通信中常见的概率分布

几个重要的分布

Gaussian Distribution

1. standard Gaussian distribution
   $$X\sim \mathcal{N}(0,1)$$

   $$f(x)=\frac{1}{\sqrt{2\pi }}{e}^{-\frac{x^2}{2}}$$

   $$\Phi (x)=\frac{1}{\sqrt{2\pi }}{\int }_{-\infty }^x{e}^{-\frac{t^2}{2}}dt$$

2. general Gaussian distribution
   $$X\sim \mathcal{N}(\mu ,{\sigma }^2)$$

   $$f(x)=\frac{1}{\sqrt{2\pi }\sigma }{e}^{-\frac{(x-\mu {)}^2}{2{\sigma }^2}}$$

   $$F(x)=\frac{1}{\sqrt{2\pi }\sigma }{\int }_{-\infty }^x{e}^{-\frac{(t-\mu {)}^2}{2{\sigma }^2}}dt$$

3. several functions related to Gaussian distribution

   $Q$函数定义为标准正态分布的右尾函数：
   $$Q(x)=\frac{1}{\sqrt{2\pi }}{\int }_x^{\infty }{e}^{-\frac{t^2}{2}}dt=1-\Phi (x)$$
   误差函数 $erf(x)$ 的物理意义是服从均值为$0$, 方差为$\frac{1}{2}$正态分布的随机变量$Y$落在区间$(-x,x)$的概率：
   $$erf(x)=\frac{2}{\sqrt{\pi }}{\int }_0^x{e}^{t^2}dt$$
   互补误差函数$erfc(x)$的物理意义即$Y$落在上述区间外的概率：
   $$erfc(x)=\frac{2}{\sqrt{\pi }}{\int }_x^{\infty }{e}^{t^2}dt$$
   显然：
   $$erfc(x)=1-erfc(x)$$

   $$Q(x)=\frac{1}{2}erfc(\frac{x}{\sqrt{2}})$$

Chi-Square Distribution

1. definition

   设$X_1,X_2,...,X_n$ i.i.d. $\sim \mathcal{N}(0,1)$，令$X={\sum }_{i=1}^n{X}_i^2$，则称$X$服从自由度为$n$的 Chi-Square distribution ${\chi }^2(n)$

2. probability density function
   $$f_n(x)=\frac{(1/2{)}^{n/2}}{\Gamma (n/2)}{x}^{n/2-1}{e}^{-x/2}$$
   上式中$k$表示自由度，$\Gamma (\cdot )$函数是阶乘函数的推广，$\Gamma (n)=(n-1)！$。当$n=2$时
   $$f(x)=\frac{1}{2}{e}^{-x/2}$$
   i.e. Chi-Square distribution 退化为均值为2的指数分布。

3. mean = $n$ and variance = $2n$

4. exponential distribution with mean $\lambda$ and variance ${\lambda }^2$ has PDF:
   $$f(x)=\frac{1}{\lambda }{e}^{-\frac{x}{\lambda }}$$

Rayleigh Distribution

$R\sim \mathrm{R}\mathrm{a}\mathrm{y}\mathrm{l}\mathrm{e}\mathrm{i}\mathrm{g}\mathrm{h}(\sigma )$ if $R=\sqrt{X^2+Y^2}$，where $X\sim N(0,{\sigma }^2)$ and $Y\sim N(0,{\sigma }^2)$ are independent normal random variables.
$$f(r)=\frac{r}{{\sigma }^2}{e}^{-\frac{r^2}{2{\sigma }^2}}$$

• mean and variance are proportional to $\sigma$ and ${\sigma }^2$ respectively.

One example where the Rayleigh distribution naturally arises is when wind velocity is analyzed in two dimensions. Assuming that each component is uncorrelated, normally distributed with equal variance, and zero mean, then the overall wind speed (vector magnitude) will be characterized by a Rayleigh distribution.

A second example of the distribution arises in the case of random complex numbers whose real and imaginary components are independently and identically distributed Gaussian with equal variance and zero mean. In that case, the absolute value of the complex number is Rayleigh-distributed.

The Chi-Square distribution with $n=2$ is equivalent to the Rayleigh Distribution with $\sigma =1$. I.e., if $R\sim \mathrm{R}\mathrm{a}\mathrm{y}\mathrm{l}\mathrm{e}\mathrm{i}\mathrm{g}\mathrm{h}(1)$, then $R^2\sim {\chi }^2(2)$. I have mentioned above that Chi-Square distribution with $n=2$ is the same as exponential distribution with mean 2, so there must be a relationship between Rayleigh distribution and exponential distribution:
$$\begin{gathered} X\sim E(\lambda ) \\ \frac{X}{\lambda /2}\sim E(2) \\ \frac{X}{\lambda /2}\sim {\chi }^2(2) \\ \sqrt{\frac{X}{\lambda /2}}\sim \mathrm{R}\mathrm{a}\mathrm{y}\mathrm{l}\mathrm{e}\mathrm{i}\mathrm{g}\mathrm{h}(1) \\ \sqrt{\mathrm{X}}\sim \mathrm{R}\mathrm{a}\mathrm{y}\mathrm{l}\mathrm{e}\mathrm{i}\mathrm{g}\mathrm{h}(\sqrt{\lambda /2}) \end{gathered}$$

Consider fading coefficient $h\sim \mathcal{C}\mathcal{N}(0,1)$, we know that $|h|$ is Rayleigh distributed with parameter ${\sigma }^2=\frac{1}{2}$ (each dimension). So the squared magnitude $r=|h{|}^2$ is exponential distributed with mean $2{\sigma }^2=1$, i.e.
$$f(r)=e^{-r}$$