论文免费来源: 《Physical Layer Network Coding with Multiple Antennas》
https://xueshu.baidu.com/usercenter/paper/show?paperid=0790a364e921183bd929563d288e23b8&site=xueshu_se&hitarticle=1
论文系统模型: 双路中继系统,两个基站分别具有1根天线,基站具有2根天线
论文当中直接通过利用和差矩阵来实现对中继接收信号的网络编码映射,但是在经过仿真之后发现通过论文当中的公式并不能获得良好的仿真性能,于是对相关的公式进行推导,并将其记录。
1、中继接收 Y = H X + N \boldsymbol{Y}=\boldsymbol{HX}+\boldsymbol{N} Y=HX+N 2、和差矩阵 D = 2 D − 1 = [ 1 1 1 − 1 ] \boldsymbol{D}=2\boldsymbol{D}^{-1}=\left[ \begin{matrix} 1& 1\\ 1& -1\\ \end{matrix} \right] D=2D−1=[111−1] 3、将中继接收信号等效处理 Y = H X + N = ( H D − 1 ) ( D X ) + N \boldsymbol{Y}=\boldsymbol{HX}+\boldsymbol{N}=\left( \boldsymbol{HD}^{-1} \right) \left( \boldsymbol{DX} \right) +\boldsymbol{N} Y=HX+N=(HD−1)(DX)+N
X ^ = D X = [ x 1 + x 2 x 1 − x 2 ] \boldsymbol{\hat{X}}=\boldsymbol{DX}=\left[ \begin{array}{c} x_1+x_2\\ x_1-x_2\\ \end{array} \right] X^=DX=[x1+x2x1−x2]
H ^ = H D − 1 \boldsymbol{\hat{H}}=\boldsymbol{HD}^{-1} H^=HD−1
G = ( H ^ H H ^ ) − 1 H ^ H \boldsymbol{G}=\left( \boldsymbol{\hat{H}}^H\boldsymbol{\hat{H}} \right) ^{-1}\boldsymbol{\hat{H}}^H G=(H^HH^)−1H^H
r = G Y \boldsymbol{r}=\boldsymbol{GY} r=GY
4、基于对数似然率(LLR)的网络编码映射 L ( x 1 ⊕ x 2 ∣ r 1 r 2 ) = p ( r 1 r 2 ∣ x 1 ⊕ x 2 = 1 ) p ( r 1 r 2 ∣ x 1 ⊕ x 2 = − 1 ) = e ( 2 σ 2 2 − 2 σ 1 2 ) cosh ( 2 r 1 σ 1 2 ) cosh ( 2 r 2 σ 2 2 ) L\left( x_1\oplus x_2|r_1r_2 \right) =\frac{p\left( r_1r_2|x_1\oplus x_2=1 \right)}{p\left( r_1r_2|x_1\oplus x_2=-1 \right)} \\ =e^{\left( \frac{2}{\sigma _{2}^{2}}-\frac{2}{\sigma _{1}^{2}} \right)}\frac{\cosh \left( \frac{2r_1}{\sigma _{1}^{2}} \right)}{\cosh \left( \frac{2r_2}{\sigma _{2}^{2}} \right)} L(x1⊕x2∣r1r2)=p(r1r2∣x1⊕x2=−1)p(r1r2∣x1⊕x2=1)=e(σ222−σ122)cosh(σ222r2)cosh(σ122r1) 其中 σ i 2 = { G G H } i , i σ 2 \sigma _{i}^{2}=\left\{ \boldsymbol{GG}^H \right\} _{i,i}\sigma _{}^{2} σi2={GGH}i,iσ2 σ 2 \sigma _{}^{2} σ2为噪声方差。 5、推导 σ i 2 \sigma _{i}^{2} σi2 由: R Y = E { Y Y H } = E { H H H } + σ 2 I \boldsymbol{R}_{\boldsymbol{Y}}=E\left\{ \boldsymbol{YY}^H \right\} =E\left\{ \boldsymbol{HH}^H \right\} +\sigma ^2\boldsymbol{I} RY=E{YYH}=E{HHH}+σ2I 与 r = x ^ + G N \boldsymbol{r}=\boldsymbol{\hat{x}}+\boldsymbol{GN} r=x^+GN。 有: R r = E { r r H } = E { ( x ^ + G N ) ( x ^ + G N ) H } = E { x ^ x ^ H } + G E { N N H } G H = D D H + G G H σ 2 \boldsymbol{R}_{\boldsymbol{r}}=E\left\{ \boldsymbol{rr}^H \right\} =E\left\{ \left( \boldsymbol{\hat{x}}+\boldsymbol{GN} \right) \left( \boldsymbol{\hat{x}}+\boldsymbol{GN} \right) ^H \right\} \\ =E\left\{ \boldsymbol{\hat{x}\hat{x}}^H \right\} +\boldsymbol{G}E\left\{ \boldsymbol{NN}^H \right\} \boldsymbol{G}^H \\ =\boldsymbol{DD}^H+\boldsymbol{GG}^H\sigma ^2 Rr=E{rrH}=E{(x^+GN)(x^+GN)H}=E{x^x^H}+GE{NNH}GH=DDH+GGHσ2 所以,从上式当中的最右边项当中的第二项可知, σ i 2 = { G G H } i , i σ 2 \sigma _{i}^{2}=\left\{ \boldsymbol{GG}^H \right\} _{i,i}\sigma _{}^{2} σi2={GGH}i,iσ2,并且通过仿真获得了与论文当中相似的仿真效果。 ps. 很奇怪还有几篇在这篇论文的基础上写出来的论文《Physical network coding for Bidirectional relay MIMO-SDM system》《 Design and Implementation of Signal Processing Unit for Two-Way Relay Node in MIMO-SDM-PNC System》,当中的公式都是 σ i 2 = { G H G } i , i σ 2 \sigma _{i}^{2}=\left\{ \boldsymbol{G}^H\boldsymbol{G} \right\} _{i,i}\sigma ^2 σi2={GHG}i,iσ2。这是怎么做到的???原论文是不是标错了???有看见的大佬希望能够指点一下,感谢!