Scribd
Upload
9 views3 pages

Fourier Transform Calculations and LTI Systems

Uploaded by

Samrudh
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
9 views3 pages

Fourier Transform Calculations and LTI Systems

Uploaded by

Samrudh
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

ELL 205: Signals And Systems

Tutorial Sheet-4

1. Calculate the Fourier transforms of:


a) (e−2(t−1) u(t − 1))
P∞+ 1)k+ δ(t − 1)
b) δ(t
c) Pk=0 α δ(t − kT ), |α| < 1
∞ −|t−2n|
d) n=−∞ e
e) x(t) as shown in Figure 1(a)

Figure 1: (a)

h i
2. Find F −1 1
(9+λ2 )(4+λ2 )

3. Given y(t) = x(t) ∗ h(t) and g(t) = x(3t) ∗ h(3t)

and given that x(t) has Fourier transform X(jω) and h(t) has Fourier transform H(jω), use
Fourier transform properties to show that g(t) has the form

g(t) = Ay(Bt)

Determine the values of A and B.


4. Let x(t) be a signal with Fourier transform X(jω). Suppose we are given the following facts:
a) x(t) is real.
b) x(t)R = 0 for t ≤ 0.
1 ∞ jωt dω = |t|e−|t| .
c) 2π −∞ ℜ{X(jω)}e
Determine a closed-form expression for x(t).
5. Show that the three LTI systems with impulse responses

h1 (t) = u(t), h2 (t) = −2δ(t) + 5e−2t u(t), and h3 (t) = 2te−t u(t),

all have the same response to x(t) = cos(t).


6. If the signal x(t) is given by

sin t sin t
∗
x(t) =
πt πt
with ∗ denoting the convolution operation, then x(t) is equal to

1
7. Consider a causal LTI system with frequency response
1
H(jω) =
jω + 3
For a particular input x(t), this system is observed to produce the output

y(t) = e−3t u(t) − e−4t u(t).

Determine x(t).
8. A causal and stable LTI system S has the frequency response
jω + 4
H(jω) = .
6 − ω 2 + 5jω
(a) Determine a differential equation relating the input x(t) and output y(t) of S.
(b) Determine the impulse response h(t) of S.
9. The input and the output of a stable and causal LTI system are related by the differential
equation
d2 y(t) dy(t)
+6 + 8y(t) = 2x(t)
dt2 dt
a) Find the impulse response of this system.
b) What is the response of this system if

x(t) = te−2t u(t)


10. Prove that the Fourier transform of
x2
f (x) = e− 2

is self-reciprocal.
h i
11. Using Convolution theorem, find F −1 1
12+7iλ−λ2
.
2
12. Find the Fourier transform of e−x . Hence find Fourier transform of
2
e−ax , a > 0
(
1 − x2 , |x| < 1
13. Find Fourier transform of f (x) =
0, |x| > 1
14. Suppose that a signal x(t) has Fourier transform X(jω). Now consider another signal g(t) whose
shape is the same as the shape of X(jω); that is,

g(t) = X(jt).

(a) Show that the Fourier transform G(jω) of g(t) has the same shape as 2πx(−t); that is, show
that
G(jω) = 2πx(−ω).
(b) Using the fact that
F{δ(t + B)} = ejBω
in conjunction with the result from part (a), show that

F{ejBt } = 2πδ(ω − B).


15. Let x(t) and y(t) be two real signals. Then the cross-correlation function of x(t) and y(t) is
defined as Z ∞
ϕxy (t) = x(t + τ )y(τ )dτ.
−∞
Similarly, we can define ϕyx (t), ϕxx (t), and ϕyy (t). [The last two of these are called the auto-
correlation functions of the signals x(t) and y(t), respectively.] Let Φxy (jω), Φyx (jω), Φxx (jω),
and Φyy (jω) denote the Fourier transforms of ϕxy (t), ϕyx (t), ϕxx (t), and ϕyy (t), respectively.

2
(a) What is the relationship between Φxy (jω) and Φyx (jω)?
(b) Find an expression for Φxy (jω) in terms of X(jω) and Y (jω).
(c) Show that Φxx (jω) is real and nonnegative for every ω.
(d) Suppose now that x(t) is the input to an LTI system with a real-valued impulse response and
with frequency response H(jω) and that y(t) is the output. Find expressions for Φxy (jω) and
Φyy (jω) in terms of Φxx (jω) and H(jω).
(e)
Let x(t) be as illustrated in Figure P4.50, and let the LTI system impulse response be

h(t) = e−at u(t), a > 0.

Compute Φxx (jω), Φxy (jω), and Φyy (jω) using the results of parts (a)-(d).

(f ) Suppose that we are given the following Fourier transform of a function ϕ(t):

ω 2 + 100
Φ(jω) = .
ω 2 + 25
Find the impulse responses of two causal, stable LTI systems that have autocorrelation functions
equal to ϕ(t). Which one of these has a causal, stable inverse?
16. Let us consider a system with a real and causal impulse response h(t) that does not have any
singularities at t = 0. We shall derive an explicit relationship between HR (jω) and HI (jω), the
real and imaginary parts of H(jω).
(a) To begin, note that since h(t) is causal, except perhaps at t = 0.

h(t) = h(t)u(t),

Now, since h(t) contains no singularities at t = 0, the Fourier transforms of both sides of above
eq. must be identical. Use this fact, together with the multiplication property, to show that
Z ∞
1 H(jη)
H(jω) = dη. (P4.48–2)
jπ −∞ ω − η

Use eq. (P4.48–2) to determine an expression for HR (jω) in terms of HI (jω) and one for HI (jω)
in terms of HR (jω).
(b) The operation
1 ∞ x(τ )
Z
y(t) = dτ (P4.48–3)
π −∞ t − τ
is called the Hilbert transform. We have just seen that the real and imaginary parts of the
transform of a real, causal impulse response h(t) can be determined from one another using the
Hilbert transform.
Now consider eq. (P4.48–3), and regard y(t) as the output of an LTI system with input x(t).
Show that the frequency response of this system is
(
−j, ω > 0,
H(jω) =
j, ω < 0.

(c) What is the Hilbert transform of the signal x(t) = cos 3t?

Common questions

Powered by AI

A Fourier transform is self-reciprocal if applying the Fourier transform twice returns the original function, up to a possible scaling factor. This property is significant as it shows symmetry in the frequency and time domains. For example, f(x) = e^(-x^2/2) is self-reciprocal because its Fourier transform has the same form: F{f(x)} = e^(-ω^2/2), demonstrating that the shape of the Gaussian function is preserved under Fourier transform.

To find the impulse response of a system described by a differential equation such as d²y/dt² + 6dy/dt + 8y = 2x(t), first find the characteristic equation and solve for its roots to determine the system's natural response. Then, apply the Laplace transform to the differential equation to find Y(s), and use inverse Laplace transform to obtain h(t), the impulse response. This may involve using convolution integrals if the input is not an impulse.

The Hilbert Transform relates the real and imaginary parts of a causal system's frequency response using the principle of causality and the Kramers-Kronig relations. Specifically, for a system with impulse response h(t), H(jω) can be expressed as HR(jω) and HI(jω). The Hilbert transform states that each can be analytically derived from the other, ensuring no information loss, by inducing a 90-degree phase shift within spectral components of the signal.

To determine a differential equation relating x(t) and y(t) for a system with frequency response H(jω) = (jω + 4)/(6 - ω^2 + 5jω), one needs to express H(jω) as a ratio of polynomials in jω. Then use inverse Laplace transforms and partial fraction decomposition to derive the corresponding differential equation. This involves equating coefficients in the polynomial to identify the constants in the differential equation.

The convolution theorem states that the Fourier transform of a convolution of two signals is the product of their Fourier transforms. To find the inverse Fourier transform of a complex expression like 1/(12 + 7iλ - λ^2), decompose the expression using partial fraction decomposition or similar techniques, and apply the inverse Fourier transform to each term separately, expressing the result as a convolution of simpler functions with known inverse transforms.

Applying a low-pass filter to a signal smooths out its auto-correlation function, emphasizing long-term correlations while reducing short-term random fluctuations. This can be analyzed in the frequency domain by considering how the multiplication by the filter's frequency response in the Fourier domain alters the spectral density, effectively reducing high-frequency noise and abrupt changes in the time domain. The resulting smoother auto-correlation reflects enhanced signal coherence.

Cross-correlation measures the similarity between two signals as a function of a time-lag applied to one of them, and for two signals, x(t) and y(τ), it is given by ϕxy(t) = ∫ x(t + τ)y(τ)dτ. Convolution, on the other hand, combines two signals so that one modifies or filters the other, defined as (x ∗ h)(t) = ∫ x(τ)h(t - τ)dτ. While similar mathematically, cross-correlation differs as it doesn't involve time-reversal of h(t) unlike convolution, relevant for understanding system outputs in LTI systems.

When both signals are passed through a linear time-invariant system, their correlation changes due to alterations in magnitude and phase introduced by the system's impulse response, affecting energy and time alignment. This results in new correlation functions, with Φxy(jω) and Φyy(jω) expressed terms of the system's frequency response H(jω) and the input's auto-correlation Φxx(jω). Practically, this implies signal distortion based on system characteristics, impacting applications like filtering and detection.

To prove that F{e^(jBt)x(t)} results in modulation in the frequency domain, use the time-shifting property of the Fourier transform. The Fourier transform of x(t)e^(jBt) represents a frequency shift, resulting in multiplying the transform of x(t) by e^(jBω), showing that F{e^(jBt)x(t)} = X(jω - B), thus demonstrating modulation. This illustrates the duality between shifts in time and modulation in frequency.

The cross-correlation function of a real signal can be negative, reflecting dissimilarity between signals at certain time shifts, whereas the auto-correlation function measures similarity of a signal with itself, which is why it is always non-negative and real, capturing energy distribution over time. Auto-correlation is symmetric and peaks at zero shift, indicating maximum similarity.

You might also like