ECE 318 - Analog & Digital Communication Systems

Fall 2018

Tutorial 1

TA: Subhajit Majhi ([email protected])

Office Hours: Mondays, Wednesdays & Thursdays from 5:30 pm to 6:30 pm (DC 2638)
Textbook: Fundamentals of Communication Systems, 2nd edition, J.G. Proakis and M. Salehi
We will review
• the Fourier Series representation of periodic signals

• the Energy-type and power-type of signals.

Fourier Series Representation of Periodic Signals

Definition
Let the signal x(t) be a periodic signal with period T0 . x(t) can be expanded in terms of complex
j2π Tn t +∞
exponential signals {e 0 }
n=−∞ .

The Fourier series representation of x(t) is:

∞
j2π Tn t
X
x(t) = xn e 0

n=−∞

where
Z α+T0
1 −j2π Tn t
xn = x(t)e 0 dt.
T0 α

Here, the integral is over one period, starting from some arbitrary α.

• The coefficients xn are called the Fourier series coefficients of the signal x(t).
1
• The quantity f0 = T0
is called the fundamental frequency of the signal x(t).

• Frequencies of the complex exponential signals are multiples of f0 . The nth multiple of f0
is called the nth harmonic. The coefficient xn determines the contribution of the complex
exponential corresponding to the nth harmonic to the magnitude of x(t).

1
Alternative Representations for Real Periodic Signals
For a real periodic signal x(t), the positive and negative coefficients are conjugates, i.e.,

x−n = x∗n .

For such signals, the Fourier series has an alternative representation, called the Trigonometric
Fourier series. It is given as:
∞ ∞
a0 X n X n
x(t) = + an cos(2π t) + bn sin(2π t)
2 n=1
T0 n=1
T0

where the Trigonometric Fourier series coefficients an and bn are defined as:
Z α+T0 Z α+T0
2 n 2 n
an = x(t) cos(2π t)dt, bn = x(t) sin(2π t)dt. (1)
T0 α T0 T0 α T0

Yet another representation of the Fourier series for real periodic signals in terms of magnitude and
phase is given next:
∞
X n
x(t) = x0 + 2 |xn | cos(2π t + θn )
n=1
T0

where
a0 1p 2 b 
n
x0 = , |xn | = an + b2n , θn = −arctan , n = 1, 2, 3, . . .
2 2 an

• In addition to being real and periodic, if the signal x(t) is even, then bn = 0.

Since x(t) is even and sin(2π Tn0 t) is odd, the product x(t) sin(2π Tn0 t) is an odd signal. Now, the
integration of an odd signal over one period is zero. Hence, bn in eqn.(1) is zero, and the series
simplifies to
∞
a0 X n
x(t) = + an cos(2π t).
2 n=1
T0

• Similarly, in addition to being real and periodic, if the signal x(t) is odd, then an = 0.

Since x(t) is now odd and cos(2π Tn0 t) even, the product x(t) cos(2π Tn0 t) is still an odd signal. More-
over, since the integration of an odd signal over one period is zero, an = 0 in eqn.(1). The series
thus simplifies to
∞
X n
x(t) = bn sin(2π t).
n=1
T0

2
Problem 1: Textbook Problem 2.39 Part 4
Determine the Fourier series expansion of the following signal
∞
X
x(t) = Λ(t − 2n)
n=−∞

where Λ(t) is a triangular signal, defined in eqn.(2) and shown in Figure 1 below

t+1 −1 ≤ t ≤ 0





Λ(t) = −t + 1 0≤t≤1 (2)



0
 otherwise

Figure 1: The triangular signal Λ(t)

The signal x(t) is formed by repeated copies of the Triangular Signal as shown in Figure 2.

Figure 2: Signal x(t) of Problem 2.39-4

Solution:
Note that the signal x(t) is a periodic signal with period T0 = 2.
First, we derive the complex Fourier series and then simplify it to the Trigonometric Fourier series.

3
We start by finding the coefficients xn , n 6= 0, as follows:
Z α+T0
1 −j2π Tn t
xn = x(t)e 0 dt
T0 α
1 1
Z
n
= Λ(t)e−j2π 2 t dt
2 −1
1 0 1 1
Z Z
−jπnt
= (t + 1)e dt + (−t + 1)e−jπnt dt
2 −1 2 0
1 h 0 −jπnt
Z Z 0 Z 1 Z 1 i
−jπnt −jπnt −jπnt
= te dt + e dt − te dt + e dt
2 −1 −1 0 0

Now,
Z b Z b
−jπnt j −jπnt b 1 b j −jπnt b
te dt = te + 2 2 e−jπnt and e−jπnt dt = e .
a πn a π n a a πn a

Hence
1 h j −jπnt 0 1 0 j −jπnt 0 j −jπnt 1 1 1 j −jπnt 1 i
xn = te + 2 2 e−jπnt + e − te − 2 2 e−jπnt + e
2 πn −1 π n −1 πn −1 πn 0 π n 0 πn 0
1 h jπn 1 1 jπn 1 1 i
= je + − e + j − jejπn − je−jπn − e−jπn + + je−jπn − j
2πn πn πn πn πn
1 h 2 1 jπn i
= − (e + e−jπn )
2πn πn πn
1 1 ejπn + e−jπn
= 2 2 − 2 2( )
π n π n 2
ejθ +e−jθ
By Euler’s Formula, cos(θ) = 2
. Therefore,
1 1
xn = − cos(πn)
π 2 n2 π 2 n2
1
= (1 − cos(πn)), n 6= 0.
π 2 n2
Finally, for n = 0, x0 can be calculated as
Z α+T0
1
x0 = x(t)e0 dt
T0 α
1 1 1 0 1 1
Z Z Z
= Λ(t)dt = (t + 1)dt + (−t + 1)dt
2 −1 2 −1 2 0
1  t2 0 0 t2 1 1

= | + t|−1 − |0 + t|0
2 2 −1 2
1  1 1  1
= − +1− +1 = .
2 2 2 2
Therefore, the Fourier series expansion of x(t) is
−1 ∞
j2π n n
X X
t
x(t) = x0 + xn e 2 + xn ej2π 2 t
n=−∞ n=1

4
 ∞
X ∞
X
−jπnt
= x0 + x−n e + xn ejπnt [note the change of indices in the first summation].
n=1 n=1

Since x(t) is real x−n = x∗n . Moreover, since x(t) is an even function xn = 1
π 2 n2
(1 − cos(πn)) is real,
x−n = x∗n = xn . Hence,
∞
X
xn e−jπnt + ejπnt


x(t) = x0 +
n=1
∞
1 X 1 −jπnt jπnt


= + (1 − cos(πn)) e + e (3)
2 n=1 π 2 n2

Furthermore, by observing that cos(nπ) = −1, for n odd, and cos(nπ) = 1, for n, even we have

 2

n odd ,
2 2
xn = π n
0
 n even .

Hence, the simplified Fourier series is

1 X 2
e−jπnt + ejπnt .


x(t) = + 2 2
2 n=1,3,5,... π n

Transformation to the Trigonometric Fourier series:

a0 1
Since x0 = 2
, the coefficient a0 = 2x0 which, in this case is a0 = 2 × 2
= 1.
an −jbn an
Next, since bn = 0 ( x(t) is an even function), xn = 2
= 2
. Thus,

2
an = (1 − cos(πn)).
π 2 n2
The Trigonometric Fourier series is thus given by
∞ ∞
a0 X n X n
x(t) = + an cos(2π t) + bn sin(2π t)
2 n=1
2 n=1
2
∞ ∞
1 X 2 X
= + (1 − cos(πn)) cos(πnt) + (0)sin(πnt)
2 n=1 π 2 n2 n=1
∞
1 X 2
= + (1 − cos(πn))cos(πnt).
2 n=1 π 2 n2

In fact, this result should be clear from eqn.(3) where using e−jπnt + ejπnt = 2 cos(πnt), would

immediately give us the Trigonometric series.

5
Problem 2: Textbook Example 2.2.2
This example can be found in page 46-48 of the textbook.

Determine the Fourier series expansion for the signal x(t) shown in Figure 3 and described by:
∞
X
x(t) = (−1)n Π(t − n)
n=−∞

where Π(t) is a rectangular signal of height 1 and duration 1 centered around t = 0.

Figure 3: Signal x(t) of Example 2.2.2

Solution:
In this case, T0 = 2. We choose α = − 21 to find the Fourier series coefficients xn :
Z α+T0
1 −j2π Tn t
xn = x(t)e 0 dt
T0 α
Z 3
1 2
= x(t)e−jπnt dt
2 − 12
Z 1 Z 3
1 2 −jπnt 1 2
= 1e dt + (−1)e−jπnt dt
2 −21 2 2 1

1 3
1 2 1 2
= e−jπnt 1 − e−jπnt 1
−j2πn −2 −j2πn 2
1 h −jπn jπn
i 1 h −j3πn −jπn
i
=− e 2 −e 2 + e 2 −e 2
j2πn j2πn
1 h −jπn jπn
i e−jπn h −jπn jπn
i
= −e 2 +e 2 + e 2 −e 2
j2πn j2πn
jπn −jπn jπn −jπn
1 h e 2 − e 2 i e−jπn h e 2 − e 2 i
= −
πn 2j πn 2j
−jπn
1 πn e πn
= sin( ) − sin( )
πn 2 πn 2
6
 1 πn
= (1 − e−jπn ) sin( ).
πn 2

Since, e−jπn = cos(πn) − j sin(πn) = cos(πn), xn simplifies to

1 πn
xn = (1 − cos(πn)) sin( ).
πn 2
Now, xn = 0 when n is even due to cos(πn) = 1.
Also, among odd n, if n = 4k + 1, we have sin( πn
2
) = sin(2πkn + π/2) = 1, and if n = 4k + 3, we
have sin( πn
2
) = sin(2πkn + 3π/2) = −1.
Therefore, xn simplifies to

2
, n = 4k + 1


 πn


xn = − πn2
, n = 4k + 3



0,
 n even

Finally, x0 can be found out as

Z 3
1 2
x0 = x(t)e0 dt
2 − 12
Z 1 Z 3
1 2 1 2
= (1)dt + (−1)dt
2 − 12 2 12
Z 1 Z 3
1 2 1 2
= 1dt − 1dt
2 − 12 2 21
1 12 1 32
= t 1− t 1
2 −2 2 2
1h1 1 i 1h3 1i
= − (− ) − −
2 2 2 2 2 2
1 1
= (1) − (1).
2 2
Therefore,

x0 = 0.

The Fourier series expansion is thus given by

∞
X
x(t) = xn ejπnt
n=−∞
2 jπt 2 j3πt 2 j5πt
= (e + e−jπt ) − (e + e−j3πt ) + (e + e−j5πt ) − ...
π 3π 5π

7
 4 4 4
= cos(πt) − cos(3πt) + cos(5πt) − ...
π 3π 5π
It can be succinctly written as
∞
4 X (−1)m
x(t) = cos[(2m + 1)πt].
π m=0 2m + 1

Energy-Type and Power-Type Signals

Definition
For any signal x(t)

• the Energy content of the signal is defined as:

Z ∞ Z T /2
2
Ex = |x(t)| dt = lim |x(t)|2 dt
−∞ T →∞ −T /2

• the Power content of the signal is defined as:

1 T /2
Z
Px = lim |x(t)|2 dt
T →∞ T −T /2

• For real signals |x(t)|2 is replaced by x2 (t).

• A signal x(t) is Energy-type if Ex < ∞. Consequently, Px = 0.

• A signal x(t) is Power-type if 0 < Px < ∞. Consequently, Ex = ∞.

• A signal cannot be both energy- and power-type because Px = 0 for energy-type signals and
Ex = ∞ for power-type signals.

• It is possible for a signal to be neither energy-type nor power-type, i.e., both Ex = ∞ and
Px = ∞.

Problem 3: Problem 2.7 Part-2

Determine whether the signal x(t) given below is energy-type or power-type. In either case, also
find its energy/power content.

x(t) = e−t cos(t). (4)

Solution:
Is x(t) is an Energy-type signal:
We can calculate the Energy content of x(t) as,
Z T /2 Z T /2
−t
Ex = lim 2
(e cos(t)) dt = lim e−2t cos2 (t)dt.
T →∞ −T /2 T →∞ −T /2

8
 R
For this, we need to solve the integral of the form I = eax cos2 (x)dx. We solve this integral first
R R
by repeatedly using the method of integration by parts ( udv = uv − vdu):
Z
I = eax cos2 (x)dx
Z
1
= cos2 (x)d(eax )
a
Z
1 ax 2 1
= e cos (x) − eax (−2 cos(x) sin(x))dx
a a

Since 2 sin(x) cos(x) = sin(2x),

Z
1 1
I = eax cos2 (x) + eax sin(2x)dx
a a
Again applying integration by parts:
Z
1 1 1
I = eax cos2 (x) + . sin(2x)d(eax )
a a a
Z
1 1 ax 1
= eax cos2 (x) + 2
e sin(2x) − 2 2 cos(2x)eax dx
a a a
Z
1 1 ax 2
= eax cos2 (x) + e sin(2x) − 2 eax cos(2x)dx
a a2 a

Since cos(2u) = 2 cos2 (u) − 1, therefore

Z
1 1 ax 2
I = eax cos2 (x) + e sin(2x) − eax (2/cos2 (x) − 1)dx
a a2 a2
Z Z
1 1 ax 4 2
= eax cos2 (x) + e sin(2x) − e cos (x)dx + 2 eax (1)dx
ax 2
a a2 a2 a
1 1 ax 4 2 ax
= eax cos2 (x) + e sin(2x) − I + 3e .
a a2 a2 a
Hence,
4 1h 2 2 i ax
I + 2 I = 2 a cos (x) + sin(2x) + e
a a a
 a2 + 4  1h 2 i ax
2
I = a cos (x) + sin(2x) + e
a2 a2 a
Therefore,
Z
ax 2 1 h 2 2 i ax
I= e cos (x)dx = a cos (x) + sin(2x) + e (5)
4 + a2 a

By using the result of eqn.(5), where a = −2 we get,

h 1 
2 2  −2t i T /2
Ex = lim − 2 cos (t) + sin(2t) + e
T →∞ 4 + (−2)2 −2 −T /2

9
 1 h T
  T
 i
= lim − 2 cos2 + sin(T ) − 1 e−T − − 2 cos2 − + sin(−T ) − 1 eT
T →∞ 8 2 2
1 h T
  T  i
− 2 cos2 + sin(T ) − 1 e−T + 2 cos2 + sin(T ) + 1 eT


= lim (6)
T →∞ 8 2 2
With the limit, the term with e−T will tend to zero.
Moreover, since (2 cos2 ( T2 ) + sin(T ) + 1) > 0, the term with eT will tend to infinity. Therefore,

Ex = ∞ (7)

Hence, the signal x(t) is not an Energy-type signal.

Is x(t) is a Power-type signal:

We can calculate the Power content of x(t) as,
Z T /2
1
Px = lim e−2t cos2 (t)dt
T →∞ T −T /2
−T h
1 e 2 T
i 1 eT h 2 T

i
= lim (−2 cos + sin(T ) − 1) + lim (2 cos + sin(T ) + 1)
8 T →∞ T 2 8 T →∞ T 2

eT
The first term with T
tends to zero. Therefore,

1 eT h T
i
Px = lim 2 cos2 + sin(T ) + 1 .
8 T →∞ T 2

Since, 2 cos2 (u) = 1 + cos(2u)

1 eT
Px = lim [1 + cos(T ) + sin(T ) + 1]
8 T →∞ T
1 eT
= lim [2 + cos(T ) + sin(T )]
8 T →∞ T

Now, cos(T ) + sin(T ) = cos(T ) + cos( π2 − T ).

By the Sum-to-Product formula: cos(α) + cos(β) = 2 cos( α+β 2
) cos( α−β
2
), we have
π π
√ π
√
cos(T ) + sin(T ) = 2 cos( 4 ) cos(T − 4 ) = 2 cos(T − 4 ) ≥ − 2
Hence
1 eT
Px = lim [2 + cos(T ) + sin(T )]
8 T →∞ T
1 eT √
≥ lim (2 − 2)
8 T →∞
√ T
2− 2 eT
= lim
8 T →∞ T

10
 T2 T3
From the Maclaurin series for the exponential term eT , i.e., eT = 1 + T + 2!
+ 3!
+ ...
√
2− 2 1 T2 
Px ≥ lim 1+T + + ... = ∞
8 T →∞ T 2
Therefore:

Px = ∞. (8)

Hence, the signal x(t)is not a Power-type signal either.

Equations (7) and (8) show that the x(t) is neither an energy-type nor a power-type signal.
Formulae to remember
Euler’s Formulae

• ejθ = cos(θ) + j sin(θ)

ejθ +e−jθ
• cos(θ) = 2

ejθ −e−jθ
• sin(θ) = 2j

11