EE 220 : Signals and Systems

Department of Electronics and Electrical Engineerng

Indian Institute of Technology, Guwahati
Mid semester examination 2022

Instructions: Solve all questions. [Total: 30 marks]

Q1. Find out whether following signals are even/odd signals [1 mark]

a.
x(t) = t2

b.
x(t) = 2n + 1

Answer: (a) Even. (b) Neither even nor odd.

Q2. Is x(t) = e|t| absolutely integrable? [1 mark]

Answer: No

Q3. What is the fundamental period of x[n] = sin[3n]? Explain your answer? [1 mark]
Answer: No. A discrete-time sinusoidal sequence is periodic if its 2π/Ω is a rational num-
ber. Otherwise, it is non-periodic.

Q4. The input x(t) - output y(t) relationships are described below for two systems. De-
termine if these are stable, causal, and time invariant systems?
R t/2
a. y(t) = −∞
x(τ )dτ
d
b. y(t) = dt
x(t) [2 marks]

Answer: a. No, No, No. Unstable: Check when x(t) = 1 for all t. Not causal: check for -ve
time instants.
b. No, Yes, yes. To check causality: use the definition

′ x(t) − x(t − h)
x (t) = lim
h→0 h
This definition is sufficient to compute the derivative. Hence, it’s a causal system. To check
stability : Try x(t) = u(t). It will result in bounded input, unbounded output.

Q5: Find the convolution between following sequences. [2 marks]

a. [2 + j, 0, 1] and [2 − j, 0, −1]

b. [1, 2, 3] and [4, 5, −1, −2]

1
 Answer: [5,0,-2j,0,-1] and [4,13,21,11,-7,-6]

Q6: Find the solution for the system described by the following equation [3 marks]
y ′′ (t) + 5y ′ (t) + 6y(t) = x(t)
given the input is x(t) = e−t u(t); and initial conditions are y(0) = −1/2 and y ′ (0) = 1/2.
Answer: Homogeneous solution:
Characteristic equation is r2 + 5r + 6 = 0. Roots = -2, -3.
Thus homogeneous solution : yh (t) = c1 e−2t + c2 e−3t

Particular solution:
Let’s assume the particular solution is of the form ce−t for t ≥ 0. Substituting this in the
given differential equation will result in:
yp (t) = (1/2)e−t for t ≥ 0
So the complete solution becomes:
y(t) = yh (t) + yp (t) = c1 e−2t + c2 e−3t + (1/2)e−t for t ≥ 0
It’s derivative: y ′ (t) = −2c1 e−2t − 3c2 e−3t − (1/2)e−t
Susbstituting initial conditions will result in: c1 = −2; c2 = 1
Hence, the complete solution:
y(t) = yh (t) + yp (t) = −2e−2t + e−3t + (1/2)e−t for t ≥ 0

Q7: Write the continuous time Fourier series coefficients for the signal [3 marks]
∞
X
x(t) = δ(t − 4m)
m=−∞

Solution : The fundamental period is T = 4. Hence, fundamental frequency ω0 = π/2.

Note that x(t) is even symmetric, so it is easy to integrate over a period that is symmetric
around the origin (at m=0). Let’s choose the interval of integration from -2 to 2 rather than 0
to 4 for fun. [It’s not going to change the answer as mentioned in the note above.] This results
in

1 2
Z
X[k] = δ(t)e−jkπt/2 dt
4 −2
1 1
= e0 = .
4 4
We could have also integrated over 2 to 6 (for m= 1). That would lead to
1 6
Z
X[k] = δ(t − 4)e−jkπt/2 dt
4 2
1 1
= e−jk2π = .
4 4

Q8: What are the convergence conditions for writing continuous time Fourier transform
representation of a function x(t)? Can we write the Fourier transform of u(t)? If yes, explain
the steps. [3 marks]

2
 1. x(t) is absolutely integrable, i.e.
Z ∞
|x(t)|dt ≤ B < ∞
−∞

2. x(t) has a finite number of maxima and minima in any finite interval

3. x(t) has a finite number of finite discontinuities in any finite interval.

Although, u[t] is not absolutely integrable, we have the following trick:

Replace u(t) = 21 [1 + sgn(t)] . which leads to U (ω) = πδ(ω) + jω
1

Q9: Find the Laplace transform of x(t) = e−at cos(ωt)u(t). What is the region of conver-
gence? [3
marks]
s+a
Answer: (s+a)2 +ω 2 ; ROC is σ > −a

Q10: Using the bilateral time-shift and differentiation properties, find the Laplace trans-
form of

d2 −3(t−2)
x(t) = 2 (e u(t − 2))
dt
[3 marks]
s2 −2s
Answer: Done in tutorial 5. s+3
e with ROC σ > −3

Q11: Derive the following for a function x(t) whose Fourier transform is X(ω):
Z ∞ Z ∞
2 1
|x(t)| dt = |X(ω)|2
−∞ 2π −∞

What does it signify? [4 marks]

Derivation: let there be two functions x(t) and y(t). Let’s consider the Fourier transform
of their product:
Z ∞
F.T.[x(t) · y(t)] = x(t)y(t)e−jωt dt
−∞
Z ∞
1 1
≡ X(ω) ∗ Y (ω) = X(ϕ)Y (ω − ϕ)dϕ
2π 2π −∞
here, we have used the property that multiplication in time domain is equivalent to convolution
in frequency domain (with a factor). Both the integrals expressed above are equivalent in all
cases. This is called Generalized Parseval’s expression. The specific expression can be written
by assuming a specific case when ω = 0 and y(t) = x∗ (t), and using x∗ (t) ←
→ X ∗ (−ϕ). Equating
both the integral expressions above gives us:
Z ∞ Z ∞
∗ 1
x(t)x (t)dt = X(ϕ)Y (−ϕ)dϕ
−∞ 2π −∞
Z ∞
1
= X(ϕ)X ∗ (ϕ)dϕ
2π −∞
Z ∞ Z ∞
2 1
≡ |x(t)| dt = |X(ω)|2 dω
−∞ 2π −∞

3
 Q12: Prove the following identity [4 marks]
Z ∞
1
δ(t) = ejωt dω
2π −∞

Solution: CTFT properties can be used to define impulse function in a new way. Let us
start with the definitions first
Z ∞
1
x(t) = X(ω)ejωt dω, (1)
2π −∞
where Z ∞
X(ω) = x(τ )e−jωτ dτ, (2)
−∞

Replace X(ω) in Eq. 1:

Z ∞ Z ∞
1
x(t) = x(τ )e−jωτ dτ ejωt dω
2π −∞ −∞

Rearranging the integrals, we get:

Z ∞ hZ ∞
1 i
x(t) = x(τ ) ejω(t−τ ) dω dτ
2π −∞ −∞
Z ∞ (3)
1
= x(τ )g(t − τ )dτ
2π −∞
where Z ∞
g(t) = ejωt dω
−∞

Now once again Eq. (3) can be observed to be equivalent to:

1
x(t) = x(t) ∗ g(t)
2π
We know from the impulse function property that this expression can be true only if
1
g(t) = δ(t)
2π
Thus, we have a new definition of the impulse function δ(t) :
Z ∞
1
δ(t) = ejωt dω
2π −∞

But since we know that δ(t) is an even function, and hence,

Z ∞
1
δ(t) = δ(−t) = e−jωt dω
2π −∞

PS. Note that different texts may have different signs of the exponential within the integral
in the above definition, which is alright. It is also intuitive to observe that both integrals
reprsent area and it is going to be same in both cases.