Communication Systems

Ref: Chapter 2 of text book)

SIGNALS AND SIGNAL SPACE

 Contents

• Signals

• Classification of Signals

• Signal Comparison: Correlation, Cross Correlation Function

• Auto Correlation Function

• Orthogonal Signal Set

• Review of Fourier Series and Fourier Transform and their Properties

2
 Signals

• A signal is a set of information or data.

• Examples
– a telephone or television signal,
– monthly sales of a corporation,
– the daily closing prices of a stock market.

• Signal is a function of a dependent variable against an

independent one.
– For our present course the independent variable is time.

• We deal exclusively with signals that are functions of time.

3
 Classifications of Signals

• Continuous-time and discrete-time signals

• Analog and digital signals

• Periodic and aperiodic signals

• Energy and power signals

• Deterministic and probabilistic signals

• Other Classifications
– even / odd signals,
– One-dimensional / Multi-dimensional

4
 Continuous-time and discrete-time
Signals
• A signal that is specified for every value of time t is a
continuous-time signal.
• A signal that is specified only at discrete values of t is a
discrete-time signals.

3 5
 Continuous-time and discrete-time
Signals (continued)
• A discrete-time signal can be obtained by sampling a
continuous-time signal.
• In some cases, it is possible to 'undo' the sampling operation.
That is, it is possible to get back the continuous-time signal
from the discrete-time signal.
• How?
• Sampling Theorem
– The sampling theorem states that if the highest frequency in the signal
spectrum is 𝐵, the signal can be reconstructed from its samples taken
at a rate not less than 2𝐵 sample per second.

6
 Analog and digital signals

• A signal whose amplitude can take

on any value in a continuous range is
an analog signal.
• A signal whose amplitude can take a
finite number of values is a digital
signal.
• The concept of analog and digital
signals is different from the concept
of continuous-time and discrete-
time signals.
• For example, we can have a digital
and continuous-time signal, or a
analog and discrete-time signal.

7
 Periodic and Aperiodic Signals

• A signal 𝑔(𝑡) is said to be periodic if for some positive

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

• A signal is aperiodic if it is not periodic.

• Some famous periodic signals:
– sin 𝑤0𝑡
– cos 𝑤0𝑡
– 𝑒 𝑗𝑤0𝑡 = cos 𝑤0𝑡 + j sin 𝑤0𝑡

• where 𝑤0 = 2𝜋 𝑇0 and 𝑇0 is the period of the function.

8
 Periodic Signal

• A periodic signal 𝑔(𝑡) can be generated by periodic extension

of any segment of 𝑔(𝑡) of duration 𝑇0.

9
 Energy and Power signal

• First, define energy.

– The signal energy 𝐸𝑔 of 𝑔(𝑡) is defined (for a real signal) as
∞
𝐸𝑔 = 𝑔2 𝑡 𝑑𝑡
−∞
– In the case of a complex valued signal 𝑔(𝑡), the energy is given by
∞ ∞
𝐸𝑔 = 𝑔∗ 𝑡 𝑔 𝑡 𝑑𝑡 = 𝑔 𝑡 2 𝑑𝑡
−∞ −∞
• A necessary condition for energy to be finite is that the
signal amplitude approaches zero as 𝑡 → ∞ otherwise the
above integral will not converge. If this condition doesn’t
hold the signal energy is infinite.
• A signal 𝑔(𝑡) is an energy signal if 𝐸𝑔 < ∞.

10
 Energy and Power signal

• A power signal must have an infinite duration.

• In case of signals with infinite energy (e.g., periodic signals), a
more meaningful measure is the signal power.
𝑇
1 2
𝑃𝑔 = lim 𝑔2 (𝑡)𝑑𝑡
𝑇→∞ 𝑇 −𝑇
2
• For a complex signal 𝑔(𝑡), the signal power is given by
𝑇
1 2
𝑃𝑔 = lim 𝑔(𝑡) 2 𝑑𝑡
𝑇→∞ 𝑇 −𝑇
2
• The square root of 𝑃𝑔 is known as the rms value of 𝑔(𝑡)
• Note: A signal cannot be an energy and a power signal at the
same time.
11
 Energy Signal
Example
• Determine the energy of the following signal.

• Signal Energy calculation

∞ 0 ∞
−𝑡
𝐸𝑔 = 𝑔2 𝑡 𝑑𝑡 = (2)2 𝑑𝑡 + (2𝑒 2 )2 𝑑𝑡 =4+4=8
−∞ −1 0

12
 Power Signal
Example

• Why it is a power signal?

– Signal does not approach 0 as 𝑡 → ∞
– Signal is periodic
• Signal power calculation?
𝑇 1
1 2
2
1 1
𝑃𝑔 = lim 𝑔(𝑡) 𝑑𝑡 = 𝑡 2 𝑑𝑡 =
𝑇→∞ 𝑇 −𝑇 2 −1 3
2
• Its rms value will be?
– Square root of 𝑃𝑔 i.e., 1 3

3 13
 Practice Problems

• Find the energy of the following signal.

• Find a suitable measure of the following signal.

14
 Deterministic and Probabilistic Signals

• A signal whose physical description is known completely is a

deterministic signal.
– Mathematical form
– Graphical form

• A signal known only in terms of probabilistic descriptions is a

random signal.
– Mean value
– Mean square value
– Distributions
– Example: Noise, message signals

15
 Other Classifications

• Even and Odd Signals

– A signal is even if 𝑥(𝑡) = 𝑥(−𝑡). One dimensional
– A signal is odd if 𝑥(𝑡) = −𝑥(−𝑡)
• Examples:
– cos(𝑡) is an even signal.
– sin(𝑡) is an odd signal.
Two dimensional
• One-dimensional & Multi-dimensional
– Speech varies as a function of time  one-dimensional
– Image intensity varies as a function of (𝑥, 𝑦) coordinates  multi-
dimensional

16
 Unit Impulse Function
• The unit impulse function or Dirac delta function is defined as
𝛿 𝑡 =0 𝑡≠0
∞
𝛿 𝑡 𝑑𝑡 = 1
−∞
• Can be visualized as a tall, narrow, rectangular pulse of unit area.

• Multiplication of a function by an impulse – Sifting Property

𝑔 𝑡 𝛿 𝑡 =𝑔 0 𝛿 𝑡
𝑔 𝑡 𝛿 𝑡−𝑇 =𝑔 𝑇 𝛿 𝑡−𝑇
∞
𝑔(𝑡)𝛿 𝑡 − 𝑇 𝑑𝑡 = 𝑔(𝑇)
−∞

17
 Unit step function
• Another useful signal is the unit step function 𝑢(𝑡), defined by

1 𝑡≥0
𝑢 𝑡 =
0 𝑡<0

• Observe that
𝑡
1 𝑡≥0
𝛿 𝜏 𝑑𝜏 = = 𝑢(𝑡)
−∞ 0 𝑡<0
• Which also results into
𝑑𝑢
= 𝛿(𝑡)
𝑑𝑡 18
 Sinusoids

• Consider the sinusoid

𝑥 𝑡 = 𝐶 cos(2𝜋𝑓0 𝑡 + 𝜃)
• 𝑓0 (measured in Hertz) is the frequency of the sinusoid and
𝑇0 = 1 𝑓0 is the period. 𝐶 is the amplitude and 𝜃 is the phase.
• Sometimes we use 𝑤0 (radians per second) to express 2𝜋𝑓0 .
• Important Identities
𝑒 ±𝑗𝑥 = cos 𝑥 ± 𝑗 sin 𝑥
1 𝑗𝑥
cos 𝑥 = 𝑒 + 𝑒 −𝑗𝑥
2
1 𝑗𝑥
sin 𝑥 = 𝑒 − 𝑒 −𝑗𝑥
2𝑗

19
 Signal Operations
Time Shifting Time Scaling Time Reversal or inversion
φ(𝑡) = g(𝑡 − 𝑇) represents “Delay” φ(𝑡) = g(2𝑡) represents φ 𝑡 = g −𝑡
“Compression”
φ(𝑡) = g(𝑡 + 𝑇) represents “Advance” φ(𝑡) = g(𝑡/2) represents Mirror image on vertical axis i.e., y-axis
“Expansion”
Note: −𝑥 𝑡 means mirror image on
horizontal axis i.e., x-axis

20
 Signals and Vectors
• Signals and vectors are closely related. For
example,
– A vector has components,
– A signal has also its components.
• g is a certain vector.
– It is specified by its magnitude or length g and
direction.
• Consider a second vector x.
• Different ways to express g in term of vector x
– g = 𝑐x + 𝑒 = 𝑐1 x + 𝑒1 = 𝑐2 x + 𝑒2
– e is the error vector
– c is the magnitude of projection of g on x
– Choose c to minimize e i.e., e = g − cx

21
 Inner product in vector spaces

• For convenience we define the dot (inner or scalar) product of two

vectors as
g, x = g x cos θ

• Therefore, x, x = x 2
• The length of the component of g along x is g cos 𝜃, but it is also
c𝑥.
c 𝑥 = g cos 𝜃
• Multiplying both sides by 𝑥
c𝑥 2
= g x cos 𝜃 = g, x

g, x g. x
c= =
x, x x2
• When g, x = 0, we say that g and x are orthogonal to each other
𝜋
i.e., geometrically, 𝜃 =
2
22
 Signals as vectors

• The same notion of inner product can be applied for signals.

• What is the useful part of this analogy?
– We can use some geometrical interpretation of vectors to understand
signals.
• Consider two (energy) signals 𝑦(𝑡) and 𝑥(𝑡).
• The inner product is defined by
∞
𝑦(𝑡), 𝑥(𝑡) = 𝑦 𝑡 𝑥 𝑡 𝑑𝑡
−∞
• For complex signals
∞
𝑦(𝑡), 𝑥(𝑡) = 𝑦 𝑡 𝑥 ∗ 𝑡 𝑑𝑡
−∞
• The two signals are orthogonal if
𝑦(𝑡), 𝑥(𝑡) = 0

23
 Signals as vectors

• If a signal g(t) is approximated in terms of another real signal

x(t) over an interval (t1, t2), then the best approximation
would be the one that minimizes the size of error e(t) (Error
energy). t t
2 2

Ee   e (t )dt   [ g (t )  cx(t )]2 dt

2

t1 t1

dEe
• Minimizing error energy would mean putting dc
0
• Which would eventually give
t2

 g (t ) x(t )dt 1 2
t

Ex t1
c 
t1
t2
g (t ) x(t )dt
 (t )dt
2
x
t1

• which is optimum value of c that minimizes the energy of the

error signal in the approximation g(t)~cx(t).
24
 Signals as vectors
Example
For a square signal g(t) shown below find the component in g(t) of the form sin 𝑡.

𝐠(𝐭) ~ 𝐜 𝐬𝐢𝐧 𝐭
Objective: Select c so that the energy of the error signal is minimum. As we already know that
this condition holds for
1 t2
c= g t x t dt
Ex t1
Let, x(t) = sin t
2π
So, Ex = 0
sin2 t dt = π
t2 π 2𝜋
1 1 1 4
c= g t sin 𝑡 dt = sin 𝑡 dt + − sin 𝑡 dt =
𝜋 t1 𝜋0 𝜋 π 𝜋
4
Therefor, g(t) ~ sin t
𝜋
25
 Energy of orthogonal signals

• If vectors 𝑥 and 𝑦 are orthogonal, and if 𝑧 = 𝑥 + 𝑦

𝑧 2 = 𝑥 2+ 𝑦 2 (Pythagorean Theorem)

• If signals 𝑥(𝑡) and 𝑦(𝑡) are orthogonal and if 𝑧(𝑡) = 𝑥(𝑡) + 𝑦(𝑡) then
𝐸𝑧 = 𝐸𝑥 + 𝐸𝑦
• Proof?

26
 Power of orthogonal signals

• The same concepts of orthogonality and inner product extend

to power signals.
• For example,
• 𝑔 𝑡 = 𝑥 𝑡 + 𝑦 𝑡 = 𝐶1 cos(𝑤1 𝑡 + 𝜃1 ) + 𝐶2 cos(𝑤2 𝑡 + 𝜃2 )
and 𝑤1 ≠ 𝑤2
2 2
𝐶1 𝐶2
𝑃𝑥 = , 𝑃𝑦 =
2 2
• The signal 𝑥(𝑡) and 𝑦(𝑡) are orthogonal: 𝑦(𝑡), 𝑥(𝑡) = 0.
Therefore,
2 2
𝐶1 𝐶2
𝑃𝑔 = 𝑃𝑥 + 𝑃𝑦 = +
2 2
27
 Signal Comparison: Correlation

• Why bother poor undergraduate students with correlation?

– Correlation is widely used in engineering.

• To design receivers in many communication systems
• To identify signals in radar systems
• For classifications.

28
 Signal Comparison: Correlation

• If vectors 𝑥 and 𝑦 are given, we have the correlation measure as

𝑦, 𝑥
𝑐𝑛 = cos 𝜃 =
𝑦 𝑥
• Clearly, −1 ≤ 𝑐𝑛 ≤ 1 and is called correlation coefficient.
• In the case of energy signals:
∞
1
𝑐𝑛 = 𝑦 𝑡 𝑥(𝑡) 𝑑𝑡
𝐸𝑦 𝐸𝑥 −∞
• Again, −1 ≤ 𝑐𝑛 ≤ 1
• Auto-Correlation
∞
𝜓𝑔 𝜏 = 𝑔 𝑡 𝑔(𝑡 + 𝜏) 𝑑𝑡
−∞
• It is a measure of the similarity of a signal with its displaced version.

29
 Best friends, worst enemies and
complete strangers
• 𝑐𝑛 = 1. Best friends. This happens when g t = K x(t) and K
is positive.
– The signals are aligned, maximum similarity.

• 𝑐𝑛 = −1. Worst Enemies. This happens when g t = K x(t)

and K is negative.
– The signals are again aligned, but in opposite directions.
– The signals understand each other, but they do not like each other.

• 𝑐𝑛 = 0. Complete Strangers. The two signals are orthogonal.

– We may view orthogonal signals as unrelated signals.

30
 Correlation Examples

• Find the correlation coefficients between:

– 𝑥 𝑡 = 𝐴0 cos 𝑤0 𝑡 and 𝑦(𝑡) = 𝐴1 sin(𝑤1 𝑡)
• 𝑐𝑥,𝑦 = 0
– 𝑥 𝑡 = 𝐴0 cos 𝑤0 𝑡 and 𝑦(𝑡) = 𝐴1 cos(𝑤1 𝑡) and 𝑤0 ≠ 𝑤1
• 𝑐𝑥,𝑦 = 0
– 𝑥 𝑡 = 𝐴0 cos 𝑤0 𝑡 and 𝑦(𝑡) = 𝐴1 cos(𝑤0 𝑡)
• 𝑐𝑥,𝑦 = 1
– 𝑥 𝑡 = 𝐴0 sin 𝑤0 𝑡 and 𝑦(𝑡) = 𝐴1 sin(𝑤1 𝑡) and 𝑤0 ≠ 𝑤1
• 𝑐𝑥,𝑦 = 0
– 𝑥 𝑡 = 𝐴0 sin 𝑤0 𝑡 and 𝑦(𝑡) = 𝐴1 sin(𝑤0 𝑡)
• 𝑐𝑥,𝑦 = 1
– 𝑥 𝑡 = 𝐴0 sin 𝑤0 𝑡 and 𝑦 𝑡 = −𝐴1 sin 𝑤0 𝑡
• 𝑐𝑥,𝑦 = −1
31
 Signal representation by
orthogonal signal sets

• Examine a way of representing a signal as a sum of orthogonal

signals.

• We know that a vector can be represented as the sum of

orthogonal vectors.

• Review the case of vectors and extend to signals.

32
 Orthogonal vector space
• Consider a three-dimensional Cartesian vector space
described by three mutually orthogonal vectors, 𝑥1 , 𝑥2 and
𝑥3 .
0 𝑚≠𝑛
𝑥𝑚 , 𝑥𝑛 =
𝑥𝑚 2 𝑚 = 𝑛
• Any three-dimensional vector can be expressed as a linear
combination of those three vectors:
𝑔 = 𝑐1 𝑥1 + 𝑐2 𝑥2 + 𝑐3 𝑥3
𝑔,𝑥𝑖
• Where 𝑐𝑖 =
𝑥𝑖 2
• In this case, we say that this set of vectors i.e., 𝑥1 , 𝑥2 , 𝑥3 is a
complete set of orthogonal vectors in 3D space.
• Such vectors are known as a basis vector.
33
 Orthogonal signal space

• Same notions of completeness extend to signals.

• A set of mutually orthogonal signals 𝑥1 𝑡 , 𝑥2 𝑡 , … , 𝑥𝑁 (𝑡) is complete if it can

represent any signal belonging to a certain space. For example:

𝑔 𝑡 ≅ 𝑐1 𝑥1 (𝑡) + 𝑐2 𝑥2 (𝑡) + ⋯ + 𝑐𝑁 𝑥𝑁 (𝑡)

𝑁

≅ 𝑐𝑛 𝑥𝑛 (𝑡)
𝑛=1

• If the approximation error is zero for any 𝑔(𝑡) then the set of signals
𝑥1 𝑡 , 𝑥2 𝑡 , … , 𝑥𝑁 (𝑡) is complete. In general, the set is complete when 𝑁 → ∞.

• This would lead to the Generalized Fourier series of 𝑔(𝑡) as

∞

𝑔 𝑡 = 𝑐𝑛 𝑥𝑛 (𝑡)
𝑛=1

34
 Parseval’s Theorem

• We have already established that the energy of the sum of

orthogonal signals is equal to the sum of their energies.

• The energy of 𝑔 𝑡 = ∞ 𝑛=1 𝑐𝑛 𝑥𝑛 (𝑡) can be expressed as the sum of

the energies of the individual orthogonal components

𝐸𝑔 = 𝑐1 2 𝐸1 + 𝑐2 2 𝐸2 + ⋯
∞

= 𝑐𝑛 2 𝐸𝑛
𝑛=1
• In vector space, the square of the length of vector is equal to the
sum of the squares of the lengths of its orthogonal components.
– Parseval’s theorem is the statement of this fact when applied to
signals.
 Trigonometric Fourier series

• Consider a signal set

{1, cos 𝑤0 𝑡, cos 2𝑤0 𝑡, … , cos 𝑛𝑤0 𝑡, … , sin 𝑤0 𝑡 , sin 2𝑤0 𝑡 , … , sin 𝑛𝑤0 𝑡}

• A sinusoid of frequency 𝑛𝑤0 is called the 𝑛𝑡ℎ harmonic of the

sinusoid, where 𝑛 is an integer.

• The sinusoid of frequency 𝑤0 is called the fundamental harmonic.

• This set is orthogonal over an interval of duration 𝑇0 = 2𝜋 𝑤0 ,

which is the period of the fundamental.

36
 Trigonometric Fourier series

• The components of the following set are orthogonal.

{1, cos 𝑤0 𝑡, cos 2𝑤0 𝑡, … , cos 𝑛𝑤0 𝑡, … , sin 𝑤0 𝑡 , sin 2𝑤0 𝑡 , … , sin 𝑛𝑤0 𝑡}

0 𝑚≠𝑛
cos 𝑛𝑤0 𝑡 cos 𝑚𝑤0 𝑡 𝑑𝑡 = 𝑇0
𝑇0 𝑚=𝑛≠0
2
0 𝑚≠𝑛
sin 𝑛𝑤0 𝑡 sin 𝑚𝑤0 𝑡 𝑑𝑡 = 𝑇0
𝑇0 𝑚=𝑛≠0
2
𝑇
sin 𝑛𝑤0 𝑡 cos 𝑚𝑤0 𝑡 𝑑𝑡 = 0 for all 𝑚 and 𝑛
0

• 𝑇0
means integral over an interval from 𝑡 = 𝑡1 to 𝑡 = 𝑡1 + 𝑇0 for
any value of 𝑡1 .
37
 Trigonometric Fourier series

• This set is also complete in 𝑇0 . That is, any signal in an interval 𝑡1 ≤

𝑡 ≤ 𝑡1 + 𝑇0 can be written as the sum of sinusoids. Or

𝑔 𝑡 = 𝑎0 + 𝑎1 cos 𝑤0 𝑡 + 𝑎2 cos 2𝑤0 𝑡+…+𝑏1 sin 𝑤0 𝑡 + 𝑏2 sin 2𝑤0 𝑡 + ⋯

∞

= 𝑎0 + 𝑎𝑛 cos 𝑛𝑤0 𝑡 + 𝑏𝑛 sin 𝑛𝑤0 𝑡

𝑛=1
• Series coefficients

𝑔 𝑡 , cos 𝑛𝑤0 𝑡
𝑎𝑛 =
cos 𝑛𝑤0 𝑡 , cos 𝑛𝑤0 𝑡

𝑔 𝑡 , sin 𝑛𝑤0 𝑡
𝑏𝑛 =
sin 𝑛𝑤0 𝑡 , sin 𝑛𝑤0 𝑡

38
 Trigonometric Fourier Coefficients

• Therefore
𝑡1 +𝑇0
𝑡1
𝑔 𝑡 cos 𝑛𝑤0 𝑡 𝑑𝑡
𝑎𝑛 = 𝑡1 +𝑇0
cos 2 𝑛𝑤 𝑡𝑑𝑡
𝑡1 0

As
𝑡1 +𝑇0 𝑡1 +𝑇0
𝑡1
cos 2 𝑛𝑤 𝑡𝑑𝑡
0 = 𝑇0 2 and 𝑡1
sin2 𝑛𝑤 𝑡𝑑𝑡
0 = 𝑇0 2

We get
𝑡1 +𝑇0
1
𝑎0 = 𝑔(𝑡)𝑑𝑡
𝑇0 𝑡1
𝑡1 +𝑇0
2
𝑎𝑛 = 𝑔(𝑡) cos 𝑛𝑤0 𝑡 𝑑𝑡 𝑛 = 1,2,3
𝑇0 𝑡1
𝑡1 +𝑇0
2
𝑏𝑛 = 𝑔(𝑡) sin 𝑛𝑤0 𝑡 𝑑𝑡 𝑛 = 1,2,3
𝑇0 𝑡1
39
 Exponential Fourier Series

• Consider a set of exponentials

𝑒 𝑗𝑛𝑤0 𝑡 𝑛 = 0, ±1, ±2, …

• The components of this set are orthogonal.

• A signal 𝑔 𝑡 can be expressed as an exponential series over

an interval 𝑇0 as follows
∞
𝑗𝑛𝑤 𝑡
1
𝑔 𝑡 = 𝐷𝑛 𝑒 0 𝐷𝑛 = 𝑔(𝑡)𝑒 −𝑗𝑛𝑤0 𝑡 𝑑𝑡
𝑇0 𝑇0
𝑛=−∞

40
 Trigonometric and exponential
Fourier series
• Trigonometric and exponential Fourier series are related. In
fact, a sinusoid in the trigonometric series can be expressed
as a sum of two exponentials using Euler's formula.
𝐶𝑛 𝑗(𝑛𝑤 𝑡+𝜃 )
𝐶𝑛 cos(𝑛𝑤0 𝑡 + 𝜃𝑛 ) = 𝑒 0 𝑛 + 𝑒 −𝑗(𝑛𝑤0 𝑡+𝜃𝑛 )
2
𝐶𝑛 𝑗𝜃 𝑗𝑛𝑤0 𝑡
𝐶𝑛 −𝑗𝜃
= 𝑒 𝑛 𝑒 + 𝑒 𝑛 𝑒 −𝑗𝑛𝑤0 𝑡
2 2
= 𝐷𝑛 𝑒 𝑗𝑛𝑤0 𝑡 + 𝐷−𝑛 𝑒 −𝑗𝑛𝑤0 𝑡
Where
𝐶𝑛 𝑗𝜃 𝐶𝑛 −𝑗𝜃
𝐷𝑛 = 𝑒 𝑛 and 𝐷−𝑛 = 𝑒 𝑛
2 2

41
 Parseval's Theorem

• Exponential Fourier series representation

∞

𝑔 𝑡 = 𝐷𝑛 𝑒 𝑗𝑛𝑤0 𝑡
𝑛=−∞

• Where
1
𝐷𝑛 = 𝑔(𝑡)𝑒 −𝑗𝑛𝑤0 𝑡 𝑑𝑡
𝑇0 𝑇0

• Power for the exponential representation is

∞

𝑃𝑔 = 𝐷𝑛 2

𝑛=−∞

42
 Example
• Find the exponential Fourier series for the following signal

43