228.223 Assignment 4. Solutions — 2012 1. Find the Fourier transform of the function
f(t) =
2−t2, |t|<1 0, otherwise You will need to use integration by parts.
F(ω) = Z ∞
−∞
f(t)e−iωtdt = Z 1
−1
(2−t2)e−iωtdt
Now Z
t2e−iωtdt=e−iωt[it2/ω+ 2t/ω2−2i/ω3]
so Z 1
−1
(2−t2)e−iωtdt=
e−iωt(2i/ω3 −2t/ω2+ (2−t2)i/ω)1
−1
=e−iω(2i/ω3−2/ω2+i/ω)−eiω(2i/ω3+2/ω2+i/ω) = 4 sinω
ω3 −4 cosω
ω2 +2 sinω ω 2. IfF(ω)is the Fourier transform off(t), show that
(a)
F(0) = Z ∞
−∞
f(t)dt
F(ω) = Z ∞
−∞
f(t)e−iωtdt so
F(0) = Z ∞
−∞
f(t)e0×tdt= Z ∞
−∞
f(t)dt
(b)
f(0) = 1 2π
Z ∞
−∞
F(ω)dω
Similarly:
f(t) = 1 2π
Z ∞
−∞
F(ω)eiωtdω
and it follows.
(c) Show that iff(t)is even,
F(ω) = 2 Z ∞
0
f(t) cos (ωt)dt
F(ω) = Z ∞
−∞
f(t)e−iωtdt= Z 0
−∞
f(t)e−iωtdt+ Z ∞
0
f(t)e−iωtdt=I1+I2
1
Puttingt=−sinI1we have I1 =
Z 0
∞
f(−s)eiωs(−ds) = Z ∞
0
f(s)eiωsds
sincef is even. SincesinI1is a dummy variable, we can replace it byt, obtaining
F(ω) =I1+I2 = Z ∞
0
f(t)eiωtdt+ Z ∞
0
f(t)e−iωtdt= Z ∞
0
f(t)[eiωt+e−iωt]dt
= 2 Z ∞
0
f(t) cos (ωt)dt
3. Letf(t) =H(t)e−2t, whereHis the Heaviside function, and
g(t) =
1, |t|<2 0, otherwise
(a) Find the convolution off(t)andg(t). Your answer will be defined piece- wise. We will find
h(t) = Z ∞
−∞
f(τ)g(t−τ)dτ
If t + 2 < 0, i.e. t < −2, there is no overlap of the functions, so the convolution is zero.
Whent−2<0andt+ 2>0, i.e. when−2< t <2, the functions overlap for0< τ < t+ 2, so
h(t) = Z t+2
0
e−2τdτ = (1−e−2(t+2))/2
Fort−2>0, i.e. t >2, the overlap is fort−2< τ < t+ 2, so h(t) =
Z t+2 t−2
e−2τdτ =e−2t(e4−e−4)/2
So
h(t) =
0, t <−2
(1−e−2(t+2))/2, −2< t <2 e−2t(e4−e−4)/2, 2< t (b) Find the Fourier transforms off(t)andg(t).
F(ω) = Z ∞
0
e−2te−iωtdt= 1 2 +iω
G(ω) = Z 2
−2
e−iωtdt=
e−iωt
−iω 2
−2
= (2 sin (2ω))/ω
2
(c) Using your results, write down the Fourier transform of the result in part (a).
It is the product:
F(ω)G(ω) = 2 sin (2ω) ω(2 +iω)
4. Find the discrete Fourier transform of the sequence {1,1,1,−1}. Then find the inverse DFT of the same sequence.
X[r] = 1 N
N−1
X
k=0
x[k]e−i2πrk/N
X[0] = (1 + 1 + 1−1)/4 = 1/2
X[1] = (1 +e−iπ/2+e−iπ−e−3iπ/2)/4 =−i/2 X[2] = (1 +e−iπ+e−2iπ−e−3iπ)/4 = 1/2 X[3] = (1 +e−3iπ/2+e−3iπ−e−9iπ/2)/4 =i/2 And
x[r] =
N−1
X
r=0
X[r]ei2πrk/N
so
x[0] = 1 + 1 + 1−1 = 2 x[1] = 1 +eiπ/2+eiπ−e3iπ/2 = 2i
x[2] = 1 +eiπ+e2iπ−e3iπ = 2 x[3] = 1 +e3iπ/2+e3iπ−e9iπ/2 =−2i
5. Write a matrix W such that X = W x, where x is a column vector with 4 values of a sequence in it and X is a column vector containing the DFT of that sequence. Use this matrix to find the DFT of{1,0,−3,−1}.
W = 1 4
1 1 1 1
1 e−iπ/2 e−iπ e−3iπ/2 1 e−iπ e−2iπ e−3iπ 1 e−3iπ/2 e−3iπ e−9iπ/2
= 1 4
1 1 1 1
1 −i −1 i
1 −1 1 −1
1 i −1 −i
X = (−3,4−i,−1,4 +i)/4
6. For the following difference equations, state their order, and whether they are homogeneous or inhomogeneous.
3
(a)
2x[n]−nx[n−1] +x[n−2] + 2n2 = 0 Order=n−(n−2) = 2. Rewrite as
2x[n]−nx[n−1] +x[n−2] = −2n2
Inhomogeneous.
(b)
(x[n+ 1]−x[n−1])/2 =x[n]
Order=n+ 1−(n−1) = 2. Rewrite as
(x[n+ 1]−x[n−1])/2−x[n] = 0
Homogeneous.
(c)
3x[n−1]
x[n−2] = n−1 n+ 1 Order=n−1−(n−2) = 1. Rewrite as
3(n+ 1)x[n−1]−(n−1)x[n−2] = 0 Homogeneous.
4