Finding the anti-Fourier’s transform (problem with integrals)

69 Views Asked by At

I am having problems finding the integrals when calculating this anti-Fourier's transform.

$$ F(\omega) = 4(\omega ^2- 2\omega +5)^{-1} $$

Then,

$$ f(t) = \frac{1}{j2 \pi} \left(\int_{-\infty }^{\infty } \dfrac{\mathrm{e}^{j \omega t}}{ \omega -2j-1} d\omega-\int_{-\infty }^{\infty } \dfrac{\mathrm{e}^{j \omega t}}{\omega +2j-1} d\omega\right)$$

How do we find this integral: $$ \int_{-\infty }^{\infty } \dfrac{\mathrm{e}^{j \omega t}}{\omega -2j-1} d\omega $$

I tried to use some online calculator but I don't know why it uses some 'weird' u-substitution:

$u=jwt+j⋅(2j−1)w, dt=-j/w du$

Then, $$e^{-j(2j-1)w}\int_{-\infty }^{\infty } \frac{e^{u}}{u} du$$

1

There are 1 best solutions below

0
On BEST ANSWER

Let's simplify things a bit. Let $F(\omega)=\frac{4}{\omega^2-2\omega+5}=\frac{4}{(\omega-1)^2+4}$ and $f(t)$ be given by

$$\begin{align} f(t)&=\mathscr{F}^{-1}\{F\}(t)\\\\ &=\frac2\pi \int_{-\infty}^\infty \frac1{(\omega-1)^2+4}e^{i\omega t}\,d\omega\\\\ &=\frac{2}{\pi}e^{it}\int_{-\infty}^\infty \frac{1}{\omega^2+4}e^{i\omega t}\,d\omega\tag1 \end{align}$$

Now the analysis boils down to evaluation of the integral on the right-hand side of $(1)$. This can be accomplished in a number of ways.


METHODOLOGY $1$: REAL ANALYSIS

Let $g(t)=\int_{-\infty}^\infty \frac{1}{\omega^2+4}e^{i\omega t}\,d\omega=2\int_0^\infty \frac{1}{\omega^2+4}\cos(\omega t)\,d\omega$. One can justify differentiating under the integral to find that

$$\begin{align} g'(t)&=2\int_0^\infty \frac{\omega}{\omega^2+4}\sin(\omega t)\,d\omega\\\\ &=2\int_0^\infty \frac{\omega^2+4-4}{\omega(\omega^2+4)}\sin(\omega t)\,d\omega\\\\ &=2\int_0^\infty \frac{\sin(\omega t)}{\omega}\,d\omega-8\int_0^\infty \frac{\sin(\omega t)}{\omega(\omega^2+4)}\,d\omega\\\\ &=\pi \text{sgn}(t)-8\int_0^\infty \int_0^t \frac{\cos(\omega t')}{\omega^2+4}\,dt'\,d\omega\tag2 \end{align}$$

For $t\ne 0$, we can differentiate $(2)$ to find that $g''(t)+4g(t)=0$, with $g(0)=\pi/2$ and $g'(0)=0$. Therefore, we find that

$$g(t)=\frac\pi2e^{-2|t|}\tag3$$

Using $(3)$ in $(1)$ reveals

$$f(t)=e^{-2|t|+it}$$



METHODOLOGY $2$: COMPLEX ANALYSIS

We can evaluate the integral on the right-hand side of $(1)$ using the residue theorem. Proceeding, we have

$$\begin{align} \int_{-\infty}^\infty \frac{e^{i\omega t}}{\omega^2 +4}\,d\omega&=2\pi i \text{Res}\left(\frac{e^{i\omega t}}{\omega^2 +4}, \omega =2i\text{sgn}(t)\right)\\\\ &=2\pi i \frac{e^{-2|t|}}{4i}\\\\ &=\frac\pi 2 e^{-2|t|}\tag4 \end{align}$$

Using $(4)$ in $(1)$ yields

$$f(t)=e^{-2|t|+it}$$

as expected!



METHODOLOGY $3$: DISTRIBUTION THEORY

Let $h_-(t)=\int_{-\infty}^\infty \frac{e^{i\omega t}}{\omega-1- i2}\,d\omega$. In distribution, we have

$$\begin{align} h_-'(t)&=\int_{-\infty}^\infty \frac{i\omega e^{i\omega t}}{\omega-1- i2}\,d\omega\\\\ &=i\underbrace{\int_{-\infty}^\infty e^{i\omega t}\,d\omega}_{=2\pi \delta(t)} +i(1+ i2)h(t) \end{align}$$

where $\delta$ is the Dirac Delta. Therefore, in distribution we find that $h_-'(t) -(-2+i)h_-(t)=i2\pi \delta(t)$. Since $\lim_{t\to \pm \infty}h_-(t)=0$, we find that

$$h_-(t)=i2\pi e^{- 2t+it}u(t)$$

Similarly, if $h_+(t)=\int_{-\infty}^\infty \frac{e^{i\omega t}}{\omega-1+ i2}\,d\omega$, we find that

$$h_+(t)=-i2\pi e^{2t+it}u(-t)$$

where $h_-(t)-h_+(t)=i2\pi e^{-2|t|+it}$.

Multiplying by $\frac{1}{i2\pi}$ recovers the expected result!