Fourier analysis to explain three

Quote: https://www.jianshu.com/p/f5a89d76eb28

Previous simply describes what is Fourier series, finally got the cycle 2\picoefficient solution Fourier series, then how to get in any period of the Fourier series it?

We look at the period of 2\pithe series expressed as a function of Fourier:
f(x)= \frac{a_0}{2}+\sum_{n=1}^\infty (a_ncos(nx) +b_nsin(nx) )
the corresponding solution is:
a_0=\frac{1}{\pi }\int_{-\pi}^\pi f(x)dx
a_n=\frac{1}{\pi }\int_{-\pi}^\pi f(x)cos(nx)dx
b_n = \ frac {1} {\ pi} \ int _ {- \ pi} ^ \ pi f (x) end (nx) dx
how to turn it into a function of any cycle of it?

In fact, here only a simple change to the meta-operation.
For chestnut:
f(x)=\sin(x)its cycle 2\pi, f (x) = f (x + 2 \ pi). We order  x=2t, then f (x) = f (2t) = f (2t + 2 \ pi) = f (2 (t + \ pi)), finishing next:
f(2(t))=f(2(t+\pi))
So is it for t been converted to a period of \pithe function.
so for the period of 2Lthe function f (t) (the convenience of calculation) is just so  x=\frac{\pi}{L}tinto the primary cycle 2\pifunction can be: the
f(t)= \frac{a_0}{2}+\sum_{n=1}^\infty (a_ncos(\frac{\pi nt}{L}) +b_nsin(\frac{\pi nt}{L}) )
same can be obtained:
\cos(x)=\cos(\frac{\pi t}{L})
\sin(x)=\sin(\frac{\pi t}{L}) )
\ Int _ {- \ pi} ^ \ pi dx = \ int _ {-} ^ L L d \ frac {\ pi t} {L}
Finally, we get:
a_0=\frac{1}{L }\int_{-L}^L f(t)dt
a_n=\frac{1}{L }\int_{-L}^L f(t)cos(\frac{\pi nt}{L})dt
b_n=\frac{1}{L }\int_{-L}^L f(t)sin(\frac{\pi nt}{L})dt

The process is simple, I have omitted, after all, life is short.

2 Fourier series in plural form

We write about Fourier series equation:
f(t)= \frac{a_0}{2}+\sum_{n=1}^\infty (a_ncos(n \omega t) +b_nsin(n \omega t) ) \\ \omega=\frac{2\pi }{T}
where T represents the periodic function, which is above 2L, the corresponding solution is:

a_0=\frac{2}{T }\int_{0}^T f(t)dt
a_n=\frac{2}{T }\int_{0}^T f(t)cos(n \omega t)dt
b_n=\frac{2}{T }\int_{0}^T f(t)sin(n \omega t)dt

Want to get the plural form of the Fourier series, we need to first understand Euler's formula.
About Euler's formula, there are many online blog, will not elaborate here, but simply the nature of Euler's formula under.
We look at the following formula:
e^{i\theta}=cos(\theta)+isin(\theta)

e^{i\theta}Can be seen as a vector in the complex plane, which is projected to the real axis cos(\theta)to the imaginary axis is the projection isin(\theta), wherein \thetathe angle between the vectors is the real axis.

 
image.png

 

Euler's formula and intuitive understanding is circular motion on the complex plane

 

 
Euler's formula .gif

随着\theta变化,e^{i\theta}就变成圆周运动了。而前面的系数a则是圆的半径,当a=1的时候就是在单位圆上做圆周运动。

而且通过欧拉公式,我们可以得到三角函数的复数形式:
\cos(\theta)=\frac{1}{2}(e^{i \theta}+e^{-i \theta}) \\ \sin(\theta)=-\frac{1}{2}i(e^{i \theta}-e^{-i \theta} )

将上面的复变三角函数替换傅里叶级数中的三角函数得到:
f(t)= \frac{a_0}{2}+\sum_{n=1}^\infty (a_n\frac{1}{2}(e^{i n \omega t}+e^{-i n \omega t}) -b_n\frac{i}{2}(e^{i n \omega t}-e^{-i n \omega t})) \\ =\frac{a_0}{2}+\sum_{n=1}^\infty ((a_n\frac{1}{2}e^{i n \omega t}+a_n\frac{1}{2}e^{-i n \omega t}) -(b_n\frac{i}{2}e^{i n \omega t}- b_n\frac{i}{2}e^{-i n \omega t})) \\ =\frac{a_0}{2}+\sum_{n=1}^\infty ((\frac{a_n-b_ni}{2}e^{i n \omega t}+\frac{a_n+b_ni}{2}e^{-i n \omega t}) ) \\ =\sum_{n=0}^0\frac{a_0}{2} e^{in\omega t}+\sum_{n=1}^\infty (\frac{a_n-b_ni}{2}e^{i n \omega t})+\sum_{n=1}^\infty(\frac{a_n+b_ni}{2}e^{-i n \omega t} )
我们令\sum_{n=1}^\infty(\frac{a_n+b_ni}{2}e^{-i n \omega t} )中的n为-n
则得到:
f(t)=\sum_{n=0}^0\frac{a_0}{2} e^{in\omega t}+\sum_{n=1}^\infty (\frac{a_n-b_n}{2}e^{i n \omega t})+\sum_{n=-\infty}^{-1}(\frac{a_{-n}+b_{-n}i}{2}e^{-i -n \omega t} )
所以可以看到n的范围变成了-\infty 到 \infty,并且每一项都有e^{i n \omega t},于是我们可以得到一个漂亮的形式:
f(t)=\sum_{n=-\infty}^{\infty}C_n e^{in\omega t}

其中C_n分为3中情况:
n=0 \quad\quad C_n=\frac{a_0}{2} \\ n=1,2,3... \quad C_n=\frac{a_n-b_ni}{2}\\ n=-1,-2,-3... \quad C_n=\frac{a_{-n}+b_{-n}i}{2}

我们将傅里叶级数之前的解带入上边
a_0=\frac{2}{T }\int_{0}^T f(t)dt \\ a_n=\frac{2}{T }\int_{0}^T f(t)cos(n \omega t)dt \\ b_n=\frac{2}{T }\int_{0}^T f(t)sin(n \omega t)dt

当n=0的时候:

n=0 \quad\quad C_n=\frac{a_0}{2}=\frac{2}{T }\int_{0}^T f(t)dt = \frac{1}{T }\int_{0}^T f(t)e^{-n \omega t}dt\\

当 n=1,2,3...的时候

\quad C_n=\frac{a_n-b_ni}{2}=\frac{1}{2}( \frac{2}{T }\int_{0}^T f(t)cos(n \omega t)dt-\frac{2i}{T }\int_{0}^T f(t)sin(n \omega t)dt)\\ = \frac{1}{T }\int_{0}^T f(t)cos(n \omega t)-isin(n \omega t)dt
这里因为cos是偶函数,sin是奇函数所以:

\frac{1}{T }\int_{0}^T f(t)cos(n \omega t)-isin(n \omega t)dt = \frac{1}{T }\int_{0}^T f(t)cos(-n \omega t)+isin(-n \omega t)dt =\frac{1}{T }\int_{0}^T f(t)e^{-in \omega t}dt

当 n=-1,-2,-3...的时候

C_n=\frac{a_{-n}+b_{-n}i}{2}= \frac{1}{2}( \frac{2}{T }\int_{0}^T f(t)cos(-n \omega t)dt+\frac{2i}{T }\int_{0}^T f(t)sin(-n \omega t)dt) \\ =\frac{1}{T }\int_{0}^T f(t)cos(-n \omega t)+isin(-n \omega t)dt =\frac{1}{T }\int_{0}^T f(t)e^{-in \omega t}dt

可以惊奇的发现,三种情况的解是一样的。所以对于任意周期函数,我们都可以写成:
f(t)=\sum_{n=-\infty}^{\infty}C_n e^{in\omega t} \\ C_n=\frac{1}{T }\int_{0}^T f(t)e^{-in \omega t}dt
但其中的每一项是什么意思呢?
还记得之前说的e^{i\theta}的本质吗?在圆上做圆周运动,那么e^{-in \omega t}也是在做周期运动了。那n\omega又是什么呢?
我们知道\omega=\frac{2\pi }{T},所以我们可以把\omega 看成是以2\pi为单位的频率(正常来讲频率是\frac{1}{T})。而系数n是就可以看成是几倍的基频,正数是逆时针运动,负数就是顺时针运动。在图形上的反应就是,频率越高,转的越快了
,但其最小公共周期是一样的。
1倍基频

 
1倍频.gif

 

 

 

10倍基频

 
10倍频.gif

那么系数C_n怎么理解呢?前面说过ae^{i\theta}的系数a是代表e^{i\theta}运动的圆半径,这里C_n是复数是不是也能这样理解呢?其实粗糙来讲是可以这样理解的。
看个图,只管的理解下把

 
gif1.gif

 

上图中红色的向量相对于蓝色的向量只是多了系数C_n=\sqrt2+i\sqrt2,所以红色向量运动的半径就是2刚好是复数\sqrt2+i\sqrt2的模长乘以1,当然除此之外,红色向量的幅角也变大了些。这些都是因为复数的乘法性质---复数相乘表现为幅角相加,模长相乘。
这下,当有人和你说傅里叶变换是把时域变换到频域上,你应该就很容易理解是什么意思了。频域就是1倍,2倍,3倍.......的\omega,而每个 \omega都有自己的幅长C_n,当把这些所有的\omega相加,就得到时域中的图像。

更加生动有趣的介绍可以参见傅里叶分析之掐死教程,我这里是从数学的角度来介绍傅里叶变换。

3 推广到非周期函数上

目前该证明的都差不多了,还有最后一个任务,就是推广到非周期函数上。对于非周期函数,我们可以看成是周期无限远的函数,那也就是周期T变成\infty的时候傅里叶级数。随则T的变大w也就不断的减小,当T趋近于 \infty的时候,w也由1w,2w,3w......变成了\Delta w,那么很自然就需要对w 做积分。

我们先看下
\Delta w=(n+1)w-nw=w=\frac{2\pi}{T} \\ \frac{1}{T} =\frac{\Delta w}{2\pi}\\

当T趋近于 \infty的时候 我们可以得到:
\int_{0}^T f(t)dt \quad -> \quad \int_{-\infty}^\infty f(t)dt\\ \sum_{n=-\infty}^{\infty}\Delta w \quad -> \quad \int_{-\infty}^\infty dw
将这些带入 傅里叶级数,并且T趋近于\infty,就得到:
f(t)=\sum_{n=-\infty}^{\infty}\frac{1}{T }\int_{0}^T f(t)e^{-in \omega t}dt e^{in\omega t} \\ =\sum_{n=-\infty}^{\infty}\frac{\Delta w}{2\pi}\int_{0}^T f(t)e^{-in \omega t}dt e^{in\omega t}\\ =\frac{1}{2\pi} \int_{-\infty}^{\infty} \int_{-\infty}^{\infty}f(t)e^{-i \omega t}dt e^{i\omega t}dw
其中画红圈的地方就是傅里叶变换

 
image.png


It is generally written as a function of, in fact, equivalent to the previous:

 

The entire formula is the inverse Fourier transform, wrote:
F^{-1}(t)=\frac{1}{2\pi} \int_{-\infty}^{\infty} \int_{-\infty}^{\infty}f(t)e^{-i \omega t}dt e^{i\omega t}dw=\frac{1}{2\pi} \int_{-\infty}^{\infty} F(w) e^{i\omega t}dw

Guess you like

Origin www.cnblogs.com/Sweepingmonk/p/11582386.html