Channel论文阅读笔记7-2之multipath interference by Jakes

信道各分量的自相关和功率谱

$$<T_{c1}{T}_{c2}>=<T_{s1}{T}_{s2}>=g(\tau )$$ $$<T_{c1}{T}_{s2}>=-<T_{s1}{T}_{c2}>=h(\tau )\text{\hspace{1em}(1.3-1)}$$
$$g(\tau )={\int }_{f_c-f_m}^{f_c+f_m}{S}_i(f)\cos (2\pi (f-f_c)\tau )df\text{\hspace{1em}(1.3-2)}$$ $$h(\tau )={\int }_{f_c-f_m}^{f_c+f_m}{S}_i(f)\sin (2\pi (f-f_c)\tau )df\text{\hspace{1em}(1.3-3)}$$
(1) 对于电场分量 $$g(\tau )=b_0{J}_0({\omega }_m\tau )$$
简单证明：代人电场的功率谱，设 $f-f_c=f_m\cos \alpha$，$$g(\tau )=\frac{3b}{2\pi }{\int }_0^{\pi }\cos ({\omega }_m\tau \cos \alpha )d\alpha =b_0{J}_0({\omega }_m\tau ).$$
(2) 对于磁场 $x$ 分量 $$g(\tau )=b_{0H}(J_0({\omega }_m\tau )+J_2({\omega }_m\tau ))$$
简单证明：代人磁场 $x$ 分量的功率谱，设 $f-f_c=f_m\cos \alpha$，$$\begin{array}{rl}g(\tau )& =\frac{3b}{2\pi }{\int }_0^{\pi }{\sin }^2\alpha \cos ({\omega }_m\tau \cos \alpha )d\alpha \\ & =\frac{3b}{4\pi }{\int }_0^{\pi }[1+(1-2{\cos }^2\alpha )]\cos ({\omega }_m\tau \cos \alpha )d\alpha \\ & =b_{0H}(J_0({\omega }_m\tau )+J_2({\omega }_m\tau ))\end{array}.$$ 最后等号是由于
$$\begin{array}{rl}& {\int }_0^{\pi }(1-2{\cos }^2\alpha )\cos ({\omega }_m\tau \cos \alpha )d\alpha \\ & =\sum _{k=0}^{\infty }(-1{)}^k\left[\frac{({\omega }_m\tau {)}^{2k}(2k-1)!!}{(2k)!(2k)!!}-\frac{2({\omega }_m\tau {)}^{2k}(2k+1)!!}{(2k)!(2k+2)!!}\right]\\ & =\sum _{k=1}^{\infty }(-1{)}^k({\omega }_m\tau /2{)}^{2k}\left[\frac{-1}{(k-1)!(k+1)!}\right]\\ & =\sum _{k=0}^{\infty }(-1{)}^k({\omega }_m\tau /2{)}^{2k+2}\left[\frac{1}{(k)!(k+2)!}\right]=J_2({\omega }_m\tau ))\end{array}$$
(3) 对于磁场 $y$ 分量 $$g(\tau )=b_{0H}(J_0({\omega }_m\tau )-J_2({\omega }_m\tau ))$$
简单证明：代人磁场 $x$ 分量的功率谱，设 $f-f_c=f_m\cos \alpha$，$$\begin{array}{rl}g(\tau )& =\frac{3b}{2\pi }{\int }_0^{\pi }{\cos }^2\alpha \cos ({\omega }_m\tau \cos \alpha )d\alpha \\ & =\frac{3b}{4\pi }{\int }_0^{\pi }[1-(1-2{\cos }^2\alpha )]\cos ({\omega }_m\tau \cos \alpha )d\alpha \\ & =b_{0H}(J_0({\omega }_m\tau )-J_2({\omega }_m\tau ))\end{array}.$$

信号的矩

$$b_n=(2\pi {)}^n{\int }_{f_c-f_m}^{f_c+f_m}{S}_i(f)(f-f_c{)}^{n}df\text{\hspace{1em}(1.3-4)}$$
对于电场分量
$$b_n=b_0{\omega }_m^n\frac{1\cdot 3\cdots (n-1)}{2\cdot 4\cdots n},b_0=1.5b=\frac{3E_0^2}{4}\text{\hspace{1em}(1.3-6)}$$
证明：对于电场分量，由（1.2-11） $$S_i(f)=\frac{3b}{{\omega }_m}{\left[1-{\left(\frac{f-fc}{f_m}\right)}^2\right]}^{-1/2}$$
$$b_n=(2\pi {)}^n{\int }_{f_c-f_m}^{f_c+f_m}\frac{3b}{{\omega }_m}{\left[1-{\left(\frac{f-fc}{f_m}\right)}^2\right]}^{-1/2}(f-f_c{)}^{n}df$$ 令 $f-f_c=f_m\cos \alpha$,
$$\begin{gathered} b_n={\omega }_m^n{\int }_{-1}^1\frac{3b}{{\omega }_m}{\sin }^{-1}\alpha {\cos }^n\alpha f_{m}d(\cos \alpha ) \\ =\frac{3b{\omega }_m^n}{2\pi }{\int }_{\pi }^0-{\cos }^n\alpha d\alpha =\frac{3b{\omega }_m^n}{2}\frac{1\cdot 3\cdots (n-1)}{2\cdot 4\cdots n} \end{gathered}$$ 证毕。

包络自相关和功率谱

窄带高斯过程包络的自相关为
$$R_r(\tau )=\frac{\pi }{2}{b}_{0}F[-\frac{1}{2},-\frac{1}{2};1;{\rho }^2(\tau )],\text{\hspace{1em}(1.3-12)}$$ 其中 $${\rho }^2(\tau )]=\frac{1}{b_0^2}[g^2(\tau )+h^2(\tau )]\text{\hspace{1em}(1.3-13)}$$
证明：
设 $x_{c1}=x_c(t),x_{s1}=x_s(t),x_{c2}=x_c(t+\tau ),x_{s2}=x_s(t+\tau )$, 随机变量 $x_{c1},x_{s1},x_{c2},x_{s2}$ 的相关阵为 [2]
$$\Lambda =\left[\begin{array}{cccc}{\sigma }_x^2& 0& R_c(\tau )& R_{cs}(\tau )\\ 0& {\sigma }_x^2& -R_{cs}(\tau )& R_c(\tau )\\ R_c(\tau )& -R_{cs}(\tau )& {\sigma }_x^2& 0\\ R_{cs}(\tau )& R_c(\tau )& 0& {\sigma }_x^2\end{array}\right]\text{\hspace{1em}(B.1)}$$
行列式
$$|\Lambda |=[{\sigma }_x^4-R_c^2(\tau )-R_{cs}^2(\tau ){]}^2\text{\hspace{1em}(B.2)}$$
相关阵的逆
$${\Lambda }^{-1}=|\Lambda {|}^{-1/2}\left[\begin{array}{cccc}{\sigma }_x^2& 0& -R_c(\tau )& -R_{cs}(\tau )\\ 0& {\sigma }_x^2& R_{cs}(\tau )& -R_c(\tau )\\ -R_c(\tau )& R_{cs}(\tau )& {\sigma }_x^2& 0\\ -R_{cs}(\tau )& -R_c(\tau )& 0& {\sigma }_x^2\end{array}\right]\text{\hspace{1em}(B.3)}$$
功率密度函数
$$\begin{gathered} p(x_{c1},x_{s1},x_{c2},x_{s2})=\frac{1}{4{\pi }^2|\Lambda {|}^{1/2}} \\ \cdot \exp \left[-\frac{1}{2}(x_{c1},x_{s1},x_{c2},x_{s2}){\Lambda }^{-1}(x_{c1},x_{s1},x_{c2},x_{s2}{)}^T\right]\text{\hspace{1em}(B.4)} \end{gathered}$$ 由 $$d{x}_{c1}d{x}_{s1}d{x}_{c2}d{x}_{s2}=r_1{r}_{2}d{r}_{1}d{r}_{2}d{\varphi }_{1}d{\varphi }_2\text{\hspace{1em}(B.5)}$$ 可得
$$\begin{gathered} p(r_1,{\theta }_1,r_2,{\theta }_2)=\frac{r_1{r}_2}{4{\pi }^2|\Lambda {|}^{1/2}} \\ \cdot \exp \left[-\frac{{\sigma }_x^2(r_1^2+r_2^2)-2R_c(\tau )r_1{r}_2\cos ({\varphi }_2-{\varphi }_1)-2R_{cs}(\tau )r_1{r}_2\sin ({\varphi }_2-{\varphi }_1)}{2|\Lambda {|}^{1/2}}\right]\text{\hspace{1em}(B.6)} \end{gathered}$$
令 $\alpha ={\varphi }_2-{\varphi }_1-{\tan }^{-1}[R_{cs}(\tau )/R_c(\tau )]$
$$\begin{array}{rl}\frac{1}{4{\pi }^2}& {\int }_0^{2\pi }d{\varphi }_1{\int }_0^{2\pi }d{\varphi }_2\exp \left\{-\frac{r_1{r}_2}{|\Lambda {|}^{1/2}}[R_c(\tau )\cos ({\varphi }_2-{\varphi }_1)+R_{cs}(\tau )\sin ({\varphi }_2-{\varphi }_1)]\right\}\\ & =\frac{1}{2\pi }{\int }_0^{2\pi }d{\varphi }_1\frac{1}{2\pi }{\int }_0^{2\pi }\exp \left\{-\frac{r_1{r}_2[R_c^2(\tau )+R_{cs}^2(\tau ){]}^{1/2}}{|\Lambda {|}^{1/2}}\cos \alpha \right\}d\alpha \\ & =I_0\left\{\frac{r_1{r}_2[R_c^2(\tau )+R_{cs}^2(\tau ){]}^{1/2}}{|\Lambda {|}^{1/2}}\right\}\end{array}\text{\hspace{1em}(B.7)}$$
$$\begin{array}{rl}{R}_r(\tau )& =<r_1,r_2>\\ & ={\int }_0^{\infty }{\int }_0^{\infty }\frac{r_1^2{r}_2^2}{|\Lambda {|}^{1/2}}{I}_0\left\{\frac{r_1{r}_2[R_c^2(\tau )+R_{cs}^2(\tau ){]}^{1/2}}{|\Lambda {|}^{1/2}}\right\}\exp \left[-\frac{{\sigma }_x^2(r_1^2+r_2^2)}{2|\Lambda {|}^{1/2}}\right]d{r}_{1}d{r}_2\\ & ={\int }_0^{\infty }{\int }_0^{\infty }\frac{r_1^2{r}_2^2}{|\Lambda {|}^{1/2}}\sum _{k=0}^{\infty }{\left\{\frac{r_1{r}_2[R_c^2(\tau )+R_{cs}^2(\tau ){]}^{1/2}}{2|\Lambda {|}^{1/2}}\right\}}^{2k}\frac{1}{k!k!}\exp \left[-\frac{{\sigma }_x^2(r_1^2+r_2^2)}{2|\Lambda {|}^{1/2}}\right]d{r}_{1}d{r}_2\\ & =\sum _{k=0}^{\infty }{\left[\frac{R_c^2(\tau )+R_{cs}^2(\tau )}{{\sigma }_x^2}\right]}^k{\left\{{\int }_0^{\infty }\frac{r_1^2}{|\Lambda {|}^{1/4}}\frac{r_1^{2k}{\sigma }_x^{2k}}{2^k|\Lambda {|}^{k/2}k!}\exp \left[-\frac{{\sigma }_x^2{r}_1^2}{2|\Lambda {|}^{1/2}}\right]d{r}_1\right\}}^2\\ & =\sum _{k=0}^{\infty }{\rho }^{2k}(\tau ){\left\{{\int }_0^{\infty }\frac{\sqrt{2}|\Lambda {|}^{1/2}}{{\sigma }_x^3}\frac{({\sigma }_x^2{r}_1^2{)}^{(k+1/2)}}{(2|\Lambda {|}^{1/2}{)}^{(k+1/2)}k!}\exp \left[-\frac{{\sigma }_x^2{r}_1^2}{2|\Lambda {|}^{1/2}}\right]d\frac{{\sigma }_x^2{r}_1^2}{2|\Lambda {|}^{1/2}}\right\}}^2\\ & =\sum _{k=0}^{\infty }{\rho }^{2k}(\tau )\frac{2|\Lambda |}{{\sigma }_x^6}\frac{{\Gamma }^2(k+3/2)}{k!k!}\end{array}\text{\hspace{1em}(B.8)}$$
注意到
$$\sum _{k=0}^{\infty }{\rho }^{2k}(\tau )\frac{2|\Lambda |}{{\sigma }_x^6}\frac{{\Gamma }^2(k+3/2)}{k!k!}=\frac{{\sigma }_x^2\pi }{2}\sum _{k=0}^{\infty }{\rho }^{2k}(\tau )(1-{\rho }^2(\tau ){)}^2{\left[\frac{(2k+1)!!}{(2k)!!}\right]}^2\text{\hspace{1em}(B.9)}$$
$$\sum _{k=0}^{\infty }{\rho }^{2k}(\tau ){\rho }^{2l}(\tau ){\left[\frac{(2k+1)!!}{(2k)!!}\right]}^2=\sum _{k=l}^{\infty }{\rho }^{2k}(\tau ){\left[\frac{(2k-2l+1)!!}{(2k-2l)!!}\right]}^2\text{\hspace{1em}(B.10)}$$
$$\sum _{k=0}^{\infty }{\left[\frac{(2k+1)!!}{(2k)!!}\right]}^2-2\sum _{k=1}^{\infty }{\left[\frac{(2k-1)!!}{(2k-2)!!}\right]}^2+\sum _{k=2}^{\infty }{\left[\frac{(2k-3)!!}{(2k-4)!!}\right]}^2=\sum _{k=0}^{\infty }{\left[\frac{(2k-3)!!}{(2k)!!}\right]}^2\text{\hspace{1em}(B.11)}$$
因此
$$(B.8)=\frac{{\sigma }_x^2\pi }{2}\sum _{k=0}^{\infty }{\rho }^{2k}(\tau ){\left[\frac{(2k-3)!!}{(2k)!!}\right]}^2=\frac{\pi b_0}{2}{}_2{F}_1\left(-\frac{1}{2},-\frac{1}{2};1;{\rho }^2(\tau )\right)\text{\hspace{1em}(B.12)}$$证毕。

根据 Gauss’s Hypergeometric Theorem [5,6]
$${}_2{F}_1(a,b;c;1)=\frac{(c-b{)}_{-a}}{(c{)}_{-a}}=\frac{\Gamma (c)\Gamma (c-a-b)}{\Gamma (c-a)\Gamma (c-b)}\text{\hspace{1em}(B.13)}$$
得
$${}_2{F}_1\left(-\frac{1}{2},-\frac{1}{2};1;1\right)=\frac{\Gamma (2)}{\Gamma (3/2)\Gamma (3/2)}=\frac{4}{\pi }\text{\hspace{1em}(B.14)}$$
在 $\tau =0$， $\rho (0)=1$，
$$R_r(0)=\frac{\pi b_0}{2}{}_2{F}_1\left(-\frac{1}{2},-\frac{1}{2};1;1\right)=2b_0\text{\hspace{1em}(1.3.14)}$$
展开（1.3-12）的超几何函数
$$R_r(\tau )=\frac{\pi b_0}{2}\left(1+\frac{{\rho }^2(\tau )}{4}+\frac{{\rho }^4(\tau )}{64}+\cdots \right)\text{\hspace{1em}(1.3.15)}$$
保留前两项 $$R_r(\tau )\doteq \frac{\pi b_0}{2}\left(1+\frac{{\rho }^2(\tau )}{4}\right)\text{\hspace{1em}(1.3.16)}$$
对于 $\tau =0$， $$R_r(0)\doteq \frac{5\pi b_0}{8}=1.964b_0,\text{\hspace{1em}(1.3.17)}$$ 误差仅 1.8%。

包络的功率谱密度可以表示为
$$\begin{array}{rl}{S}_e(f)& =\frac{\pi b_0}{2}{\int }_{-\infty }^{\infty }\left(1+\frac{{\rho }^2(\tau )}{4}\right)e^{-i\omega \tau }d\tau \\ & =\frac{\pi b_0}{2}\delta (f)+\frac{\pi }{8b_0}{\int }_{-\infty }^{\infty }\left(g^2(\tau )+h^2(\tau )\right)e^{-i\omega \tau }d\tau \end{array}\text{\hspace{1em}(1.3.19)}$$

Ref

1. Microwave Mobile Communicatiaons, Ch1.
2. W. B. Davenport, Jr., and W. L. Root, An Introduction to the Theory of Randum Signals and Noise, McGraw-Hill, New York, 1958.
3. https://archive.lib.msu.edu/crcmath/math/math/e/e078.htm
4. https://mathworld.wolfram.com/PochhammerSymbol.html
5. https://mathworld.wolfram.com/GausssHypergeometricTheorem.html
6. https://archive.lib.msu.edu/crcmath/math/math/g/g054.htm