高斯白噪声的统计表示

实高斯白噪声向量

零均值的实高斯白噪声向量$\vec{x}\left(t\right)=[x_1\left(t\right),\cdots ,x_m\left(t\right){]}^{\mathrm{T}}$的各个元素为相互统计不相关的实高斯白噪声过程。若这些高斯白噪声具有相同的方差${\sigma }^2$，则有$$c_{x_i,y_i}=r_{x_i,y_i}=\{\begin{array}{ll}{\sigma }^2,& \text{if }i=j\\ 0,& \text{if }i\ne j\end{array}$$式中，$c_{x_i,y_i}=E\left\{[x_i(t)-{\mu }_{x_i}][y_j(t)-{\mu }_{y_j}{]}^{\ast }\right\}$是随机变量$x_i(t)$和$y_i(t)$之间的互协方差，$r_{x_i,y_i}=E[x_i(t)y_j^{\ast }(t)]$是随机变量$x_i(t)$和$y_i(t)$之间的互相关。于是，实高斯白噪声向量的自协方差矩阵$${\mathbf{C}}_x={\mathbf{R}}_x=\mathrm{E}\left\{\vec{x}\left(t\right){\vec{x}}^{\mathrm{T}}\left(t\right)\right\}=\left[\begin{array}{ccc}{r}_{x_1,x_1}& \cdots & r_{x_1,x_m}\\ ⋮& \ddots & ⋮\\ r_{x_m,x_1}& \cdots & r_{x_m,x_m}\end{array}\right]={\sigma }^2\mathbf{I}$$因此，实高斯白噪声的统计表示为 E { x ⃗ ( t ) } = 0 \mathrm{E}\left \{\vec x\left ( t \right ) \right \} =\bf 0 E{x (t)}=0$$\mathrm{E}\left\{\vec{x}\left(t\right){\vec{x}}^{\mathrm{T}}\left(t\right)\right\}={\sigma }^2\mathbf{I}$$

复高斯白噪声向量

复高斯随机向量$\vec{x}\left(t\right)=\left[\begin{array}{c}{x}_1\left(t\right)\\ ⋮\\ x_m\left(t\right)\end{array}\right]$的每个元素都是复高斯白噪声，它们彼此统计不相关。若他们都具有零均值和相同的方差${\sigma }^2$，则意味着每一个复高斯白噪声过程的实部$x_{\mathrm{R}k}\left(t\right)$和虚部$x_{\mathrm{I}k}\left(t\right)$是两个相互统计独立的实高斯白噪声过程，它们具有相同的方差。因此$x_k\left(t\right)$为零均值和方差${\sigma }^2$的高斯白噪声过程，也意味着 E { x R k ( t ) } = 0 ,   E { x I k ( t ) } = 0 \mathrm{E} \left \{ x_{\mathrm{R} k} \left ( t \right ) \right \} =0,\text{ } \mathrm{E} \left \{ x_{\mathrm{I} k} \left ( t \right ) \right \} =0 E{xRk​(t)}=0, E{xIk​(t)}=0$$\mathrm{E}\left\{x_{\mathrm{R}k}^2\left(t\right)\right\}=\mathrm{E}\left\{x_{\mathrm{I}k}^2\left(t\right)\right\}=\frac{1}{2}{\sigma }^2$$ E { x R k ( t ) x I k ( t ) } = 0 \mathrm{E} \left \{ x_{\mathrm{R} k} \left ( t \right ) x_{\mathrm{I} k} \left ( t \right )\right \}=0 E{xRk​(t)xIk​(t)}=0 E { x k ( t ) x k ∗ ( t ) } = E { x R k 2 ( t ) } + E { x I k 2 ( t ) } = σ 2 \mathrm{E} \left \{ x_{k} \left ( t \right ) x_{k}^{*} \left ( t \right )\right \}=\mathrm{E} \left \{ x_{\mathrm{R} k}^{2} \left ( t \right ) \right \} +\mathrm{E} \left \{ x_{\mathrm{I} k}^{2} \left ( t \right ) \right \}=\sigma ^{2} E{xk​(t)xk∗​(t)}=E{xRk2​(t)}+E{xIk2​(t)}=σ2由上述条件可知 E { x k 2 ( t ) } = E { [ x R k ( t ) + j x I k ( t ) ] 2 } = E { x R k 2 ( t ) } − E { x I k 2 ( t ) } + j 2 E { x R k ( t ) x I k ( t ) } = 1 2 σ 2 − 1 2 σ 2 + 0 = 0 \begin{aligned} \mathrm{E}\left\{x_{k}^{2}(t)\right\} &=\mathrm{E}\left\{\left[x_{\mathrm{R} k}(t)+\mathrm{j} x_{\mathrm{I} k}(t)\right]^{2}\right\} \\ &=\mathrm{E}\left\{x_{\mathrm{R} k}^{2}(t)\right\}-\mathrm{E}\left\{x_{\mathrm{I} k}^{2}(t)\right\}+\mathrm{j} 2 \mathrm{E}\left\{x_{\mathrm{R} k}(t) x_{\mathrm{I} k}(t)\right\} \\ &=\frac{1}{2} \sigma^{2}-\frac{1}{2} \sigma^{2}+0=0 \end{aligned} E{xk2​(t)}​=E{[xRk​(t)+jxIk​(t)]2}=E{xRk2​(t)}−E{xIk2​(t)}+j2E{xRk​(t)xIk​(t)}=21​σ2−21​σ2+0=0​由于$x_1\left(t\right),\cdots ,x_m\left(t\right)$是$m$个彼此不相关的高斯白噪声过程，故 E { x i ( t ) x j ( t ) } = 0 , E { x i ( t ) x j ∗ ( t ) } = 0 , i ≠ j \mathrm{E}\left\{x_{i}(t) x_{j}(t)\right\}=0, \quad \mathrm{E}\left\{x_{i}(t) x_{j}^{*}(t)\right\}=0, \quad i \neq j E{xi​(t)xj​(t)}=0,E{xi​(t)xj∗​(t)}=0,i​=j综合以上条件，即可得到复高斯白噪声向量$\vec{x}(t)$的统计表示为 E { x ⃗ ( t ) } = 0 E { x ⃗ ( t ) x ⃗ H ( t ) } = σ 2 I E { x ⃗ ( t ) x ⃗ T ( t ) } = O \begin{aligned} \mathrm{E}\{\vec{x}(t)\} &=\mathbf{0} \\ \mathrm{E}\left\{\vec{x}(t) \vec{x}^{\mathrm{H}}(t)\right\} &=\sigma^{2} \boldsymbol{I} \\ \mathrm{E}\left\{\vec{x}(t) \vec{x}^{\mathrm{T}}(t)\right\} &=\bf{O} \end{aligned} E{x (t)}E{x (t)x H(t)}E{x (t)x T(t)}​=0=σ2I=O​