification of signals

• Energy Signal A i l i h fi i i ll d i l

Dr. Sarmad Ullah Khan

A signal with finite energy is called energy signal

• Power Signal A signal with finite power is called power signal

35

Classification of signals

• Remarks:

Dr. Sarmad Ullah Khan

A signal with finite energy has zero power.

A signal can be either energy signal or power signal, notboth.

Every signal in daily life is energy signal, NOT powersignal

Power signal in practice is not possible because of infiniteduration and infinite energy

36

Page 19: Chapter 2

2/9/2014

Communication Systems 19

Classification of signals

• Deterministic Signal:

Dr. Sarmad Ullah Khan

A signal whose physical description is know completely,either mathematically or graphically is called deterministicsignal

• Random Signal:

A signal which is known in terms of probabilisticdescription such as mean value, mean squared value anddistribution

37

Classification of signals

• Casual Signal:

Dr. Sarmad Ullah Khan

A signal which is zero prior to zero time

Signal amplitude A=0 for T = -t

• Non Casual Signal:

A signal which is zero after zero time

Signal amplitude A=0 for T = +t

38

Page 20: Chapter 2

2/9/2014

Communication Systems 20

Classification of signals

• Right sided and Left sided Signal:A i ht id d i l i f t < T d l ft id d i l

Dr. Sarmad Ullah Khan

A right-sided signal is zero for t < T and a left-sided signalis zero for t > T where T can be positive or negative.

39

Classification of signals

• Even and Odd Signal:E i l ( ) d dd i l ( ) d fi d

Dr. Sarmad Ullah Khan

Even signals xe(t) and odd signals xo(t) are defined as

xe(t) = xe(−t) and xo(−t) = −xo(t).

40

Page 21: Chapter 2

2/9/2014

Communication Systems 21

Classification of signals

• Even and Odd Signal:If h i l i i i d f i If h

Dr. Sarmad Ullah Khan

If the signal is even, it is composed of cosine waves. If thesignal is odd, it is composed out of sine waves. If thesignal is neither even nor odd, it is composed of both sineand cosine waves.

41

Outlines

• Signals and systems• Size of signal

Dr. Sarmad Ullah Khan

Size of signal• Classification of signals• Signal operations• The unit impulse function• Signals versus Vectors• Correlation

O th l i l• Orthogonal signals• Trigonometric Fourier Series• Exponential Fourier Series

42

Page 22: Chapter 2

2/9/2014

Communication Systems 22

Signal Operation

Dr. Sarmad Ullah Khan

Time Shifting

A signal g(t) is said to be time shifted if g(t) isdelayed or advanced by time T

If signal g(t) is delayed by time T, then

43

If signal g(t) is advanced by time T, then

Signal Operation

Time Scaling

Dr. Sarmad Ullah Khan

The compression or expansion of signal g(t) intime is known as Time Scaling

Compression

g(t)=g(at)

Expansion

g(t)=g(t/a)

44

Page 23: Chapter 2

2/9/2014

Communication Systems 23

Signal Operation

Time Inversion

Dr. Sarmad Ullah Khan

In time inversion, signal g(t) is multiplied by afactor a = -1 in time domain

g(t) = g(at) if a=1

Time inverted signal

g(t) = g(at) if a=-1

45

Signal Operation

• Example: 2.4

F th i l (t) h i th fi B l k t h

Dr. Sarmad Ullah Khan

• For the signal g(t), shown in the fig. Below , sketch g(-t)

46

Page 24: Chapter 2

2/9/2014

Communication Systems 24

Outlines

• Signals and systems• Size of signal

Dr. Sarmad Ullah Khan

Size of signal• Classification of signals• Signal operations• The unit impulse function• Signals versus Vectors• Correlation

O th l i l• Orthogonal signals• Trigonometric Fourier Series• Exponential Fourier Series

47

Unit Impulse Signal

• The Dirac delta function or unit impulse or oftenreferred to as the delta function is the function that

Dr. Sarmad Ullah Khan

referred to as the delta function, is the function thatdefines the idea of a unit impulse in continuous-time

• It is infinitesimally narrow, infinitely tall, yetintegrates to one

• simplest way to visualize this as a rectangular pulsep y g pfrom a -D/2 to a +D/2 with a height of 1/D

• The impulse function is often written as

48

Page 25: Chapter 2

2/9/2014

Communication Systems 25

Unit Impulse Signal

Dr. Sarmad Ullah Khan

=0 for al t ≠ 0

49

Unit Impulse Signal

• Multiplication of a Function by an Impulse

Dr. Sarmad Ullah Khan

If a function g(t) is multiplied by impulse function we getimpulse value of g(t)

50

Page 26: Chapter 2

2/9/2014

Communication Systems 26

Unit Impulse Signal

Dr. Sarmad Ullah Khan

51

Unit Impulse Signal

Dr. Sarmad Ullah Khan

52

Page 27: Chapter 2

2/9/2014

Communication Systems 27

Unit Impulse Signal

• Unit Step Function u(t)

Dr. Sarmad Ullah Khan

53

Outlines

• Signals and systems• Size of signal

Dr. Sarmad Ullah Khan

Size of signal• Classification of signals• Signal operations• The unit impulse function• Signals versus Vectors• Correlation

O th l i l• Orthogonal signals• Trigonometric Fourier Series• Exponential Fourier Series

54

Page 28: Chapter 2

2/9/2014

Communication Systems 28

Signals versus Vectors

• A Vector can be represented as a sum of itscomponents

Dr. Sarmad Ullah Khan

components

• A Signal can also be represented as a sum of its• A Signal can also be represented as a sum of itscomponents

55

Signals versus Vectors

• A Signal defined over a finite number of time instantscan be written as a Vector

Dr. Sarmad Ullah Khan

can be written as a Vector• Consider a signal g(t) defined over interval [a,b]• Uniformly divide interval [a,b] in N points

T1 = a, T2 = a+ϵ, T3 = a+2ϵ, …. TN = a+(N-1)ϵ

• WhereStep Sizeϵ =

56

Page 29: Chapter 2

2/9/2014

Communication Systems 29

Signals versus Vectors

• Signal vector g can be written a N-dimensional vector

Dr. Sarmad Ullah Khan

• Signal vector g grows as N increases

g = [g(t1) g(t2) … g(tN)]

• Signal transforms into continuous time signal g(t)

57

Signals versus Vectors

• Signal transforms into continuous time signal g(t)

Dr. Sarmad Ullah Khan

• Continuous time signals are straightforwardgeneralization of finite dimension vectors

• Vector properties can be applied to signals

58

Page 30: Chapter 2

2/9/2014

Communication Systems 30

Signals versus Vectors

• A vector is represented by bold-face type

Dr. Sarmad Ullah Khan

• Specified by its magnitude and its direction

• For example, Vector x have magnitude | x | andVector g have magnitude | g |

• Inner product (dot or scalar) of two real valuedvectors ‘g’ and ‘x’ is

59

Signals versus Vectors

• Component of a Vector

Dr. Sarmad Ullah Khan

Consider two vectors ‘x’ and ‘g’

‘cx’ (projection) is component of ‘g’ along ‘x’

What is the mathematical significance of a vector alonganother vector?

g = cx + eg

However, this is not a unique way of vector decomposition

60

Page 31: Chapter 2

2/9/2014

Communication Systems 31

Signals versus Vectors

• Component of a Vector

Oth t ‘ ’ i

Dr. Sarmad Ullah Khan

Other ways to express ‘g’ is

g is represented in terms of x plus another vectorwhich is called the error vector e

61

Signals versus Vectors

• Component of a Vector

Dr. Sarmad Ullah Khan

If we approximate

e = g – cx

Geometrically component of g along x is

Hence

62

Page 32: Chapter 2

2/9/2014

Communication Systems 32

Signals versus Vectors

• Component of a Vector

Dr. Sarmad Ullah Khan

Based on definition of inner product, multiply both side by|x|

63

Signals versus Vectors

• Component of a Vector

Dr. Sarmad Ullah Khan

If g and x are orthogonal, then

64

Page 33: Chapter 2

2/9/2014

Communication Systems 33

Signals versus Vectors

• Component of Signal

Dr. Sarmad Ullah Khan

Vector component and orthogonality can be extended tocontinuous time signals

Consider approximating a real signal g(t) in terms ofanother real signal x(t)

And

65

Signals versus Vectors

• Component of Signal

Dr. Sarmad Ullah Khan

As energy is one possible measure of signal size.

To minimize the effect of error signal we need to minimize its size-----which is its energy over the interval [t1 , t2]

66

Page 34: Chapter 2

2/9/2014

Communication Systems 34

Signals versus Vectors

• Component of Signal

Dr. Sarmad Ullah Khan

But ‘e’ is a function of ‘c’, not ‘t’, hence

67

Signals versus Vectors

• Component of Signal

Dr. Sarmad Ullah Khan

68

Page 35: Chapter 2

2/9/2014

Communication Systems 35

Signals versus Vectors

• Component of Signal

Dr. Sarmad Ullah Khan

69

Signals versus Vectors

• Component of Signal

Dr. Sarmad Ullah Khan

Two signals g(t) and x(t) are said to be orthogonal if thereis zero contribution from one signal to other signal

For N dimensional vectors ‘g’ and ‘x’

70

Page 36: Chapter 2

2/9/2014

Communication Systems 36

For the square signal g(t), find the component of g(t) of the form sint or in other words approximate g(t) in terms of sint

Example 2.5

Dr. Sarmad Ullah Khan

sint or in other words approximate g(t) in terms of sint

tctg sin)( ≅ π20 ≥≤ t

71

ttx sin)( = and

Example 2.5

Dr. Sarmad Ullah Khan

)(

From equation for signals

dttxtg

Ec

t

tx=2

1

)()(1

72

πππ

π π

π

π 4sinsin

1sin)(

1

0

22

=

−+== tdttdttdttgc

o

ttg sin4

)(π

Page 37: Chapter 2

2/9/2014

Communication Systems 37

Signals versus Vectors

• Orthogonality in Complex Signals

Dr. Sarmad Ullah Khan

For complex function g(t), its approximation by anothercomplex function x(t) over finite interval

‘c’ and ‘e’ are complex functions

73

Signals versus Vectors

• Orthogonality in Complex Signals

Dr. Sarmad Ullah Khan

Energy of a complex signal x(t) over finite interval

Choose ‘c’ such that it reduces ‘Ee’

74

Page 38: Chapter 2

2/9/2014

Communication Systems 38

Signals versus Vectors

• Orthogonality in Complex Signals

Dr. Sarmad Ullah Khan

We know that

After certain manipulations

( )( ) ∗∗∗∗ ++=++=+ uvvuvuvuvuvu222

222 2222

1

2

1

2

1

)()(1

)()(1

)( ∗∗ −+−=t

tx

x

t

tx

t

t

e dttxtgE

EcdttxtgE

dttgE

dttxtgE

ct

tx

)()(1 2

1

∗= 75

Signals versus Vectors

• Orthogonality in Complex Signals

Dr. Sarmad Ullah Khan

So, two complex functions are orthogonal over an interval,if

0)()( 21

2

1

=∗ dttxtxt

t

0)()( 21

2

1

= ∗ dttxtxt

t

or

76

Page 39: Chapter 2

2/9/2014

Communication Systems 39

Signals versus Vectors

• Energy of the Sum of Orthogonal Signals

Dr. Sarmad Ullah Khan

Sum of the two orthogonal vectors is equal to the sum ofthe lengths of the squared of two vectors. z = x+y then

Sum of the energy of two orthogonal signals is equal to the

222yxz +=

gy g g qsum of the energy of the two signals. If x(t) and y(t) areorthogonal signals over the interval, and if

z(t) = x(t)+ y(t) then

yxz EEE += 77

Outlines

• Signals and systems• Size of signal

Dr. Sarmad Ullah Khan

Size of signal• Classification of signals• Signal operations• The unit impulse function• Signals versus Vectors• Correlation

O th l i l• Orthogonal signals• Trigonometric Fourier Series• Exponential Fourier Series

78

Page 40: Chapter 2

2/9/2014

Communication Systems 40

Correlation of Signals

• Correlation addresses the question: “to what degree issignal A similar to signal B”

Dr. Sarmad Ullah Khan

signal A similar to signal B

• Two vectors ‘g’ and ‘x’ are similar if ‘g’ has a largecomponent along ‘x’

• If ‘c’ has a large value, then the two vectors will besimilar

‘ ’ ld b id d h i i f• ‘c’ could be considered the quantitative measure ofsimilarity between ‘g’ and ‘x’But such a measure could be defective. The amount of similarity should be independent of the lengths of g and x 79

Correlation of Signals

• Doubling g should not change the similarity between gand x

Dr. Sarmad Ullah Khan

and x

• Similarity between the vectors is indicated by angle

However:

Doubling g doubles the value of c

Doubling x halves the value of cc is faulty measure for similarity

y y gbetween the vectors.

• The smaller the angle , the largest is the similarity, andvice versa

• Thus, a suitable measure would be , given by80

Page 41: Chapter 2

2/9/2014

Communication Systems 41

Correlation of Signals

• Where

Dr. Sarmad Ullah Khan

• This similarity measure is known as correlationco-efficient

Independent of the lengths of g and x

• And

81

Correlation of Signals

• Same arguments for defining a similarity index(correlation co efficient) for signals

Dr. Sarmad Ullah Khan

(correlation co-efficient) for signals

• Consider signals over the entire time interval

• To establish a similarity index independent ofenergies (sizes) of g(t) and x(t), normalize c bynormalizing the two signals to have unit energiesg g g

82

Page 42: Chapter 2

2/9/2014

Communication Systems 42

Correlation of Signals

• Best Friends

Dr. Sarmad Ullah Khan

• Opposite personalities (Enemies)

• Complete Strangers

83

Example 2.6

Find the correlation co-efficient between the pulse x(t) and the pulses

nc6,5,4,3,2,1,)( == itgi

Dr. Sarmad Ullah Khan

p i

84

5)(5

0

5

0

2 === dtdttxEx5

1=gE

155

1 5

0

= dtcndttxtgEE

cxg

n ∞

∞−

= )()(1

Similarly

Maximum possible similarity

Page 43: Chapter 2

2/9/2014

Communication Systems 43

Example 2.6 (cont…)Dr. Sarmad Ullah Khan

5)(5

0

5

0

2 === dtdttxEx25.1

2 =gE

85

1)5.0(525.1

1 5

0

= dtcndttxtg

EEc

xg

n ∞

∞−

= )()(1

Maximum possible similarity……independent of amplitude

Example 2.6 (cont…)Dr. Sarmad Ullah Khan

5)(55

2 === dtdttxEx 51

=gESimilarly

86

00 1g

1)1)(1(55

1 5

0

−=−×

= dtcndttxtgEE

cxg

n ∞

∞−

= )()(1

y

Page 44: Chapter 2

2/9/2014

Communication Systems 44

Example 2.6(cont…)Dr. Sarmad Ullah Khan

)1(2

1)( 22

2

aTT

atT

at ea

dtedteE −−− −===

5)(5

0

5

0

2 === dtdttxEx

87

200 a

5

1=a 5=T

1617.24 =gE961.0

1617.25

1 5

0

5 =×

= −

dtect

n

Here

Reaching Maximum similarity

Outlines

• Signals and systems• Size of signal

Dr. Sarmad Ullah Khan

Size of signal• Classification of signals• Signal operations• The unit impulse function• Signals versus Vectors• Correlation

O th l i l• Orthogonal signals• Trigonometric Fourier Series• Exponential Fourier Series

88

Page 45: Chapter 2

2/9/2014

Communication Systems 45

Orthogonal Signal Sets

• A signal can be represented as a sum of orthogonalset of signals

Dr. Sarmad Ullah Khan

set of signals

• Orthogonal set of signals form a basis for specificsignal space

• For example, a vector is represented as a sum oforthogonal set of vectors

I f di f• It forms a coordinate system for vector space

89

Orthogonal Signal Sets

• Orthogonal Vector SpaceC t i i d ib d b th t ll th l

Dr. Sarmad Ullah Khan

Cartesian space is described by three mutually orthogonalvectors x1, x2, and x3

If a three dimensional vector g is approximated by twoorthogonal vectors x1 and x2, then

And

90

Page 46: Chapter 2

2/9/2014

Communication Systems 46

Orthogonal Signal Sets

• Orthogonal Vector SpaceIf th di i l t i t d b th

Dr. Sarmad Ullah Khan

If a three dimensional vector g is represented by threeorthogonal vectors x1 , x2 and x3, then

And e = 0 in this case

x1,x2 and x3 is complete set of orthogonal space, No x4 exist

91

Orthogonal Signal Sets

• Orthogonal Vector SpaceTh th l t ll d b i t

Dr. Sarmad Ullah Khan

These orthogonal vectors are called basis vectors

Complete set of vectors is called complete orthogonal basisof a vector

A set of vector {xi} is mutually{ i} y

orthogonal if

92

Page 47: Chapter 2

2/9/2014

Communication Systems 47

Orthogonal Signal Sets

• Orthogonal Signal SpaceLik t th lit f i l t (t) (t)

Dr. Sarmad Ullah Khan

Like vector, orthogonality of signal set x1(t), x2(t), …..xN(t) over time interval [t1, t2] is defined as

If all signal energies are equal En = 1 then set is normalizedand is called an orthogonal set

An orthogonal set can be normalized by dividing xN(t) by

93

Orthogonal Signal Sets

• Orthogonal Signal SpaceN i l (t) [t t ] b t f N th l

Dr. Sarmad Ullah Khan

Now signal g(t) over [t1, t2] by a set of N-orthogonalsignals x1(t), x2(t), ….. xN(t) is

Energy of error signal e(t) can be minimized if

94

Page 48: Chapter 2

2/9/2014

Communication Systems 48

Outlines

• Signals and systems• Size of signal

Dr. Sarmad Ullah Khan

Size of signal• Classification of signals• Signal operations• The unit impulse function• Signals versus Vectors• Correlation

O th l i l• Orthogonal signals• Trigonometric Fourier Series• Exponential Fourier Series

95

Trigonometric Fourier Series

• Like vector, signal can be represented as a sum of itsorthogonal signal (Basis signals)

Dr. Sarmad Ullah Khan

g g ( g )• There are number of such basis signals e.g. trigonometric

function, exponential function, Walsh function, Besselfunction, Legendre polynomial, Laguerre functions,Jaccobi polynomial

• Consider a periodic signal of period T0

• Consider a signal set• Consider a signal set

{1+Cos w0t+Cos 2w0t+……Cos nw0t…. Sin w0t+Sin 2w0t……Sin nw0t….}

96

Page 49: Chapter 2

2/9/2014

Communication Systems 49

Trigonometric Fourier Series

• nw0 is called the nth harmonic of sinusoid of angularfrequency w where n is an integer

Dr. Sarmad Ullah Khan

frequency w0 where n is an integer

• A sinusoid of frequency w0 is called the fundamentaltone/anchor of the set

97

Trigonometric Fourier Series

• This set is orthogonal over any interval of duration because:

oo wT π2=

Dr. Sarmad Ullah Khan

because:

=2

0coscos

oTo

oo Ttdtmwtnwomn

mn

≠=≠

mn≠

=0

sinsin oo Ttdtmwtnwomn ≠=

2

oTo

oo T

0cossin = tdtmwtnwTo

oofor all n and m

and

98

Page 50: Chapter 2

2/9/2014

Communication Systems 50

Trigonometric Fourier Series

• The trigonometric set is a complete set.

Dr. Sarmad Ullah Khan

• Each signal g(t) can be described by a trigonometric Fourier series over the interval To :

...2sinsin 21 +++ twbtwb oo

...2coscos)( 21 +++= twatwaatg ooooTttt +≤≤ 11

=

++=1

sincos)(n

onono tnwbtnwaatg oTttt +≤≤ 11

or

on T

wπ2=

99

Trigonometric Fourier Series

• We determine the Fourier co-efficient as:

Dr. Sarmad Ullah Khan

+

+

=o

o

Tt

t o

o

Tt

tn

tdtnw

tdtnwtgC

1

1

1

1

2cos

cos)(

doTt

+1

)(1

,......3,2,1=n 100

dttgT

ato=1

)(1

0

tdtnwtgT

a o

Tt

ton

o

cos)(2 1

1

+

=

tdtnwtgT

b o

Tt

ton

o

sin)(2 1

1

+

=

Page 51: Chapter 2

2/9/2014

Communication Systems 51

Compact Trigonometric Fourier Series

• Consider trigonometric Fourier series

Dr. Sarmad Ullah Khan

• It contains sine and cosine terms of the samefrequency. We can represents the above equation in ai l t f th f i th

...2sinsin 21 +++ twbtwb oo

...2coscos)( 21 +++= twatwaatg ooo oTttt +≤≤ 11

single term of the same frequency using thetrigonometry identity

)cos(sincos nononon tnwCtnwbtnwa θ+=+

22nnn baC +=

−= −

n

nn a

b1tanθ

oo aC =

oTttt +≤≤ 11 101∞

=

++=1

0 )cos()(n

non tnwCCtg θ

Example 2.7

Find the compact trigonometric Fourier series for the following function

Dr. Sarmad Ullah Khan

102

Page 52: Chapter 2

2/9/2014

Communication Systems 52

Example 2.7Solution:We are required to represent g(t) by the trigonometric Fourier series over the interval andπ≤≤ t0 π=T

Dr. Sarmad Ullah Khan

series over the interval and π≤≤ t0 πoT

sec22 radT

wo

o == π

T i i f f F i i

103

Trigonometric form of Fourier series:

??,?,0 nn baa

ntbntaatg nn

no 2sin2cos)(1

++= ∞

=

π≤≤ t0

Example 2.7

5001 2 ==

−dtea

Dr. Sarmad Ullah Khan

50.00

0 == dteaπ

+==

20

2

161

2504.02cos

2

ndtntea

t

n

π

π

+==

20

2

161

8504.02sin

2

n

nntdteb

t

n

π

π

22nnn

oo

baC

aC

+=

=

104

0

Compact Fourier series is given by

)cos()(1

0 non

n tnwCCtg θ++= ∞

=π≤≤ t0

Page 53: Chapter 2

2/9/2014

Communication Systems 53

Example 2.7

2644

504.0

2n

aC oo ==

Dr. Sarmad Ullah Khan

( ) )161

2(504.0

)161(

64

161

4504.0

22222

22

nn

n

nbaC nnn

+=

++

+=+=

( ) nna

b

n

nn 4tan4tantan 11 −=−=

−= −−θ

( )4t22

50405040)( 1−∞

tt π≤≤ t0

105

( )

.......)42.868cos(063.0)24.856cos(084.0

)87.824cos(25.1)96.752cos(244.0504.0

4tan2cos161

504.0504.0)( 1

12

+−+−+−+−+=

−+

+==

oo

oo

n

tt

tt

nntn

tg π≤≤ t0

π≤≤ t0

Example 2.7

n 0 1 2 3 4 5 6 7

Dr. Sarmad Ullah Khan

Cn 0.504 0.244 0.125 0.084 0.063 0.054 0.042 0.063

Өn 0 -75.96 -82.87 -85.24 -86.42 -87.14 -87.61 -87.95

l d d h f f h

106

Amplitudes and phases for first seven harmonics

Page 54: Chapter 2

2/9/2014

Communication Systems 54

Trigonometric Fourier Series

• Periodicity of the trigonometric Fourier series

Dr. Sarmad Ullah Khan

The co-efficient of the of the Fourier series are calculatedfor the interval [t1, t1+T0]

=

=

+++=+

++=

100

1

])([cos()(

)cos()(

nnono

nnono

TtnwCCTt

tnwCCt

θφ

θφ for all t

)(

)cos(

)2cos(

1

1

t

nwtCC

nnwtCC

no

n

no

no

n

no

φ

θ

θπ

=

++=

+++=

=

=

for all t

107

Trigonometric Fourier Series

• Periodicity of the trigonometric Fourier series

Dr. Sarmad Ullah Khan

tdtnwtgT

a o

Ton

o

2

cos)(2 =

dttgT

aTo

o

o

)(1 =

n= 1,2,3,……

tdtnwtgT

b o

Ton

o

sin)(2 = n= 1,2,3,……

108

Page 55: Chapter 2

2/9/2014

Communication Systems 55

Trigonometric Fourier Series

• Fourier Spectrum

Dr. Sarmad Ullah Khan

Consider the compact Fourier series

This equation can represents a periodic signal g(t) of

f i

)cos()(1

0 non

n tnwCCtg θ++= ∞

=

d )(frequencies:

Amplitudes:

Phases:

oooo nwwwwdc ,.....,3,2,),(0

nCCCCC ,......,3210 ,,,

nθθθθ ,.....,,,0 321

109

Trigonometric Fourier Series

• Fourier SpectrumF d i d i ti f )(

Dr. Sarmad Ullah Khan

nc vs w (Amplitude spectrum)

wvsθ (phase spectrum)

Frequency domain description of )( tφ

Time domain description of )( tφ

110

Page 56: Chapter 2

2/9/2014

Communication Systems 56

Example 2.8

Dr. Sarmad Ullah Khan

Find the compact Fourier series for the periodic square wave w(t) shown in figure and sketch amplitude and phase spectrum( ) g p p p

F i i

111

=

++=1

sincos)(n

onono tnwbtnwaatw

Fourier series:

2

11 4

4

0 == dtT

a

o

o

T

To

W(t)=1 only over (-To/4, To/4) andw(t)=0 over the remaining segmentdttgT

aoTt

to+

=1

1

)(1

0

Example 2.8

Dr. Sarmad Ullah Khan

== 2

sin2

cos2 4 π

πn

ndttnw

Ta

oT

on

− 24

πnToTo

=

π

π

n

n2

2

0

...15,11,7,3

...13,9,5,1

=

=

n

n

evenn

112

πn ,,,

0sin2 4

4

== ntdtT

b

o

o

T

Ton 0= nb

All the sine terms are zero

Page 57: Chapter 2

2/9/2014

Communication Systems 57

Example 2.8

Dr. Sarmad Ullah Khan

+−+−+= ....7cos

7

15cos

5

13cos

3

1cos

2

2

1)( twtwtwtwtw ooooπ 7532 π

The series is already in compact form as there are no sine terms

Except the alternating harmonics have negative amplitudes

The negative sign can be accommodated by a phase of radians asπ)cos(cos π−=− xx

113

Series can be expressed as:

++−++−++= ....9cos

9

1)7cos(

7

15cos

5

1)3cos(

3

1cos

2

2

1)( twtwtwtwtwtw ooooo ππ

π

Example 2.8

Dr. Sarmad Ullah Khan

2

1=oC

=πn

C n 2

0

oddn

evenn

−=

πθ

0n

for all n 3,5,7,11,15,…..

for all n = 3,5,7,11,15,…..≠

We could plot amplitude and phase spectra using these values….

In this special case if we allow Cn to take negative values we do not need a phase of to account for sign.π−

114

Means all phases are zero, so only amplitude spectrum is enough

Page 58: Chapter 2

2/9/2014

Communication Systems 58

Example 2.8

Dr. Sarmad Ullah Khan

Consider figure

)5.0)((2)( −= twtwo

115

+−+−= ....7cos

7

15cos

5

13cos

3

1cos

4)( twtwtwtwtw ooooπ

Outlines

• Signals and systems• Size of signal

Dr. Sarmad Ullah Khan

Size of signal• Classification of signals• Signal operations• The unit impulse function• Signals versus Vectors• Correlation

O th l i l• Orthogonal signals• Trigonometric Fourier Series• Exponential Fourier Series

116

Page 59: Chapter 2

2/9/2014

Communication Systems 59

Exponential Fourier Series

• According to Euler’s theorem, each Sin function canbe represented as a sum of ejwt and e-jwt

Dr. Sarmad Ullah Khan

be represented as a sum of ejwt and e jwt

• Also a set of exponentials ejnwt is orthogonal over anytime interval T=2*pi/w

A i l (t) l b t d• A signal g(t) can also be represented as anexponential Fourier series over an interval T0

117

Exponential Fourier Series

Dr. Sarmad Ullah Khan

• Where the coefficient Dn can be calculated as

• Exponential Fourier series is an another form oftrigonometric Fourier series

118

Page 60: Chapter 2

2/9/2014

Communication Systems 60

Exponential Fourier Series

• Exponential Fourier series is an another form oftrigonometric Fourier series

Dr. Sarmad Ullah Khan

trigonometric Fourier series

119

Exponential Fourier Series

• The compact trigonometric Fourier series of aperiodic signal g(t) is given by

Dr. Sarmad Ullah Khan

periodic signal g(t) is given by

120

Page 61: Chapter 2

2/9/2014

Communication Systems 61

Exponential Fourier Series

• Exponential Fourier Spectra

Dr. Sarmad Ullah Khan

To draw Dn we need to find its spectra

As Dn is a complex quantity having real and imaginaryvalue, thus we need two plots (real and imaginary parts ORamplitude and angle of Dn)

Amplitude and phase is prefered because of its closeti ith th di t fconnection with the corresponding components of

trigonometric Fourier series

We plot |Dn| vs ω and ∟Dn vs ω

121

Exponential Fourier Series

• Exponential Fourier Spectra

Dr. Sarmad Ullah Khan

Comparing exponential with trigonometric Fourierspectrum yields

For real periodic signal, Dn and D n are conjugate, thusp g , n -n j g ,

122

Page 62: Chapter 2

2/9/2014

Communication Systems 62

Exponential Fourier Series

• Exponential Fourier Spectra

Dr. Sarmad Ullah Khan

123

Exponential Fourier Series

• Exponential Fourier Spectra

Dr. Sarmad Ullah Khan

sec22 radT

wo

o == π

−∞=

=n

ntjn eDt 2)(ϕ

dteedtetT

D ntjtntj

To

no

2

0

22 1)(

1 −−− ==π

πϕ

== +−

π)2

2

1(1

dtetn

π=oT

124

π 0

nj 41

504.0

+=

Page 63: Chapter 2

2/9/2014

Communication Systems 63

Exponential Fourier Series

• Exponential Fourier Spectra

Dr. Sarmad Ullah Khan

sec22 radT

wo

o == π

π=oTntj

n

enj

t 2

41

1504.0)(

−∞= +=ϕ

and

+

++

++

++

=111

...121

1

81

1

41

11

504.0

642 tjtjtj ej

ej

ej

Dn are complex

Dn and D-n are conjugates

125

+

++

−+−

−−− ...

121

1

81

1

41

1 642 tjtjtj ej

ej

ej

Exponential Fourier Series

• Exponential Fourier Spectra

Dr. Sarmad Ullah Khan

sec22 radT

wo

o == π

π=oTnnn CDD2

1== −

nnD θ=< nnD θ−=< −and

thusnj

nn eDD θ= njnn eDD θ−

− =

126o

o

j

j

o

ej

D

ej

D

D

96.751

96.751

122.041

504.0

122.041

504.0

504.0

−−

=

+

=

=

o

o

D

D

96.75

96.75

1

1

=<

−=<

Page 64: Chapter 2

2/9/2014

Communication Systems 64

Exponential Fourier Series

• Exponential Fourier Spectra

Dr. Sarmad Ullah Khan

sec22 radT

wo

o == π

π=oT

o

o

j

j

ej

D

ej

D

87.822

87.822

625.081

504.0

625.081

504.0

−−

=

+

=

o

o

D

D

87.82

87.82

1

1

=<

−=<

127

And so on….

Exponential Fourier Series

• Exponential Fourier Spectra

Dr. Sarmad Ullah Khan

128