一个讯号s(t)，自相关函数为
R
(
τ
)
=
∫
−
∞
∞
s
(
t
)
s
∗
(
t
−
τ
)
d
t
{\displaystyle R(\tau )=\int _{-\infty }^{\infty }s(t)s^{*}(t-\tau )\,dt}
R
(
τ
)
{\displaystyle R(\tau )}
R
(
t
,
τ
)
{\displaystyle R(t,\tau )}
P
(
t
,
ω
)
=
∫
−
∞
∞
R
(
t
,
τ
)
e
−
j
ω
τ
d
τ
{\displaystyle P(t,\omega )=\int _{-\infty }^{\infty }R(t,\tau )e^{-j\omega \tau }\,d\tau }
R
(
t
,
τ
)
=
s
(
t
+
τ
2
)
s
∗
(
t
−
τ
2
)
{\displaystyle R(t,\tau )=s\left(t+{\frac {\tau }{2}}\right)s^{*}\left(t-{\frac {\tau }{2}}\right)}
R
(
t
,
τ
)
{\displaystyle R(t,\tau )}
对称模糊函数(symmetric ambiguity function,SAF)
S
A
F
s
(
θ
,
τ
)
=
1
2
π
∫
−
∞
∞
s
(
t
+
τ
2
)
s
∗
(
t
−
τ
2
)
e
j
θ
t
d
t
{\displaystyle SAF_{s}(\theta ,\tau )={\frac {1}{2\pi }}\int _{-\infty }^{\infty }s\left(t+{\frac {\tau }{2}}\right)s^{*}\left(t-{\frac {\tau }{2}}\right)e^{j\theta t}\,dt}
S
A
F
s
(
θ
,
τ
)
{\displaystyle SAF_{s}(\theta ,\tau )}
∫
−
∞
∞
S
A
F
s
(
θ
,
τ
)
e
−
j
θ
t
d
θ
=
s
(
t
+
τ
2
)
s
∗
(
t
−
τ
2
)
{\displaystyle \int _{-\infty }^{\infty }SAF_{s}(\theta ,\tau )e^{-j\theta t}\,d\theta =s\left(t+{\frac {\tau }{2}}\right)s^{*}\left(t-{\frac {\tau }{2}}\right)}
W
D
s
(
t
,
ω
)
=
∫
−
∞
∞
∫
−
∞
∞
S
A
F
s
(
θ
,
τ
)
e
−
j
(
ω
τ
+
θ
t
)
d
θ
d
τ
{\displaystyle WD_{s}(t,\omega )=\int _{-\infty }^{\infty }\int _{-\infty }^{\infty }SAF_{s}(\theta ,\tau )e^{-j(\omega \tau +\theta t)}\,d\theta \,d\tau }
对称模糊函数 做两次傅立叶变换可以得到Wigner distribution
一个讯号为两个Gaussian函数的和:
s
(
t
)
=
∑
i
=
1
2
s
i
(
t
)
=
∑
i
=
1
2
α
π
4
e
−
α
2
(
t
−
t
i
)
2
+
j
ω
i
t
{\displaystyle s(t)=\sum _{i=1}^{2}s_{i}(t)=\sum _{i=1}^{2}{\sqrt[{4}]{\frac {\alpha }{\pi }}}e^{-{\tfrac {\alpha }{2}}(t-t_{i})^{2}+j\omega _{i}t}}
⇒
S
A
F
s
(
θ
,
τ
)
=
∑
i
=
1
2
S
A
F
s
i
(
θ
,
τ
)
+
S
A
F
s
1
,
s
2
(
θ
,
τ
)
+
S
A
F
s
2
,
s
1
(
θ
,
τ
)
{\displaystyle \Rightarrow SAF_{s}(\theta ,\tau )=\sum _{i=1}^{2}SAF_{si}(\theta ,\tau )+SAF_{s1,s2}(\theta ,\tau )+SAF_{s2,s1}(\theta ,\tau )}
其中
S
A
F
s
1
(
θ
,
τ
)
,
S
A
F
s
2
(
θ
,
τ
)
{\displaystyle SAF_{s1}(\theta ,\tau ),SAF_{s2}(\theta ,\tau )}
S
A
F
s
1
,
s
2
(
θ
,
τ
)
{\displaystyle SAF_{s1,s2}(\theta ,\tau )}
(
t
1
−
t
2
,
ω
1
−
ω
2
)
{\displaystyle (t_{1}-t_{2},\omega _{1}-\omega _{2})}
S
A
F
s
2
,
s
1
(
θ
,
τ
)
{\displaystyle SAF_{s2,s1}(\theta ,\tau )}
S
A
F
s
1
,
s
2
(
θ
,
τ
)
{\displaystyle SAF_{s1,s2}(\theta ,\tau )}
(
t
2
−
t
1
,
ω
2
−
ω
1
)
{\displaystyle (t_{2}-t_{1},\omega _{2}-\omega _{1})}
S
A
F
s
1
,
s
2
(
θ
,
τ
)
=
e
−
1
4
α
(
θ
−
ω
d
)
2
+
α
4
(
τ
−
t
d
)
2
e
j
(
ω
u
τ
−
θ
t
u
+
ω
d
t
u
)
{\displaystyle SAF_{s1,s2}(\theta ,\tau )=e^{-{\tfrac {1}{4\alpha }}(\theta -\omega _{d})^{2}+{\tfrac {\alpha }{4}}(\tau -t_{d})^{2}}e^{j(\omega _{u}\tau -\theta t_{u}+\omega _{d}t_{u})}}
t
u
=
t
1
+
t
2
2
{\displaystyle t_{u}={\tfrac {t_{1}+t_{2}}{2}}}
ω
u
=
ω
1
+
ω
2
2
{\displaystyle \omega _{u}={\tfrac {\omega _{1}+\omega _{2}}{2}}}
t
d
=
t
1
−
t
2
{\displaystyle t_{d}=t_{1}-t_{2}}
ω
d
=
ω
1
−
ω
2
{\displaystyle \omega _{d}=\omega _{1}-\omega _{2}}
编辑
从范例中得知一项重要事实，即为，在模糊域(ambiguity domain)中的auto-term总是集中在原点(0,0)，而cross-term总是在远离原点处，所以可以用一个2D lowpass filter在模糊域中抑制cross-term的干扰，如下:
∫
−
∞
∞
∫
−
∞
∞
S
A
F
s
(
θ
,
τ
)
Φ
(
θ
,
τ
)
e
−
j
(
θ
t
+
ω
τ
)
d
θ
d
τ
{\displaystyle \int _{-\infty }^{\infty }\int _{-\infty }^{\infty }SAF_{s}(\theta ,\tau )\Phi (\theta ,\tau )e^{-j(\theta t+\omega \tau )}\,d\theta \,d\tau }
Φ
(
θ
,
τ
)
{\displaystyle \Phi (\theta ,\tau )}
如果
∫
−
∞
∞
∫
−
∞
∞
Φ
(
θ
,
τ
)
d
θ
d
τ
=
ϕ
(
t
,
ω
)
{\displaystyle \int _{-\infty }^{\infty }\int _{-\infty }^{\infty }\Phi (\theta ,\tau )\,d\theta \,d\tau =\phi (t,\omega )}
∫
−
∞
∞
∫
−
∞
∞
S
A
F
s
(
θ
,
τ
)
Φ
(
θ
,
τ
)
e
−
j
(
θ
t
+
ω
τ
)
d
θ
d
τ
{\displaystyle \int _{-\infty }^{\infty }\int _{-\infty }^{\infty }SAF_{s}(\theta ,\tau )\Phi (\theta ,\tau )e^{-j(\theta t+\omega \tau )}\,d\theta \,d\tau }
=
∫
−
∞
∞
∫
−
∞
∞
ϕ
(
x
,
y
)
W
D
s
(
t
−
x
,
ω
−
y
)
d
x
d
y
=
S
W
D
(
t
,
ω
)
{\displaystyle =\int _{-\infty }^{\infty }\int _{-\infty }^{\infty }\phi (x,y)WD_{s}(t-x,\omega -y)\,dx\,dy=SWD(t,\omega )}
其中SWD为smoothed Wigner distribution 通常
Φ
(
θ
,
τ
)
{\displaystyle \Phi (\theta ,\tau )}
ϕ
(
t
,
ω
)
{\displaystyle \phi (t,\omega )}
W
D
s
1
,
s
2
(
t
,
ω
)
=
A
W
D
(
t
,
ω
)
e
j
φ
W
D
(
t
,
ω
)
{\displaystyle WD_{s1,s2}(t,\omega )=A_{WD}(t,\omega )e^{j\varphi _{WD}(t,\omega )}}
S
A
F
s
1
,
s
2
(
θ
,
τ
)
=
A
S
A
F
(
θ
,
τ
)
e
j
φ
S
A
F
(
θ
,
τ
)
{\displaystyle SAF_{s1,s2}(\theta ,\tau )=A_{SAF}(\theta ,\tau )e^{j\varphi _{SAF}(\theta ,\tau )}}
∂
∂
θ
φ
S
A
F
(
θ
,
τ
)
=
−
t
u
{\displaystyle {\frac {\partial }{\partial \theta }}\varphi _{SAF}(\theta ,\tau )=-t_{u}}
∂
∂
τ
φ
S
A
F
(
θ
,
τ
)
=
ω
u
{\displaystyle {\frac {\partial }{\partial \tau }}\varphi _{SAF}(\theta ,\tau )=\omega _{u}}
对称模糊函数 的相位做偏微分，会等于Wigner分布的时频(time-frequency)中心。
∂
∂
ω
φ
W
D
(
t
,
ω
)
=
t
d
{\displaystyle {\frac {\partial }{\partial \omega }}\varphi _{WD}(t,\omega )=t_{d}}
∂
∂
t
φ
W
D
(
t
,
ω
)
=
ω
d
{\displaystyle {\frac {\partial }{\partial t}}\varphi _{WD}(t,\omega )=\omega _{d}}
ω
1
=
ω
2
=
ω
0
{\displaystyle \omega _{1}=\omega _{2}=\omega _{0}}
S
A
F
s
1
,
s
2
(
θ
,
τ
)
=
e
−
[
1
4
α
θ
2
+
α
4
(
τ
−
t
d
)
2
]
e
j
(
ω
0
τ
−
θ
t
u
)
{\displaystyle SAF_{s1,s2}(\theta ,\tau )=e^{-[{\tfrac {1}{4\alpha }}\theta ^{2}+{\tfrac {\alpha }{4}}(\tau -t_{d})^{2}]}e^{j(\omega _{0}\tau -\theta t_{u})}}
τ
{\displaystyle \tau }
t
1
=
t
2
=
t
0
{\displaystyle t_{1}=t_{2}=t_{0}}
S
A
F
s
1
,
s
2
(
θ
,
τ
)
=
e
−
[
1
4
α
(
θ
−
ω
d
)
2
+
α
4
τ
2
]
e
j
[
ω
u
τ
−
(
θ
−
ω
d
)
t
0
]
{\displaystyle SAF_{s1,s2}(\theta ,\tau )=e^{-[{\tfrac {1}{4\alpha }}(\theta -\omega _{d})^{2}+{\tfrac {\alpha }{4}}\tau ^{2}]}e^{j[\omega _{u}\tau -(\theta -\omega _{d})t_{0}]}}
θ
{\displaystyle \theta }
![{\displaystyle s(t)=\sum _{i=1}^{2}s_{i}(t)=\sum _{i=1}^{2}{\sqrt[{4}]{\frac {\alpha }{\pi }}}e^{-{\tfrac {\alpha }{2}}(t-t_{i})^{2}+j\omega _{i}t}}](https://wikimedia.org/api/rest_v1/media/math/render/svg/907a08b77bcea5e7266dbd0eb1d1de9bb5411f5e)
![{\displaystyle SAF_{s1,s2}(\theta ,\tau )=e^{-[{\tfrac {1}{4\alpha }}\theta ^{2}+{\tfrac {\alpha }{4}}(\tau -t_{d})^{2}]}e^{j(\omega _{0}\tau -\theta t_{u})}}](https://wikimedia.org/api/rest_v1/media/math/render/svg/c4a0f47686d64b8b1e6d995dbb1a6371d6e6fcf7)
![{\displaystyle SAF_{s1,s2}(\theta ,\tau )=e^{-[{\tfrac {1}{4\alpha }}(\theta -\omega _{d})^{2}+{\tfrac {\alpha }{4}}\tau ^{2}]}e^{j[\omega _{u}\tau -(\theta -\omega _{d})t_{0}]}}](https://wikimedia.org/api/rest_v1/media/math/render/svg/6e199399019265863b4e5fde588b8b946f148e07)
