文章目录
傅里叶变换表示形式
假设任意一个周期函数能否表示成为一系列正余弦函数的和:
f
(
t
)
=
A
0
+
∑
n
=
1
∞
A
n
s
i
n
(
n
ω
0
t
+
φ
n
)
其
中
:
A
n
s
i
n
(
n
ω
0
t
+
φ
n
)
=
A
n
s
i
n
φ
n
c
o
s
n
ω
0
t
+
A
n
c
o
s
φ
n
s
i
n
n
ω
0
t
\begin{aligned} &f(t)=A_0+\sum_{n=1}^{\infty}A_n sin(n\omega_0 t+\varphi_n)\\ 其中:&A_n sin(n\omega_0 t+\varphi_n)=A_nsin\varphi_ncosn\omega_0 t+A_ncos\varphi_nsinn\omega_0 t \\ \end{aligned}
其中:f(t)=A0+n=1∑∞Ansin(nω0t+φn)Ansin(nω0t+φn)=Ansinφncosnω0t+Ancosφnsinnω0t为计算的方便和表达的整洁,令:
a
0
2
=
A
0
,
a
n
=
A
n
s
i
n
φ
n
,
b
n
=
A
n
c
o
s
φ
n
,
ω
0
t
=
x
⟹
f
(
t
)
=
a
0
2
+
∑
n
=
1
∞
(
a
n
c
o
s
n
x
+
b
n
s
i
n
n
x
)
(1)
\begin{aligned} &\frac{a_0}{2}=A_0,a_n = A_nsin\varphi_n, b_n=A_ncos \varphi_n, \omega_0 t=x\\ \color {blue}\Longrightarrow &\color {blue}f(t)=\frac{a_0}{2}+\sum_{n=1}^{\infty}(a_n cos\ nx+b_nsin\ nx)\tag{1} \end{aligned}
⟹2a0=A0,an=Ansinφn,bn=Ancosφn,ω0t=xf(t)=2a0+n=1∑∞(ancos nx+bnsin nx)(1)
如何计算参数 a 0 , a n , b n a_0,a_n,b_n a0,an,bn?
高次三角函数积分公式:
∫
0
π
2
cos
n
(
x
)
d
x
=
∫
0
π
2
s
i
n
n
(
x
)
d
x
=
{
(
n
−
1
)
n
×
(
n
−
3
)
(
n
−
2
)
×
…
×
4
5
×
2
3
n为奇数
(
n
−
1
)
n
×
(
n
−
3
)
(
n
−
2
)
×
…
×
3
4
×
1
2
×
π
2
n为偶数
\begin{aligned} &\int^{\frac{π}{2}}_0\cos^n(x)dx=\int_0^{\frac \pi 2}sin^n(x)dx\\ &= \left\{\begin{aligned} &\frac{(n-1)}n\times\frac{(n-3)}{(n-2)}\times…\times\frac45\times\frac23\qquad\qquad\ \color{blue}\text{n为奇数}\\ &\frac{(n-1)}n\times\frac{(n-3)}{(n-2)}\times…\times\frac34\times\frac12\times\fracπ2\qquad \color{blue}\text{n为偶数} \end{aligned}\right . \end{aligned}
∫02πcosn(x)dx=∫02πsinn(x)dx=⎩⎪⎪⎪⎨⎪⎪⎪⎧n(n−1)×(n−2)(n−3)×…×54×32 n为奇数n(n−1)×(n−2)(n−3)×…×43×21×2πn为偶数
根据三角函数的正交性有:
∫
−
π
π
c
o
s
n
x
d
x
=
0
(
n
=
1
,
2
,
3
,
…
)
∫
−
π
π
s
i
n
n
x
d
x
=
0
(
n
=
1
,
2
,
3
,
…
)
∫
−
π
π
s
i
n
k
x
c
o
s
n
x
d
x
=
0
(
n
=
1
,
2
,
3
,
…
)
∫
−
π
π
c
o
s
k
x
c
o
s
n
x
d
x
=
0
(
n
=
1
,
2
,
3
,
…
,
k
≠
n
)
∫
−
π
π
s
i
n
k
x
s
i
n
n
x
d
x
=
0
(
n
=
1
,
2
,
3
,
…
,
k
≠
n
)
\begin{aligned} &\int_{-\pi}^{\pi}cos\ nxdx = 0 (n=1,2,3,\dots)\\ &\int_{-\pi}^{\pi}sin\ nxdx = 0 (n=1,2,3,\dots)\\ &\int_{-\pi}^{\pi}sin\ kx \ cos\ nx dx = 0 (n=1,2,3,\dots)\\ &\int_{-\pi}^{\pi}cos\ kx\ cos\ nxdx = 0 (n=1,2,3,\dots,k\neq n)\\ &\int_{-\pi}^{\pi}sin\ kx\ sin\ nxdx = 0 (n=1,2,3,\dots,k\neq n) \end{aligned}
∫−ππcos nxdx=0(n=1,2,3,…)∫−ππsin nxdx=0(n=1,2,3,…)∫−ππsin kx cos nxdx=0(n=1,2,3,…)∫−ππcos kx cos nxdx=0(n=1,2,3,…,k=n)∫−ππsin kx sin nxdx=0(n=1,2,3,…,k=n)
1、对(1)进行积分得到
a
0
\color {blue}a_0
a0:
∫
−
π
π
f
(
x
)
d
x
=
∫
−
π
π
a
0
2
d
x
+
∑
k
=
1
∞
[
a
n
∫
−
π
π
c
o
s
n
x
d
x
+
b
n
∫
−
π
π
s
i
n
n
x
d
x
]
⏟
0
⟹
a
0
=
1
π
∫
−
π
π
f
(
x
)
d
x
\begin{aligned} &\int_{-\pi}^{\pi}f(x)dx=\int_{-\pi}^{\pi}\frac{a_0}{2}dx+\underbrace {\sum_{k=1}^{\infty}\left[a_n \int_{-\pi}^{\pi}cos\ nx\ dx+b_n \int_{-\pi}^{\pi}sin\ nx\ dx \right]}\\ &\qquad\qquad\qquad\qquad\qquad\qquad\qquad\qquad\qquad\qquad\ \ \text{0}\\ &\color{blue}\Longrightarrow a_0=\frac1\pi\int_{-\pi}^{\pi}f(x)dx \end{aligned}
∫−ππf(x)dx=∫−ππ2a0dx+
k=1∑∞[an∫−ππcos nx dx+bn∫−ππsin nx dx] 0⟹a0=π1∫−ππf(x)dx
2、原函数乘以
c
o
s
(
k
x
)
cos(kx)
cos(kx)再进行积分得到
a
n
\color {blue}a_n
an:
∫
−
π
π
f
(
x
)
c
o
s
n
x
d
x
=
a
0
2
∫
−
π
π
c
o
s
k
x
d
x
+
∑
k
=
1
∞
[
a
n
∫
−
π
π
c
o
s
k
x
c
o
s
n
x
d
x
+
b
n
∫
−
π
π
s
i
n
n
x
c
o
s
k
x
d
x
]
⟹
a
n
=
1
π
∫
−
π
π
f
(
x
)
c
o
s
n
x
d
x
(
n
=
1
,
2
,
3
,
…
)
\begin{aligned} \int_{-\pi}^{\pi}f(x)cos\ &nx\ dx=\frac{a_0}{2} \int_{-\pi}^{\pi}cos\ kx\ dx+\sum_{k=1}^{\infty}\left[a_n \int_{-\pi}^{\pi}cos\ kx\ cos\ nx\ dx+b_n \int_{-\pi}^{\pi}sin\ nx\ cos\ kx\ dx \right]\\ &\color{blue}\Longrightarrow a_n=\frac{1}{\pi}\int_{-\pi}^{\pi}f(x)cos\ nx\ dx\ \ (n=1,2,3,\dots) \end{aligned}
∫−ππf(x)cos nx dx=2a0∫−ππcos kx dx+k=1∑∞[an∫−ππcos kx cos nx dx+bn∫−ππsin nx cos kx dx]⟹an=π1∫−ππf(x)cos nx dx (n=1,2,3,…)
3、原函数乘以
s
i
n
(
k
x
)
sin(kx)
sin(kx)再进行积分得到
b
n
\color {blue}b_n
bn:
∫
−
π
π
f
(
x
)
s
i
n
n
x
d
x
=
a
0
2
∫
−
π
π
s
i
n
k
x
d
x
+
∑
k
=
1
∞
[
a
n
∫
−
π
π
c
o
s
n
x
s
i
n
k
x
d
x
+
b
n
∫
−
π
π
s
i
n
k
x
s
i
n
n
x
d
x
]
⟹
b
n
=
1
π
∫
−
π
π
f
(
x
)
s
i
n
n
x
d
x
(
n
=
1
,
2
,
3
,
…
)
\begin{aligned} \int_{-\pi}^{\pi}f(x)sin\ &nx\ dx=\frac{a_0}{2} \int_{-\pi}^{\pi}sin\ kx\ dx+\sum_{k=1}^{\infty}\left[a_n \int_{-\pi}^{\pi}cos\ nx\ sin\ kx\ dx+b_n \int_{-\pi}^{\pi}sin\ kx\ sin\ nx\ dx \right]\\ &\color{blue}\Longrightarrow b_n=\frac{1}{\pi}\int_{-\pi}^{\pi}f(x)sin\ nxdx\ \ (n=1,2,3,\dots) \end{aligned}
∫−ππf(x)sin nx dx=2a0∫−ππsin kx dx+k=1∑∞[an∫−ππcos nx sin kx dx+bn∫−ππsin kx sin nx dx]⟹bn=π1∫−ππf(x)sin nxdx (n=1,2,3,…)
什么条件下函数可以展开成傅里叶级数?
定理(收敛定理,狄利克雷(Dirichlet)充分条件): 设 f ( x ) f(x) f(x)是周期为 2 π 2\pi 2π的周期函数,如果它满足:
- 在一个周期内连续或只有有限个第一类间断点
- 在个周期内至多只有有限个极值点
则 f ( x ) f(x) f(x)的傅里叶级数收敛,并且
- 当 x x x是 f ( x ) f(x) f(x)的连续点时,级数收敛于 f ( x ) f(x) f(x);
- 当 x x x是 f ( x ) f(x) f(x)的间断点时,级数收敛于: 1 2 [ f ( x − ) + f ( x + ) ] \ \frac12[f(x^-)+f(x^+)] 21[f(x−)+f(x+)]
对于周期不是 2 π 2\pi 2π的函数呢?
对周期为
2
l
2l
2l 的函数
f
(
x
)
f(x)
f(x),令:
z
=
2
π
x
2
l
\large {z=\frac{2πx}{2l}}
z=2l2πx,得到:
f
(
z
)
f(z)
f(z)的周期为
2
π
2\pi
2π,所以有:
f
(
x
)
=
a
0
2
+
∑
n
=
1
∞
(
a
n
c
o
s
n
π
x
l
+
b
n
s
i
n
n
π
x
l
)
⟹
a
0
=
1
l
∫
−
l
l
f
(
x
)
d
x
⟹
a
n
=
1
l
∫
−
l
l
f
(
x
)
c
o
s
n
π
x
l
d
x
⟹
b
n
=
1
l
∫
−
l
l
f
(
x
)
s
i
n
n
π
x
l
d
x
\begin{aligned} f(x)&=\frac{a_0}{2}+\sum_{n=1}^{\infty}(a_n cos\frac{n\pi x}{l}+b_n sin\frac{n\pi x}{l}) \\ \color{blue}\Longrightarrow a_0&=\frac1l\int_{-l}^{l}f(x)dx\\ \color{blue}\Longrightarrow a_n &= \frac{1}{l}\int_{-l}^{l}f(x)cos\frac{n\pi x}{l}dx\\ \color{blue}\Longrightarrow b_n &= \frac{1}{l}\int_{-l}^{l}f(x)sin\frac{n\pi x}{l}dx \end{aligned}
f(x)⟹a0⟹an⟹bn=2a0+n=1∑∞(ancoslnπx+bnsinlnπx)=l1∫−llf(x)dx=l1∫−llf(x)coslnπxdx=l1∫−llf(x)sinlnπxdx
由欧拉公式:
e
i
θ
=
c
o
s
θ
+
j
s
i
n
θ
,
e
−
j
θ
=
c
o
s
θ
−
j
s
i
n
θ
e^{i\theta}=cos\theta+j\ sin\theta,e^{-j\theta}=cos\theta-j\ sin\theta
eiθ=cosθ+j sinθ,e−jθ=cosθ−j sinθ 可推出傅里叶级数的复数表达式:
f
(
x
)
=
a
0
2
+
∑
n
=
1
∞
(
a
n
c
o
s
n
π
x
l
+
b
n
s
i
n
n
π
x
l
)
=
a
0
2
+
∑
n
=
1
∞
(
a
n
[
1
2
(
e
j
n
π
x
l
+
e
−
j
n
π
x
l
)
]
−
b
n
[
j
2
(
e
j
n
π
x
l
−
e
−
j
n
π
x
l
)
]
)
=
a
0
2
⏟
+
∑
n
=
1
∞
(
a
n
−
j
b
n
2
⏟
e
j
n
π
x
l
+
a
n
+
j
b
n
2
⏟
e
−
j
n
π
x
l
)
c
0
c
n
c
−
n
c
0
=
1
2
l
∫
−
l
l
f
(
x
)
d
x
c
n
=
a
n
−
j
b
n
2
=
1
2
[
1
l
∫
−
l
l
f
(
x
)
c
o
s
n
π
x
l
d
x
−
j
l
∫
−
l
l
f
(
x
)
s
i
n
n
π
x
l
d
x
]
=
1
2
l
∫
−
l
l
(
f
(
x
)
c
o
s
n
π
x
l
−
j
s
i
n
n
π
x
l
)
d
x
=
1
2
l
∫
−
l
l
f
(
x
)
e
−
j
n
π
x
l
d
x
c
−
n
=
a
n
−
j
b
n
2
=
1
2
[
1
l
∫
−
l
l
f
(
x
)
c
o
s
n
π
x
l
d
x
+
j
l
∫
−
l
l
f
(
x
)
s
i
n
n
π
x
l
d
x
]
=
1
2
l
∫
−
l
l
(
f
(
x
)
c
o
s
n
π
x
l
+
j
s
i
n
n
π
x
l
)
d
x
=
1
2
l
∫
−
l
l
f
(
x
)
e
j
n
π
x
l
d
x
⟹
f
(
x
)
=
∑
n
=
−
∞
+
∞
c
n
e
j
n
π
x
l
(
n
=
0
,
±
1
,
±
2
,
…
)
其
中
c
n
=
1
2
l
∫
−
l
l
f
(
x
)
e
−
j
n
π
x
l
d
x
(
n
=
0
,
±
1
,
±
2
,
…
)
\begin{aligned} f(x)&=\frac{a_0}{2}+\sum_{n=1}^{\infty}(a_n cos\frac{n\pi x}{l}+b_n sin\frac{n\pi x}{l}) \\ &=\frac{a_0}{2}+\sum_{n=1}^{\infty}(a_n\left[\frac{1}{2}(e^{j\frac{n\pi x}{l}}+e^{-j\frac{n\pi x}{l}})\right]-b_n\left[\frac{j}{2}(e^{j\frac{n\pi x}{l}}-e^{-j\frac{n\pi x}{l}})\right])\\ &=\underbrace\frac{a_0}{2}+\sum^\infty_{n=1}\left(\underbrace\frac{a_n-jb_n}{2}e^{j\frac{n\pi x}{l}}+\underbrace\frac{a_n+jb_n}{2}e^{-j\frac{n\pi x}{l}}\right)\\ &\Large\quad\ c_0\qquad\quad\ \ c_{n}\qquad\qquad c_{-n}\\[6pt] \Large c_{0}&=\frac1{2l}\int_{-l}^{l}f(x)dx\\[10pt] \Large c_{n}&=\frac{a_n-jb_n}{2}\\ &=\frac12 \left[\frac{1}{l}\int_{-l}^{l}f(x)cos\frac{n\pi x}{l}dx-\frac{j }{l}\int_{-l}^{l}f(x)sin\frac{n\pi x}{l}dx\right]\\ &=\frac1{2l} \int_{-l}^{l}\left(f(x)cos\frac{n\pi x}{l}-j \ sin\frac{n\pi x}{l}\right)dx\\ &=\frac1{2l} \int_{-l}^{l}f(x)e^{-j\frac{n\pi x}{l}}dx\\[12pt] \Large c_{-n}&=\frac{a_n-jb_n}{2}\\ &=\frac12 \left[\frac{1}{l}\int_{-l}^{l}f(x)cos\frac{n\pi x}{l}dx+\frac{j }{l}\int_{-l}^{l}f(x)sin\frac{n\pi x}{l}dx\right]\\ &=\frac1{2l} \int_{-l}^{l}\left(f(x)cos\frac{n\pi x}{l}+j \ sin\frac{n\pi x}{l}\right)dx\\ &=\frac1{2l} \int_{-l}^{l}f(x)e^{j\frac{n\pi x}{l}}dx\\[12pt] \color{blue}\Longrightarrow f(x)&= \sum_{n=-\infty}^{+\infty}c_ne^{j\frac{n\pi x}{l}}\ \ (n=0,\pm 1,\pm 2,\dots)\\ 其中\Large c_n&=\frac1{2l} \int_{-l}^{l}f(x)e^{-j\frac{n\pi x}{l}}dx\ \ (n=0,\pm 1,\pm 2,\dots) \end{aligned}
f(x)c0cnc−n⟹f(x)其中cn=2a0+n=1∑∞(ancoslnπx+bnsinlnπx)=2a0+n=1∑∞(an[21(ejlnπx+e−jlnπx)]−bn[2j(ejlnπx−e−jlnπx)])=
2a0+n=1∑∞(
2an−jbnejlnπx+
2an+jbne−jlnπx) c0 cnc−n=2l1∫−llf(x)dx=2an−jbn=21[l1∫−llf(x)coslnπxdx−lj∫−llf(x)sinlnπxdx]=2l1∫−ll(f(x)coslnπx−j sinlnπx)dx=2l1∫−llf(x)e−jlnπxdx=2an−jbn=21[l1∫−llf(x)coslnπxdx+lj∫−llf(x)sinlnπxdx]=2l1∫−ll(f(x)coslnπx+j sinlnπx)dx=2l1∫−llf(x)ejlnπxdx=n=−∞∑+∞cnejlnπx (n=0,±1,±2,…)=2l1∫−llf(x)e−jlnπxdx (n=0,±1,±2,…)
傅里叶变换
令
ω
n
=
n
π
l
\omega_n=\frac{n\pi}{l}
ωn=lnπ,则:
f
(
x
)
=
∑
n
=
−
∞
+
∞
c
n
e
j
ω
n
x
,
c
n
=
1
2
l
∫
−
l
l
f
(
x
)
e
−
j
ω
n
x
d
x
\begin{aligned} f(x)&= \sum_{n=-\infty}^{+\infty}c_ne^{j\omega_n x},\quad c_n=\frac1{2l} \int_{-l}^{l}f(x)e^{-j\omega_n x}dx \end{aligned}
f(x)=n=−∞∑+∞cnejωnx,cn=2l1∫−llf(x)e−jωnxdx
令
F
n
=
c
n
∗
2
l
F_n=c_n*2l
Fn=cn∗2l,则:
F
n
=
∫
−
l
l
f
(
x
)
e
−
j
ω
n
x
d
x
\begin{aligned} F_n&=\int_{-l}^{l}f(x)e^{-j\omega_n x}dx \end{aligned}
Fn=∫−llf(x)e−jωnxdx
令
Δ
ω
=
π
l
\Delta\omega=\frac\pi l
Δω=lπ,则:
f
(
x
)
=
1
2
π
∑
n
=
−
∞
+
∞
F
n
e
j
ω
n
x
Δ
ω
\begin{aligned} f(x)&=\frac1{2\pi} \sum_{n=-\infty}^{+\infty}F_ne^{j\omega_n x}\Delta\omega \end{aligned}
f(x)=2π1n=−∞∑+∞FnejωnxΔω
当
l
→
∞
,
Δ
ω
→
0
l\rightarrow\infty,\Delta\omega \rightarrow 0
l→∞,Δω→0,离散求和变成连续函数积分
Δ
→
d
ω
,
F
n
→
F
(
ω
)
\Delta\rightarrow d\omega,F_n\rightarrow F(\omega)
Δ→dω,Fn→F(ω):
f
(
x
)
=
1
2
π
∫
−
∞
+
∞
F
(
ω
)
e
j
ω
x
d
ω
=
∫
−
∞
+
∞
F
(
f
)
e
j
2
π
f
x
d
f
F
(
ω
)
=
∫
−
∞
+
∞
f
(
x
)
e
−
j
ω
x
d
x
\begin{aligned} \color{blue}f(x)&=\frac1{2\pi} \int_{-\infty}^{+\infty}F(\omega)e^{j\omega x}d\omega=\int_{-\infty}^{+\infty}F(f)e^{j2\pi fx}df \\[10pt] \color{blue}F(\omega)&=\int_{-\infty}^{+\infty}f(x)e^{-j\omega x}dx\\ \end{aligned}
f(x)F(ω)=2π1∫−∞+∞F(ω)ejωxdω=∫−∞+∞F(f)ej2πfxdf=∫−∞+∞f(x)e−jωxdx其中
f
=
2
π
ω
f=\frac{2\pi} \omega
f=ω2π,即频率
能量谱密度
能量信号
【
3
】
^{【3】}
【3】
f
(
x
)
f(x)
f(x) 满足绝对可积:
W
=
∫
−
∞
+
∞
f
2
(
x
)
d
x
<
∞
W = \int_{-\infty}^{+\infty}f^2(x)dx < \infty
W=∫−∞+∞f2(x)dx<∞,其傅里叶变换
F
(
ω
)
\color{blue}F(ω)
F(ω) 称为信号
f
(
x
)
\color{blue}f(x)
f(x) 的频谱函数,频谱函数的模称为
f
(
x
)
f(x)
f(x) 的振幅频谱,函数满足 Parseval 定理:
W
=
∫
−
∞
+
∞
∣
f
(
x
)
∣
2
d
x
=
1
2
π
∫
−
∞
+
∞
∣
F
(
ω
)
∣
2
d
ω
=
∫
−
∞
+
∞
∣
F
(
2
π
f
)
∣
2
d
f
W = \int_{-\infty}^{+\infty}|f(x)|^2dx = \frac{1}{2\pi}\int_{-\infty}^{+\infty}|F(\omega)|^2d\omega = \int_{-\infty}^{+\infty}|F(2\pi f)|^2df
W=∫−∞+∞∣f(x)∣2dx=2π1∫−∞+∞∣F(ω)∣2dω=∫−∞+∞∣F(2πf)∣2df
定义能量型信号的能量谱密度为:
E
(
ω
)
=
∣
F
(
ω
)
∣
2
E(\omega) = |F(\omega)|^2
E(ω)=∣F(ω)∣2,则有;
W
=
1
2
π
∫
−
∞
+
∞
E
(
ω
)
d
ω
=
∫
−
∞
+
∞
E
(
2
π
f
)
d
f
W = \frac{1}{2\pi} \int_{-\infty}^{+\infty}E(\omega)d\omega= \int_{-\infty}^{+\infty}E(2\pi f)df
W=2π1∫−∞+∞E(ω)dω=∫−∞+∞E(2πf)df
说明时域和频域的能量守恒
功率谱密度
功率信号
【
3
】
^{【3】}
【3】
f
(
x
)
f(x)
f(x) 在其周期内满足:
P
=
lim
l
→
∞
1
2
l
∫
−
T
T
f
2
(
x
)
d
x
<
∞
P = \lim_{l \to \infty} \frac{1}{2l}\int_{-T}^{T}f^2(x)dx < \infty
P=l→∞lim2l1∫−TTf2(x)dx<∞
截取一个周期样本构造函数
f
l
(
x
)
,
l
f_l(x),l
fl(x),l 作为周期符号 :
f
l
(
x
)
=
{
f
(
x
)
,
∣
x
∣
≤
l
0
,
∣
x
∣
>
l
f_l(x) = \begin{cases} f(x), |x| \le l \\ 0, |x| > l \end{cases}
fl(x)={f(x),∣x∣≤l0,∣x∣>l
则该样本功率:
P
l
=
∫
−
∞
+
∞
f
l
2
(
x
)
d
x
P_{l} = \int_{-\infty}^{+\infty}f^2_l(x)dx
Pl=∫−∞+∞fl2(x)dx
由 Parseval 定理:
P
l
=
∫
−
∞
+
∞
f
x
2
(
x
)
d
x
=
1
2
π
∫
−
∞
+
∞
∣
F
(
ω
,
l
)
∣
2
d
ω
P_{l} = \int_{-\infty}^{+\infty}f_x^2(x)dx = \frac{1}{2\pi}\int_{-\infty}^{+\infty}|F(\omega, l)|^2d\omega
Pl=∫−∞+∞fx2(x)dx=2π1∫−∞+∞∣F(ω,l)∣2dω
两边同时除以周期
2
l
2l
2l有:
1
2
l
∫
−
l
l
f
2
(
x
)
d
x
=
1
4
π
l
∫
−
∞
+
∞
∣
F
(
ω
,
l
)
∣
2
d
ω
\frac{1}{2l}\int_{-l}^{l}f^2(x)dx = \frac{1}{4\pi l}\int_{-\infty}^{+\infty}|F(\omega, l)|^2d\omega
2l1∫−llf2(x)dx=4πl1∫−∞+∞∣F(ω,l)∣2dω
令
l
l
l 趋于无穷,因为
∣
x
∣
>
l
|x| > l
∣x∣>l 时,
f
l
(
t
)
=
0
f_l(t) =0
fl(t)=0,所以在
(
−
∞
,
+
∞
)
(-\infty, +\infty)
(−∞,+∞) 上该样本平均功率为:
P
=
lim
l
→
∞
1
2
l
∫
−
l
l
f
2
(
x
)
d
x
=
1
2
π
∫
−
∞
+
∞
lim
l
→
∞
1
2
l
∣
F
(
ω
,
l
)
∣
2
d
ω
\begin{aligned} P &= \lim_{l \to \infty}\frac{1}{2l}\int_{-l}^{l}f^2(x)dx \\[10pt] &= \frac{1}{2\pi} \int_{-\infty}^{+\infty} \lim_{l \to \infty} \frac{1}{2l}|F(\omega, l)|^2d\omega \end{aligned}
P=l→∞lim2l1∫−llf2(x)dx=2π1∫−∞+∞l→∞lim2l1∣F(ω,l)∣2dω
功率型信号的平均功率谱密度,简称功率谱密度
S
(
ω
)
S(\omega)
S(ω),则该样本函数功率谱密度和功率分别为:
S
(
ω
)
=
lim
l
→
∞
1
2
l
∣
F
(
ω
,
l
)
∣
2
⟹
P
=
1
2
π
∫
−
∞
+
∞
S
(
ω
)
d
ω
\begin{aligned} S(\omega) &= \lim_{l \to \infty}\frac{1}{2l}|F(\omega, l)|^2\\[8pt] \color{blue}\Longrightarrow P&= \frac{1}{2\pi}\int_{-\infty}^{+\infty}S(\omega)d\omega \end{aligned}
S(ω)⟹P=l→∞lim2l1∣F(ω,l)∣2=2π1∫−∞+∞S(ω)dω
在平稳随机过程中,需要对所有样本函数取统计平均:
【
7
】
^{【7】}
【7】
S
(
ω
)
=
lim
l
→
∞
E
{
1
2
l
∣
F
(
ω
,
l
)
∣
2
}
⟹
P
=
1
2
π
∫
−
∞
+
∞
S
(
ω
)
d
ω
\begin{aligned} S(\omega) &= \lim_{l \to \infty} E\left\{\frac{1}{2l}|F(\omega, l)|^2\right\}\\[8pt] \color{blue}\Longrightarrow P&= \frac{1}{2\pi}\int_{-\infty}^{+\infty}S(\omega)d\omega \end{aligned}
S(ω)⟹P=l→∞limE{2l1∣F(ω,l)∣2}=2π1∫−∞+∞S(ω)dω
参考文献
【1】狄利克雷定理的证明
【2】傅里叶变换、{能量,功率,互}谱密度、白噪声随机过程
【3】自相关、互相关函数学习笔记
【4】Fourier Series
【5】Derivation of Fourier Series
【6】From Fourier Series to Fourier Transform
【7】第五章 随机过程的功率谱密度
【8】如何理解傅里叶变换公式?
【9】一小时学会快速傅里叶变换(Fast Fourier Transform)