EE357 e_Lecture 1 A Review

Review of Signals and Systems

1. Signals
A voltage or a current waveform in an electrical circuit is a signal. A signal
can be represented, for example, as x(t) which is the signal value at a time
instant t. In some cases, x(t) is deterministic where it is described in functional
form. In some other cases, x(t) is random where it is described using a
probability density function.
Energy and Power of a Continuous Time Deterministic Signal:
For an electrical signal x(t), the instantaneous power dissipated in a unit
resistor is
x 2 (t )
p(t )   x 2 (t )
1
(normal way of representing energy (normalized to 1  resistor))

Energy dissipated over a time interval of T is

T /2 T /2

E  p(t )dt  x (t )dt

2

T / 2 T / 2

Average power dissipated by the signal during this interval is

1 T /2 2
P  x (t )dt
T T / 2
Some signals, such as periodic signals, have infinite energy over time but their
average power is finite. Some other signals, which last for a short time, have
finite energy but their average is very small.

1.1Energy Signals
x(t) is an energy signal if and only if it has nonzero but finite energy (0 < E <
) for all time where
Lim T /2 

E T   x (t )dt   x (t )dt  finite

2 2

T / 2 

1.2Power Signals
x(t) is a power signal if and only if it has finite but nonzero power (0 < P <
) for all time time where
Lim 1 T /2 2
P T   x (t )dt  finite, nonzero
T T / 2

Prepared By K D R Jagath Kumara

EE357 e_Lecture 1 A Review

Under this classification, an energy signal has finite energy but zero average
power whereas a power signal has a finite average power but infinite energy.

1.3Fourier Series
A periodic signal is an example for a power signal. Such a signal may be
represented using a sum of sinusoids known as the Fourier series.
Theorem: Let x(t) be a periodic function with a period of T and x(t) and x’(t)
are piecewise continuous on the interval –T/2  t  T/2. Then, x(t) has a
Fourier series of the form
a 
x(t )  0   an cos(n 0 t )  bn sin(n 0 t )
2 n 1

2
where the fundamental frequency  0  and
T
2 T /2 2 T /2 2 T /2
a0   x(t )dt ; an  T T/x2 (t ) cos(n0 t )dt ; bn  T T/x2 (t ) sin(n 0 t )dt
T T / 2
The RHS converges to all the points where x(t) is continuous and to [x(t+0) +
x(t-0)/2] at jump discontinuities.
Dirichlet Conditions: In general, if x(t) has a Fourier series it must satisfy
a T

(i) x(t) is absolutely integrable over any period, that is  x(t ) dt   for any
a

a.
(ii) x(t) has only finite number of maxima or minima over any period.
(iii) x(t) has only a finite number of discontinuities over any period.
Exponential Fourier Series: The sum of sinusoids in the trigonometric
Fourier series can be represented using a sum of exponentials.
Theorem: A periodic function x(t) can be represented as an infinite sum of
complex exponentials of the form

x(t )  c
k  
k
e jk t ;    t   where
0

1 T /2
ck   x(t )e  jk t dt ; k  0 , 1, 2 ,..
T T / 2
0

Prepared By K D R Jagath Kumara

EE357 e_Lecture 1 A Review

Note that all the frequencies which form the spectra of x(t) contains integer
multiples of 0.

Example: Rectangular pulse sequence where T = 2.

x(t)
1
t
-2 -1 -0.5 0 0.5 1 2
Dirichlet conditions: Integrable over a period; maximas and minimas = 3;
discontinuities = 2.
2
0    rad/s
T

x( t )  c
k  
k
e jkt ;    t  

1 T /2 11 1 0.5
c0   x
T T / 2
( t )dt  
2 1
x( t )dt   1.dt  0.5
2 0.5
1 T /2 1 0.5  jkt 1 1
ck   x ( t ) e jk t
dt 0
 1.e dt  e  jkt
0.5
 e  jk / 2  e jk / 2 
T T / 2 2 0.5  j 2 k  0.5
 j 2 k
1
ck  sin( k / 2 ); k  1, 2 ,..
k

0 ; k even

ck   1
 k ( 1)
( k 1) / 2
; k odd


1
x(t )  0.5   k ( 1) e jkt
( k 1 ) / 2

k  
k odd

0 ; k even

ck   1 and c k  0 0 or 180 0 for k  1,  3,..
 k ; k odd

Prepared By K D R Jagath Kumara

EE357 e_Lecture 1 A Review

|Ck|

1/ 1/
1/3 1/3
1/5 1/5

-5 -4 -3 -2 - 0  2 3 4 5 

2
Frequency Spectra (Amplitude Spectrum) of a rectangular pulse sequence)

Parseval’s Theorem: Let x(t) be a periodic signal with period T. The average
power normalized to a 1  resistor
1 T /2 2 


2
P  x
T T / 2
( t )dt 
k  
ck

Power Spectral Density (PSD): The distribution of power of a power signal

x(t) in the frequency domain is defined as


 c ( f
2
Gx ( f )  k
 kf 0 )
k  

Note that the delta functions indicate where each frequency component occurs.

1.4Fourier Transform
The frequency information of non-periodic signals could be obtained using
the Fourier transform. An energy signal is an example for such a signal which
exists over a finite time interval only and has a finite energy.
The Fourier transform X() of x(t), which is an energy signal, is defined to be


X ( )   x(t )e  jt dt ;      


where  is a continuous frequency variable.

A signal x(t) will have a Fourier transform in the ordinary sense if, within any
finite interval of time, x(t)


(i) is integrable (  x(t ) dt   ).



Prepared By K D R Jagath Kumara

EE357 e_Lecture 1 A Review

(ii) has finite number of maxima and minima.

(iii) has finite number of discontinuities.
(Note: The signals that can be physically generated, satisfy these conditions.)
Examples:
1) x(t) = 1, constant for all t: F[1] cannot be evaluated using first principles.
Hence, F[1] does not exist in the ordinary sense.

2) x(t) = u(t): Again, F[u(t)] does not exist in the ordinary sense. However, a
generalized transform could be found by using a property of Fourier
transform. This is
1
F[u(t )]    ( )
j

 
3) x(t) = p(t): F[ p (t )]  P( )   sin c  
 2 

P()
P(t)
1 

t 
-/2 0 /2 -6/ -4/ -2/ 0 2/ 4/ 6/

Inverse Fourier Transform:

It is possible to obtain the time domain version x(t) of the signal by taking the
inverse Fourier transform of the corresponding frequency domain signal X()
where
1 
x(t )   X ( )e jt d
2 
Note: 1/2 could be included in both transforms. e-jt and ejt could be
interchanged in the two transforms.

Prepared By K D R Jagath Kumara

EE357 e_Lecture 1 A Review

Some Fourier Transform Pairs:

1
 (t )  1 , e bt u(t )  ; b  0, e j t  2 (  0 ) ,
0

b  j
 
p (t )   sin c  
 2 
If x(t )  X( ) , then

x(t  c )  X( )e  jc

1 
x( at )  X   ; a > 0
a a
j
x(t ) sin 0 t  [ X(  0 )  X(  0 )]
2
Further, if v(t )  V ( ) , then

x(t )  v(t )  X ( )V ( )

1
x(t ).v(t )  [ X( )  V ( )]
2
Parseval’s Theorem:

1 
2
If x(t )  X( ) then  x (t )dt   X( ) d
2

 2 

The energy of a signal could be found either in the time domain or in the
frequency domain.
Energy Spectral Density (ESD): From Parseval’s theorem, the energy Ex of
a real valued signal is the area under the function |𝑋(𝜔)| . This implies that
|𝑋(𝜔)| is the energy per unit bandwidth in J/Hz. Hence, the energy
spectral density,
1 2 2
 x ( )  X( ) and  x ( f )  X( f )
2

Prepared By K D R Jagath Kumara

EE357 e_Lecture 1 A Review

1.5Autocorrelation
Energy Signals:
Autocorrelation matches a signal with a delayed version of itself. For a real
valued energy signal x(t), the autocorrelation is defined as


R x ( )   x(t )x(t   )dt ;      



where  is the time shift.

Rx() has these properties.
(i) R x ( )  R x (  ) , symmetrical about  = 0.
(ii) R x ( )  R x ( 0 ) for all . Maximum at the origin (Rx(0)) is the signal
energy.
(iii) R x ( )   x ( ) , Fourier transform of the autocorrelation is the ESD.

Power Signals:
The autocorrelation function of a power signal x(t) is defined as
1 T /2
Lim
Rx ( )  T   x(t )x(t   )dt ;      
T T / 2
If the power signal is periodic with period T0, time averaging can be performed
over a single period as
1 T0 / 2

R x ( )   x(t )x(t   )dt ;     

T0 T0 / 2

Note that
(i) R x ( )  R x (  ) , symmetrical about  = 0.
(ii) R x ( )  R x ( 0 ) for all . Maximum at the origin (Rx(0)) is the signal
power.
(iii) R x ( )  G x ( ) , Fourier transform of the autocorrelation is the PSD.

Prepared By K D R Jagath Kumara