ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM
TECHNIQUES
1 "ANALYSIS OF NONLINEAR EFFECTS IN FOURIER TRANSFORM TECHNIQUES"
Problem 1. Consider a nonlinear signal x(t) = sin(2πt) + sin(4πt). Compute the Fourier
transform X(f)of x(t).
Solution 1. To compute the Fourier transform of x(t), we use the definition of the Fourier
transform:
X(f) = Z∞
−∞
x(t)e−j2πftdt
a) For the first frequency component sin(2πt):
X1(f) = Z∞
−∞
sin(2πt)e−j2πftdt
We can use the property of Fourier transform of sin(at):
F{sin(at)}=j
2[δ(f−a)−δ(f+a)]
Hence, for sin(2πt), we have:
X1(f) = j
2[δ(f−2) −δ(f+ 2)]
b) For the second frequency component sin(4πt): Similarly, using the property of Fourier trans-
form of sin(at):
X2(f) = j
2[δ(f−4) −δ(f+ 4)]
c) Therefore, the Fourier transform of the nonlinear signal x(t) = sin(2πt) + sin(4πt)is:
X(f) = X1(f) + X2(f) = j
2[δ(f−2) −δ(f+ 2) + δ(f−4) −δ(f+ 4)]
This completes the computation of the Fourier transform of x(t)containing nonlinear frequency
components.
2 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider the periodic function f(x) = sin(2πx) + cos(4πx).
a) Calculate the Fourier series coefficients anand bnfor the function f(x).
b) Determine the value of the constant term in the Fourier series of f(x).
c) Find the representation of f(x)in its Fourier series form.
Solution 1.
a) We can calculate the Fourier series coefficients anand bnusing the following formulas:
an=1
LZL
−L
f(x) cos 2πnx
Ldx
bn=1
LZL
−L
f(x) sin 2πnx
Ldx
where Lis the period of the function.
For f(x) = sin(2πx) + cos(4πx), the period is L= 1. Thus, we have:
an=Z1
−1
(sin(2πx) + cos(4πx)) cos(2πnx)dx
bn=Z1
−1
(sin(2πx) + cos(4πx)) sin(2πnx)dx
Calculating these integrals will give us the Fourier series coefficients anand bn.
b) To find the constant term in the Fourier series of f(x), we need to determine a0, which is
given by:
a0=1
LZL
−L
f(x)dx
Substitute L= 1 and integral bounds −1to 1to find the value of a0.
c) The representation of f(x)in its Fourier series form is given by:
f(x)≈a0
2+
∞
X
n=1 ancos 2πnx
L+bnsin 2πnx
L
Substitute the calculated values of a0,an, and bninto this equation to obtain the Fourier series
representation of f(x).
3 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider the function f(t) = sin(3t) + cos(5t).
a) Find the Fourier transform of f(t).
b) Determine the frequency components present in f(t).
Solution 1.
a) The Fourier transform of a function f(t)is defined as:
F(ω) = Z∞
−∞
f(t)e−iωtdt
Applying this definition to our function f(t) = sin(3t) + cos(5t), we have:
F(ω) = Z∞
−∞
(sin(3t) + cos(5t))e−iωtdt
=Z∞
−∞
sin(3t)e−iωtdt +Z∞
−∞
cos(5t)e−iωtdt
We can solve these integrals using the properties of the Fourier transform.
b) The frequency components present in f(t)are the frequencies of the sine and cosine func-
tions, which are 3and 5respectively.
Therefore, the frequency components present in f(t)are 3and 5.
4 "ANALYZING THE CONVERGENCE AND DIVERGENCE OF FOURIER SERIES REPRE-
SENTATIONS"
Problem 4. Consider the function f(x) = x2defined on the interval [−π, π].
a) Find the Fourier cosine series representation of f(x).
b) Determine the convergence of the Fourier cosine series of f(x).
c) Calculate the value at which the Fourier cosine series of f(x)converges to f(x).
Solution 4.
a) To find the Fourier cosine series representation of f(x) = x2, we need to calculate the
coefficients anusing the formula:
an=2
πZπ
0
x2cos(nx)dx.
Calculating the integral gives:
an=2
πZπ
0
x2cos(nx)dx
=2
π2(nx sin(nx)+(n2x2−2) cos(nx))
n2
π
0
=2
π2nπ sin(nπ)
n2+n2π2−2
n2
= 2 (−1)n
n+π2−2
n2.
Therefore, the Fourier cosine series representation of f(x)is:
f(x) = π2−2
2+ 2
∞
X
n=1 (−1)n
n+π2−2
n2cos(nx).
b) To determine the convergence of the Fourier cosine series of f(x), we need to check if f(x)is
piecewise smooth on [−π, π]. Since f(x) = x2is continuous and has derivative f′(x) = 2xexisting
for all x∈[−π, π], the Fourier cosine series converges to f(x).
c) The Fourier cosine series converges to f(x)for x∈(−π, π), so the value at which the Fourier
cosine series of f(x)converges to f(x)is x= 0.
5 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider a square wave signal x(t)given by:
x(t) = (1if 0≤t < T/2
−1if T/2≤t<T
where Tis the period of the signal.
a) Calculate the Fourier transform X(ω)of x(t).
b) Determine the value of X(0).
c) Find the power spectral density SX(ω)of the signal x(t).
Solution 1.
a) To calculate the Fourier transform X(ω)of the given square wave signal x(t), we apply the
definition of Fourier transform:
X(ω) = Z∞
−∞
x(t)e−jωtdt
Splitting the integral based on x(t):
X(ω) = ZT/2
0
e−jωtdt −ZT
T/2
e−jωtdt
Solving the integrals, we get:
X(ω) = 1−e−jωT /2
jω −e−jωT /2−e−jωT
jω
Simplifying further gives the Fourier transform as:
X(ω) = 2
jω (1 −cos(ωT /2))
b) To find the value of X(0), we substitute ω= 0 in the Fourier transform:
X(0) = 2
0(1 −cos(0)) = 0
c) The power spectral density SX(ω)is given by:
SX(ω) = |X(ω)|2
Substitute the Fourier transform X(ω)obtained earlier to get:
SX(ω) =
2
jω (1 −cos(ωT /2))
2
=4(1 −cos(ωT/2))2
ω2
6 CALCULATING FOURIER TRANSFORMS OF DISTRIBUTIONS
Problem 1. Compute the Fourier transform of the distribution T=δ(x−3) + δ(x+ 4), where δ
denotes the Dirac delta distribution.
Solution 1. The Fourier transform of a distribution Tis defined as
F(δ)(ϕ) = δ(F(ϕ))
where ϕis a test function.
a) Let’s first compute the Fourier transform of δ(x−3) using the formula:
δ(x−a)F
−→ e−2πiaξ
Thus, the Fourier transform of δ(x−3) is
F(δ(x−3)) = e−6πiξ
b) Next, we compute the Fourier transform of δ(x+ 4) which is given by:
δ(x+a)F
−→ e2πiaξ
Therefore, the Fourier transform of δ(x+ 4) is
F(δ(x+ 4)) = e8πiξ
c) Finally, computing the Fourier transform of the given distribution T:
F(T)(ϕ) = δ(F(ϕ))
=δ(e−6πiξ) + δ(e8πiξ )
=ϕ(6) + ϕ(−8)
Therefore, the Fourier transform of Tis ϕ(6) + ϕ(−8).
Problem 2. Determine the Fourier transform of the distribution T=x2·sin(2πx).
Solution 2. The Fourier transform of a distribution Tis given by:
F(xn·ϕ)(ξ)=(−i)n·dnF(ϕ)
dξn
where ϕis a test function.
We need to compute the Fourier transform of x2·sin(2πx).
a) First, we find the Fourier transforms of x2and sin(2πx)individually.
The Fourier transform of x2can be calculated using the formula:
F(x2)(ξ) = (−i)2·d2
dξ2δ
This simplifies to:
F(x2)(ξ) = −δ′′(ξ)
b) Next, we find the Fourier transform of sin(2πx):
F(sin(αx)) = α
α2−ξ2
Applying this to sin(2πx)gives:
F(sin(2πx)) = 2π
(2π)2−ξ2
c) Now, we calculate the Fourier transform of x2·sin(2πx)by convolving the individual trans-
forms:
F(x2·sin(2πx))(ξ) = −δ′′(ξ)∗2π
(2π)2−ξ2
=−2π
(2π)2−ξ2
Therefore, the Fourier transform of x2·sin(2πx)is −2π
(2π)2−ξ2.
7 "ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES"
Problem 1. Consider the function f(x) = cos(2πx) + 3 sin(4πx).
a) Find the period of f(x).
b) Compute the Fourier series of f(x).
Solution 1.
a) To find the period of f(x), we need to consider the period of each component function
cos(2πx)and sin(4πx)separately.
The period of cos(2πx)is given by T=2π
2π= 1.
Similarly, the period of sin(4πx)is T=2π
4π=1
2.
Since the period of f(x)is the least common multiple of the periods of its component functions,
the period of f(x)is T=lcm(1,1
2)=1.
Thus, the period of f(x)is 1.
b) The Fourier series of f(x)can be found by expressing it in the form:
f(x) = a0+
∞
X
n=1 ancos 2nπx
T+bnsin 2nπx
T
where Tis the period of f(x).
Since we found that T= 1 in part (a), we have:
f(x) = a0+
∞
X
n=1
[ancos(2nπx) + bnsin(2nπx)]
The coefficients can be calculated using the formulas:
a0=1
TZT
2
−T
2
f(x)dx
an=2
TZT
2
−T
2
f(x) cos 2nπx
Tdx
bn=2
TZT
2
−T
2
f(x) sin 2nπx
Tdx
Since T= 1, these formulas simplify to:
a0=Z0.5
−0.5
f(x)dx
an= 2 Z0.5
−0.5
f(x) cos(2nπx)dx
bn= 2 Z0.5
−0.5
f(x) sin(2nπx)dx
Plugging in the function f(x)and computing the integrals will give the coefficients for the Fourier
series of f(x).
8 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider a continuous-time signal x(t)in a given bandwidth, represented by its
Fourier transform X(f). The signal is sampled with a sampling rate of Fs= 200 Hz. Determine if
this sampling rate satisfies the Nyquist criterion for this signal. Justify your answer.
Solution 1.
Given that the signal is sampled with a sampling rate of Fs= 200 Hz, we need to check if this
sampling rate satisfies the Nyquist criterion.
The Nyquist theorem states that in order to perfectly reconstruct a signal from its samples, the
sampling rate must be at least twice the maximum frequency present in the signal. In other words,
Fs≥2B, where Bis the bandwidth of the signal.
In this case, since Fs= 200 Hz, we need to determine the bandwidth of the signal B. If X(f)is
non-zero only up to BHz, then the condition is satisfied.
If the signal’s Fourier transform X(f)has energy up to BHz, then the condition Fs≥2Bis
satisfied and the sampling rate of 200 Hz meets the Nyquist criterion for this signal.
Therefore, the sampling rate of 200 Hz satisfies the Nyquist criterion for the given continuous-
time signal.
This problem highlighted the importance of the Nyquist criterion in determining the appropriate
sampling rate for a signal in order to avoid aliasing.
9 "ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES"
Problem 1. Consider a non-stationary signal x(t) = sin(2πt2).
a) Find the Fourier transform X(ω)of the signal.
b) Determine the time-frequency representation of the signal using the Short-Time Fourier
Transform (STFT) with a Gaussian window function g(t) = e−t2.
Solution 1.
a) To find the Fourier transform X(ω)of the signal x(t) = sin(2πt2), we use the definition of the
Fourier transform:
X(ω) = F{x(t)}=Z∞
−∞
x(t)e−jωtdt
Substitute x(t) = sin(2πt2)into the integral:
X(ω) = Z∞
−∞
sin(2πt2)e−jωtdt
This integral does not have a closed-form solution, but can be numerically evaluated.
b) The Short-Time Fourier Transform (STFT) of a signal x(t)with a window function g(t)is given
by:
X(ω, τ) = Z∞
−∞
x(t)g(t−τ)e−jωtdt
For the given signal x(t) = sin(2πt2)and Gaussian window function g(t) = e−t2, we can com-
pute the STFT X(ω, τ)by sliding the window along the time axis and computing the Fourier trans-
form for each windowed segment.
The time-frequency representation of the signal using STFT provides information about how
the frequency content of the signal evolves over time, which is particularly useful for analyzing
non-stationary signals like x(t) = sin(2πt2).
10 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider a function f(x) = (xif 0≤x≤π
0otherwise
a) Compute the Fourier sine series of f(x).
b) Find the sum of the series.
Solution 1.
a) The Fourier sine series for a function f(x)defined on 0≤x≤Lis given by:
f(x)∼
∞
X
n=1
Bnsin nπx
L
where Bn=2
LRL
0f(x) sin nπx
Ldx.
For the function f(x)given in the problem, L=πand f(x) = xfor 0≤x≤π. Thus, we need
to compute Bn:
Bn=2
πZπ
0
xsin(nx)dx
=2
π−x
ncos(nx)
π
0
+1
nZπ
0
cos(nx)dx
=2
π0−(−1)nπ
n+1
n2sin(nx)
π
0
= 0 + 2(−1)n
n2+2(−1)n
πn
=2(−1)n
n21 + 1
πn
Therefore, the Fourier sine series of f(x)is:
f(x)∼
∞
X
n=1
2(−1)n
n21 + 1
πnsin(nx)
b) To find the sum of the series, we note that the Fourier sine series of an odd function converges
to f(x−)+f(x+)
2.
For the given function f(x) = x,f(x−) = xand f(x+)=0(at the jump discontinuity x= 0).
Thus, the sum of the series is:
x+ 0
2=x
2
Therefore, the sum of the Fourier sine series of f(x)is x
2.
11 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider a function f(t) = cos(2πt)in the time domain. Compute the Fourier
transform F(ω)of f(t), where F(ω)is defined as:
F(ω) = Z∞
−∞
f(t)e−iωtdt
Solution 1. To compute the Fourier transform F(ω)of f(t) = cos(2πt), we substitute f(t)into
the definition of the Fourier transform:
F(ω) = Z∞
−∞
cos(2πt)e−iωtdt
Using Euler’s formula eix = cos(x) + isin(x), we can express the integrand as:
cos(2πt)e−iωt =1
2(e2πit +e−2πit)1
2i(eiωt −e−iωt)
Multiplying and simplifying, the integrand becomes:
1
4i(e(2πi−ωi)t−e(−2πi+ωi)t+e(2πi+ωi)t−e(−2πi−ωi)t)
Now, we can integrate term by term:
F(ω) = 1
4i"e(2πi−ωi)t
2πi −ωi −e(−2πi+ωi)t
−2πi +ωi +e(2πi+ωi)t
2πi +ωi −e(−2πi−ωi)t
−2πi −ωi #
∞
−∞
Given that the oscillatory terms will go to zero at ±∞ due to periodicity, we are left with evalu-
ating the residues at the poles ω=±2π. This gives us the final result:
F(ω) = π(δ(ω−2π) + δ(ω+ 2π))
where δ(ω)is the Dirac delta function.
So, the Fourier transform of f(t) = cos(2πt)is F(ω) = π(δ(ω−2π) + δ(ω+ 2π)).
12 NONLINEAR SIGNAL PROCESSING USING HARMONIC ANALYSIS AND FOURIER TRANS-
FORM
Problem 1.
Consider a nonlinear signal x(t) = cos(2πt) + cos(4πt). Determine the Fourier series represen-
tation of x(t)over the interval [−T, T ], where Tis the period of the signal.
Solution 1.
Given signal x(t) = cos(2πt) + cos(4πt), we first need to find the period of the signal by calcu-
lating the least common multiple of the periods of the individual cosine components.
The period of cos(2πt)is T1= 1 and the period of cos(4πt)is T2=1
2.
The period of the combined signal is T=LCM(T1, T2) = LCM(1,1
2) = 1.
Therefore, the Fourier series representation of x(t)over the interval [−1,1] is:
x(t) = a0+
∞
X
n=1
(ancos(nπt) + bnsin(nπt))
where an=2
TRT
−Tx(t) cos(nπt)dt and bn=2
TRT
−Tx(t) sin(nπt)dt.
Calculating the coefficients:
a0=1
TZ1
−1
x(t)dt =1
1Z1
−1
(cos(2πt) + cos(4πt))dt
an=2
TZ1
−1
x(t) cos(nπt)dt =2
1Z1
−1
(cos(2πt) + cos(4πt)) cos(nπt)dt
bn=2
TZ1
−1
x(t) sin(nπt)dt =2
1Z1
−1
(cos(2πt) + cos(4πt)) sin(nπt)dt
By performing the integrations, we can find the Fourier series representation of x(t)over the
interval [−1,1].
13 "WAVELET ANALYSIS AND APPLICATIONS IN SIGNAL PROCESSING"
Problem 13. Consider a signal f(t) = sin(4t) + sin(8t). Perform the following wavelet analysis
using the Morlet wavelet:
ψ(t) = eiω0te−t2/2
where ω0= 5.
a) Compute the continuous wavelet transform (CWT) of f(t).
b) Sketch the scalogram of the CWT.
c) Determine the dominant frequency component(s) present in the signal.
Solution 13.
a) To compute the CWT of f(t), we use the formula:
Wf(a, b) = Z∞
−∞
f(t)·¯
ψt−b
adt,
where ais the dilation parameter and bis the translation parameter.
For this problem, f(t) = sin(4t) + sin(8t)and ψ(t) = ei·5te−t2/2.
The CWT is given by:
Wf(a, b) = Z∞
−∞
[sin(4t) + sin(8t)] ·e−i·5(t−b)/ae−(t−b)2/(2a2)dt.
This integral can be evaluated by numerical methods or special software such as MATLAB to
find the values of Wf(a, b).
b) The scalogram is a visual representation of the CWT. It shows the magnitude of the CWT
coefficients at each scale aand time b. To sketch the scalogram, we plot the magnitude of |Wf(a, b)|
at different aand bvalues.
c) By examining the scalogram, we can identify the dominant frequency components by look-
ing at the peaks in the magnitude of the CWT coefficients. In this case, the dominant frequency
components correspond to the frequencies that have the highest magnitude values in the CWT.
14 "APPLICATIONS OF WAVELET TRANSFORM IN SIGNAL PROCESSING"
Problem 14. Consider a signal f(t)with the following data:
f(t) = (1,if 0≤t≤1
0,otherwise
Let’s analyze this signal using the Haar wavelet transform.
a) Calculate the Haar wavelet coefficients ck,m corresponding to f(t).
b) Determine the wavelet transform of f(t).
c) Reconstruct the original signal using the inverse Haar wavelet transform from the coefficients
found in part (a).
Solution 14.
a) The Haar wavelet coefficients ck,m can be calculated as:
ck,m =Z∞
−∞
f(t)ψk,m(t)dt
Here, ψk,m(t)is the Haar wavelet function. Since the Haar wavelet is a piecewise constant
function, we have:
ψk,m(t) = (2k
2,if 2km≤t < 2k(m+ 1)
0,otherwise
Given f(t), we can calculate the coefficients by integrating over the intervals where ψk,m(t)is
non-zero. For the Haar wavelet, we only need to consider the intervals [0,1) and [1,2).
For the interval [0,1):
c0,0=Z1
0
f(t)ψ0,0(t)dt
c0,0=Z1
0
(1) ·(20
2)dt
c0,0=Z1
0
dt = 1
For the interval [1,2):
c1,0=Z2
1
f(t)ψ1,0(t)dt
c1,0=Z2
1
(0) ·(21
2)dt
c1,0= 0
Therefore, the Haar wavelet coefficients for f(t)are c0,0= 1 and c1,0= 0.
b) The wavelet transform of f(t)is given by:
Wf(a, b) = X
k,m
ck,m ·ψk,m(t)
Since c1,0= 0, the wavelet transform simplifies to:
Wf(a, b) = c0,0·ψ0,0(t)=1·20
2= 1
c) Reconstructing the original signal using the inverse Haar wavelet transform involves summing
the scaled and shifted Haar wavelets:
f(t) = X
k,m
ck,m ·ψk,m(t)
Substitute the coefficients c0,0= 1 and c1,0= 0 into the formula:
f(t) = 1 ·ψ0,0(t) = 1 ·20
2= 1
Therefore, the reconstructed signal using the inverse Haar wavelet transform is a constant
signal f(t)=1.
15 "APPLICATION OF WAVELET TRANSFORM IN SIGNAL DENOISING AND COMPRES-
SION"
Problem 15. Consider a signal f(t)defined as:
f(t) = 2 sin(4πt) + 5 cos(8πt) + 3 sin(12πt)
a) Using the wavelet transform, determine the frequency components present in the signal
f(t). b) If we apply a wavelet transform and find that the wavelet coefficients corresponding to the
frequency component 4πare very small compared to the coefficients at other frequencies, explain
why this might indicate the presence of noise in the signal. c) How can wavelet transforms be used
in denoising this signal?
Solution 15.
a) To determine the frequency components present in the signal f(t), we need to apply the
wavelet transform. Let’s consider the Haar wavelet transform for simplicity. The Haar wavelet
transform of a signal f(t)is given by:
W(a, b) = Z∞
−∞
f(t)ψa,b(t)dt
Where ψa,b(t)is the Haar wavelet defined as:
ψa,b(t) =
1if a≤t<a+1
2
−1if a+1
2≤t<a+ 1
0otherwise
Let’s calculate the wavelet coefficients for the given signal f(t)at different scales a:
W(a, b) = Z∞
−∞
f(t)ψa,b(t)dt
For a= 1, we have:
W(1, b) = Z∞
−∞
(2 sin(4πt) + 5 cos(8πt) + 3 sin(12πt))ψ1,b(t)dt
Similarly, we can calculate W(2, b),W(3, b), and so on to determine the frequency components
present in the signal f(t).
b) If the wavelet coefficients corresponding to the frequency component 4πare very small com-
pared to the coefficients at other frequencies, it might indicate the presence of noise in the signal.
This is because in the absence of noise, the coefficients for the true signal frequencies would be
much larger compared to any noise present in the signal.
c) Wavelet transforms can be used in denoising the signal by thresholding the wavelet coeffi-
cients. By setting small coefficients below a certain threshold to zero, we can eliminate noise while
retaining the important frequency components of the signal. This is a common technique in signal
denoising using wavelet transforms.
16 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 17. Consider a non-stationary signal given by f(t) = sin(2πt2).
a) Determine the Fourier transform of the signal f(t).
b) Compute the time-frequency representation of f(t)using the Short-Time Fourier Transform
(STFT) with a Gaussian window function g(t)given by g(t) = e−t2.
Solution 17.
a) The Fourier transform of a signal f(t)is defined as
F(ω) = Z∞
−∞
f(t)e−iωtdt.
For the given signal f(t) = sin(2πt2), we have
F(ω) = Z∞
−∞
sin(2πt2)e−iωtdt.
This particular integral does not have a closed-form solution, so we would need to evaluate it using
numerical methods or approximation techniques.
b) The Short-Time Fourier Transform (STFT) of a signal f(t)using a window function g(t)is
defined as
ST F T (t, ω) = Z∞
−∞
f(τ)g(τ−t)e−iωτ dτ,
where the overline denotes complex conjugation. For a Gaussian window function g(t) = e−t2, the
STFT of the signal f(t) = sin(2πt2)can be computed as
ST F T (t, ω) = Z∞
−∞
sin(2πτ2)e−(τ−t)2e−iωτ dτ.
We can evaluate this integral numerically to compute the time-frequency representation of the
signal using the given Gaussian window function.
17 "CONVERGENCE AND DIVERGENCE OF FOURIER SERIES FOR PIECEWISE FUNC-
TIONS"
Problem 17. Consider the function f(x)defined as:
f(x) = (x0≤x<π
−x−π < x < 0
a) Find the Fourier series of f(x).
b) Determine the conditions for which the Fourier series of f(x)converges.
Solution 17.
a) To find the Fourier series of f(x), we need to compute the coefficients anand bn:
an=1
πZπ
−π
f(x) cos(nx)dx
bn=1
πZπ
−π
f(x) sin(nx)dx
Let’s start with the calculation of an:
an=1
πZ0
−π
(−x) cos(nx)dx +Zπ
0
xcos(nx)dx
Solving these integrals and simplifying yields:
an=2((−1)n−1)
πn2
Now, let’s find bn:
bn=1
πZ0
−π
(−x) sin(nx)dx +Zπ
0
xsin(nx)dx
After solving and simplifying the integrals, we get:
bn= 0
The Fourier series of f(x)is then given by:
f(x)∼2
π1
12−1
32cos(x) + 1
52cos(2x)− · · ·
b) The Fourier series of f(x)converges if the function f(x)satisfies the Dirichlet conditions. In
this case, since f(x)is piecewise continuous and has a finite number of discontinuities within a
period of 2π, the Fourier series of f(x)converges everywhere.
18 "NON-LINEAR FOURIER TRANSFORM TECHNIQUES FOR SIGNAL ANALYSIS"
Problem 1. Consider a signal f(t) = sin(2πt) + 2 cos(4πt).
a) Compute the Fourier Transform of f(t).
b) Find the frequency components present in the signal f(t).
c) Determine the amplitude and phase of each frequency component.
Solution 1.
a) The Fourier Transform of a signal f(t)is defined as:
F(ω) = Z∞
−∞
f(t)e−jωtdt
Applying the definition to the given signal f(t), we have:
F(ω) = Z∞
−∞
[sin(2πt) + 2 cos(4πt)]e−jωtdt
=Z∞
−∞
sin(2πt)e−jωtdt + 2 Z∞
−∞
cos(4πt)e−jωtdt
Solving the integrals, we get:
F(ω) = π(δ(ω−2π)−δ(ω+ 2π)) + π(δ(ω−4π) + δ(ω+ 4π))
b) The frequency components present in the signal f(t)are 2πand 4π.
c) To determine the amplitude and phase of each frequency component, we can use the prop-
erties of the Dirac delta function in the Fourier Transform. The amplitude is the coefficient in front
of the delta function and the phase is given by the sign of the delta function. Thus, we have:
For 2πcomponent: Amplitude = πand Phase = (δ(ω−2π)−δ(ω+ 2π)) = 0
For 4πcomponent: Amplitude = πand Phase = (δ(ω−4π) + δ(ω+ 4π)) = 0
19 "ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES"
Problem 1. Consider a continuous-time signal x(t)defined as:
x(t) = 2 cos(3πt) + 3 sin(6πt)
Find the Fourier transform X(ω)of the signal x(t).
Solution 1.
The Fourier transform of a continuous-time signal x(t)is defined as:
X(ω) = Z∞
−∞
x(t)e−jωtdt
Plugging in the expression for x(t), we have:
X(ω) = Z∞
−∞
(2 cos(3πt) + 3 sin(6πt))e−jωtdt
Expanding the trigonometric functions using Euler’s formula, we get:
X(ω) = Z∞
−∞
2(ej3πt +e−j3πt
2) + 3(ej6πt −e−j6πt
2j)e−jωtdt
X(ω) = Z∞
−∞
(ej(3π−ω)t+e−j(3π+ω)t−3
2j(ej(6π−ω)t−e−j(6π+ω)t))dt
Remembering that the Fourier transform has a property with delta functions, we find:
X(ω) = 2π(δ(ω−3π) + δ(ω+ 3π)) −3π(δ(ω−6π) + δ(ω+ 6π))
Thus, the Fourier transform of the signal x(t)is:
X(ω) = 2π(δ(ω−3π) + δ(ω+ 3π)) −3π(δ(ω−6π) + δ(ω+ 6π))
Certainly! Here is a numerical problem in the realm of Advanced Harmonic Analysis and Fourier
Transform Techniques with a step-by-step solution:
20 "ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES"
Problem 21. Consider a non-stationary signal f(t)defined as:
f(t) = (sin(2πt) 0 ≤t < 1
0otherwise
Calculate the Fourier transform F(ω)of the signal f(t).
Solution 21. To find the Fourier transform F(ω)of the given signal f(t), we use the definition
of the Fourier transform:
F(ω) = Z∞
−∞
f(t)e−iωtdt
Given that f(t)is defined piecewise, we split the integral into two parts:
F(ω) = Z1
0
sin(2πt)e−iωtdt +Z∞
1
0·e−iωtdt
Solving the first integral, we can proceed to calculate the Fourier transform. Let’s solve the
integral as follows:
F(ω) = Z1
0
sin(2πt)e−iωtdt
Applying Euler’s formula to express the sine function in terms of complex exponentials, we have:
F(ω) = Z1
0
ei2πt −e−i2πt
2ie−iωtdt
F(ω) = 1
2iZ1
0
ei(2π−ω)tdt −Z1
0
e−i(2π+ω)tdt
F(ω) = 1
2i"ei(2π−ω)t
i(2π−ω)
1
0−e−i(2π+ω)t
−i(2π+ω)
1
0#
F(ω) = 1
2i"ei(2π−ω)−1
i(2π−ω)−e−i(2π+ω)−1
−i(2π+ω)#
This integral evaluation gives us the Fourier transform F(ω).
a) We can calculate the Fourier series coefficients anand bnusing the following formulas:
an=1
LZL
−L
f(x) cos 2πnx
Ldx
bn=1
LZL
−L
f(x) sin 2πnx
Ldx
where Lis the period of the function.
For f(x) = sin(2πx) + cos(4πx), the period is L= 1. Thus, we have:
an=Z1
−1
(sin(2πx) + cos(4πx)) cos(2πnx)dx
bn=Z1
−1
(sin(2πx) + cos(4πx)) sin(2πnx)dx
Calculating these integrals will give us the Fourier series coefficients anand bn.
b) To find the constant term in the Fourier series of f(x), we need to determine a0, which is
given by:
a0=1
LZL
−L
f(x)dx
Substitute L= 1 and integral bounds −1to 1to find the value of a0.
c) The representation of f(x)in its Fourier series form is given by:
f(x)≈a0
2+
∞
X
n=1 ancos 2πnx
L+bnsin 2πnx
L
Substitute the calculated values of a0,an, and bninto this equation to obtain the Fourier series
representation of f(x).
3 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider the function f(t) = sin(3t) + cos(5t).
a) Find the Fourier transform of f(t).
b) Determine the frequency components present in f(t).
Solution 1.
a) The Fourier transform of a function f(t)is defined as:
F(ω) = Z∞
−∞
f(t)e−iωtdt
Applying this definition to our function f(t) = sin(3t) + cos(5t), we have:
F(ω) = Z∞
−∞
(sin(3t) + cos(5t))e−iωtdt
=Z∞
−∞
sin(3t)e−iωtdt +Z∞
−∞
cos(5t)e−iωtdt
We can solve these integrals using the properties of the Fourier transform.
b) The frequency components present in f(t)are the frequencies of the sine and cosine func-
tions, which are 3and 5respectively.
Therefore, the frequency components present in f(t)are 3and 5.
4 "ANALYZING THE CONVERGENCE AND DIVERGENCE OF FOURIER SERIES REPRE-
SENTATIONS"
Problem 4. Consider the function f(x) = x2defined on the interval [−π, π].
a) Find the Fourier cosine series representation of f(x).
b) Determine the convergence of the Fourier cosine series of f(x).
c) Calculate the value at which the Fourier cosine series of f(x)converges to f(x).
Solution 4.
a) To find the Fourier cosine series representation of f(x) = x2, we need to calculate the
coefficients anusing the formula:
an=2
πZπ
0
x2cos(nx)dx.
Calculating the integral gives:
an=2
πZπ
0
x2cos(nx)dx
=2
π2(nx sin(nx)+(n2x2−2) cos(nx))
n2
π
0
=2
π2nπ sin(nπ)
n2+n2π2−2
n2
= 2 (−1)n
n+π2−2
n2.
Therefore, the Fourier cosine series representation of f(x)is:
f(x) = π2−2
2+ 2
∞
X
n=1 (−1)n
n+π2−2
n2cos(nx).
b) To determine the convergence of the Fourier cosine series of f(x), we need to check if f(x)is
piecewise smooth on [−π, π]. Since f(x) = x2is continuous and has derivative f′(x) = 2xexisting
for all x∈[−π, π], the Fourier cosine series converges to f(x).
c) The Fourier cosine series converges to f(x)for x∈(−π, π), so the value at which the Fourier
cosine series of f(x)converges to f(x)is x= 0.
5 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider a square wave signal x(t)given by:
x(t) = (1if 0≤t < T/2
−1if T/2≤t<T
where Tis the period of the signal.
a) Calculate the Fourier transform X(ω)of x(t).
b) Determine the value of X(0).
c) Find the power spectral density SX(ω)of the signal x(t).
Solution 1.
a) To calculate the Fourier transform X(ω)of the given square wave signal x(t), we apply the
definition of Fourier transform:
X(ω) = Z∞
−∞
x(t)e−jωtdt
Splitting the integral based on x(t):
X(ω) = ZT/2
0
e−jωtdt −ZT
T/2
e−jωtdt
Solving the integrals, we get:
X(ω) = 1−e−jωT/2
jω −e−jωT /2−e−jωT
jω
Simplifying further gives the Fourier transform as:
X(ω) = 2
jω (1 −cos(ωT /2))
b) To find the value of X(0), we substitute ω= 0 in the Fourier transform:
X(0) = 2
0(1 −cos(0)) = 0
c) The power spectral density SX(ω)is given by:
SX(ω) = |X(ω)|2
Substitute the Fourier transform X(ω)obtained earlier to get:
SX(ω) =
2
jω (1 −cos(ωT /2))
2
=4(1 −cos(ωT/2))2
ω2
6 CALCULATING FOURIER TRANSFORMS OF DISTRIBUTIONS
Problem 1. Compute the Fourier transform of the distribution T=δ(x−3) + δ(x+ 4), where δ
denotes the Dirac delta distribution.
Solution 1. The Fourier transform of a distribution Tis defined as
F(δ)(ϕ) = δ(F(ϕ))
where ϕis a test function.
a) Let’s first compute the Fourier transform of δ(x−3) using the formula:
δ(x−a)F
−→ e−2πiaξ
Thus, the Fourier transform of δ(x−3) is
F(δ(x−3)) = e−6πiξ
b) Next, we compute the Fourier transform of δ(x+ 4) which is given by:
δ(x+a)F
−→ e2πiaξ
Therefore, the Fourier transform of δ(x+ 4) is
F(δ(x+ 4)) = e8πiξ
c) Finally, computing the Fourier transform of the given distribution T:
F(T)(ϕ) = δ(F(ϕ))
=δ(e−6πiξ) + δ(e8πiξ )
=ϕ(6) + ϕ(−8)
Therefore, the Fourier transform of Tis ϕ(6) + ϕ(−8).
Problem 2. Determine the Fourier transform of the distribution T=x2·sin(2πx).
Solution 2. The Fourier transform of a distribution Tis given by:
F(xn·ϕ)(ξ)=(−i)n·dnF(ϕ)
dξn
where ϕis a test function.
We need to compute the Fourier transform of x2·sin(2πx).
a) First, we find the Fourier transforms of x2and sin(2πx)individually.
The Fourier transform of x2can be calculated using the formula:
F(x2)(ξ) = (−i)2·d2
dξ2δ
This simplifies to:
F(x2)(ξ) = −δ′′(ξ)
b) Next, we find the Fourier transform of sin(2πx):
F(sin(αx)) = α
α2−ξ2
Applying this to sin(2πx)gives:
F(sin(2πx)) = 2π
(2π)2−ξ2
c) Now, we calculate the Fourier transform of x2·sin(2πx)by convolving the individual trans-
forms:
F(x2·sin(2πx))(ξ) = −δ′′(ξ)∗2π
(2π)2−ξ2
=−2π
(2π)2−ξ2
Therefore, the Fourier transform of x2·sin(2πx)is −2π
(2π)2−ξ2.
7 "ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES"
Problem 1. Consider the function f(x) = cos(2πx) + 3 sin(4πx).
a) Find the period of f(x).
b) Compute the Fourier series of f(x).
Solution 1.
a) To find the period of f(x), we need to consider the period of each component function
cos(2πx)and sin(4πx)separately.
The period of cos(2πx)is given by T=2π
2π= 1.
Similarly, the period of sin(4πx)is T=2π
4π=1
2.
Since the period of f(x)is the least common multiple of the periods of its component functions,
the period of f(x)is T=lcm(1,1
2)=1.
Thus, the period of f(x)is 1.
b) The Fourier series of f(x)can be found by expressing it in the form:
f(x) = a0+
∞
X
n=1 ancos 2nπx
T+bnsin 2nπx
T
where Tis the period of f(x).
Since we found that T= 1 in part (a), we have:
f(x) = a0+
∞
X
n=1
[ancos(2nπx) + bnsin(2nπx)]
The coefficients can be calculated using the formulas:
a0=1
TZT
2
−T
2
f(x)dx
an=2
TZT
2
−T
2
f(x) cos 2nπx
Tdx
bn=2
TZT
2
−T
2
f(x) sin 2nπx
Tdx
Since T= 1, these formulas simplify to:
a0=Z0.5
−0.5
f(x)dx
an= 2 Z0.5
−0.5
f(x) cos(2nπx)dx
bn= 2 Z0.5
−0.5
f(x) sin(2nπx)dx
Plugging in the function f(x)and computing the integrals will give the coefficients for the Fourier
series of f(x).
8 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider a continuous-time signal x(t)in a given bandwidth, represented by its
Fourier transform X(f). The signal is sampled with a sampling rate of Fs= 200 Hz. Determine if
this sampling rate satisfies the Nyquist criterion for this signal. Justify your answer.
Solution 1.
Given that the signal is sampled with a sampling rate of Fs= 200 Hz, we need to check if this
sampling rate satisfies the Nyquist criterion.
The Nyquist theorem states that in order to perfectly reconstruct a signal from its samples, the
sampling rate must be at least twice the maximum frequency present in the signal. In other words,
Fs≥2B, where Bis the bandwidth of the signal.
In this case, since Fs= 200 Hz, we need to determine the bandwidth of the signal B. If X(f)is
non-zero only up to BHz, then the condition is satisfied.
If the signal’s Fourier transform X(f)has energy up to BHz, then the condition Fs≥2Bis
satisfied and the sampling rate of 200 Hz meets the Nyquist criterion for this signal.
Therefore, the sampling rate of 200 Hz satisfies the Nyquist criterion for the given continuous-
time signal.
This problem highlighted the importance of the Nyquist criterion in determining the appropriate
sampling rate for a signal in order to avoid aliasing.
9 "ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES"
Problem 1. Consider a non-stationary signal x(t) = sin(2πt2).
a) Find the Fourier transform X(ω)of the signal.
b) Determine the time-frequency representation of the signal using the Short-Time Fourier
Transform (STFT) with a Gaussian window function g(t) = e−t2.
Solution 1.
a) To find the Fourier transform X(ω)of the signal x(t) = sin(2πt2), we use the definition of the
Fourier transform:
X(ω) = F{x(t)}=Z∞
−∞
x(t)e−jωtdt
Substitute x(t) = sin(2πt2)into the integral:
X(ω) = Z∞
−∞
sin(2πt2)e−jωtdt
This integral does not have a closed-form solution, but can be numerically evaluated.
b) The Short-Time Fourier Transform (STFT) of a signal x(t)with a window function g(t)is given
by:
X(ω, τ) = Z∞
−∞
x(t)g(t−τ)e−jωtdt
For the given signal x(t) = sin(2πt2)and Gaussian window function g(t) = e−t2, we can com-
pute the STFT X(ω, τ)by sliding the window along the time axis and computing the Fourier trans-
form for each windowed segment.
The time-frequency representation of the signal using STFT provides information about how
the frequency content of the signal evolves over time, which is particularly useful for analyzing
non-stationary signals like x(t) = sin(2πt2).
10 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider a function f(x) = (xif 0≤x≤π
0otherwise
a) Compute the Fourier sine series of f(x).
b) Find the sum of the series.
Solution 1.
a) The Fourier sine series for a function f(x)defined on 0≤x≤Lis given by:
f(x)∼
∞
X
n=1
Bnsin nπx
L
where Bn=2
LRL
0f(x) sin nπx
Ldx.
For the function f(x)given in the problem, L=πand f(x) = xfor 0≤x≤π. Thus, we need
to compute Bn:
Bn=2
πZπ
0
xsin(nx)dx
=2
π−x
ncos(nx)
π
0
+1
nZπ
0
cos(nx)dx
=2
π0−(−1)nπ
n+1
n2sin(nx)
π
0
= 0 + 2(−1)n
n2+2(−1)n
πn
=2(−1)n
n21 + 1
πn
Therefore, the Fourier sine series of f(x)is:
f(x)∼
∞
X
n=1
2(−1)n
n21 + 1
πnsin(nx)
b) To find the sum of the series, we note that the Fourier sine series of an odd function converges
to f(x−)+f(x+)
2.
For the given function f(x) = x,f(x−) = xand f(x+)=0(at the jump discontinuity x= 0).
Thus, the sum of the series is:
x+ 0
2=x
2
Therefore, the sum of the Fourier sine series of f(x)is x
2.
11 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider a function f(t) = cos(2πt)in the time domain. Compute the Fourier
transform F(ω)of f(t), where F(ω)is defined as:
F(ω) = Z∞
−∞
f(t)e−iωtdt
Solution 1. To compute the Fourier transform F(ω)of f(t) = cos(2πt), we substitute f(t)into
the definition of the Fourier transform:
F(ω) = Z∞
−∞
cos(2πt)e−iωtdt
Using Euler’s formula eix = cos(x) + isin(x), we can express the integrand as:
cos(2πt)e−iωt =1
2(e2πit +e−2πit)1
2i(eiωt −e−iωt)
Multiplying and simplifying, the integrand becomes:
1
4i(e(2πi−ωi)t−e(−2πi+ωi)t+e(2πi+ωi)t−e(−2πi−ωi)t)
Now, we can integrate term by term:
F(ω) = 1
4i"e(2πi−ωi)t
2πi −ωi −e(−2πi+ωi)t
−2πi +ωi +e(2πi+ωi)t
2πi +ωi −e(−2πi−ωi)t
−2πi −ωi #
∞
−∞
Given that the oscillatory terms will go to zero at ±∞ due to periodicity, we are left with evalu-
ating the residues at the poles ω=±2π. This gives us the final result:
F(ω) = π(δ(ω−2π) + δ(ω+ 2π))
where δ(ω)is the Dirac delta function.
So, the Fourier transform of f(t) = cos(2πt)is F(ω) = π(δ(ω−2π) + δ(ω+ 2π)).
12 NONLINEAR SIGNAL PROCESSING USING HARMONIC ANALYSIS AND FOURIER TRANS-
FORM
Problem 1.
Consider a nonlinear signal x(t) = cos(2πt) + cos(4πt). Determine the Fourier series represen-
tation of x(t)over the interval [−T, T ], where Tis the period of the signal.
Solution 1.
Given signal x(t) = cos(2πt) + cos(4πt), we first need to find the period of the signal by calcu-
lating the least common multiple of the periods of the individual cosine components.
The period of cos(2πt)is T1= 1 and the period of cos(4πt)is T2=1
2.
The period of the combined signal is T=LCM(T1, T2) = LCM(1,1
2) = 1.
Therefore, the Fourier series representation of x(t)over the interval [−1,1] is:
x(t) = a0+
∞
X
n=1
(ancos(nπt) + bnsin(nπt))
where an=2
TRT
−Tx(t) cos(nπt)dt and bn=2
TRT
−Tx(t) sin(nπt)dt.
Calculating the coefficients:
a0=1
TZ1
−1
x(t)dt =1
1Z1
−1
(cos(2πt) + cos(4πt))dt
an=2
TZ1
−1
x(t) cos(nπt)dt =2
1Z1
−1
(cos(2πt) + cos(4πt)) cos(nπt)dt
bn=2
TZ1
−1
x(t) sin(nπt)dt =2
1Z1
−1
(cos(2πt) + cos(4πt)) sin(nπt)dt
By performing the integrations, we can find the Fourier series representation of x(t)over the
interval [−1,1].
13 "WAVELET ANALYSIS AND APPLICATIONS IN SIGNAL PROCESSING"
Problem 13. Consider a signal f(t) = sin(4t) + sin(8t). Perform the following wavelet analysis
using the Morlet wavelet:
ψ(t) = eiω0te−t2/2
where ω0= 5.
a) Compute the continuous wavelet transform (CWT) of f(t).
b) Sketch the scalogram of the CWT.
c) Determine the dominant frequency component(s) present in the signal.
Solution 13.
a) To compute the CWT of f(t), we use the formula:
Wf(a, b) = Z∞
−∞
f(t)·¯
ψt−b
adt,
where ais the dilation parameter and bis the translation parameter.
For this problem, f(t) = sin(4t) + sin(8t)and ψ(t) = ei·5te−t2/2.
The CWT is given by:
Wf(a, b) = Z∞
−∞
[sin(4t) + sin(8t)] ·e−i·5(t−b)/ae−(t−b)2/(2a2)dt.
This integral can be evaluated by numerical methods or special software such as MATLAB to
find the values of Wf(a, b).
b) The scalogram is a visual representation of the CWT. It shows the magnitude of the CWT
coefficients at each scale aand time b. To sketch the scalogram, we plot the magnitude of |Wf(a, b)|
at different aand bvalues.
c) By examining the scalogram, we can identify the dominant frequency components by look-
ing at the peaks in the magnitude of the CWT coefficients. In this case, the dominant frequency
components correspond to the frequencies that have the highest magnitude values in the CWT.
14 "APPLICATIONS OF WAVELET TRANSFORM IN SIGNAL PROCESSING"
Problem 14. Consider a signal f(t)with the following data:
f(t) = (1,if 0≤t≤1
0,otherwise
Let’s analyze this signal using the Haar wavelet transform.
a) Calculate the Haar wavelet coefficients ck,m corresponding to f(t).
b) Determine the wavelet transform of f(t).
c) Reconstruct the original signal using the inverse Haar wavelet transform from the coefficients
found in part (a).
Solution 14.
a) The Haar wavelet coefficients ck,m can be calculated as:
ck,m =Z∞
−∞
f(t)ψk,m(t)dt
Here, ψk,m(t)is the Haar wavelet function. Since the Haar wavelet is a piecewise constant
function, we have:
ψk,m(t) = (2k
2,if 2km≤t < 2k(m+ 1)
0,otherwise
Given f(t), we can calculate the coefficients by integrating over the intervals where ψk,m(t)is
non-zero. For the Haar wavelet, we only need to consider the intervals [0,1) and [1,2).
For the interval [0,1):
c0,0=Z1
0
f(t)ψ0,0(t)dt
c0,0=Z1
0
(1) ·(20
2)dt
c0,0=Z1
0
dt = 1
For the interval [1,2):
c1,0=Z2
1
f(t)ψ1,0(t)dt
c1,0=Z2
1
(0) ·(21
2)dt
c1,0= 0
Therefore, the Haar wavelet coefficients for f(t)are c0,0= 1 and c1,0= 0.
b) The wavelet transform of f(t)is given by:
Wf(a, b) = X
k,m
ck,m ·ψk,m(t)
Since c1,0= 0, the wavelet transform simplifies to:
Wf(a, b) = c0,0·ψ0,0(t)=1·20
2= 1
c) Reconstructing the original signal using the inverse Haar wavelet transform involves summing
the scaled and shifted Haar wavelets:
f(t) = X
k,m
ck,m ·ψk,m(t)
Substitute the coefficients c0,0= 1 and c1,0= 0 into the formula:
f(t) = 1 ·ψ0,0(t) = 1 ·20
2= 1
Therefore, the reconstructed signal using the inverse Haar wavelet transform is a constant
signal f(t)=1.
15 "APPLICATION OF WAVELET TRANSFORM IN SIGNAL DENOISING AND COMPRES-
SION"
Problem 15. Consider a signal f(t)defined as:
f(t) = 2 sin(4πt) + 5 cos(8πt) + 3 sin(12πt)
a) Using the wavelet transform, determine the frequency components present in the signal
f(t). b) If we apply a wavelet transform and find that the wavelet coefficients corresponding to the
frequency component 4πare very small compared to the coefficients at other frequencies, explain
why this might indicate the presence of noise in the signal. c) How can wavelet transforms be used
in denoising this signal?
Solution 15.
a) To determine the frequency components present in the signal f(t), we need to apply the
wavelet transform. Let’s consider the Haar wavelet transform for simplicity. The Haar wavelet
transform of a signal f(t)is given by:
W(a, b) = Z∞
−∞
f(t)ψa,b(t)dt
Where ψa,b(t)is the Haar wavelet defined as:
ψa,b(t) =
1if a≤t<a+1
2
−1if a+1
2≤t<a+ 1
0otherwise
Let’s calculate the wavelet coefficients for the given signal f(t)at different scales a:
W(a, b) = Z∞
−∞
f(t)ψa,b(t)dt
For a= 1, we have:
W(1, b) = Z∞
−∞
(2 sin(4πt) + 5 cos(8πt) + 3 sin(12πt))ψ1,b(t)dt
Similarly, we can calculate W(2, b),W(3, b), and so on to determine the frequency components
present in the signal f(t).
b) If the wavelet coefficients corresponding to the frequency component 4πare very small com-
pared to the coefficients at other frequencies, it might indicate the presence of noise in the signal.
This is because in the absence of noise, the coefficients for the true signal frequencies would be
much larger compared to any noise present in the signal.
c) Wavelet transforms can be used in denoising the signal by thresholding the wavelet coeffi-
cients. By setting small coefficients below a certain threshold to zero, we can eliminate noise while
retaining the important frequency components of the signal. This is a common technique in signal
denoising using wavelet transforms.
16 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 17. Consider a non-stationary signal given by f(t) = sin(2πt2).
a) Determine the Fourier transform of the signal f(t).
b) Compute the time-frequency representation of f(t)using the Short-Time Fourier Transform
(STFT) with a Gaussian window function g(t)given by g(t) = e−t2.
Solution 17.
a) The Fourier transform of a signal f(t)is defined as
F(ω) = Z∞
−∞
f(t)e−iωtdt.
For the given signal f(t) = sin(2πt2), we have
F(ω) = Z∞
−∞
sin(2πt2)e−iωtdt.
This particular integral does not have a closed-form solution, so we would need to evaluate it using
numerical methods or approximation techniques.
b) The Short-Time Fourier Transform (STFT) of a signal f(t)using a window function g(t)is
defined as
ST F T (t, ω) = Z∞
−∞
f(τ)g(τ−t)e−iωτ dτ,
where the overline denotes complex conjugation. For a Gaussian window function g(t) = e−t2, the
STFT of the signal f(t) = sin(2πt2)can be computed as
ST F T (t, ω) = Z∞
−∞
sin(2πτ2)e−(τ−t)2e−iωτ dτ.
We can evaluate this integral numerically to compute the time-frequency representation of the
signal using the given Gaussian window function.
17 "CONVERGENCE AND DIVERGENCE OF FOURIER SERIES FOR PIECEWISE FUNC-
TIONS"
Problem 17. Consider the function f(x)defined as:
f(x) = (x0≤x<π
−x−π < x < 0
a) Find the Fourier series of f(x).
b) Determine the conditions for which the Fourier series of f(x)converges.
Solution 17.
a) To find the Fourier series of f(x), we need to compute the coefficients anand bn:
an=1
πZπ
−π
f(x) cos(nx)dx
bn=1
πZπ
−π
f(x) sin(nx)dx
Let’s start with the calculation of an:
an=1
πZ0
−π
(−x) cos(nx)dx +Zπ
0
xcos(nx)dx
Solving these integrals and simplifying yields:
an=2((−1)n−1)
πn2
Now, let’s find bn:
bn=1
πZ0
−π
(−x) sin(nx)dx +Zπ
0
xsin(nx)dx
After solving and simplifying the integrals, we get:
bn= 0
The Fourier series of f(x)is then given by:
f(x)∼2
π1
12−1
32cos(x) + 1
52cos(2x)− · · ·
b) The Fourier series of f(x)converges if the function f(x)satisfies the Dirichlet conditions. In
this case, since f(x)is piecewise continuous and has a finite number of discontinuities within a
period of 2π, the Fourier series of f(x)converges everywhere.
18 "NON-LINEAR FOURIER TRANSFORM TECHNIQUES FOR SIGNAL ANALYSIS"
Problem 1. Consider a signal f(t) = sin(2πt) + 2 cos(4πt).
a) Compute the Fourier Transform of f(t).
b) Find the frequency components present in the signal f(t).
c) Determine the amplitude and phase of each frequency component.
Solution 1.
a) The Fourier Transform of a signal f(t)is defined as:
F(ω) = Z∞
−∞
f(t)e−jωtdt
Applying the definition to the given signal f(t), we have:
F(ω) = Z∞
−∞
[sin(2πt) + 2 cos(4πt)]e−jωtdt
=Z∞
−∞
sin(2πt)e−jωtdt + 2 Z∞
−∞
cos(4πt)e−jωtdt
Solving the integrals, we get:
F(ω) = π(δ(ω−2π)−δ(ω+ 2π)) + π(δ(ω−4π) + δ(ω+ 4π))
b) The frequency components present in the signal f(t)are 2πand 4π.
c) To determine the amplitude and phase of each frequency component, we can use the prop-
erties of the Dirac delta function in the Fourier Transform. The amplitude is the coefficient in front
of the delta function and the phase is given by the sign of the delta function. Thus, we have:
For 2πcomponent: Amplitude = πand Phase = (δ(ω−2π)−δ(ω+ 2π)) = 0
For 4πcomponent: Amplitude = πand Phase = (δ(ω−4π) + δ(ω+ 4π)) = 0
19 "ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES"
Problem 1. Consider a continuous-time signal x(t)defined as:
x(t) = 2 cos(3πt) + 3 sin(6πt)
Find the Fourier transform X(ω)of the signal x(t).
Solution 1.
The Fourier transform of a continuous-time signal x(t)is defined as:
X(ω) = Z∞
−∞
x(t)e−jωtdt
Plugging in the expression for x(t), we have:
X(ω) = Z∞
−∞
(2 cos(3πt) + 3 sin(6πt))e−jωtdt
Expanding the trigonometric functions using Euler’s formula, we get:
X(ω) = Z∞
−∞
2(ej3πt +e−j3πt
2) + 3(ej6πt −e−j6πt
2j)e−jωtdt
X(ω) = Z∞
−∞
(ej(3π−ω)t+e−j(3π+ω)t−3
2j(ej(6π−ω)t−e−j(6π+ω)t))dt
Remembering that the Fourier transform has a property with delta functions, we find:
X(ω) = 2π(δ(ω−3π) + δ(ω+ 3π)) −3π(δ(ω−6π) + δ(ω+ 6π))
Thus, the Fourier transform of the signal x(t)is:
X(ω) = 2π(δ(ω−3π) + δ(ω+ 3π)) −3π(δ(ω−6π) + δ(ω+ 6π))
Certainly! Here is a numerical problem in the realm of Advanced Harmonic Analysis and Fourier
Transform Techniques with a step-by-step solution:
20 "ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES"
Problem 21. Consider a non-stationary signal f(t)defined as:
f(t) = (sin(2πt) 0 ≤t < 1
0otherwise
Calculate the Fourier transform F(ω)of the signal f(t).
Solution 21. To find the Fourier transform F(ω)of the given signal f(t), we use the definition
of the Fourier transform:
F(ω) = Z∞
−∞
f(t)e−iωtdt
Given that f(t)is defined piecewise, we split the integral into two parts:
F(ω) = Z1
0
sin(2πt)e−iωtdt +Z∞
1
0·e−iωtdt
Solving the first integral, we can proceed to calculate the Fourier transform. Let’s solve the
integral as follows:
F(ω) = Z1
0
sin(2πt)e−iωtdt
Applying Euler’s formula to express the sine function in terms of complex exponentials, we have:
F(ω) = Z1
0
ei2πt −e−i2πt
2ie−iωtdt
F(ω) = 1
2iZ1
0
ei(2π−ω)tdt −Z1
0
e−i(2π+ω)tdt
F(ω) = 1
2i"ei(2π−ω)t
i(2π−ω)
1
0−e−i(2π+ω)t
−i(2π+ω)
1
0#
F(ω) = 1
2i"ei(2π−ω)−1
i(2π−ω)−e−i(2π+ω)−1
−i(2π+ω)#
This integral evaluation gives us the Fourier transform F(ω).
a) We can calculate the Fourier series coefficients anand bnusing the following formulas:
an=1
LZL
−L
f(x) cos 2πnx
Ldx
bn=1
LZL
−L
f(x) sin 2πnx
Ldx
where Lis the period of the function.
For f(x) = sin(2πx) + cos(4πx), the period is L= 1. Thus, we have:
an=Z1
−1
(sin(2πx) + cos(4πx)) cos(2πnx)dx
bn=Z1
−1
(sin(2πx) + cos(4πx)) sin(2πnx)dx
Calculating these integrals will give us the Fourier series coefficients anand bn.
b) To find the constant term in the Fourier series of f(x), we need to determine a0, which is
given by:
a0=1
LZL
−L
f(x)dx
Substitute L= 1 and integral bounds −1to 1to find the value of a0.
c) The representation of f(x)in its Fourier series form is given by:
f(x)≈a0
2+
∞
X
n=1 ancos 2πnx
L+bnsin 2πnx
L
Substitute the calculated values of a0,an, and bninto this equation to obtain the Fourier series
representation of f(x).
3 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider the function f(t) = sin(3t) + cos(5t).
a) Find the Fourier transform of f(t).
b) Determine the frequency components present in f(t).
Solution 1.
a) The Fourier transform of a function f(t)is defined as:
F(ω) = Z∞
−∞
f(t)e−iωtdt
Applying this definition to our function f(t) = sin(3t) + cos(5t), we have:
F(ω) = Z∞
−∞
(sin(3t) + cos(5t))e−iωtdt
=Z∞
−∞
sin(3t)e−iωtdt +Z∞
−∞
cos(5t)e−iωtdt
We can solve these integrals using the properties of the Fourier transform.
b) The frequency components present in f(t)are the frequencies of the sine and cosine func-
tions, which are 3and 5respectively.
Therefore, the frequency components present in f(t)are 3and 5.
4 "ANALYZING THE CONVERGENCE AND DIVERGENCE OF FOURIER SERIES REPRE-
SENTATIONS"
Problem 4. Consider the function f(x) = x2defined on the interval [−π, π].
a) Find the Fourier cosine series representation of f(x).
b) Determine the convergence of the Fourier cosine series of f(x).
c) Calculate the value at which the Fourier cosine series of f(x)converges to f(x).
Solution 4.
a) To find the Fourier cosine series representation of f(x) = x2, we need to calculate the
coefficients anusing the formula:
an=2
πZπ
0
x2cos(nx)dx.
Calculating the integral gives:
an=2
πZπ
0
x2cos(nx)dx
=2
π2(nx sin(nx)+(n2x2−2) cos(nx))
n2
π
0
=2
π2nπ sin(nπ)
n2+n2π2−2
n2
= 2 (−1)n
n+π2−2
n2.
Therefore, the Fourier cosine series representation of f(x)is:
f(x) = π2−2
2+ 2
∞
X
n=1 (−1)n
n+π2−2
n2cos(nx).
b) To determine the convergence of the Fourier cosine series of f(x), we need to check if f(x)is
piecewise smooth on [−π, π]. Since f(x) = x2is continuous and has derivative f′(x) = 2xexisting
for all x∈[−π, π], the Fourier cosine series converges to f(x).
c) The Fourier cosine series converges to f(x)for x∈(−π, π), so the value at which the Fourier
cosine series of f(x)converges to f(x)is x= 0.
5 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider a square wave signal x(t)given by:
x(t) = (1if 0≤t < T/2
−1if T/2≤t<T
where Tis the period of the signal.
a) Calculate the Fourier transform X(ω)of x(t).
b) Determine the value of X(0).
c) Find the power spectral density SX(ω)of the signal x(t).
Solution 1.
a) To calculate the Fourier transform X(ω)of the given square wave signal x(t), we apply the
definition of Fourier transform:
X(ω) = Z∞
−∞
x(t)e−jωtdt
Splitting the integral based on x(t):
X(ω) = ZT/2
0
e−jωtdt −ZT
T/2
e−jωtdt
Solving the integrals, we get:
X(ω) = 1−e−jωT/2
jω −e−jωT /2−e−jωT
jω
Simplifying further gives the Fourier transform as:
X(ω) = 2
jω (1 −cos(ωT /2))
b) To find the value of X(0), we substitute ω= 0 in the Fourier transform:
X(0) = 2
0(1 −cos(0)) = 0
c) The power spectral density SX(ω)is given by:
SX(ω) = |X(ω)|2
Substitute the Fourier transform X(ω)obtained earlier to get:
SX(ω) =
2
jω (1 −cos(ωT /2))
2
=4(1 −cos(ωT/2))2
ω2
6 CALCULATING FOURIER TRANSFORMS OF DISTRIBUTIONS
Problem 1. Compute the Fourier transform of the distribution T=δ(x−3) + δ(x+ 4), where δ
denotes the Dirac delta distribution.
Solution 1. The Fourier transform of a distribution Tis defined as
F(δ)(ϕ) = δ(F(ϕ))
where ϕis a test function.
a) Let’s first compute the Fourier transform of δ(x−3) using the formula:
δ(x−a)F
−→ e−2πiaξ
Thus, the Fourier transform of δ(x−3) is
F(δ(x−3)) = e−6πiξ
b) Next, we compute the Fourier transform of δ(x+ 4) which is given by:
δ(x+a)F
−→ e2πiaξ
Therefore, the Fourier transform of δ(x+ 4) is
F(δ(x+ 4)) = e8πiξ
c) Finally, computing the Fourier transform of the given distribution T:
F(T)(ϕ) = δ(F(ϕ))
=δ(e−6πiξ) + δ(e8πiξ )
=ϕ(6) + ϕ(−8)
Therefore, the Fourier transform of Tis ϕ(6) + ϕ(−8).
Problem 2. Determine the Fourier transform of the distribution T=x2·sin(2πx).
Solution 2. The Fourier transform of a distribution Tis given by:
F(xn·ϕ)(ξ)=(−i)n·dnF(ϕ)
dξn
where ϕis a test function.
We need to compute the Fourier transform of x2·sin(2πx).
a) First, we find the Fourier transforms of x2and sin(2πx)individually.
The Fourier transform of x2can be calculated using the formula:
F(x2)(ξ) = (−i)2·d2
dξ2δ
This simplifies to:
F(x2)(ξ) = −δ′′(ξ)
b) Next, we find the Fourier transform of sin(2πx):
F(sin(αx)) = α
α2−ξ2
Applying this to sin(2πx)gives:
F(sin(2πx)) = 2π
(2π)2−ξ2
c) Now, we calculate the Fourier transform of x2·sin(2πx)by convolving the individual trans-
forms:
F(x2·sin(2πx))(ξ) = −δ′′(ξ)∗2π
(2π)2−ξ2
=−2π
(2π)2−ξ2
Therefore, the Fourier transform of x2·sin(2πx)is −2π
(2π)2−ξ2.
7 "ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES"
Problem 1. Consider the function f(x) = cos(2πx) + 3 sin(4πx).
a) Find the period of f(x).
b) Compute the Fourier series of f(x).
Solution 1.
a) To find the period of f(x), we need to consider the period of each component function
cos(2πx)and sin(4πx)separately.
The period of cos(2πx)is given by T=2π
2π= 1.
Similarly, the period of sin(4πx)is T=2π
4π=1
2.
Since the period of f(x)is the least common multiple of the periods of its component functions,
the period of f(x)is T=lcm(1,1
2)=1.
Thus, the period of f(x)is 1.
b) The Fourier series of f(x)can be found by expressing it in the form:
f(x) = a0+
∞
X
n=1 ancos 2nπx
T+bnsin 2nπx
T
where Tis the period of f(x).
Since we found that T= 1 in part (a), we have:
f(x) = a0+
∞
X
n=1
[ancos(2nπx) + bnsin(2nπx)]
The coefficients can be calculated using the formulas:
a0=1
TZT
2
−T
2
f(x)dx
an=2
TZT
2
−T
2
f(x) cos 2nπx
Tdx
bn=2
TZT
2
−T
2
f(x) sin 2nπx
Tdx
Since T= 1, these formulas simplify to:
a0=Z0.5
−0.5
f(x)dx
an= 2 Z0.5
−0.5
f(x) cos(2nπx)dx
bn= 2 Z0.5
−0.5
f(x) sin(2nπx)dx
Plugging in the function f(x)and computing the integrals will give the coefficients for the Fourier
series of f(x).
8 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider a continuous-time signal x(t)in a given bandwidth, represented by its
Fourier transform X(f). The signal is sampled with a sampling rate of Fs= 200 Hz. Determine if
this sampling rate satisfies the Nyquist criterion for this signal. Justify your answer.
Solution 1.
Given that the signal is sampled with a sampling rate of Fs= 200 Hz, we need to check if this
sampling rate satisfies the Nyquist criterion.
The Nyquist theorem states that in order to perfectly reconstruct a signal from its samples, the
sampling rate must be at least twice the maximum frequency present in the signal. In other words,
Fs≥2B, where Bis the bandwidth of the signal.
In this case, since Fs= 200 Hz, we need to determine the bandwidth of the signal B. If X(f)is
non-zero only up to BHz, then the condition is satisfied.
If the signal’s Fourier transform X(f)has energy up to BHz, then the condition Fs≥2Bis
satisfied and the sampling rate of 200 Hz meets the Nyquist criterion for this signal.
Therefore, the sampling rate of 200 Hz satisfies the Nyquist criterion for the given continuous-
time signal.
This problem highlighted the importance of the Nyquist criterion in determining the appropriate
sampling rate for a signal in order to avoid aliasing.
9 "ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES"
Problem 1. Consider a non-stationary signal x(t) = sin(2πt2).
a) Find the Fourier transform X(ω)of the signal.
b) Determine the time-frequency representation of the signal using the Short-Time Fourier
Transform (STFT) with a Gaussian window function g(t) = e−t2.
Solution 1.
a) To find the Fourier transform X(ω)of the signal x(t) = sin(2πt2), we use the definition of the
Fourier transform:
X(ω) = F{x(t)}=Z∞
−∞
x(t)e−jωtdt
Substitute x(t) = sin(2πt2)into the integral:
X(ω) = Z∞
−∞
sin(2πt2)e−jωtdt
This integral does not have a closed-form solution, but can be numerically evaluated.
b) The Short-Time Fourier Transform (STFT) of a signal x(t)with a window function g(t)is given
by:
X(ω, τ) = Z∞
−∞
x(t)g(t−τ)e−jωtdt
For the given signal x(t) = sin(2πt2)and Gaussian window function g(t) = e−t2, we can com-
pute the STFT X(ω, τ)by sliding the window along the time axis and computing the Fourier trans-
form for each windowed segment.
The time-frequency representation of the signal using STFT provides information about how
the frequency content of the signal evolves over time, which is particularly useful for analyzing
non-stationary signals like x(t) = sin(2πt2).
10 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider a function f(x) = (xif 0≤x≤π
0otherwise
a) Compute the Fourier sine series of f(x).
b) Find the sum of the series.
Solution 1.
a) The Fourier sine series for a function f(x)defined on 0≤x≤Lis given by:
f(x)∼
∞
X
n=1
Bnsin nπx
L
where Bn=2
LRL
0f(x) sin nπx
Ldx.
For the function f(x)given in the problem, L=πand f(x) = xfor 0≤x≤π. Thus, we need
to compute Bn:
Bn=2
πZπ
0
xsin(nx)dx
=2
π−x
ncos(nx)
π
0
+1
nZπ
0
cos(nx)dx
=2
π0−(−1)nπ
n+1
n2sin(nx)
π
0
= 0 + 2(−1)n
n2+2(−1)n
πn
=2(−1)n
n21 + 1
πn
Therefore, the Fourier sine series of f(x)is:
f(x)∼
∞
X
n=1
2(−1)n
n21 + 1
πnsin(nx)
b) To find the sum of the series, we note that the Fourier sine series of an odd function converges
to f(x−)+f(x+)
2.
For the given function f(x) = x,f(x−) = xand f(x+)=0(at the jump discontinuity x= 0).
Thus, the sum of the series is:
x+ 0
2=x
2
Therefore, the sum of the Fourier sine series of f(x)is x
2.
11 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider a function f(t) = cos(2πt)in the time domain. Compute the Fourier
transform F(ω)of f(t), where F(ω)is defined as:
F(ω) = Z∞
−∞
f(t)e−iωtdt
Solution 1. To compute the Fourier transform F(ω)of f(t) = cos(2πt), we substitute f(t)into
the definition of the Fourier transform:
F(ω) = Z∞
−∞
cos(2πt)e−iωtdt
Using Euler’s formula eix = cos(x) + isin(x), we can express the integrand as:
cos(2πt)e−iωt =1
2(e2πit +e−2πit)1
2i(eiωt −e−iωt)
Multiplying and simplifying, the integrand becomes:
1
4i(e(2πi−ωi)t−e(−2πi+ωi)t+e(2πi+ωi)t−e(−2πi−ωi)t)
Now, we can integrate term by term:
F(ω) = 1
4i"e(2πi−ωi)t
2πi −ωi −e(−2πi+ωi)t
−2πi +ωi +e(2πi+ωi)t
2πi +ωi −e(−2πi−ωi)t
−2πi −ωi #
∞
−∞
Given that the oscillatory terms will go to zero at ±∞ due to periodicity, we are left with evalu-
ating the residues at the poles ω=±2π. This gives us the final result:
F(ω) = π(δ(ω−2π) + δ(ω+ 2π))
where δ(ω)is the Dirac delta function.
So, the Fourier transform of f(t) = cos(2πt)is F(ω) = π(δ(ω−2π) + δ(ω+ 2π)).
12 NONLINEAR SIGNAL PROCESSING USING HARMONIC ANALYSIS AND FOURIER TRANS-
FORM
Problem 1.
Consider a nonlinear signal x(t) = cos(2πt) + cos(4πt). Determine the Fourier series represen-
tation of x(t)over the interval [−T, T ], where Tis the period of the signal.
Solution 1.
Given signal x(t) = cos(2πt) + cos(4πt), we first need to find the period of the signal by calcu-
lating the least common multiple of the periods of the individual cosine components.
The period of cos(2πt)is T1= 1 and the period of cos(4πt)is T2=1
2.
The period of the combined signal is T=LCM(T1, T2) = LCM(1,1
2) = 1.
Therefore, the Fourier series representation of x(t)over the interval [−1,1] is:
x(t) = a0+
∞
X
n=1
(ancos(nπt) + bnsin(nπt))
where an=2
TRT
−Tx(t) cos(nπt)dt and bn=2
TRT
−Tx(t) sin(nπt)dt.
Calculating the coefficients:
a0=1
TZ1
−1
x(t)dt =1
1Z1
−1
(cos(2πt) + cos(4πt))dt
an=2
TZ1
−1
x(t) cos(nπt)dt =2
1Z1
−1
(cos(2πt) + cos(4πt)) cos(nπt)dt
bn=2
TZ1
−1
x(t) sin(nπt)dt =2
1Z1
−1
(cos(2πt) + cos(4πt)) sin(nπt)dt
By performing the integrations, we can find the Fourier series representation of x(t)over the
interval [−1,1].
13 "WAVELET ANALYSIS AND APPLICATIONS IN SIGNAL PROCESSING"
Problem 13. Consider a signal f(t) = sin(4t) + sin(8t). Perform the following wavelet analysis
using the Morlet wavelet:
ψ(t) = eiω0te−t2/2
where ω0= 5.
a) Compute the continuous wavelet transform (CWT) of f(t).
b) Sketch the scalogram of the CWT.
c) Determine the dominant frequency component(s) present in the signal.
Solution 13.
a) To compute the CWT of f(t), we use the formula:
Wf(a, b) = Z∞
−∞
f(t)·¯
ψt−b
adt,
where ais the dilation parameter and bis the translation parameter.
For this problem, f(t) = sin(4t) + sin(8t)and ψ(t) = ei·5te−t2/2.
The CWT is given by:
Wf(a, b) = Z∞
−∞
[sin(4t) + sin(8t)] ·e−i·5(t−b)/ae−(t−b)2/(2a2)dt.
This integral can be evaluated by numerical methods or special software such as MATLAB to
find the values of Wf(a, b).
b) The scalogram is a visual representation of the CWT. It shows the magnitude of the CWT
coefficients at each scale aand time b. To sketch the scalogram, we plot the magnitude of |Wf(a, b)|
at different aand bvalues.
c) By examining the scalogram, we can identify the dominant frequency components by look-
ing at the peaks in the magnitude of the CWT coefficients. In this case, the dominant frequency
components correspond to the frequencies that have the highest magnitude values in the CWT.
14 "APPLICATIONS OF WAVELET TRANSFORM IN SIGNAL PROCESSING"
Problem 14. Consider a signal f(t)with the following data:
f(t) = (1,if 0≤t≤1
0,otherwise
Let’s analyze this signal using the Haar wavelet transform.
a) Calculate the Haar wavelet coefficients ck,m corresponding to f(t).
b) Determine the wavelet transform of f(t).
c) Reconstruct the original signal using the inverse Haar wavelet transform from the coefficients
found in part (a).
Solution 14.
a) The Haar wavelet coefficients ck,m can be calculated as:
ck,m =Z∞
−∞
f(t)ψk,m(t)dt
Here, ψk,m(t)is the Haar wavelet function. Since the Haar wavelet is a piecewise constant
function, we have:
ψk,m(t) = (2k
2,if 2km≤t < 2k(m+ 1)
0,otherwise
Given f(t), we can calculate the coefficients by integrating over the intervals where ψk,m(t)is
non-zero. For the Haar wavelet, we only need to consider the intervals [0,1) and [1,2).
For the interval [0,1):
c0,0=Z1
0
f(t)ψ0,0(t)dt
c0,0=Z1
0
(1) ·(20
2)dt
c0,0=Z1
0
dt = 1
For the interval [1,2):
c1,0=Z2
1
f(t)ψ1,0(t)dt
c1,0=Z2
1
(0) ·(21
2)dt
c1,0= 0
Therefore, the Haar wavelet coefficients for f(t)are c0,0= 1 and c1,0= 0.
b) The wavelet transform of f(t)is given by:
Wf(a, b) = X
k,m
ck,m ·ψk,m(t)
Since c1,0= 0, the wavelet transform simplifies to:
Wf(a, b) = c0,0·ψ0,0(t)=1·20
2= 1
c) Reconstructing the original signal using the inverse Haar wavelet transform involves summing
the scaled and shifted Haar wavelets:
f(t) = X
k,m
ck,m ·ψk,m(t)
Substitute the coefficients c0,0= 1 and c1,0= 0 into the formula:
f(t) = 1 ·ψ0,0(t) = 1 ·20
2= 1
Therefore, the reconstructed signal using the inverse Haar wavelet transform is a constant
signal f(t)=1.
15 "APPLICATION OF WAVELET TRANSFORM IN SIGNAL DENOISING AND COMPRES-
SION"
Problem 15. Consider a signal f(t)defined as:
f(t) = 2 sin(4πt) + 5 cos(8πt) + 3 sin(12πt)
a) Using the wavelet transform, determine the frequency components present in the signal
f(t). b) If we apply a wavelet transform and find that the wavelet coefficients corresponding to the
frequency component 4πare very small compared to the coefficients at other frequencies, explain
why this might indicate the presence of noise in the signal. c) How can wavelet transforms be used
in denoising this signal?
Solution 15.
a) To determine the frequency components present in the signal f(t), we need to apply the
wavelet transform. Let’s consider the Haar wavelet transform for simplicity. The Haar wavelet
transform of a signal f(t)is given by:
W(a, b) = Z∞
−∞
f(t)ψa,b(t)dt
Where ψa,b(t)is the Haar wavelet defined as:
ψa,b(t) =
1if a≤t<a+1
2
−1if a+1
2≤t<a+ 1
0otherwise
Let’s calculate the wavelet coefficients for the given signal f(t)at different scales a:
W(a, b) = Z∞
−∞
f(t)ψa,b(t)dt
For a= 1, we have:
W(1, b) = Z∞
−∞
(2 sin(4πt) + 5 cos(8πt) + 3 sin(12πt))ψ1,b(t)dt
Similarly, we can calculate W(2, b),W(3, b), and so on to determine the frequency components
present in the signal f(t).
b) If the wavelet coefficients corresponding to the frequency component 4πare very small com-
pared to the coefficients at other frequencies, it might indicate the presence of noise in the signal.
This is because in the absence of noise, the coefficients for the true signal frequencies would be
much larger compared to any noise present in the signal.
c) Wavelet transforms can be used in denoising the signal by thresholding the wavelet coeffi-
cients. By setting small coefficients below a certain threshold to zero, we can eliminate noise while
retaining the important frequency components of the signal. This is a common technique in signal
denoising using wavelet transforms.
16 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 17. Consider a non-stationary signal given by f(t) = sin(2πt2).
a) Determine the Fourier transform of the signal f(t).
b) Compute the time-frequency representation of f(t)using the Short-Time Fourier Transform
(STFT) with a Gaussian window function g(t)given by g(t) = e−t2.
Solution 17.
a) The Fourier transform of a signal f(t)is defined as
F(ω) = Z∞
−∞
f(t)e−iωtdt.
For the given signal f(t) = sin(2πt2), we have
F(ω) = Z∞
−∞
sin(2πt2)e−iωtdt.
This particular integral does not have a closed-form solution, so we would need to evaluate it using
numerical methods or approximation techniques.
b) The Short-Time Fourier Transform (STFT) of a signal f(t)using a window function g(t)is
defined as
ST F T (t, ω) = Z∞
−∞
f(τ)g(τ−t)e−iωτ dτ,
where the overline denotes complex conjugation. For a Gaussian window function g(t) = e−t2, the
STFT of the signal f(t) = sin(2πt2)can be computed as
ST F T (t, ω) = Z∞
−∞
sin(2πτ2)e−(τ−t)2e−iωτ dτ.
We can evaluate this integral numerically to compute the time-frequency representation of the
signal using the given Gaussian window function.
17 "CONVERGENCE AND DIVERGENCE OF FOURIER SERIES FOR PIECEWISE FUNC-
TIONS"
Problem 17. Consider the function f(x)defined as:
f(x) = (x0≤x<π
−x−π < x < 0
a) Find the Fourier series of f(x).
b) Determine the conditions for which the Fourier series of f(x)converges.
Solution 17.
a) To find the Fourier series of f(x), we need to compute the coefficients anand bn:
an=1
πZπ
−π
f(x) cos(nx)dx
bn=1
πZπ
−π
f(x) sin(nx)dx
Let’s start with the calculation of an:
an=1
πZ0
−π
(−x) cos(nx)dx +Zπ
0
xcos(nx)dx
Solving these integrals and simplifying yields:
an=2((−1)n−1)
πn2
Now, let’s find bn:
bn=1
πZ0
−π
(−x) sin(nx)dx +Zπ
0
xsin(nx)dx
After solving and simplifying the integrals, we get:
bn= 0
The Fourier series of f(x)is then given by:
f(x)∼2
π1
12−1
32cos(x) + 1
52cos(2x)− · · ·
b) The Fourier series of f(x)converges if the function f(x)satisfies the Dirichlet conditions. In
this case, since f(x)is piecewise continuous and has a finite number of discontinuities within a
period of 2π, the Fourier series of f(x)converges everywhere.
18 "NON-LINEAR FOURIER TRANSFORM TECHNIQUES FOR SIGNAL ANALYSIS"
Problem 1. Consider a signal f(t) = sin(2πt) + 2 cos(4πt).
a) Compute the Fourier Transform of f(t).
b) Find the frequency components present in the signal f(t).
c) Determine the amplitude and phase of each frequency component.
Solution 1.
a) The Fourier Transform of a signal f(t)is defined as:
F(ω) = Z∞
−∞
f(t)e−jωtdt
Applying the definition to the given signal f(t), we have:
F(ω) = Z∞
−∞
[sin(2πt) + 2 cos(4πt)]e−jωtdt
=Z∞
−∞
sin(2πt)e−jωtdt + 2 Z∞
−∞
cos(4πt)e−jωtdt
Solving the integrals, we get:
F(ω) = π(δ(ω−2π)−δ(ω+ 2π)) + π(δ(ω−4π) + δ(ω+ 4π))
b) The frequency components present in the signal f(t)are 2πand 4π.
c) To determine the amplitude and phase of each frequency component, we can use the prop-
erties of the Dirac delta function in the Fourier Transform. The amplitude is the coefficient in front
of the delta function and the phase is given by the sign of the delta function. Thus, we have:
For 2πcomponent: Amplitude = πand Phase = (δ(ω−2π)−δ(ω+ 2π)) = 0
For 4πcomponent: Amplitude = πand Phase = (δ(ω−4π) + δ(ω+ 4π)) = 0
19 "ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES"
Problem 1. Consider a continuous-time signal x(t)defined as:
x(t) = 2 cos(3πt) + 3 sin(6πt)
Find the Fourier transform X(ω)of the signal x(t).
Solution 1.
The Fourier transform of a continuous-time signal x(t)is defined as:
X(ω) = Z∞
−∞
x(t)e−jωtdt
Plugging in the expression for x(t), we have:
X(ω) = Z∞
−∞
(2 cos(3πt) + 3 sin(6πt))e−jωtdt
Expanding the trigonometric functions using Euler’s formula, we get:
X(ω) = Z∞
−∞
2(ej3πt +e−j3πt
2) + 3(ej6πt −e−j6πt
2j)e−jωtdt
X(ω) = Z∞
−∞
(ej(3π−ω)t+e−j(3π+ω)t−3
2j(ej(6π−ω)t−e−j(6π+ω)t))dt
Remembering that the Fourier transform has a property with delta functions, we find:
X(ω) = 2π(δ(ω−3π) + δ(ω+ 3π)) −3π(δ(ω−6π) + δ(ω+ 6π))
Thus, the Fourier transform of the signal x(t)is:
X(ω) = 2π(δ(ω−3π) + δ(ω+ 3π)) −3π(δ(ω−6π) + δ(ω+ 6π))
Certainly! Here is a numerical problem in the realm of Advanced Harmonic Analysis and Fourier
Transform Techniques with a step-by-step solution:
20 "ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES"
Problem 21. Consider a non-stationary signal f(t)defined as:
f(t) = (sin(2πt) 0 ≤t < 1
0otherwise
Calculate the Fourier transform F(ω)of the signal f(t).
Solution 21. To find the Fourier transform F(ω)of the given signal f(t), we use the definition
of the Fourier transform:
F(ω) = Z∞
−∞
f(t)e−iωtdt
Given that f(t)is defined piecewise, we split the integral into two parts:
F(ω) = Z1
0
sin(2πt)e−iωtdt +Z∞
1
0·e−iωtdt
Solving the first integral, we can proceed to calculate the Fourier transform. Let’s solve the
integral as follows:
F(ω) = Z1
0
sin(2πt)e−iωtdt
Applying Euler’s formula to express the sine function in terms of complex exponentials, we have:
F(ω) = Z1
0
ei2πt −e−i2πt
2ie−iωtdt
F(ω) = 1
2iZ1
0
ei(2π−ω)tdt −Z1
0
e−i(2π+ω)tdt
F(ω) = 1
2i"ei(2π−ω)t
i(2π−ω)
1
0−e−i(2π+ω)t
−i(2π+ω)
1
0#
F(ω) = 1
2i"ei(2π−ω)−1
i(2π−ω)−e−i(2π+ω)−1
−i(2π+ω)#
This integral evaluation gives us the Fourier transform F(ω).
a) We can calculate the Fourier series coefficients anand bnusing the following formulas:
an=1
LZL
−L
f(x) cos 2πnx
Ldx
bn=1
LZL
−L
f(x) sin 2πnx
Ldx
where Lis the period of the function.
For f(x) = sin(2πx) + cos(4πx), the period is L= 1. Thus, we have:
an=Z1
−1
(sin(2πx) + cos(4πx)) cos(2πnx)dx
bn=Z1
−1
(sin(2πx) + cos(4πx)) sin(2πnx)dx
Calculating these integrals will give us the Fourier series coefficients anand bn.
b) To find the constant term in the Fourier series of f(x), we need to determine a0, which is
given by:
a0=1
LZL
−L
f(x)dx
Substitute L= 1 and integral bounds −1to 1to find the value of a0.
c) The representation of f(x)in its Fourier series form is given by:
f(x)≈a0
2+
∞
X
n=1 ancos 2πnx
L+bnsin 2πnx
L
Substitute the calculated values of a0,an, and bninto this equation to obtain the Fourier series
representation of f(x).
3 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider the function f(t) = sin(3t) + cos(5t).
a) Find the Fourier transform of f(t).
b) Determine the frequency components present in f(t).
Solution 1.
a) The Fourier transform of a function f(t)is defined as:
F(ω) = Z∞
−∞
f(t)e−iωtdt
Applying this definition to our function f(t) = sin(3t) + cos(5t), we have:
F(ω) = Z∞
−∞
(sin(3t) + cos(5t))e−iωtdt
=Z∞
−∞
sin(3t)e−iωtdt +Z∞
−∞
cos(5t)e−iωtdt
We can solve these integrals using the properties of the Fourier transform.
b) The frequency components present in f(t)are the frequencies of the sine and cosine func-
tions, which are 3and 5respectively.
Therefore, the frequency components present in f(t)are 3and 5.
4 "ANALYZING THE CONVERGENCE AND DIVERGENCE OF FOURIER SERIES REPRE-
SENTATIONS"
Problem 4. Consider the function f(x) = x2defined on the interval [−π, π].
a) Find the Fourier cosine series representation of f(x).
b) Determine the convergence of the Fourier cosine series of f(x).
c) Calculate the value at which the Fourier cosine series of f(x)converges to f(x).
Solution 4.
a) To find the Fourier cosine series representation of f(x) = x2, we need to calculate the
coefficients anusing the formula:
an=2
πZπ
0
x2cos(nx)dx.
Calculating the integral gives:
an=2
πZπ
0
x2cos(nx)dx
=2
π2(nx sin(nx)+(n2x2−2) cos(nx))
n2
π
0
=2
π2nπ sin(nπ)
n2+n2π2−2
n2
= 2 (−1)n
n+π2−2
n2.
Therefore, the Fourier cosine series representation of f(x)is:
f(x) = π2−2
2+ 2
∞
X
n=1 (−1)n
n+π2−2
n2cos(nx).
b) To determine the convergence of the Fourier cosine series of f(x), we need to check if f(x)is
piecewise smooth on [−π, π]. Since f(x) = x2is continuous and has derivative f′(x) = 2xexisting
for all x∈[−π, π], the Fourier cosine series converges to f(x).
c) The Fourier cosine series converges to f(x)for x∈(−π, π), so the value at which the Fourier
cosine series of f(x)converges to f(x)is x= 0.
5 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider a square wave signal x(t)given by:
x(t) = (1if 0≤t < T/2
−1if T/2≤t<T
where Tis the period of the signal.
a) Calculate the Fourier transform X(ω)of x(t).
b) Determine the value of X(0).
c) Find the power spectral density SX(ω)of the signal x(t).
Solution 1.
a) To calculate the Fourier transform X(ω)of the given square wave signal x(t), we apply the
definition of Fourier transform:
X(ω) = Z∞
−∞
x(t)e−jωtdt
Splitting the integral based on x(t):
X(ω) = ZT/2
0
e−jωtdt −ZT
T/2
e−jωtdt
Solving the integrals, we get:
X(ω) = 1−e−jωT/2
jω −e−jωT /2−e−jωT
jω
Simplifying further gives the Fourier transform as:
X(ω) = 2
jω (1 −cos(ωT /2))
b) To find the value of X(0), we substitute ω= 0 in the Fourier transform:
X(0) = 2
0(1 −cos(0)) = 0
c) The power spectral density SX(ω)is given by:
SX(ω) = |X(ω)|2
Substitute the Fourier transform X(ω)obtained earlier to get:
SX(ω) =
2
jω (1 −cos(ωT /2))
2
=4(1 −cos(ωT/2))2
ω2
6 CALCULATING FOURIER TRANSFORMS OF DISTRIBUTIONS
Problem 1. Compute the Fourier transform of the distribution T=δ(x−3) + δ(x+ 4), where δ
denotes the Dirac delta distribution.
Solution 1. The Fourier transform of a distribution Tis defined as
F(δ)(ϕ) = δ(F(ϕ))
where ϕis a test function.
a) Let’s first compute the Fourier transform of δ(x−3) using the formula:
δ(x−a)F
−→ e−2πiaξ
Thus, the Fourier transform of δ(x−3) is
F(δ(x−3)) = e−6πiξ
b) Next, we compute the Fourier transform of δ(x+ 4) which is given by:
δ(x+a)F
−→ e2πiaξ
Therefore, the Fourier transform of δ(x+ 4) is
F(δ(x+ 4)) = e8πiξ
c) Finally, computing the Fourier transform of the given distribution T:
F(T)(ϕ) = δ(F(ϕ))
=δ(e−6πiξ) + δ(e8πiξ )
=ϕ(6) + ϕ(−8)
Therefore, the Fourier transform of Tis ϕ(6) + ϕ(−8).
Problem 2. Determine the Fourier transform of the distribution T=x2·sin(2πx).
Solution 2. The Fourier transform of a distribution Tis given by:
F(xn·ϕ)(ξ)=(−i)n·dnF(ϕ)
dξn
where ϕis a test function.
We need to compute the Fourier transform of x2·sin(2πx).
a) First, we find the Fourier transforms of x2and sin(2πx)individually.
The Fourier transform of x2can be calculated using the formula:
F(x2)(ξ) = (−i)2·d2
dξ2δ
This simplifies to:
F(x2)(ξ) = −δ′′(ξ)
b) Next, we find the Fourier transform of sin(2πx):
F(sin(αx)) = α
α2−ξ2
Applying this to sin(2πx)gives:
F(sin(2πx)) = 2π
(2π)2−ξ2
c) Now, we calculate the Fourier transform of x2·sin(2πx)by convolving the individual trans-
forms:
F(x2·sin(2πx))(ξ) = −δ′′(ξ)∗2π
(2π)2−ξ2
=−2π
(2π)2−ξ2
Therefore, the Fourier transform of x2·sin(2πx)is −2π
(2π)2−ξ2.
7 "ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES"
Problem 1. Consider the function f(x) = cos(2πx) + 3 sin(4πx).
a) Find the period of f(x).
b) Compute the Fourier series of f(x).
Solution 1.
a) To find the period of f(x), we need to consider the period of each component function
cos(2πx)and sin(4πx)separately.
The period of cos(2πx)is given by T=2π
2π= 1.
Similarly, the period of sin(4πx)is T=2π
4π=1
2.
Since the period of f(x)is the least common multiple of the periods of its component functions,
the period of f(x)is T=lcm(1,1
2)=1.
Thus, the period of f(x)is 1.
b) The Fourier series of f(x)can be found by expressing it in the form:
f(x) = a0+
∞
X
n=1 ancos 2nπx
T+bnsin 2nπx
T
where Tis the period of f(x).
Since we found that T= 1 in part (a), we have:
f(x) = a0+
∞
X
n=1
[ancos(2nπx) + bnsin(2nπx)]
The coefficients can be calculated using the formulas:
a0=1
TZT
2
−T
2
f(x)dx
an=2
TZT
2
−T
2
f(x) cos 2nπx
Tdx
bn=2
TZT
2
−T
2
f(x) sin 2nπx
Tdx
Since T= 1, these formulas simplify to:
a0=Z0.5
−0.5
f(x)dx
an= 2 Z0.5
−0.5
f(x) cos(2nπx)dx
bn= 2 Z0.5
−0.5
f(x) sin(2nπx)dx
Plugging in the function f(x)and computing the integrals will give the coefficients for the Fourier
series of f(x).
8 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider a continuous-time signal x(t)in a given bandwidth, represented by its
Fourier transform X(f). The signal is sampled with a sampling rate of Fs= 200 Hz. Determine if
this sampling rate satisfies the Nyquist criterion for this signal. Justify your answer.
Solution 1.
Given that the signal is sampled with a sampling rate of Fs= 200 Hz, we need to check if this
sampling rate satisfies the Nyquist criterion.
The Nyquist theorem states that in order to perfectly reconstruct a signal from its samples, the
sampling rate must be at least twice the maximum frequency present in the signal. In other words,
Fs≥2B, where Bis the bandwidth of the signal.
In this case, since Fs= 200 Hz, we need to determine the bandwidth of the signal B. If X(f)is
non-zero only up to BHz, then the condition is satisfied.
If the signal’s Fourier transform X(f)has energy up to BHz, then the condition Fs≥2Bis
satisfied and the sampling rate of 200 Hz meets the Nyquist criterion for this signal.
Therefore, the sampling rate of 200 Hz satisfies the Nyquist criterion for the given continuous-
time signal.
This problem highlighted the importance of the Nyquist criterion in determining the appropriate
sampling rate for a signal in order to avoid aliasing.
9 "ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES"
Problem 1. Consider a non-stationary signal x(t) = sin(2πt2).
a) Find the Fourier transform X(ω)of the signal.
b) Determine the time-frequency representation of the signal using the Short-Time Fourier
Transform (STFT) with a Gaussian window function g(t) = e−t2.
Solution 1.
a) To find the Fourier transform X(ω)of the signal x(t) = sin(2πt2), we use the definition of the
Fourier transform:
X(ω) = F{x(t)}=Z∞
−∞
x(t)e−jωtdt
Substitute x(t) = sin(2πt2)into the integral:
X(ω) = Z∞
−∞
sin(2πt2)e−jωtdt
This integral does not have a closed-form solution, but can be numerically evaluated.
b) The Short-Time Fourier Transform (STFT) of a signal x(t)with a window function g(t)is given
by:
X(ω, τ) = Z∞
−∞
x(t)g(t−τ)e−jωtdt
For the given signal x(t) = sin(2πt2)and Gaussian window function g(t) = e−t2, we can com-
pute the STFT X(ω, τ)by sliding the window along the time axis and computing the Fourier trans-
form for each windowed segment.
The time-frequency representation of the signal using STFT provides information about how
the frequency content of the signal evolves over time, which is particularly useful for analyzing
non-stationary signals like x(t) = sin(2πt2).
10 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider a function f(x) = (xif 0≤x≤π
0otherwise
a) Compute the Fourier sine series of f(x).
b) Find the sum of the series.
Solution 1.
a) The Fourier sine series for a function f(x)defined on 0≤x≤Lis given by:
f(x)∼
∞
X
n=1
Bnsin nπx
L
where Bn=2
LRL
0f(x) sin nπx
Ldx.
For the function f(x)given in the problem, L=πand f(x) = xfor 0≤x≤π. Thus, we need
to compute Bn:
Bn=2
πZπ
0
xsin(nx)dx
=2
π−x
ncos(nx)
π
0
+1
nZπ
0
cos(nx)dx
=2
π0−(−1)nπ
n+1
n2sin(nx)
π
0
= 0 + 2(−1)n
n2+2(−1)n
πn
=2(−1)n
n21 + 1
πn
Therefore, the Fourier sine series of f(x)is:
f(x)∼
∞
X
n=1
2(−1)n
n21 + 1
πnsin(nx)
b) To find the sum of the series, we note that the Fourier sine series of an odd function converges
to f(x−)+f(x+)
2.
For the given function f(x) = x,f(x−) = xand f(x+)=0(at the jump discontinuity x= 0).
Thus, the sum of the series is:
x+ 0
2=x
2
Therefore, the sum of the Fourier sine series of f(x)is x
2.
11 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider a function f(t) = cos(2πt)in the time domain. Compute the Fourier
transform F(ω)of f(t), where F(ω)is defined as:
F(ω) = Z∞
−∞
f(t)e−iωtdt
Solution 1. To compute the Fourier transform F(ω)of f(t) = cos(2πt), we substitute f(t)into
the definition of the Fourier transform:
F(ω) = Z∞
−∞
cos(2πt)e−iωtdt
Using Euler’s formula eix = cos(x) + isin(x), we can express the integrand as:
cos(2πt)e−iωt =1
2(e2πit +e−2πit)1
2i(eiωt −e−iωt)
Multiplying and simplifying, the integrand becomes:
1
4i(e(2πi−ωi)t−e(−2πi+ωi)t+e(2πi+ωi)t−e(−2πi−ωi)t)
Now, we can integrate term by term:
F(ω) = 1
4i"e(2πi−ωi)t
2πi −ωi −e(−2πi+ωi)t
−2πi +ωi +e(2πi+ωi)t
2πi +ωi −e(−2πi−ωi)t
−2πi −ωi #
∞
−∞
Given that the oscillatory terms will go to zero at ±∞ due to periodicity, we are left with evalu-
ating the residues at the poles ω=±2π. This gives us the final result:
F(ω) = π(δ(ω−2π) + δ(ω+ 2π))
where δ(ω)is the Dirac delta function.
So, the Fourier transform of f(t) = cos(2πt)is F(ω) = π(δ(ω−2π) + δ(ω+ 2π)).
12 NONLINEAR SIGNAL PROCESSING USING HARMONIC ANALYSIS AND FOURIER TRANS-
FORM
Problem 1.
Consider a nonlinear signal x(t) = cos(2πt) + cos(4πt). Determine the Fourier series represen-
tation of x(t)over the interval [−T, T ], where Tis the period of the signal.
Solution 1.
Given signal x(t) = cos(2πt) + cos(4πt), we first need to find the period of the signal by calcu-
lating the least common multiple of the periods of the individual cosine components.
The period of cos(2πt)is T1= 1 and the period of cos(4πt)is T2=1
2.
The period of the combined signal is T=LCM(T1, T2) = LCM(1,1
2) = 1.
Therefore, the Fourier series representation of x(t)over the interval [−1,1] is:
x(t) = a0+
∞
X
n=1
(ancos(nπt) + bnsin(nπt))
where an=2
TRT
−Tx(t) cos(nπt)dt and bn=2
TRT
−Tx(t) sin(nπt)dt.
Calculating the coefficients:
a0=1
TZ1
−1
x(t)dt =1
1Z1
−1
(cos(2πt) + cos(4πt))dt
an=2
TZ1
−1
x(t) cos(nπt)dt =2
1Z1
−1
(cos(2πt) + cos(4πt)) cos(nπt)dt
bn=2
TZ1
−1
x(t) sin(nπt)dt =2
1Z1
−1
(cos(2πt) + cos(4πt)) sin(nπt)dt
By performing the integrations, we can find the Fourier series representation of x(t)over the
interval [−1,1].
13 "WAVELET ANALYSIS AND APPLICATIONS IN SIGNAL PROCESSING"
Problem 13. Consider a signal f(t) = sin(4t) + sin(8t). Perform the following wavelet analysis
using the Morlet wavelet:
ψ(t) = eiω0te−t2/2
where ω0= 5.
a) Compute the continuous wavelet transform (CWT) of f(t).
b) Sketch the scalogram of the CWT.
c) Determine the dominant frequency component(s) present in the signal.
Solution 13.
a) To compute the CWT of f(t), we use the formula:
Wf(a, b) = Z∞
−∞
f(t)·¯
ψt−b
adt,
where ais the dilation parameter and bis the translation parameter.
For this problem, f(t) = sin(4t) + sin(8t)and ψ(t) = ei·5te−t2/2.
The CWT is given by:
Wf(a, b) = Z∞
−∞
[sin(4t) + sin(8t)] ·e−i·5(t−b)/ae−(t−b)2/(2a2)dt.
This integral can be evaluated by numerical methods or special software such as MATLAB to
find the values of Wf(a, b).
b) The scalogram is a visual representation of the CWT. It shows the magnitude of the CWT
coefficients at each scale aand time b. To sketch the scalogram, we plot the magnitude of |Wf(a, b)|
at different aand bvalues.
c) By examining the scalogram, we can identify the dominant frequency components by look-
ing at the peaks in the magnitude of the CWT coefficients. In this case, the dominant frequency
components correspond to the frequencies that have the highest magnitude values in the CWT.
14 "APPLICATIONS OF WAVELET TRANSFORM IN SIGNAL PROCESSING"
Problem 14. Consider a signal f(t)with the following data:
f(t) = (1,if 0≤t≤1
0,otherwise
Let’s analyze this signal using the Haar wavelet transform.
a) Calculate the Haar wavelet coefficients ck,m corresponding to f(t).
b) Determine the wavelet transform of f(t).
c) Reconstruct the original signal using the inverse Haar wavelet transform from the coefficients
found in part (a).
Solution 14.
a) The Haar wavelet coefficients ck,m can be calculated as:
ck,m =Z∞
−∞
f(t)ψk,m(t)dt
Here, ψk,m(t)is the Haar wavelet function. Since the Haar wavelet is a piecewise constant
function, we have:
ψk,m(t) = (2k
2,if 2km≤t < 2k(m+ 1)
0,otherwise
Given f(t), we can calculate the coefficients by integrating over the intervals where ψk,m(t)is
non-zero. For the Haar wavelet, we only need to consider the intervals [0,1) and [1,2).
For the interval [0,1):
c0,0=Z1
0
f(t)ψ0,0(t)dt
c0,0=Z1
0
(1) ·(20
2)dt
c0,0=Z1
0
dt = 1
For the interval [1,2):
c1,0=Z2
1
f(t)ψ1,0(t)dt
c1,0=Z2
1
(0) ·(21
2)dt
c1,0= 0
Therefore, the Haar wavelet coefficients for f(t)are c0,0= 1 and c1,0= 0.
b) The wavelet transform of f(t)is given by:
Wf(a, b) = X
k,m
ck,m ·ψk,m(t)
Since c1,0= 0, the wavelet transform simplifies to:
Wf(a, b) = c0,0·ψ0,0(t)=1·20
2= 1
c) Reconstructing the original signal using the inverse Haar wavelet transform involves summing
the scaled and shifted Haar wavelets:
f(t) = X
k,m
ck,m ·ψk,m(t)
Substitute the coefficients c0,0= 1 and c1,0= 0 into the formula:
f(t) = 1 ·ψ0,0(t) = 1 ·20
2= 1
Therefore, the reconstructed signal using the inverse Haar wavelet transform is a constant
signal f(t)=1.
15 "APPLICATION OF WAVELET TRANSFORM IN SIGNAL DENOISING AND COMPRES-
SION"
Problem 15. Consider a signal f(t)defined as:
f(t) = 2 sin(4πt) + 5 cos(8πt) + 3 sin(12πt)
a) Using the wavelet transform, determine the frequency components present in the signal
f(t). b) If we apply a wavelet transform and find that the wavelet coefficients corresponding to the
frequency component 4πare very small compared to the coefficients at other frequencies, explain
why this might indicate the presence of noise in the signal. c) How can wavelet transforms be used
in denoising this signal?
Solution 15.
a) To determine the frequency components present in the signal f(t), we need to apply the
wavelet transform. Let’s consider the Haar wavelet transform for simplicity. The Haar wavelet
transform of a signal f(t)is given by:
W(a, b) = Z∞
−∞
f(t)ψa,b(t)dt
Where ψa,b(t)is the Haar wavelet defined as:
ψa,b(t) =
1if a≤t<a+1
2
−1if a+1
2≤t<a+ 1
0otherwise
Let’s calculate the wavelet coefficients for the given signal f(t)at different scales a:
W(a, b) = Z∞
−∞
f(t)ψa,b(t)dt
For a= 1, we have:
W(1, b) = Z∞
−∞
(2 sin(4πt) + 5 cos(8πt) + 3 sin(12πt))ψ1,b(t)dt
Similarly, we can calculate W(2, b),W(3, b), and so on to determine the frequency components
present in the signal f(t).
b) If the wavelet coefficients corresponding to the frequency component 4πare very small com-
pared to the coefficients at other frequencies, it might indicate the presence of noise in the signal.
This is because in the absence of noise, the coefficients for the true signal frequencies would be
much larger compared to any noise present in the signal.
c) Wavelet transforms can be used in denoising the signal by thresholding the wavelet coeffi-
cients. By setting small coefficients below a certain threshold to zero, we can eliminate noise while
retaining the important frequency components of the signal. This is a common technique in signal
denoising using wavelet transforms.
16 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 17. Consider a non-stationary signal given by f(t) = sin(2πt2).
a) Determine the Fourier transform of the signal f(t).
b) Compute the time-frequency representation of f(t)using the Short-Time Fourier Transform
(STFT) with a Gaussian window function g(t)given by g(t) = e−t2.
Solution 17.
a) The Fourier transform of a signal f(t)is defined as
F(ω) = Z∞
−∞
f(t)e−iωtdt.
For the given signal f(t) = sin(2πt2), we have
F(ω) = Z∞
−∞
sin(2πt2)e−iωtdt.
This particular integral does not have a closed-form solution, so we would need to evaluate it using
numerical methods or approximation techniques.
b) The Short-Time Fourier Transform (STFT) of a signal f(t)using a window function g(t)is
defined as
ST F T (t, ω) = Z∞
−∞
f(τ)g(τ−t)e−iωτ dτ,
where the overline denotes complex conjugation. For a Gaussian window function g(t) = e−t2, the
STFT of the signal f(t) = sin(2πt2)can be computed as
ST F T (t, ω) = Z∞
−∞
sin(2πτ2)e−(τ−t)2e−iωτ dτ.
We can evaluate this integral numerically to compute the time-frequency representation of the
signal using the given Gaussian window function.
17 "CONVERGENCE AND DIVERGENCE OF FOURIER SERIES FOR PIECEWISE FUNC-
TIONS"
Problem 17. Consider the function f(x)defined as:
f(x) = (x0≤x<π
−x−π < x < 0
a) Find the Fourier series of f(x).
b) Determine the conditions for which the Fourier series of f(x)converges.
Solution 17.
a) To find the Fourier series of f(x), we need to compute the coefficients anand bn:
an=1
πZπ
−π
f(x) cos(nx)dx
bn=1
πZπ
−π
f(x) sin(nx)dx
Let’s start with the calculation of an:
an=1
πZ0
−π
(−x) cos(nx)dx +Zπ
0
xcos(nx)dx
Solving these integrals and simplifying yields:
an=2((−1)n−1)
πn2
Now, let’s find bn:
bn=1
πZ0
−π
(−x) sin(nx)dx +Zπ
0
xsin(nx)dx
After solving and simplifying the integrals, we get:
bn= 0
The Fourier series of f(x)is then given by:
f(x)∼2
π1
12−1
32cos(x) + 1
52cos(2x)− · · ·
b) The Fourier series of f(x)converges if the function f(x)satisfies the Dirichlet conditions. In
this case, since f(x)is piecewise continuous and has a finite number of discontinuities within a
period of 2π, the Fourier series of f(x)converges everywhere.
18 "NON-LINEAR FOURIER TRANSFORM TECHNIQUES FOR SIGNAL ANALYSIS"
Problem 1. Consider a signal f(t) = sin(2πt) + 2 cos(4πt).
a) Compute the Fourier Transform of f(t).
b) Find the frequency components present in the signal f(t).
c) Determine the amplitude and phase of each frequency component.
Solution 1.
a) The Fourier Transform of a signal f(t)is defined as:
F(ω) = Z∞
−∞
f(t)e−jωtdt
Applying the definition to the given signal f(t), we have:
F(ω) = Z∞
−∞
[sin(2πt) + 2 cos(4πt)]e−jωtdt
=Z∞
−∞
sin(2πt)e−jωtdt + 2 Z∞
−∞
cos(4πt)e−jωtdt
Solving the integrals, we get:
F(ω) = π(δ(ω−2π)−δ(ω+ 2π)) + π(δ(ω−4π) + δ(ω+ 4π))
b) The frequency components present in the signal f(t)are 2πand 4π.
c) To determine the amplitude and phase of each frequency component, we can use the prop-
erties of the Dirac delta function in the Fourier Transform. The amplitude is the coefficient in front
of the delta function and the phase is given by the sign of the delta function. Thus, we have:
For 2πcomponent: Amplitude = πand Phase = (δ(ω−2π)−δ(ω+ 2π)) = 0
For 4πcomponent: Amplitude = πand Phase = (δ(ω−4π) + δ(ω+ 4π)) = 0
19 "ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES"
Problem 1. Consider a continuous-time signal x(t)defined as:
x(t) = 2 cos(3πt) + 3 sin(6πt)
Find the Fourier transform X(ω)of the signal x(t).
Solution 1.
The Fourier transform of a continuous-time signal x(t)is defined as:
X(ω) = Z∞
−∞
x(t)e−jωtdt
Plugging in the expression for x(t), we have:
X(ω) = Z∞
−∞
(2 cos(3πt) + 3 sin(6πt))e−jωtdt
Expanding the trigonometric functions using Euler’s formula, we get:
X(ω) = Z∞
−∞
2(ej3πt +e−j3πt
2) + 3(ej6πt −e−j6πt
2j)e−jωtdt
X(ω) = Z∞
−∞
(ej(3π−ω)t+e−j(3π+ω)t−3
2j(ej(6π−ω)t−e−j(6π+ω)t))dt
Remembering that the Fourier transform has a property with delta functions, we find:
X(ω) = 2π(δ(ω−3π) + δ(ω+ 3π)) −3π(δ(ω−6π) + δ(ω+ 6π))
Thus, the Fourier transform of the signal x(t)is:
X(ω) = 2π(δ(ω−3π) + δ(ω+ 3π)) −3π(δ(ω−6π) + δ(ω+ 6π))
Certainly! Here is a numerical problem in the realm of Advanced Harmonic Analysis and Fourier
Transform Techniques with a step-by-step solution:
20 "ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES"
Problem 21. Consider a non-stationary signal f(t)defined as:
f(t) = (sin(2πt) 0 ≤t < 1
0otherwise
Calculate the Fourier transform F(ω)of the signal f(t).
Solution 21. To find the Fourier transform F(ω)of the given signal f(t), we use the definition
of the Fourier transform:
F(ω) = Z∞
−∞
f(t)e−iωtdt
Given that f(t)is defined piecewise, we split the integral into two parts:
F(ω) = Z1
0
sin(2πt)e−iωtdt +Z∞
1
0·e−iωtdt
Solving the first integral, we can proceed to calculate the Fourier transform. Let’s solve the
integral as follows:
F(ω) = Z1
0
sin(2πt)e−iωtdt
Applying Euler’s formula to express the sine function in terms of complex exponentials, we have:
F(ω) = Z1
0
ei2πt −e−i2πt
2ie−iωtdt
F(ω) = 1
2iZ1
0
ei(2π−ω)tdt −Z1
0
e−i(2π+ω)tdt
F(ω) = 1
2i"ei(2π−ω)t
i(2π−ω)
1
0−e−i(2π+ω)t
−i(2π+ω)
1
0#
F(ω) = 1
2i"ei(2π−ω)−1
i(2π−ω)−e−i(2π+ω)−1
−i(2π+ω)#
This integral evaluation gives us the Fourier transform F(ω).
a) We can calculate the Fourier series coefficients anand bnusing the following formulas:
an=1
LZL
−L
f(x) cos 2πnx
Ldx
bn=1
LZL
−L
f(x) sin 2πnx
Ldx
where Lis the period of the function.
For f(x) = sin(2πx) + cos(4πx), the period is L= 1. Thus, we have:
an=Z1
−1
(sin(2πx) + cos(4πx)) cos(2πnx)dx
bn=Z1
−1
(sin(2πx) + cos(4πx)) sin(2πnx)dx
Calculating these integrals will give us the Fourier series coefficients anand bn.
b) To find the constant term in the Fourier series of f(x), we need to determine a0, which is
given by:
a0=1
LZL
−L
f(x)dx
Substitute L= 1 and integral bounds −1to 1to find the value of a0.
c) The representation of f(x)in its Fourier series form is given by:
f(x)≈a0
2+
∞
X
n=1 ancos 2πnx
L+bnsin 2πnx
L
Substitute the calculated values of a0,an, and bninto this equation to obtain the Fourier series
representation of f(x).
3 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider the function f(t) = sin(3t) + cos(5t).
a) Find the Fourier transform of f(t).
b) Determine the frequency components present in f(t).
Solution 1.
a) The Fourier transform of a function f(t)is defined as:
F(ω) = Z∞
−∞
f(t)e−iωtdt
Applying this definition to our function f(t) = sin(3t) + cos(5t), we have:
F(ω) = Z∞
−∞
(sin(3t) + cos(5t))e−iωtdt
=Z∞
−∞
sin(3t)e−iωtdt +Z∞
−∞
cos(5t)e−iωtdt
We can solve these integrals using the properties of the Fourier transform.
b) The frequency components present in f(t)are the frequencies of the sine and cosine func-
tions, which are 3and 5respectively.
Therefore, the frequency components present in f(t)are 3and 5.
4 "ANALYZING THE CONVERGENCE AND DIVERGENCE OF FOURIER SERIES REPRE-
SENTATIONS"
Problem 4. Consider the function f(x) = x2defined on the interval [−π, π].
a) Find the Fourier cosine series representation of f(x).
b) Determine the convergence of the Fourier cosine series of f(x).
c) Calculate the value at which the Fourier cosine series of f(x)converges to f(x).
Solution 4.
a) To find the Fourier cosine series representation of f(x) = x2, we need to calculate the
coefficients anusing the formula:
an=2
πZπ
0
x2cos(nx)dx.
Calculating the integral gives:
an=2
πZπ
0
x2cos(nx)dx
=2
π2(nx sin(nx)+(n2x2−2) cos(nx))
n2
π
0
=2
π2nπ sin(nπ)
n2+n2π2−2
n2
= 2 (−1)n
n+π2−2
n2.
Therefore, the Fourier cosine series representation of f(x)is:
f(x) = π2−2
2+ 2
∞
X
n=1 (−1)n
n+π2−2
n2cos(nx).
b) To determine the convergence of the Fourier cosine series of f(x), we need to check if f(x)is
piecewise smooth on [−π, π]. Since f(x) = x2is continuous and has derivative f′(x) = 2xexisting
for all x∈[−π, π], the Fourier cosine series converges to f(x).
c) The Fourier cosine series converges to f(x)for x∈(−π, π), so the value at which the Fourier
cosine series of f(x)converges to f(x)is x= 0.
5 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider a square wave signal x(t)given by:
x(t) = (1if 0≤t < T/2
−1if T/2≤t<T
where Tis the period of the signal.
a) Calculate the Fourier transform X(ω)of x(t).
b) Determine the value of X(0).
c) Find the power spectral density SX(ω)of the signal x(t).
Solution 1.
a) To calculate the Fourier transform X(ω)of the given square wave signal x(t), we apply the
definition of Fourier transform:
X(ω) = Z∞
−∞
x(t)e−jωtdt
Splitting the integral based on x(t):
X(ω) = ZT/2
0
e−jωtdt −ZT
T/2
e−jωtdt
Solving the integrals, we get:
X(ω) = 1−e−jωT/2
jω −e−jωT /2−e−jωT
jω
Simplifying further gives the Fourier transform as:
X(ω) = 2
jω (1 −cos(ωT /2))
b) To find the value of X(0), we substitute ω= 0 in the Fourier transform:
X(0) = 2
0(1 −cos(0)) = 0
c) The power spectral density SX(ω)is given by:
SX(ω) = |X(ω)|2
Substitute the Fourier transform X(ω)obtained earlier to get:
SX(ω) =
2
jω (1 −cos(ωT /2))
2
=4(1 −cos(ωT/2))2
ω2
6 CALCULATING FOURIER TRANSFORMS OF DISTRIBUTIONS
Problem 1. Compute the Fourier transform of the distribution T=δ(x−3) + δ(x+ 4), where δ
denotes the Dirac delta distribution.
Solution 1. The Fourier transform of a distribution Tis defined as
F(δ)(ϕ) = δ(F(ϕ))
where ϕis a test function.
a) Let’s first compute the Fourier transform of δ(x−3) using the formula:
δ(x−a)F
−→ e−2πiaξ
Thus, the Fourier transform of δ(x−3) is
F(δ(x−3)) = e−6πiξ
b) Next, we compute the Fourier transform of δ(x+ 4) which is given by:
δ(x+a)F
−→ e2πiaξ
Therefore, the Fourier transform of δ(x+ 4) is
F(δ(x+ 4)) = e8πiξ
c) Finally, computing the Fourier transform of the given distribution T:
F(T)(ϕ) = δ(F(ϕ))
=δ(e−6πiξ) + δ(e8πiξ )
=ϕ(6) + ϕ(−8)
Therefore, the Fourier transform of Tis ϕ(6) + ϕ(−8).
Problem 2. Determine the Fourier transform of the distribution T=x2·sin(2πx).
Solution 2. The Fourier transform of a distribution Tis given by:
F(xn·ϕ)(ξ)=(−i)n·dnF(ϕ)
dξn
where ϕis a test function.
We need to compute the Fourier transform of x2·sin(2πx).
a) First, we find the Fourier transforms of x2and sin(2πx)individually.
The Fourier transform of x2can be calculated using the formula:
F(x2)(ξ) = (−i)2·d2
dξ2δ
This simplifies to:
F(x2)(ξ) = −δ′′(ξ)
b) Next, we find the Fourier transform of sin(2πx):
F(sin(αx)) = α
α2−ξ2
Applying this to sin(2πx)gives:
F(sin(2πx)) = 2π
(2π)2−ξ2
c) Now, we calculate the Fourier transform of x2·sin(2πx)by convolving the individual trans-
forms:
F(x2·sin(2πx))(ξ) = −δ′′(ξ)∗2π
(2π)2−ξ2
=−2π
(2π)2−ξ2
Therefore, the Fourier transform of x2·sin(2πx)is −2π
(2π)2−ξ2.
7 "ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES"
Problem 1. Consider the function f(x) = cos(2πx) + 3 sin(4πx).
a) Find the period of f(x).
b) Compute the Fourier series of f(x).
Solution 1.
a) To find the period of f(x), we need to consider the period of each component function
cos(2πx)and sin(4πx)separately.
The period of cos(2πx)is given by T=2π
2π= 1.
Similarly, the period of sin(4πx)is T=2π
4π=1
2.
Since the period of f(x)is the least common multiple of the periods of its component functions,
the period of f(x)is T=lcm(1,1
2)=1.
Thus, the period of f(x)is 1.
b) The Fourier series of f(x)can be found by expressing it in the form:
f(x) = a0+
∞
X
n=1 ancos 2nπx
T+bnsin 2nπx
T
where Tis the period of f(x).
Since we found that T= 1 in part (a), we have:
f(x) = a0+
∞
X
n=1
[ancos(2nπx) + bnsin(2nπx)]
The coefficients can be calculated using the formulas:
a0=1
TZT
2
−T
2
f(x)dx
an=2
TZT
2
−T
2
f(x) cos 2nπx
Tdx
bn=2
TZT
2
−T
2
f(x) sin 2nπx
Tdx
Since T= 1, these formulas simplify to:
a0=Z0.5
−0.5
f(x)dx
an= 2 Z0.5
−0.5
f(x) cos(2nπx)dx
bn= 2 Z0.5
−0.5
f(x) sin(2nπx)dx
Plugging in the function f(x)and computing the integrals will give the coefficients for the Fourier
series of f(x).
8 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider a continuous-time signal x(t)in a given bandwidth, represented by its
Fourier transform X(f). The signal is sampled with a sampling rate of Fs= 200 Hz. Determine if
this sampling rate satisfies the Nyquist criterion for this signal. Justify your answer.
Solution 1.
Given that the signal is sampled with a sampling rate of Fs= 200 Hz, we need to check if this
sampling rate satisfies the Nyquist criterion.
The Nyquist theorem states that in order to perfectly reconstruct a signal from its samples, the
sampling rate must be at least twice the maximum frequency present in the signal. In other words,
Fs≥2B, where Bis the bandwidth of the signal.
In this case, since Fs= 200 Hz, we need to determine the bandwidth of the signal B. If X(f)is
non-zero only up to BHz, then the condition is satisfied.
If the signal’s Fourier transform X(f)has energy up to BHz, then the condition Fs≥2Bis
satisfied and the sampling rate of 200 Hz meets the Nyquist criterion for this signal.
Therefore, the sampling rate of 200 Hz satisfies the Nyquist criterion for the given continuous-
time signal.
This problem highlighted the importance of the Nyquist criterion in determining the appropriate
sampling rate for a signal in order to avoid aliasing.
9 "ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES"
Problem 1. Consider a non-stationary signal x(t) = sin(2πt2).
a) Find the Fourier transform X(ω)of the signal.
b) Determine the time-frequency representation of the signal using the Short-Time Fourier
Transform (STFT) with a Gaussian window function g(t) = e−t2.
Solution 1.
a) To find the Fourier transform X(ω)of the signal x(t) = sin(2πt2), we use the definition of the
Fourier transform:
X(ω) = F{x(t)}=Z∞
−∞
x(t)e−jωtdt
Substitute x(t) = sin(2πt2)into the integral:
X(ω) = Z∞
−∞
sin(2πt2)e−jωtdt
This integral does not have a closed-form solution, but can be numerically evaluated.
b) The Short-Time Fourier Transform (STFT) of a signal x(t)with a window function g(t)is given
by:
X(ω, τ) = Z∞
−∞
x(t)g(t−τ)e−jωtdt
For the given signal x(t) = sin(2πt2)and Gaussian window function g(t) = e−t2, we can com-
pute the STFT X(ω, τ)by sliding the window along the time axis and computing the Fourier trans-
form for each windowed segment.
The time-frequency representation of the signal using STFT provides information about how
the frequency content of the signal evolves over time, which is particularly useful for analyzing
non-stationary signals like x(t) = sin(2πt2).
10 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider a function f(x) = (xif 0≤x≤π
0otherwise
a) Compute the Fourier sine series of f(x).
b) Find the sum of the series.
Solution 1.
a) The Fourier sine series for a function f(x)defined on 0≤x≤Lis given by:
f(x)∼
∞
X
n=1
Bnsin nπx
L
where Bn=2
LRL
0f(x) sin nπx
Ldx.
For the function f(x)given in the problem, L=πand f(x) = xfor 0≤x≤π. Thus, we need
to compute Bn:
Bn=2
πZπ
0
xsin(nx)dx
=2
π−x
ncos(nx)
π
0
+1
nZπ
0
cos(nx)dx
=2
π0−(−1)nπ
n+1
n2sin(nx)
π
0
= 0 + 2(−1)n
n2+2(−1)n
πn
=2(−1)n
n21 + 1
πn
Therefore, the Fourier sine series of f(x)is:
f(x)∼
∞
X
n=1
2(−1)n
n21 + 1
πnsin(nx)
b) To find the sum of the series, we note that the Fourier sine series of an odd function converges
to f(x−)+f(x+)
2.
For the given function f(x) = x,f(x−) = xand f(x+)=0(at the jump discontinuity x= 0).
Thus, the sum of the series is:
x+ 0
2=x
2
Therefore, the sum of the Fourier sine series of f(x)is x
2.
11 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider a function f(t) = cos(2πt)in the time domain. Compute the Fourier
transform F(ω)of f(t), where F(ω)is defined as:
F(ω) = Z∞
−∞
f(t)e−iωtdt
Solution 1. To compute the Fourier transform F(ω)of f(t) = cos(2πt), we substitute f(t)into
the definition of the Fourier transform:
F(ω) = Z∞
−∞
cos(2πt)e−iωtdt
Using Euler’s formula eix = cos(x) + isin(x), we can express the integrand as:
cos(2πt)e−iωt =1
2(e2πit +e−2πit)1
2i(eiωt −e−iωt)
Multiplying and simplifying, the integrand becomes:
1
4i(e(2πi−ωi)t−e(−2πi+ωi)t+e(2πi+ωi)t−e(−2πi−ωi)t)
Now, we can integrate term by term:
F(ω) = 1
4i"e(2πi−ωi)t
2πi −ωi −e(−2πi+ωi)t
−2πi +ωi +e(2πi+ωi)t
2πi +ωi −e(−2πi−ωi)t
−2πi −ωi #
∞
−∞
Given that the oscillatory terms will go to zero at ±∞ due to periodicity, we are left with evalu-
ating the residues at the poles ω=±2π. This gives us the final result:
F(ω) = π(δ(ω−2π) + δ(ω+ 2π))
where δ(ω)is the Dirac delta function.
So, the Fourier transform of f(t) = cos(2πt)is F(ω) = π(δ(ω−2π) + δ(ω+ 2π)).
12 NONLINEAR SIGNAL PROCESSING USING HARMONIC ANALYSIS AND FOURIER TRANS-
FORM
Problem 1.
Consider a nonlinear signal x(t) = cos(2πt) + cos(4πt). Determine the Fourier series represen-
tation of x(t)over the interval [−T, T ], where Tis the period of the signal.
Solution 1.
Given signal x(t) = cos(2πt) + cos(4πt), we first need to find the period of the signal by calcu-
lating the least common multiple of the periods of the individual cosine components.
The period of cos(2πt)is T1= 1 and the period of cos(4πt)is T2=1
2.
The period of the combined signal is T=LCM(T1, T2) = LCM(1,1
2) = 1.
Therefore, the Fourier series representation of x(t)over the interval [−1,1] is:
x(t) = a0+
∞
X
n=1
(ancos(nπt) + bnsin(nπt))
where an=2
TRT
−Tx(t) cos(nπt)dt and bn=2
TRT
−Tx(t) sin(nπt)dt.
Calculating the coefficients:
a0=1
TZ1
−1
x(t)dt =1
1Z1
−1
(cos(2πt) + cos(4πt))dt
an=2
TZ1
−1
x(t) cos(nπt)dt =2
1Z1
−1
(cos(2πt) + cos(4πt)) cos(nπt)dt
bn=2
TZ1
−1
x(t) sin(nπt)dt =2
1Z1
−1
(cos(2πt) + cos(4πt)) sin(nπt)dt
By performing the integrations, we can find the Fourier series representation of x(t)over the
interval [−1,1].
13 "WAVELET ANALYSIS AND APPLICATIONS IN SIGNAL PROCESSING"
Problem 13. Consider a signal f(t) = sin(4t) + sin(8t). Perform the following wavelet analysis
using the Morlet wavelet:
ψ(t) = eiω0te−t2/2
where ω0= 5.
a) Compute the continuous wavelet transform (CWT) of f(t).
b) Sketch the scalogram of the CWT.
c) Determine the dominant frequency component(s) present in the signal.
Solution 13.
a) To compute the CWT of f(t), we use the formula:
Wf(a, b) = Z∞
−∞
f(t)·¯
ψt−b
adt,
where ais the dilation parameter and bis the translation parameter.
For this problem, f(t) = sin(4t) + sin(8t)and ψ(t) = ei·5te−t2/2.
The CWT is given by:
Wf(a, b) = Z∞
−∞
[sin(4t) + sin(8t)] ·e−i·5(t−b)/ae−(t−b)2/(2a2)dt.
This integral can be evaluated by numerical methods or special software such as MATLAB to
find the values of Wf(a, b).
b) The scalogram is a visual representation of the CWT. It shows the magnitude of the CWT
coefficients at each scale aand time b. To sketch the scalogram, we plot the magnitude of |Wf(a, b)|
at different aand bvalues.
c) By examining the scalogram, we can identify the dominant frequency components by look-
ing at the peaks in the magnitude of the CWT coefficients. In this case, the dominant frequency
components correspond to the frequencies that have the highest magnitude values in the CWT.
14 "APPLICATIONS OF WAVELET TRANSFORM IN SIGNAL PROCESSING"
Problem 14. Consider a signal f(t)with the following data:
f(t) = (1,if 0≤t≤1
0,otherwise
Let’s analyze this signal using the Haar wavelet transform.
a) Calculate the Haar wavelet coefficients ck,m corresponding to f(t).
b) Determine the wavelet transform of f(t).
c) Reconstruct the original signal using the inverse Haar wavelet transform from the coefficients
found in part (a).
Solution 14.
a) The Haar wavelet coefficients ck,m can be calculated as:
ck,m =Z∞
−∞
f(t)ψk,m(t)dt
Here, ψk,m(t)is the Haar wavelet function. Since the Haar wavelet is a piecewise constant
function, we have:
ψk,m(t) = (2k
2,if 2km≤t < 2k(m+ 1)
0,otherwise
Given f(t), we can calculate the coefficients by integrating over the intervals where ψk,m(t)is
non-zero. For the Haar wavelet, we only need to consider the intervals [0,1) and [1,2).
For the interval [0,1):
c0,0=Z1
0
f(t)ψ0,0(t)dt
c0,0=Z1
0
(1) ·(20
2)dt
c0,0=Z1
0
dt = 1
For the interval [1,2):
c1,0=Z2
1
f(t)ψ1,0(t)dt
c1,0=Z2
1
(0) ·(21
2)dt
c1,0= 0
Therefore, the Haar wavelet coefficients for f(t)are c0,0= 1 and c1,0= 0.
b) The wavelet transform of f(t)is given by:
Wf(a, b) = X
k,m
ck,m ·ψk,m(t)
Since c1,0= 0, the wavelet transform simplifies to:
Wf(a, b) = c0,0·ψ0,0(t)=1·20
2= 1
c) Reconstructing the original signal using the inverse Haar wavelet transform involves summing
the scaled and shifted Haar wavelets:
f(t) = X
k,m
ck,m ·ψk,m(t)
Substitute the coefficients c0,0= 1 and c1,0= 0 into the formula:
f(t) = 1 ·ψ0,0(t) = 1 ·20
2= 1
Therefore, the reconstructed signal using the inverse Haar wavelet transform is a constant
signal f(t)=1.
15 "APPLICATION OF WAVELET TRANSFORM IN SIGNAL DENOISING AND COMPRES-
SION"
Problem 15. Consider a signal f(t)defined as:
f(t) = 2 sin(4πt) + 5 cos(8πt) + 3 sin(12πt)
a) Using the wavelet transform, determine the frequency components present in the signal
f(t). b) If we apply a wavelet transform and find that the wavelet coefficients corresponding to the
frequency component 4πare very small compared to the coefficients at other frequencies, explain
why this might indicate the presence of noise in the signal. c) How can wavelet transforms be used
in denoising this signal?
Solution 15.
a) To determine the frequency components present in the signal f(t), we need to apply the
wavelet transform. Let’s consider the Haar wavelet transform for simplicity. The Haar wavelet
transform of a signal f(t)is given by:
W(a, b) = Z∞
−∞
f(t)ψa,b(t)dt
Where ψa,b(t)is the Haar wavelet defined as:
ψa,b(t) =
1if a≤t<a+1
2
−1if a+1
2≤t<a+ 1
0otherwise
Let’s calculate the wavelet coefficients for the given signal f(t)at different scales a:
W(a, b) = Z∞
−∞
f(t)ψa,b(t)dt
For a= 1, we have:
W(1, b) = Z∞
−∞
(2 sin(4πt) + 5 cos(8πt) + 3 sin(12πt))ψ1,b(t)dt
Similarly, we can calculate W(2, b),W(3, b), and so on to determine the frequency components
present in the signal f(t).
b) If the wavelet coefficients corresponding to the frequency component 4πare very small com-
pared to the coefficients at other frequencies, it might indicate the presence of noise in the signal.
This is because in the absence of noise, the coefficients for the true signal frequencies would be
much larger compared to any noise present in the signal.
c) Wavelet transforms can be used in denoising the signal by thresholding the wavelet coeffi-
cients. By setting small coefficients below a certain threshold to zero, we can eliminate noise while
retaining the important frequency components of the signal. This is a common technique in signal
denoising using wavelet transforms.
16 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 17. Consider a non-stationary signal given by f(t) = sin(2πt2).
a) Determine the Fourier transform of the signal f(t).
b) Compute the time-frequency representation of f(t)using the Short-Time Fourier Transform
(STFT) with a Gaussian window function g(t)given by g(t) = e−t2.
Solution 17.
a) The Fourier transform of a signal f(t)is defined as
F(ω) = Z∞
−∞
f(t)e−iωtdt.
For the given signal f(t) = sin(2πt2), we have
F(ω) = Z∞
−∞
sin(2πt2)e−iωtdt.
This particular integral does not have a closed-form solution, so we would need to evaluate it using
numerical methods or approximation techniques.
b) The Short-Time Fourier Transform (STFT) of a signal f(t)using a window function g(t)is
defined as
ST F T (t, ω) = Z∞
−∞
f(τ)g(τ−t)e−iωτ dτ,
where the overline denotes complex conjugation. For a Gaussian window function g(t) = e−t2, the
STFT of the signal f(t) = sin(2πt2)can be computed as
ST F T (t, ω) = Z∞
−∞
sin(2πτ2)e−(τ−t)2e−iωτ dτ.
We can evaluate this integral numerically to compute the time-frequency representation of the
signal using the given Gaussian window function.
17 "CONVERGENCE AND DIVERGENCE OF FOURIER SERIES FOR PIECEWISE FUNC-
TIONS"
Problem 17. Consider the function f(x)defined as:
f(x) = (x0≤x<π
−x−π < x < 0
a) Find the Fourier series of f(x).
b) Determine the conditions for which the Fourier series of f(x)converges.
Solution 17.
a) To find the Fourier series of f(x), we need to compute the coefficients anand bn:
an=1
πZπ
−π
f(x) cos(nx)dx
bn=1
πZπ
−π
f(x) sin(nx)dx
Let’s start with the calculation of an:
an=1
πZ0
−π
(−x) cos(nx)dx +Zπ
0
xcos(nx)dx
Solving these integrals and simplifying yields:
an=2((−1)n−1)
πn2
Now, let’s find bn:
bn=1
πZ0
−π
(−x) sin(nx)dx +Zπ
0
xsin(nx)dx
After solving and simplifying the integrals, we get:
bn= 0
The Fourier series of f(x)is then given by:
f(x)∼2
π1
12−1
32cos(x) + 1
52cos(2x)− · · ·
b) The Fourier series of f(x)converges if the function f(x)satisfies the Dirichlet conditions. In
this case, since f(x)is piecewise continuous and has a finite number of discontinuities within a
period of 2π, the Fourier series of f(x)converges everywhere.
18 "NON-LINEAR FOURIER TRANSFORM TECHNIQUES FOR SIGNAL ANALYSIS"
Problem 1. Consider a signal f(t) = sin(2πt) + 2 cos(4πt).
a) Compute the Fourier Transform of f(t).
b) Find the frequency components present in the signal f(t).
c) Determine the amplitude and phase of each frequency component.
Solution 1.
a) The Fourier Transform of a signal f(t)is defined as:
F(ω) = Z∞
−∞
f(t)e−jωtdt
Applying the definition to the given signal f(t), we have:
F(ω) = Z∞
−∞
[sin(2πt) + 2 cos(4πt)]e−jωtdt
=Z∞
−∞
sin(2πt)e−jωtdt + 2 Z∞
−∞
cos(4πt)e−jωtdt
Solving the integrals, we get:
F(ω) = π(δ(ω−2π)−δ(ω+ 2π)) + π(δ(ω−4π) + δ(ω+ 4π))
b) The frequency components present in the signal f(t)are 2πand 4π.
c) To determine the amplitude and phase of each frequency component, we can use the prop-
erties of the Dirac delta function in the Fourier Transform. The amplitude is the coefficient in front
of the delta function and the phase is given by the sign of the delta function. Thus, we have:
For 2πcomponent: Amplitude = πand Phase = (δ(ω−2π)−δ(ω+ 2π)) = 0
For 4πcomponent: Amplitude = πand Phase = (δ(ω−4π) + δ(ω+ 4π)) = 0
19 "ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES"
Problem 1. Consider a continuous-time signal x(t)defined as:
x(t) = 2 cos(3πt) + 3 sin(6πt)
Find the Fourier transform X(ω)of the signal x(t).
Solution 1.
The Fourier transform of a continuous-time signal x(t)is defined as:
X(ω) = Z∞
−∞
x(t)e−jωtdt
Plugging in the expression for x(t), we have:
X(ω) = Z∞
−∞
(2 cos(3πt) + 3 sin(6πt))e−jωtdt
Expanding the trigonometric functions using Euler’s formula, we get:
X(ω) = Z∞
−∞
2(ej3πt +e−j3πt
2) + 3(ej6πt −e−j6πt
2j)e−jωtdt
X(ω) = Z∞
−∞
(ej(3π−ω)t+e−j(3π+ω)t−3
2j(ej(6π−ω)t−e−j(6π+ω)t))dt
Remembering that the Fourier transform has a property with delta functions, we find:
X(ω) = 2π(δ(ω−3π) + δ(ω+ 3π)) −3π(δ(ω−6π) + δ(ω+ 6π))
Thus, the Fourier transform of the signal x(t)is:
X(ω) = 2π(δ(ω−3π) + δ(ω+ 3π)) −3π(δ(ω−6π) + δ(ω+ 6π))
Certainly! Here is a numerical problem in the realm of Advanced Harmonic Analysis and Fourier
Transform Techniques with a step-by-step solution:
20 "ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES"
Problem 21. Consider a non-stationary signal f(t)defined as:
f(t) = (sin(2πt) 0 ≤t < 1
0otherwise
Calculate the Fourier transform F(ω)of the signal f(t).
Solution 21. To find the Fourier transform F(ω)of the given signal f(t), we use the definition
of the Fourier transform:
F(ω) = Z∞
−∞
f(t)e−iωtdt
Given that f(t)is defined piecewise, we split the integral into two parts:
F(ω) = Z1
0
sin(2πt)e−iωtdt +Z∞
1
0·e−iωtdt
Solving the first integral, we can proceed to calculate the Fourier transform. Let’s solve the
integral as follows:
F(ω) = Z1
0
sin(2πt)e−iωtdt
Applying Euler’s formula to express the sine function in terms of complex exponentials, we have:
F(ω) = Z1
0
ei2πt −e−i2πt
2ie−iωtdt
F(ω) = 1
2iZ1
0
ei(2π−ω)tdt −Z1
0
e−i(2π+ω)tdt
F(ω) = 1
2i"ei(2π−ω)t
i(2π−ω)
1
0−e−i(2π+ω)t
−i(2π+ω)
1
0#
F(ω) = 1
2i"ei(2π−ω)−1
i(2π−ω)−e−i(2π+ω)−1
−i(2π+ω)#
This integral evaluation gives us the Fourier transform F(ω).
a) We can calculate the Fourier series coefficients anand bnusing the following formulas:
an=1
LZL
−L
f(x) cos 2πnx
Ldx
bn=1
LZL
−L
f(x) sin 2πnx
Ldx
where Lis the period of the function.
For f(x) = sin(2πx) + cos(4πx), the period is L= 1. Thus, we have:
an=Z1
−1
(sin(2πx) + cos(4πx)) cos(2πnx)dx
bn=Z1
−1
(sin(2πx) + cos(4πx)) sin(2πnx)dx
Calculating these integrals will give us the Fourier series coefficients anand bn.
b) To find the constant term in the Fourier series of f(x), we need to determine a0, which is
given by:
a0=1
LZL
−L
f(x)dx
Substitute L= 1 and integral bounds −1to 1to find the value of a0.
c) The representation of f(x)in its Fourier series form is given by:
f(x)≈a0
2+
∞
X
n=1 ancos 2πnx
L+bnsin 2πnx
L
Substitute the calculated values of a0,an, and bninto this equation to obtain the Fourier series
representation of f(x).
3 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider the function f(t) = sin(3t) + cos(5t).
a) Find the Fourier transform of f(t).
b) Determine the frequency components present in f(t).
Solution 1.
a) The Fourier transform of a function f(t)is defined as:
F(ω) = Z∞
−∞
f(t)e−iωtdt
Applying this definition to our function f(t) = sin(3t) + cos(5t), we have:
F(ω) = Z∞
−∞
(sin(3t) + cos(5t))e−iωtdt
=Z∞
−∞
sin(3t)e−iωtdt +Z∞
−∞
cos(5t)e−iωtdt
We can solve these integrals using the properties of the Fourier transform.
b) The frequency components present in f(t)are the frequencies of the sine and cosine func-
tions, which are 3and 5respectively.
Therefore, the frequency components present in f(t)are 3and 5.
4 "ANALYZING THE CONVERGENCE AND DIVERGENCE OF FOURIER SERIES REPRE-
SENTATIONS"
Problem 4. Consider the function f(x) = x2defined on the interval [−π, π].
a) Find the Fourier cosine series representation of f(x).
b) Determine the convergence of the Fourier cosine series of f(x).
c) Calculate the value at which the Fourier cosine series of f(x)converges to f(x).
Solution 4.
a) To find the Fourier cosine series representation of f(x) = x2, we need to calculate the
coefficients anusing the formula:
an=2
πZπ
0
x2cos(nx)dx.
Calculating the integral gives:
an=2
πZπ
0
x2cos(nx)dx
=2
π2(nx sin(nx)+(n2x2−2) cos(nx))
n2
π
0
=2
π2nπ sin(nπ)
n2+n2π2−2
n2
= 2 (−1)n
n+π2−2
n2.
Therefore, the Fourier cosine series representation of f(x)is:
f(x) = π2−2
2+ 2
∞
X
n=1 (−1)n
n+π2−2
n2cos(nx).
b) To determine the convergence of the Fourier cosine series of f(x), we need to check if f(x)is
piecewise smooth on [−π, π]. Since f(x) = x2is continuous and has derivative f′(x) = 2xexisting
for all x∈[−π, π], the Fourier cosine series converges to f(x).
c) The Fourier cosine series converges to f(x)for x∈(−π, π), so the value at which the Fourier
cosine series of f(x)converges to f(x)is x= 0.
5 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider a square wave signal x(t)given by:
x(t) = (1if 0≤t < T/2
−1if T/2≤t<T
where Tis the period of the signal.
a) Calculate the Fourier transform X(ω)of x(t).
b) Determine the value of X(0).
c) Find the power spectral density SX(ω)of the signal x(t).
Solution 1.
a) To calculate the Fourier transform X(ω)of the given square wave signal x(t), we apply the
definition of Fourier transform:
X(ω) = Z∞
−∞
x(t)e−jωtdt
Splitting the integral based on x(t):
X(ω) = ZT/2
0
e−jωtdt −ZT
T/2
e−jωtdt
Solving the integrals, we get:
X(ω) = 1−e−jωT/2
jω −e−jωT /2−e−jωT
jω
Simplifying further gives the Fourier transform as:
X(ω) = 2
jω (1 −cos(ωT /2))
b) To find the value of X(0), we substitute ω= 0 in the Fourier transform:
X(0) = 2
0(1 −cos(0)) = 0
c) The power spectral density SX(ω)is given by:
SX(ω) = |X(ω)|2
Substitute the Fourier transform X(ω)obtained earlier to get:
SX(ω) =
2
jω (1 −cos(ωT /2))
2
=4(1 −cos(ωT/2))2
ω2
6 CALCULATING FOURIER TRANSFORMS OF DISTRIBUTIONS
Problem 1. Compute the Fourier transform of the distribution T=δ(x−3) + δ(x+ 4), where δ
denotes the Dirac delta distribution.
Solution 1. The Fourier transform of a distribution Tis defined as
F(δ)(ϕ) = δ(F(ϕ))
where ϕis a test function.
a) Let’s first compute the Fourier transform of δ(x−3) using the formula:
δ(x−a)F
−→ e−2πiaξ
Thus, the Fourier transform of δ(x−3) is
F(δ(x−3)) = e−6πiξ
b) Next, we compute the Fourier transform of δ(x+ 4) which is given by:
δ(x+a)F
−→ e2πiaξ
Therefore, the Fourier transform of δ(x+ 4) is
F(δ(x+ 4)) = e8πiξ
c) Finally, computing the Fourier transform of the given distribution T:
F(T)(ϕ) = δ(F(ϕ))
=δ(e−6πiξ) + δ(e8πiξ )
=ϕ(6) + ϕ(−8)
Therefore, the Fourier transform of Tis ϕ(6) + ϕ(−8).
Problem 2. Determine the Fourier transform of the distribution T=x2·sin(2πx).
Solution 2. The Fourier transform of a distribution Tis given by:
F(xn·ϕ)(ξ)=(−i)n·dnF(ϕ)
dξn
where ϕis a test function.
We need to compute the Fourier transform of x2·sin(2πx).
a) First, we find the Fourier transforms of x2and sin(2πx)individually.
The Fourier transform of x2can be calculated using the formula:
F(x2)(ξ) = (−i)2·d2
dξ2δ
This simplifies to:
F(x2)(ξ) = −δ′′(ξ)
b) Next, we find the Fourier transform of sin(2πx):
F(sin(αx)) = α
α2−ξ2
Applying this to sin(2πx)gives:
F(sin(2πx)) = 2π
(2π)2−ξ2
c) Now, we calculate the Fourier transform of x2·sin(2πx)by convolving the individual trans-
forms:
F(x2·sin(2πx))(ξ) = −δ′′(ξ)∗2π
(2π)2−ξ2
=−2π
(2π)2−ξ2
Therefore, the Fourier transform of x2·sin(2πx)is −2π
(2π)2−ξ2.
7 "ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES"
Problem 1. Consider the function f(x) = cos(2πx) + 3 sin(4πx).
a) Find the period of f(x).
b) Compute the Fourier series of f(x).
Solution 1.
a) To find the period of f(x), we need to consider the period of each component function
cos(2πx)and sin(4πx)separately.
The period of cos(2πx)is given by T=2π
2π= 1.
Similarly, the period of sin(4πx)is T=2π
4π=1
2.
Since the period of f(x)is the least common multiple of the periods of its component functions,
the period of f(x)is T=lcm(1,1
2)=1.
Thus, the period of f(x)is 1.
b) The Fourier series of f(x)can be found by expressing it in the form:
f(x) = a0+
∞
X
n=1 ancos 2nπx
T+bnsin 2nπx
T
where Tis the period of f(x).
Since we found that T= 1 in part (a), we have:
f(x) = a0+
∞
X
n=1
[ancos(2nπx) + bnsin(2nπx)]
The coefficients can be calculated using the formulas:
a0=1
TZT
2
−T
2
f(x)dx
an=2
TZT
2
−T
2
f(x) cos 2nπx
Tdx
bn=2
TZT
2
−T
2
f(x) sin 2nπx
Tdx
Since T= 1, these formulas simplify to:
a0=Z0.5
−0.5
f(x)dx
an= 2 Z0.5
−0.5
f(x) cos(2nπx)dx
bn= 2 Z0.5
−0.5
f(x) sin(2nπx)dx
Plugging in the function f(x)and computing the integrals will give the coefficients for the Fourier
series of f(x).
8 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider a continuous-time signal x(t)in a given bandwidth, represented by its
Fourier transform X(f). The signal is sampled with a sampling rate of Fs= 200 Hz. Determine if
this sampling rate satisfies the Nyquist criterion for this signal. Justify your answer.
Solution 1.
Given that the signal is sampled with a sampling rate of Fs= 200 Hz, we need to check if this
sampling rate satisfies the Nyquist criterion.
The Nyquist theorem states that in order to perfectly reconstruct a signal from its samples, the
sampling rate must be at least twice the maximum frequency present in the signal. In other words,
Fs≥2B, where Bis the bandwidth of the signal.
In this case, since Fs= 200 Hz, we need to determine the bandwidth of the signal B. If X(f)is
non-zero only up to BHz, then the condition is satisfied.
If the signal’s Fourier transform X(f)has energy up to BHz, then the condition Fs≥2Bis
satisfied and the sampling rate of 200 Hz meets the Nyquist criterion for this signal.
Therefore, the sampling rate of 200 Hz satisfies the Nyquist criterion for the given continuous-
time signal.
This problem highlighted the importance of the Nyquist criterion in determining the appropriate
sampling rate for a signal in order to avoid aliasing.
9 "ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES"
Problem 1. Consider a non-stationary signal x(t) = sin(2πt2).
a) Find the Fourier transform X(ω)of the signal.
b) Determine the time-frequency representation of the signal using the Short-Time Fourier
Transform (STFT) with a Gaussian window function g(t) = e−t2.
Solution 1.
a) To find the Fourier transform X(ω)of the signal x(t) = sin(2πt2), we use the definition of the
Fourier transform:
X(ω) = F{x(t)}=Z∞
−∞
x(t)e−jωtdt
Substitute x(t) = sin(2πt2)into the integral:
X(ω) = Z∞
−∞
sin(2πt2)e−jωtdt
This integral does not have a closed-form solution, but can be numerically evaluated.
b) The Short-Time Fourier Transform (STFT) of a signal x(t)with a window function g(t)is given
by:
X(ω, τ) = Z∞
−∞
x(t)g(t−τ)e−jωtdt
For the given signal x(t) = sin(2πt2)and Gaussian window function g(t) = e−t2, we can com-
pute the STFT X(ω, τ)by sliding the window along the time axis and computing the Fourier trans-
form for each windowed segment.
The time-frequency representation of the signal using STFT provides information about how
the frequency content of the signal evolves over time, which is particularly useful for analyzing
non-stationary signals like x(t) = sin(2πt2).
10 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider a function f(x) = (xif 0≤x≤π
0otherwise
a) Compute the Fourier sine series of f(x).
b) Find the sum of the series.
Solution 1.
a) The Fourier sine series for a function f(x)defined on 0≤x≤Lis given by:
f(x)∼
∞
X
n=1
Bnsin nπx
L
where Bn=2
LRL
0f(x) sin nπx
Ldx.
For the function f(x)given in the problem, L=πand f(x) = xfor 0≤x≤π. Thus, we need
to compute Bn:
Bn=2
πZπ
0
xsin(nx)dx
=2
π−x
ncos(nx)
π
0
+1
nZπ
0
cos(nx)dx
=2
π0−(−1)nπ
n+1
n2sin(nx)
π
0
= 0 + 2(−1)n
n2+2(−1)n
πn
=2(−1)n
n21 + 1
πn
Therefore, the Fourier sine series of f(x)is:
f(x)∼
∞
X
n=1
2(−1)n
n21 + 1
πnsin(nx)
b) To find the sum of the series, we note that the Fourier sine series of an odd function converges
to f(x−)+f(x+)
2.
For the given function f(x) = x,f(x−) = xand f(x+)=0(at the jump discontinuity x= 0).
Thus, the sum of the series is:
x+ 0
2=x
2
Therefore, the sum of the Fourier sine series of f(x)is x
2.
11 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider a function f(t) = cos(2πt)in the time domain. Compute the Fourier
transform F(ω)of f(t), where F(ω)is defined as:
F(ω) = Z∞
−∞
f(t)e−iωtdt
Solution 1. To compute the Fourier transform F(ω)of f(t) = cos(2πt), we substitute f(t)into
the definition of the Fourier transform:
F(ω) = Z∞
−∞
cos(2πt)e−iωtdt
Using Euler’s formula eix = cos(x) + isin(x), we can express the integrand as:
cos(2πt)e−iωt =1
2(e2πit +e−2πit)1
2i(eiωt −e−iωt)
Multiplying and simplifying, the integrand becomes:
1
4i(e(2πi−ωi)t−e(−2πi+ωi)t+e(2πi+ωi)t−e(−2πi−ωi)t)
Now, we can integrate term by term:
F(ω) = 1
4i"e(2πi−ωi)t
2πi −ωi −e(−2πi+ωi)t
−2πi +ωi +e(2πi+ωi)t
2πi +ωi −e(−2πi−ωi)t
−2πi −ωi #
∞
−∞
Given that the oscillatory terms will go to zero at ±∞ due to periodicity, we are left with evalu-
ating the residues at the poles ω=±2π. This gives us the final result:
F(ω) = π(δ(ω−2π) + δ(ω+ 2π))
where δ(ω)is the Dirac delta function.
So, the Fourier transform of f(t) = cos(2πt)is F(ω) = π(δ(ω−2π) + δ(ω+ 2π)).
12 NONLINEAR SIGNAL PROCESSING USING HARMONIC ANALYSIS AND FOURIER TRANS-
FORM
Problem 1.
Consider a nonlinear signal x(t) = cos(2πt) + cos(4πt). Determine the Fourier series represen-
tation of x(t)over the interval [−T, T ], where Tis the period of the signal.
Solution 1.
Given signal x(t) = cos(2πt) + cos(4πt), we first need to find the period of the signal by calcu-
lating the least common multiple of the periods of the individual cosine components.
The period of cos(2πt)is T1= 1 and the period of cos(4πt)is T2=1
2.
The period of the combined signal is T=LCM(T1, T2) = LCM(1,1
2) = 1.
Therefore, the Fourier series representation of x(t)over the interval [−1,1] is:
x(t) = a0+
∞
X
n=1
(ancos(nπt) + bnsin(nπt))
where an=2
TRT
−Tx(t) cos(nπt)dt and bn=2
TRT
−Tx(t) sin(nπt)dt.
Calculating the coefficients:
a0=1
TZ1
−1
x(t)dt =1
1Z1
−1
(cos(2πt) + cos(4πt))dt
an=2
TZ1
−1
x(t) cos(nπt)dt =2
1Z1
−1
(cos(2πt) + cos(4πt)) cos(nπt)dt
bn=2
TZ1
−1
x(t) sin(nπt)dt =2
1Z1
−1
(cos(2πt) + cos(4πt)) sin(nπt)dt
By performing the integrations, we can find the Fourier series representation of x(t)over the
interval [−1,1].
13 "WAVELET ANALYSIS AND APPLICATIONS IN SIGNAL PROCESSING"
Problem 13. Consider a signal f(t) = sin(4t) + sin(8t). Perform the following wavelet analysis
using the Morlet wavelet:
ψ(t) = eiω0te−t2/2
where ω0= 5.
a) Compute the continuous wavelet transform (CWT) of f(t).
b) Sketch the scalogram of the CWT.
c) Determine the dominant frequency component(s) present in the signal.
Solution 13.
a) To compute the CWT of f(t), we use the formula:
Wf(a, b) = Z∞
−∞
f(t)·¯
ψt−b
adt,
where ais the dilation parameter and bis the translation parameter.
For this problem, f(t) = sin(4t) + sin(8t)and ψ(t) = ei·5te−t2/2.
The CWT is given by:
Wf(a, b) = Z∞
−∞
[sin(4t) + sin(8t)] ·e−i·5(t−b)/ae−(t−b)2/(2a2)dt.
This integral can be evaluated by numerical methods or special software such as MATLAB to
find the values of Wf(a, b).
b) The scalogram is a visual representation of the CWT. It shows the magnitude of the CWT
coefficients at each scale aand time b. To sketch the scalogram, we plot the magnitude of |Wf(a, b)|
at different aand bvalues.
c) By examining the scalogram, we can identify the dominant frequency components by look-
ing at the peaks in the magnitude of the CWT coefficients. In this case, the dominant frequency
components correspond to the frequencies that have the highest magnitude values in the CWT.
14 "APPLICATIONS OF WAVELET TRANSFORM IN SIGNAL PROCESSING"
Problem 14. Consider a signal f(t)with the following data:
f(t) = (1,if 0≤t≤1
0,otherwise
Let’s analyze this signal using the Haar wavelet transform.
a) Calculate the Haar wavelet coefficients ck,m corresponding to f(t).
b) Determine the wavelet transform of f(t).
c) Reconstruct the original signal using the inverse Haar wavelet transform from the coefficients
found in part (a).
Solution 14.
a) The Haar wavelet coefficients ck,m can be calculated as:
ck,m =Z∞
−∞
f(t)ψk,m(t)dt
Here, ψk,m(t)is the Haar wavelet function. Since the Haar wavelet is a piecewise constant
function, we have:
ψk,m(t) = (2k
2,if 2km≤t < 2k(m+ 1)
0,otherwise
Given f(t), we can calculate the coefficients by integrating over the intervals where ψk,m(t)is
non-zero. For the Haar wavelet, we only need to consider the intervals [0,1) and [1,2).
For the interval [0,1):
c0,0=Z1
0
f(t)ψ0,0(t)dt
c0,0=Z1
0
(1) ·(20
2)dt
c0,0=Z1
0
dt = 1
For the interval [1,2):
c1,0=Z2
1
f(t)ψ1,0(t)dt
c1,0=Z2
1
(0) ·(21
2)dt
c1,0= 0
Therefore, the Haar wavelet coefficients for f(t)are c0,0= 1 and c1,0= 0.
b) The wavelet transform of f(t)is given by:
Wf(a, b) = X
k,m
ck,m ·ψk,m(t)
Since c1,0= 0, the wavelet transform simplifies to:
Wf(a, b) = c0,0·ψ0,0(t)=1·20
2= 1
c) Reconstructing the original signal using the inverse Haar wavelet transform involves summing
the scaled and shifted Haar wavelets:
f(t) = X
k,m
ck,m ·ψk,m(t)
Substitute the coefficients c0,0= 1 and c1,0= 0 into the formula:
f(t) = 1 ·ψ0,0(t) = 1 ·20
2= 1
Therefore, the reconstructed signal using the inverse Haar wavelet transform is a constant
signal f(t)=1.
15 "APPLICATION OF WAVELET TRANSFORM IN SIGNAL DENOISING AND COMPRES-
SION"
Problem 15. Consider a signal f(t)defined as:
f(t) = 2 sin(4πt) + 5 cos(8πt) + 3 sin(12πt)
a) Using the wavelet transform, determine the frequency components present in the signal
f(t). b) If we apply a wavelet transform and find that the wavelet coefficients corresponding to the
frequency component 4πare very small compared to the coefficients at other frequencies, explain
why this might indicate the presence of noise in the signal. c) How can wavelet transforms be used
in denoising this signal?
Solution 15.
a) To determine the frequency components present in the signal f(t), we need to apply the
wavelet transform. Let’s consider the Haar wavelet transform for simplicity. The Haar wavelet
transform of a signal f(t)is given by:
W(a, b) = Z∞
−∞
f(t)ψa,b(t)dt
Where ψa,b(t)is the Haar wavelet defined as:
ψa,b(t) =
1if a≤t<a+1
2
−1if a+1
2≤t<a+ 1
0otherwise
Let’s calculate the wavelet coefficients for the given signal f(t)at different scales a:
W(a, b) = Z∞
−∞
f(t)ψa,b(t)dt
For a= 1, we have:
W(1, b) = Z∞
−∞
(2 sin(4πt) + 5 cos(8πt) + 3 sin(12πt))ψ1,b(t)dt
Similarly, we can calculate W(2, b),W(3, b), and so on to determine the frequency components
present in the signal f(t).
b) If the wavelet coefficients corresponding to the frequency component 4πare very small com-
pared to the coefficients at other frequencies, it might indicate the presence of noise in the signal.
This is because in the absence of noise, the coefficients for the true signal frequencies would be
much larger compared to any noise present in the signal.
c) Wavelet transforms can be used in denoising the signal by thresholding the wavelet coeffi-
cients. By setting small coefficients below a certain threshold to zero, we can eliminate noise while
retaining the important frequency components of the signal. This is a common technique in signal
denoising using wavelet transforms.
16 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 17. Consider a non-stationary signal given by f(t) = sin(2πt2).
a) Determine the Fourier transform of the signal f(t).
b) Compute the time-frequency representation of f(t)using the Short-Time Fourier Transform
(STFT) with a Gaussian window function g(t)given by g(t) = e−t2.
Solution 17.
a) The Fourier transform of a signal f(t)is defined as
F(ω) = Z∞
−∞
f(t)e−iωtdt.
For the given signal f(t) = sin(2πt2), we have
F(ω) = Z∞
−∞
sin(2πt2)e−iωtdt.
This particular integral does not have a closed-form solution, so we would need to evaluate it using
numerical methods or approximation techniques.
b) The Short-Time Fourier Transform (STFT) of a signal f(t)using a window function g(t)is
defined as
ST F T (t, ω) = Z∞
−∞
f(τ)g(τ−t)e−iωτ dτ,
where the overline denotes complex conjugation. For a Gaussian window function g(t) = e−t2, the
STFT of the signal f(t) = sin(2πt2)can be computed as
ST F T (t, ω) = Z∞
−∞
sin(2πτ2)e−(τ−t)2e−iωτ dτ.
We can evaluate this integral numerically to compute the time-frequency representation of the
signal using the given Gaussian window function.
17 "CONVERGENCE AND DIVERGENCE OF FOURIER SERIES FOR PIECEWISE FUNC-
TIONS"
Problem 17. Consider the function f(x)defined as:
f(x) = (x0≤x<π
−x−π < x < 0
a) Find the Fourier series of f(x).
b) Determine the conditions for which the Fourier series of f(x)converges.
Solution 17.
a) To find the Fourier series of f(x), we need to compute the coefficients anand bn:
an=1
πZπ
−π
f(x) cos(nx)dx
bn=1
πZπ
−π
f(x) sin(nx)dx
Let’s start with the calculation of an:
an=1
πZ0
−π
(−x) cos(nx)dx +Zπ
0
xcos(nx)dx
Solving these integrals and simplifying yields:
an=2((−1)n−1)
πn2
Now, let’s find bn:
bn=1
πZ0
−π
(−x) sin(nx)dx +Zπ
0
xsin(nx)dx
After solving and simplifying the integrals, we get:
bn= 0
The Fourier series of f(x)is then given by:
f(x)∼2
π1
12−1
32cos(x) + 1
52cos(2x)− · · ·
b) The Fourier series of f(x)converges if the function f(x)satisfies the Dirichlet conditions. In
this case, since f(x)is piecewise continuous and has a finite number of discontinuities within a
period of 2π, the Fourier series of f(x)converges everywhere.
18 "NON-LINEAR FOURIER TRANSFORM TECHNIQUES FOR SIGNAL ANALYSIS"
Problem 1. Consider a signal f(t) = sin(2πt) + 2 cos(4πt).
a) Compute the Fourier Transform of f(t).
b) Find the frequency components present in the signal f(t).
c) Determine the amplitude and phase of each frequency component.
Solution 1.
a) The Fourier Transform of a signal f(t)is defined as:
F(ω) = Z∞
−∞
f(t)e−jωtdt
Applying the definition to the given signal f(t), we have:
F(ω) = Z∞
−∞
[sin(2πt) + 2 cos(4πt)]e−jωtdt
=Z∞
−∞
sin(2πt)e−jωtdt + 2 Z∞
−∞
cos(4πt)e−jωtdt
Solving the integrals, we get:
F(ω) = π(δ(ω−2π)−δ(ω+ 2π)) + π(δ(ω−4π) + δ(ω+ 4π))
b) The frequency components present in the signal f(t)are 2πand 4π.
c) To determine the amplitude and phase of each frequency component, we can use the prop-
erties of the Dirac delta function in the Fourier Transform. The amplitude is the coefficient in front
of the delta function and the phase is given by the sign of the delta function. Thus, we have:
For 2πcomponent: Amplitude = πand Phase = (δ(ω−2π)−δ(ω+ 2π)) = 0
For 4πcomponent: Amplitude = πand Phase = (δ(ω−4π) + δ(ω+ 4π)) = 0
19 "ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES"
Problem 1. Consider a continuous-time signal x(t)defined as:
x(t) = 2 cos(3πt) + 3 sin(6πt)
Find the Fourier transform X(ω)of the signal x(t).
Solution 1.
The Fourier transform of a continuous-time signal x(t)is defined as:
X(ω) = Z∞
−∞
x(t)e−jωtdt
Plugging in the expression for x(t), we have:
X(ω) = Z∞
−∞
(2 cos(3πt) + 3 sin(6πt))e−jωtdt
Expanding the trigonometric functions using Euler’s formula, we get:
X(ω) = Z∞
−∞
2(ej3πt +e−j3πt
2) + 3(ej6πt −e−j6πt
2j)e−jωtdt
X(ω) = Z∞
−∞
(ej(3π−ω)t+e−j(3π+ω)t−3
2j(ej(6π−ω)t−e−j(6π+ω)t))dt
Remembering that the Fourier transform has a property with delta functions, we find:
X(ω) = 2π(δ(ω−3π) + δ(ω+ 3π)) −3π(δ(ω−6π) + δ(ω+ 6π))
Thus, the Fourier transform of the signal x(t)is:
X(ω) = 2π(δ(ω−3π) + δ(ω+ 3π)) −3π(δ(ω−6π) + δ(ω+ 6π))
Certainly! Here is a numerical problem in the realm of Advanced Harmonic Analysis and Fourier
Transform Techniques with a step-by-step solution:
20 "ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES"
Problem 21. Consider a non-stationary signal f(t)defined as:
f(t) = (sin(2πt) 0 ≤t < 1
0otherwise
Calculate the Fourier transform F(ω)of the signal f(t).
Solution 21. To find the Fourier transform F(ω)of the given signal f(t), we use the definition
of the Fourier transform:
F(ω) = Z∞
−∞
f(t)e−iωtdt
Given that f(t)is defined piecewise, we split the integral into two parts:
F(ω) = Z1
0
sin(2πt)e−iωtdt +Z∞
1
0·e−iωtdt
Solving the first integral, we can proceed to calculate the Fourier transform. Let’s solve the
integral as follows:
F(ω) = Z1
0
sin(2πt)e−iωtdt
Applying Euler’s formula to express the sine function in terms of complex exponentials, we have:
F(ω) = Z1
0
ei2πt −e−i2πt
2ie−iωtdt
F(ω) = 1
2iZ1
0
ei(2π−ω)tdt −Z1
0
e−i(2π+ω)tdt
F(ω) = 1
2i"ei(2π−ω)t
i(2π−ω)
1
0−e−i(2π+ω)t
−i(2π+ω)
1
0#
F(ω) = 1
2i"ei(2π−ω)−1
i(2π−ω)−e−i(2π+ω)−1
−i(2π+ω)#
This integral evaluation gives us the Fourier transform F(ω).
a) We can calculate the Fourier series coefficients anand bnusing the following formulas:
an=1
LZL
−L
f(x) cos 2πnx
Ldx
bn=1
LZL
−L
f(x) sin 2πnx
Ldx
where Lis the period of the function.
For f(x) = sin(2πx) + cos(4πx), the period is L= 1. Thus, we have:
an=Z1
−1
(sin(2πx) + cos(4πx)) cos(2πnx)dx
bn=Z1
−1
(sin(2πx) + cos(4πx)) sin(2πnx)dx
Calculating these integrals will give us the Fourier series coefficients anand bn.
b) To find the constant term in the Fourier series of f(x), we need to determine a0, which is
given by:
a0=1
LZL
−L
f(x)dx
Substitute L= 1 and integral bounds −1to 1to find the value of a0.
c) The representation of f(x)in its Fourier series form is given by:
f(x)≈a0
2+
∞
X
n=1 ancos 2πnx
L+bnsin 2πnx
L
Substitute the calculated values of a0,an, and bninto this equation to obtain the Fourier series
representation of f(x).
3 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider the function f(t) = sin(3t) + cos(5t).
a) Find the Fourier transform of f(t).
b) Determine the frequency components present in f(t).
Solution 1.
a) The Fourier transform of a function f(t)is defined as:
F(ω) = Z∞
−∞
f(t)e−iωtdt
Applying this definition to our function f(t) = sin(3t) + cos(5t), we have:
F(ω) = Z∞
−∞
(sin(3t) + cos(5t))e−iωtdt
=Z∞
−∞
sin(3t)e−iωtdt +Z∞
−∞
cos(5t)e−iωtdt
We can solve these integrals using the properties of the Fourier transform.
b) The frequency components present in f(t)are the frequencies of the sine and cosine func-
tions, which are 3and 5respectively.
Therefore, the frequency components present in f(t)are 3and 5.
4 "ANALYZING THE CONVERGENCE AND DIVERGENCE OF FOURIER SERIES REPRE-
SENTATIONS"
Problem 4. Consider the function f(x) = x2defined on the interval [−π, π].
a) Find the Fourier cosine series representation of f(x).
b) Determine the convergence of the Fourier cosine series of f(x).
c) Calculate the value at which the Fourier cosine series of f(x)converges to f(x).
Solution 4.
a) To find the Fourier cosine series representation of f(x) = x2, we need to calculate the
coefficients anusing the formula:
an=2
πZπ
0
x2cos(nx)dx.
Calculating the integral gives:
an=2
πZπ
0
x2cos(nx)dx
=2
π2(nx sin(nx)+(n2x2−2) cos(nx))
n2
π
0
=2
π2nπ sin(nπ)
n2+n2π2−2
n2
= 2 (−1)n
n+π2−2
n2.
Therefore, the Fourier cosine series representation of f(x)is:
f(x) = π2−2
2+ 2
∞
X
n=1 (−1)n
n+π2−2
n2cos(nx).
b) To determine the convergence of the Fourier cosine series of f(x), we need to check if f(x)is
piecewise smooth on [−π, π]. Since f(x) = x2is continuous and has derivative f′(x) = 2xexisting
for all x∈[−π, π], the Fourier cosine series converges to f(x).
c) The Fourier cosine series converges to f(x)for x∈(−π, π), so the value at which the Fourier
cosine series of f(x)converges to f(x)is x= 0.
5 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider a square wave signal x(t)given by:
x(t) = (1if 0≤t < T/2
−1if T/2≤t<T
where Tis the period of the signal.
a) Calculate the Fourier transform X(ω)of x(t).
b) Determine the value of X(0).
c) Find the power spectral density SX(ω)of the signal x(t).
Solution 1.
a) To calculate the Fourier transform X(ω)of the given square wave signal x(t), we apply the
definition of Fourier transform:
X(ω) = Z∞
−∞
x(t)e−jωtdt
Splitting the integral based on x(t):
X(ω) = ZT/2
0
e−jωtdt −ZT
T/2
e−jωtdt
Solving the integrals, we get:
X(ω) = 1−e−jωT/2
jω −e−jωT /2−e−jωT
jω
Simplifying further gives the Fourier transform as:
X(ω) = 2
jω (1 −cos(ωT /2))
b) To find the value of X(0), we substitute ω= 0 in the Fourier transform:
X(0) = 2
0(1 −cos(0)) = 0
c) The power spectral density SX(ω)is given by:
SX(ω) = |X(ω)|2
Substitute the Fourier transform X(ω)obtained earlier to get:
SX(ω) =
2
jω (1 −cos(ωT /2))
2
=4(1 −cos(ωT/2))2
ω2
6 CALCULATING FOURIER TRANSFORMS OF DISTRIBUTIONS
Problem 1. Compute the Fourier transform of the distribution T=δ(x−3) + δ(x+ 4), where δ
denotes the Dirac delta distribution.
Solution 1. The Fourier transform of a distribution Tis defined as
F(δ)(ϕ) = δ(F(ϕ))
where ϕis a test function.
a) Let’s first compute the Fourier transform of δ(x−3) using the formula:
δ(x−a)F
−→ e−2πiaξ
Thus, the Fourier transform of δ(x−3) is
F(δ(x−3)) = e−6πiξ
b) Next, we compute the Fourier transform of δ(x+ 4) which is given by:
δ(x+a)F
−→ e2πiaξ
Therefore, the Fourier transform of δ(x+ 4) is
F(δ(x+ 4)) = e8πiξ
c) Finally, computing the Fourier transform of the given distribution T:
F(T)(ϕ) = δ(F(ϕ))
=δ(e−6πiξ) + δ(e8πiξ )
=ϕ(6) + ϕ(−8)
Therefore, the Fourier transform of Tis ϕ(6) + ϕ(−8).
Problem 2. Determine the Fourier transform of the distribution T=x2·sin(2πx).
Solution 2. The Fourier transform of a distribution Tis given by:
F(xn·ϕ)(ξ)=(−i)n·dnF(ϕ)
dξn
where ϕis a test function.
We need to compute the Fourier transform of x2·sin(2πx).
a) First, we find the Fourier transforms of x2and sin(2πx)individually.
The Fourier transform of x2can be calculated using the formula:
F(x2)(ξ) = (−i)2·d2
dξ2δ
This simplifies to:
F(x2)(ξ) = −δ′′(ξ)
b) Next, we find the Fourier transform of sin(2πx):
F(sin(αx)) = α
α2−ξ2
Applying this to sin(2πx)gives:
F(sin(2πx)) = 2π
(2π)2−ξ2
c) Now, we calculate the Fourier transform of x2·sin(2πx)by convolving the individual trans-
forms:
F(x2·sin(2πx))(ξ) = −δ′′(ξ)∗2π
(2π)2−ξ2
=−2π
(2π)2−ξ2
Therefore, the Fourier transform of x2·sin(2πx)is −2π
(2π)2−ξ2.
7 "ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES"
Problem 1. Consider the function f(x) = cos(2πx) + 3 sin(4πx).
a) Find the period of f(x).
b) Compute the Fourier series of f(x).
Solution 1.
a) To find the period of f(x), we need to consider the period of each component function
cos(2πx)and sin(4πx)separately.
The period of cos(2πx)is given by T=2π
2π= 1.
Similarly, the period of sin(4πx)is T=2π
4π=1
2.
Since the period of f(x)is the least common multiple of the periods of its component functions,
the period of f(x)is T=lcm(1,1
2)=1.
Thus, the period of f(x)is 1.
b) The Fourier series of f(x)can be found by expressing it in the form:
f(x) = a0+
∞
X
n=1 ancos 2nπx
T+bnsin 2nπx
T
where Tis the period of f(x).
Since we found that T= 1 in part (a), we have:
f(x) = a0+
∞
X
n=1
[ancos(2nπx) + bnsin(2nπx)]
The coefficients can be calculated using the formulas:
a0=1
TZT
2
−T
2
f(x)dx
an=2
TZT
2
−T
2
f(x) cos 2nπx
Tdx
bn=2
TZT
2
−T
2
f(x) sin 2nπx
Tdx
Since T= 1, these formulas simplify to:
a0=Z0.5
−0.5
f(x)dx
an= 2 Z0.5
−0.5
f(x) cos(2nπx)dx
bn= 2 Z0.5
−0.5
f(x) sin(2nπx)dx
Plugging in the function f(x)and computing the integrals will give the coefficients for the Fourier
series of f(x).
8 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider a continuous-time signal x(t)in a given bandwidth, represented by its
Fourier transform X(f). The signal is sampled with a sampling rate of Fs= 200 Hz. Determine if
this sampling rate satisfies the Nyquist criterion for this signal. Justify your answer.
Solution 1.
Given that the signal is sampled with a sampling rate of Fs= 200 Hz, we need to check if this
sampling rate satisfies the Nyquist criterion.
The Nyquist theorem states that in order to perfectly reconstruct a signal from its samples, the
sampling rate must be at least twice the maximum frequency present in the signal. In other words,
Fs≥2B, where Bis the bandwidth of the signal.
In this case, since Fs= 200 Hz, we need to determine the bandwidth of the signal B. If X(f)is
non-zero only up to BHz, then the condition is satisfied.
If the signal’s Fourier transform X(f)has energy up to BHz, then the condition Fs≥2Bis
satisfied and the sampling rate of 200 Hz meets the Nyquist criterion for this signal.
Therefore, the sampling rate of 200 Hz satisfies the Nyquist criterion for the given continuous-
time signal.
This problem highlighted the importance of the Nyquist criterion in determining the appropriate
sampling rate for a signal in order to avoid aliasing.
9 "ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES"
Problem 1. Consider a non-stationary signal x(t) = sin(2πt2).
a) Find the Fourier transform X(ω)of the signal.
b) Determine the time-frequency representation of the signal using the Short-Time Fourier
Transform (STFT) with a Gaussian window function g(t) = e−t2.
Solution 1.
a) To find the Fourier transform X(ω)of the signal x(t) = sin(2πt2), we use the definition of the
Fourier transform:
X(ω) = F{x(t)}=Z∞
−∞
x(t)e−jωtdt
Substitute x(t) = sin(2πt2)into the integral:
X(ω) = Z∞
−∞
sin(2πt2)e−jωtdt
This integral does not have a closed-form solution, but can be numerically evaluated.
b) The Short-Time Fourier Transform (STFT) of a signal x(t)with a window function g(t)is given
by:
X(ω, τ) = Z∞
−∞
x(t)g(t−τ)e−jωtdt
For the given signal x(t) = sin(2πt2)and Gaussian window function g(t) = e−t2, we can com-
pute the STFT X(ω, τ)by sliding the window along the time axis and computing the Fourier trans-
form for each windowed segment.
The time-frequency representation of the signal using STFT provides information about how
the frequency content of the signal evolves over time, which is particularly useful for analyzing
non-stationary signals like x(t) = sin(2πt2).
10 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider a function f(x) = (xif 0≤x≤π
0otherwise
a) Compute the Fourier sine series of f(x).
b) Find the sum of the series.
Solution 1.
a) The Fourier sine series for a function f(x)defined on 0≤x≤Lis given by:
f(x)∼
∞
X
n=1
Bnsin nπx
L
where Bn=2
LRL
0f(x) sin nπx
Ldx.
For the function f(x)given in the problem, L=πand f(x) = xfor 0≤x≤π. Thus, we need
to compute Bn:
Bn=2
πZπ
0
xsin(nx)dx
=2
π−x
ncos(nx)
π
0
+1
nZπ
0
cos(nx)dx
=2
π0−(−1)nπ
n+1
n2sin(nx)
π
0
= 0 + 2(−1)n
n2+2(−1)n
πn
=2(−1)n
n21 + 1
πn
Therefore, the Fourier sine series of f(x)is:
f(x)∼
∞
X
n=1
2(−1)n
n21 + 1
πnsin(nx)
b) To find the sum of the series, we note that the Fourier sine series of an odd function converges
to f(x−)+f(x+)
2.
For the given function f(x) = x,f(x−) = xand f(x+)=0(at the jump discontinuity x= 0).
Thus, the sum of the series is:
x+ 0
2=x
2
Therefore, the sum of the Fourier sine series of f(x)is x
2.
11 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 1. Consider a function f(t) = cos(2πt)in the time domain. Compute the Fourier
transform F(ω)of f(t), where F(ω)is defined as:
F(ω) = Z∞
−∞
f(t)e−iωtdt
Solution 1. To compute the Fourier transform F(ω)of f(t) = cos(2πt), we substitute f(t)into
the definition of the Fourier transform:
F(ω) = Z∞
−∞
cos(2πt)e−iωtdt
Using Euler’s formula eix = cos(x) + isin(x), we can express the integrand as:
cos(2πt)e−iωt =1
2(e2πit +e−2πit)1
2i(eiωt −e−iωt)
Multiplying and simplifying, the integrand becomes:
1
4i(e(2πi−ωi)t−e(−2πi+ωi)t+e(2πi+ωi)t−e(−2πi−ωi)t)
Now, we can integrate term by term:
F(ω) = 1
4i"e(2πi−ωi)t
2πi −ωi −e(−2πi+ωi)t
−2πi +ωi +e(2πi+ωi)t
2πi +ωi −e(−2πi−ωi)t
−2πi −ωi #
∞
−∞
Given that the oscillatory terms will go to zero at ±∞ due to periodicity, we are left with evalu-
ating the residues at the poles ω=±2π. This gives us the final result:
F(ω) = π(δ(ω−2π) + δ(ω+ 2π))
where δ(ω)is the Dirac delta function.
So, the Fourier transform of f(t) = cos(2πt)is F(ω) = π(δ(ω−2π) + δ(ω+ 2π)).
12 NONLINEAR SIGNAL PROCESSING USING HARMONIC ANALYSIS AND FOURIER TRANS-
FORM
Problem 1.
Consider a nonlinear signal x(t) = cos(2πt) + cos(4πt). Determine the Fourier series represen-
tation of x(t)over the interval [−T, T ], where Tis the period of the signal.
Solution 1.
Given signal x(t) = cos(2πt) + cos(4πt), we first need to find the period of the signal by calcu-
lating the least common multiple of the periods of the individual cosine components.
The period of cos(2πt)is T1= 1 and the period of cos(4πt)is T2=1
2.
The period of the combined signal is T=LCM(T1, T2) = LCM(1,1
2) = 1.
Therefore, the Fourier series representation of x(t)over the interval [−1,1] is:
x(t) = a0+
∞
X
n=1
(ancos(nπt) + bnsin(nπt))
where an=2
TRT
−Tx(t) cos(nπt)dt and bn=2
TRT
−Tx(t) sin(nπt)dt.
Calculating the coefficients:
a0=1
TZ1
−1
x(t)dt =1
1Z1
−1
(cos(2πt) + cos(4πt))dt
an=2
TZ1
−1
x(t) cos(nπt)dt =2
1Z1
−1
(cos(2πt) + cos(4πt)) cos(nπt)dt
bn=2
TZ1
−1
x(t) sin(nπt)dt =2
1Z1
−1
(cos(2πt) + cos(4πt)) sin(nπt)dt
By performing the integrations, we can find the Fourier series representation of x(t)over the
interval [−1,1].
13 "WAVELET ANALYSIS AND APPLICATIONS IN SIGNAL PROCESSING"
Problem 13. Consider a signal f(t) = sin(4t) + sin(8t). Perform the following wavelet analysis
using the Morlet wavelet:
ψ(t) = eiω0te−t2/2
where ω0= 5.
a) Compute the continuous wavelet transform (CWT) of f(t).
b) Sketch the scalogram of the CWT.
c) Determine the dominant frequency component(s) present in the signal.
Solution 13.
a) To compute the CWT of f(t), we use the formula:
Wf(a, b) = Z∞
−∞
f(t)·¯
ψt−b
adt,
where ais the dilation parameter and bis the translation parameter.
For this problem, f(t) = sin(4t) + sin(8t)and ψ(t) = ei·5te−t2/2.
The CWT is given by:
Wf(a, b) = Z∞
−∞
[sin(4t) + sin(8t)] ·e−i·5(t−b)/ae−(t−b)2/(2a2)dt.
This integral can be evaluated by numerical methods or special software such as MATLAB to
find the values of Wf(a, b).
b) The scalogram is a visual representation of the CWT. It shows the magnitude of the CWT
coefficients at each scale aand time b. To sketch the scalogram, we plot the magnitude of |Wf(a, b)|
at different aand bvalues.
c) By examining the scalogram, we can identify the dominant frequency components by look-
ing at the peaks in the magnitude of the CWT coefficients. In this case, the dominant frequency
components correspond to the frequencies that have the highest magnitude values in the CWT.
14 "APPLICATIONS OF WAVELET TRANSFORM IN SIGNAL PROCESSING"
Problem 14. Consider a signal f(t)with the following data:
f(t) = (1,if 0≤t≤1
0,otherwise
Let’s analyze this signal using the Haar wavelet transform.
a) Calculate the Haar wavelet coefficients ck,m corresponding to f(t).
b) Determine the wavelet transform of f(t).
c) Reconstruct the original signal using the inverse Haar wavelet transform from the coefficients
found in part (a).
Solution 14.
a) The Haar wavelet coefficients ck,m can be calculated as:
ck,m =Z∞
−∞
f(t)ψk,m(t)dt
Here, ψk,m(t)is the Haar wavelet function. Since the Haar wavelet is a piecewise constant
function, we have:
ψk,m(t) = (2k
2,if 2km≤t < 2k(m+ 1)
0,otherwise
Given f(t), we can calculate the coefficients by integrating over the intervals where ψk,m(t)is
non-zero. For the Haar wavelet, we only need to consider the intervals [0,1) and [1,2).
For the interval [0,1):
c0,0=Z1
0
f(t)ψ0,0(t)dt
c0,0=Z1
0
(1) ·(20
2)dt
c0,0=Z1
0
dt = 1
For the interval [1,2):
c1,0=Z2
1
f(t)ψ1,0(t)dt
c1,0=Z2
1
(0) ·(21
2)dt
c1,0= 0
Therefore, the Haar wavelet coefficients for f(t)are c0,0= 1 and c1,0= 0.
b) The wavelet transform of f(t)is given by:
Wf(a, b) = X
k,m
ck,m ·ψk,m(t)
Since c1,0= 0, the wavelet transform simplifies to:
Wf(a, b) = c0,0·ψ0,0(t)=1·20
2= 1
c) Reconstructing the original signal using the inverse Haar wavelet transform involves summing
the scaled and shifted Haar wavelets:
f(t) = X
k,m
ck,m ·ψk,m(t)
Substitute the coefficients c0,0= 1 and c1,0= 0 into the formula:
f(t) = 1 ·ψ0,0(t) = 1 ·20
2= 1
Therefore, the reconstructed signal using the inverse Haar wavelet transform is a constant
signal f(t)=1.
15 "APPLICATION OF WAVELET TRANSFORM IN SIGNAL DENOISING AND COMPRES-
SION"
Problem 15. Consider a signal f(t)defined as:
f(t) = 2 sin(4πt) + 5 cos(8πt) + 3 sin(12πt)
a) Using the wavelet transform, determine the frequency components present in the signal
f(t). b) If we apply a wavelet transform and find that the wavelet coefficients corresponding to the
frequency component 4πare very small compared to the coefficients at other frequencies, explain
why this might indicate the presence of noise in the signal. c) How can wavelet transforms be used
in denoising this signal?
Solution 15.
a) To determine the frequency components present in the signal f(t), we need to apply the
wavelet transform. Let’s consider the Haar wavelet transform for simplicity. The Haar wavelet
transform of a signal f(t)is given by:
W(a, b) = Z∞
−∞
f(t)ψa,b(t)dt
Where ψa,b(t)is the Haar wavelet defined as:
ψa,b(t) =
1if a≤t<a+1
2
−1if a+1
2≤t<a+ 1
0otherwise
Let’s calculate the wavelet coefficients for the given signal f(t)at different scales a:
W(a, b) = Z∞
−∞
f(t)ψa,b(t)dt
For a= 1, we have:
W(1, b) = Z∞
−∞
(2 sin(4πt) + 5 cos(8πt) + 3 sin(12πt))ψ1,b(t)dt
Similarly, we can calculate W(2, b),W(3, b), and so on to determine the frequency components
present in the signal f(t).
b) If the wavelet coefficients corresponding to the frequency component 4πare very small com-
pared to the coefficients at other frequencies, it might indicate the presence of noise in the signal.
This is because in the absence of noise, the coefficients for the true signal frequencies would be
much larger compared to any noise present in the signal.
c) Wavelet transforms can be used in denoising the signal by thresholding the wavelet coeffi-
cients. By setting small coefficients below a certain threshold to zero, we can eliminate noise while
retaining the important frequency components of the signal. This is a common technique in signal
denoising using wavelet transforms.
16 ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES
Problem 17. Consider a non-stationary signal given by f(t) = sin(2πt2).
a) Determine the Fourier transform of the signal f(t).
b) Compute the time-frequency representation of f(t)using the Short-Time Fourier Transform
(STFT) with a Gaussian window function g(t)given by g(t) = e−t2.
Solution 17.
a) The Fourier transform of a signal f(t)is defined as
F(ω) = Z∞
−∞
f(t)e−iωtdt.
For the given signal f(t) = sin(2πt2), we have
F(ω) = Z∞
−∞
sin(2πt2)e−iωtdt.
This particular integral does not have a closed-form solution, so we would need to evaluate it using
numerical methods or approximation techniques.
b) The Short-Time Fourier Transform (STFT) of a signal f(t)using a window function g(t)is
defined as
ST F T (t, ω) = Z∞
−∞
f(τ)g(τ−t)e−iωτ dτ,
where the overline denotes complex conjugation. For a Gaussian window function g(t) = e−t2, the
STFT of the signal f(t) = sin(2πt2)can be computed as
ST F T (t, ω) = Z∞
−∞
sin(2πτ2)e−(τ−t)2e−iωτ dτ.
We can evaluate this integral numerically to compute the time-frequency representation of the
signal using the given Gaussian window function.
17 "CONVERGENCE AND DIVERGENCE OF FOURIER SERIES FOR PIECEWISE FUNC-
TIONS"
Problem 17. Consider the function f(x)defined as:
f(x) = (x0≤x<π
−x−π < x < 0
a) Find the Fourier series of f(x).
b) Determine the conditions for which the Fourier series of f(x)converges.
Solution 17.
a) To find the Fourier series of f(x), we need to compute the coefficients anand bn:
an=1
πZπ
−π
f(x) cos(nx)dx
bn=1
πZπ
−π
f(x) sin(nx)dx
Let’s start with the calculation of an:
an=1
πZ0
−π
(−x) cos(nx)dx +Zπ
0
xcos(nx)dx
Solving these integrals and simplifying yields:
an=2((−1)n−1)
πn2
Now, let’s find bn:
bn=1
πZ0
−π
(−x) sin(nx)dx +Zπ
0
xsin(nx)dx
After solving and simplifying the integrals, we get:
bn= 0
The Fourier series of f(x)is then given by:
f(x)∼2
π1
12−1
32cos(x) + 1
52cos(2x)− · · ·
b) The Fourier series of f(x)converges if the function f(x)satisfies the Dirichlet conditions. In
this case, since f(x)is piecewise continuous and has a finite number of discontinuities within a
period of 2π, the Fourier series of f(x)converges everywhere.
18 "NON-LINEAR FOURIER TRANSFORM TECHNIQUES FOR SIGNAL ANALYSIS"
Problem 1. Consider a signal f(t) = sin(2πt) + 2 cos(4πt).
a) Compute the Fourier Transform of f(t).
b) Find the frequency components present in the signal f(t).
c) Determine the amplitude and phase of each frequency component.
Solution 1.
a) The Fourier Transform of a signal f(t)is defined as:
F(ω) = Z∞
−∞
f(t)e−jωtdt
Applying the definition to the given signal f(t), we have:
F(ω) = Z∞
−∞
[sin(2πt) + 2 cos(4πt)]e−jωtdt
=Z∞
−∞
sin(2πt)e−jωtdt + 2 Z∞
−∞
cos(4πt)e−jωtdt
Solving the integrals, we get:
F(ω) = π(δ(ω−2π)−δ(ω+ 2π)) + π(δ(ω−4π) + δ(ω+ 4π))
b) The frequency components present in the signal f(t)are 2πand 4π.
c) To determine the amplitude and phase of each frequency component, we can use the prop-
erties of the Dirac delta function in the Fourier Transform. The amplitude is the coefficient in front
of the delta function and the phase is given by the sign of the delta function. Thus, we have:
For 2πcomponent: Amplitude = πand Phase = (δ(ω−2π)−δ(ω+ 2π)) = 0
For 4πcomponent: Amplitude = πand Phase = (δ(ω−4π) + δ(ω+ 4π)) = 0
19 "ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES"
Problem 1. Consider a continuous-time signal x(t)defined as:
x(t) = 2 cos(3πt) + 3 sin(6πt)
Find the Fourier transform X(ω)of the signal x(t).
Solution 1.
The Fourier transform of a continuous-time signal x(t)is defined as:
X(ω) = Z∞
−∞
x(t)e−jωtdt
Plugging in the expression for x(t), we have:
X(ω) = Z∞
−∞
(2 cos(3πt) + 3 sin(6πt))e−jωtdt
Expanding the trigonometric functions using Euler’s formula, we get:
X(ω) = Z∞
−∞
2(ej3πt +e−j3πt
2) + 3(ej6πt −e−j6πt
2j)e−jωtdt
X(ω) = Z∞
−∞
(ej(3π−ω)t+e−j(3π+ω)t−3
2j(ej(6π−ω)t−e−j(6π+ω)t))dt
Remembering that the Fourier transform has a property with delta functions, we find:
X(ω) = 2π(δ(ω−3π) + δ(ω+ 3π)) −3π(δ(ω−6π) + δ(ω+ 6π))
Thus, the Fourier transform of the signal x(t)is:
X(ω) = 2π(δ(ω−3π) + δ(ω+ 3π)) −3π(δ(ω−6π) + δ(ω+ 6π))
Certainly! Here is a numerical problem in the realm of Advanced Harmonic Analysis and Fourier
Transform Techniques with a step-by-step solution:
20 "ADVANCED HARMONIC ANALYSIS AND FOURIER TRANSFORM TECHNIQUES"
Problem 21. Consider a non-stationary signal f(t)defined as:
f(t) = (sin(2πt) 0 ≤t < 1
0otherwise
Calculate the Fourier transform F(ω)of the signal f(t).
Solution 21. To find the Fourier transform F(ω)of the given signal f(t), we use the definition
of the Fourier transform:
F(ω) = Z∞
−∞
f(t)e−iωtdt
Given that f(t)is defined piecewise, we split the integral into two parts:
F(ω) = Z1
0
sin(2πt)e−iωtdt +Z∞
1
0·e−iωtdt
Solving the first integral, we can proceed to calculate the Fourier transform. Let’s solve the
integral as follows:
F(ω) = Z1
0
sin(2πt)e−iωtdt
Applying Euler’s formula to express the sine function in terms of complex exponentials, we have:
F(ω) = Z1
0
ei2πt −e−i2πt
2ie−iωtdt
F(ω) = 1
2iZ1
0
ei(2π−ω)tdt −Z1
0
e−i(2π+ω)tdt
F(ω) = 1
2i"ei(2π−ω)t
i(2π−ω)
1
0−e−i(2π+ω)t
−i(2π+ω)
1
0#
F(ω) = 1
2i"ei(2π−ω)−1
i(2π−ω)−e−i(2π+ω)−1
−i(2π+ω)#
This integral evaluation gives us the Fourier transform F(ω).
