信道化原型结构:
多相信道化的应用是为了减少信道化实现所需的资源,在计算结果不变的情况下,大大降低了系统的资源消耗(计算量降低),以下为推到过程。
D
D
D为抽取因子,信道数量为
K
K
K,如果抽取因子
D
=
K
D=K
D=K,那么第
k
k
k道的信号输出
y
k
(
m
)
y_k(m)
yk(m)为:
y
k
(
m
)
=
[
x
(
n
)
⋅
e
−
j
ω
k
n
]
∗
h
(
n
)
∣
n
=
m
D
=
∑
i
=
−
∞
+
∞
[
x
(
n
−
i
)
e
−
j
ω
k
(
n
−
i
)
h
(
i
)
]
∣
n
=
m
D
=
∑
i
=
−
∞
+
∞
[
x
(
m
D
−
i
)
e
−
j
ω
k
(
m
D
−
i
)
h
(
i
)
]
,
令
i
=
j
D
+
p
=
∑
p
=
0
D
−
1
∑
j
=
−
∞
+
∞
[
x
(
m
D
−
j
D
−
p
)
e
−
j
ω
k
(
m
D
−
j
D
−
p
)
h
(
j
D
+
p
)
]
令
,
x
p
(
m
)
=
x
(
m
D
−
p
)
,
h
p
(
m
)
=
h
(
m
D
+
p
)
=
∑
p
=
0
D
−
1
∑
j
=
−
∞
+
∞
[
x
p
(
m
−
j
)
e
−
j
ω
k
(
m
D
−
j
D
−
p
)
h
p
(
j
)
]
=
∑
p
=
0
D
−
1
[
∑
j
=
−
∞
+
∞
[
x
p
(
m
−
j
)
e
−
j
ω
k
(
m
D
−
j
D
)
h
p
(
j
)
]
e
j
ω
k
p
令
,
s
p
(
m
)
=
∑
j
=
−
∞
+
∞
[
x
p
(
m
−
j
)
e
−
j
ω
k
(
m
D
−
j
D
)
h
p
(
j
)
]
=
x
p
(
m
)
e
−
j
ω
k
m
D
∗
h
p
(
m
)
=
∑
p
=
0
D
−
1
[
s
p
(
m
)
]
e
j
ω
k
p
\begin{aligned} y_k(m)&=[x(n) \cdot e^{-j \omega_kn}]*h(n)|_{n=mD} \\ &= \sum_{i = -\infty}^{+\infty }[x(n-i)e^{-j \omega_k(n-i)}h(i)]|_{n=mD}\\ &= \sum_{i = -\infty}^{+\infty }[x(mD-i)e^{-j \omega_k(mD-i)}h(i)],令i = jD+p\\ &= \sum_{p=0}^{D-1}\sum_{j = -\infty}^{+\infty }[x(mD-jD-p)e^{-j \omega_k(mD-jD-p)}h(jD+p)]\\ &令,x_p(m) = x(mD-p),h_p(m) = h(mD+p)\\ &= \sum_{p=0}^{D-1}\sum_{j = -\infty}^{+\infty }[x_p(m-j)e^{-j \omega_k(mD-jD-p)}h_p(j)]\\ &= \sum_{p=0}^{D-1} [ \sum_{j = -\infty}^{+\infty }[x_p(m-j)e^{-j \omega_k(mD-jD)}h_p(j)]e^{j \omega_kp}\\ &令,s_p(m)=\sum_{j = -\infty}^{+\infty }[x_p(m-j)e^{-j \omega_k(mD-jD)}h_p(j)]=x_p(m)e^{-j \omega_kmD}*h_p(m)\\ &= \sum_{p=0}^{D-1} [s_p(m)]e^{j \omega_kp} \end{aligned}
yk(m)=[x(n)⋅e−jωkn]∗h(n)∣n=mD=i=−∞∑+∞[x(n−i)e−jωk(n−i)h(i)]∣n=mD=i=−∞∑+∞[x(mD−i)e−jωk(mD−i)h(i)],令i=jD+p=p=0∑D−1j=−∞∑+∞[x(mD−jD−p)e−jωk(mD−jD−p)h(jD+p)]令,xp(m)=x(mD−p),hp(m)=h(mD+p)=p=0∑D−1j=−∞∑+∞[xp(m−j)e−jωk(mD−jD−p)hp(j)]=p=0∑D−1[j=−∞∑+∞[xp(m−j)e−jωk(mD−jD)hp(j)]ejωkp令,sp(m)=j=−∞∑+∞[xp(m−j)e−jωk(mD−jD)hp(j)]=xp(m)e−jωkmD∗hp(m)=p=0∑D−1[sp(m)]ejωkp
ω
k
\omega_k
ωk的取值决定每路信道输出的中心频率,也决定了信道的进一步化解。假如
ω
k
=
2
π
k
/
D
\omega_k=2\pi k/D
ωk=2πk/D,那么:
y
k
(
m
)
=
∑
p
=
0
D
−
1
[
s
p
(
m
)
]
e
j
ω
k
p
=
I
D
F
T
[
s
p
(
m
)
]
,
I
D
F
T
为
逆
离
散
傅
里
叶
变
化
=
I
D
F
T
[
x
p
(
m
)
∗
h
p
(
m
)
]
\begin{aligned} y_k(m)&=\sum_{p=0}^{D-1} [s_p(m)]e^{j \omega_kp}\\ &=IDFT[s_p(m)],IDFT为逆离散傅里叶变化\\ &=IDFT[x_p(m)*h_p(m)]\\ \end{aligned}
yk(m)=p=0∑D−1[sp(m)]ejωkp=IDFT[sp(m)],IDFT为逆离散傅里叶变化=IDFT[xp(m)∗hp(m)]
推到之后的高效信道化结构: