傅里叶变换的性质

傅里叶变换的基本性质

1. 对称性

\(F(\omega)=\mathscr{F}[f(t)]\),那么\(\mathscr{F}[F(t)]=2\pi f(-\omega)\)

证明:

\[\begin{aligned} f(t)&=\frac{1}{2\pi}\int_{-\infty}^{\infty}F(\omega)e^{j\omega t}d\omega \\ f(-t)&=\frac{1}{2\pi}\int_{-\infty}^{\infty}F(\omega)e^{-j\omega t}d\omega \\ 2\pi f(-\omega)&=\int_{-\infty}^{\infty}F(t)e^{-j\omega t}dt \end{aligned} \tag{1} \]

显然上式就是傅里叶正变换的定义形式。

2. 线性(叠加性)

\(\mathscr{F}[f_{i}(t)]=F_{i}(\omega)\ (i=1,2,\cdots,n)\),则

\[\mathscr{F}[\sum_{i=1}^{n}a_{i}f_{i}(t)]=\sum_{i=1}^{n}a_{i}F_{i}(\omega) \tag{2} \]

3. 奇偶虚实性

一般情况下,\(F(\omega)\)是复函数,因而可以把它表示成模与相位或者实部与虚部两部分,即

\[F(\omega)=|F(\omega)|e^{j\varphi(\omega)}=R(\omega)+jX(\omega) \]

\[\left \{ \begin{aligned} |F(\omega)| &= \sqrt{R^{2}(\omega)+X^{2}(\omega)} \\ \varphi(\omega) &= \text{arctan}[\frac{X(\omega}{R(\omega)}] \end{aligned} \tag{3a} \right. \]

(1) \(f(t)\)是实函数

\[\begin{aligned} F(\omega) &= \int_{-\infty}^{\infty}f(t)e^{-j\omega t}dt \\ &= \int_{-\infty}^{\infty}f(t)cos(\omega t)dt-j\int_{-\infty}^{\infty}f(t)sin(\omega t)dt \end{aligned} \]

在这种情况下,显然

\[\left \{ \begin{aligned} R(\omega)&=\int_{-\infty}^{\infty}f(t)cos(\omega t)dt \\ X(\omega) &= -\int_{-\infty}^{\infty}f(t)sin(\omega t)dt \end{aligned} \tag{3b} \right.\]

此时,\(R(-\omega)=\int_{-\infty}^{\infty}f(t)cos(-\omega t)dt=R(\omega)\)为偶函数,\(X(-\omega)= \int_{-\infty}^{\infty}f(t)sin(-\omega t)dt=-X(\omega)\)为奇函数。在此基础上,根据式\((3a)\)可以得到\(|F(\omega)|\)为偶函数,\(\varphi(\omega)\)为奇函数。也就是说,实函数的傅里叶变换的幅度谱是偶函数,相位谱是奇函数。

(2) \(f(t)\)是虚函数

\(f(t)=jg(t)\),则

\[\left \{ \begin{aligned} R(\omega)&=\int_{-\infty}^{\infty}g(t)sin(\omega t)dt \\ X(\omega) &= \int_{-\infty}^{\infty}g(t)cos(\omega t)dt \end{aligned} \right.\]

在这种情况下,\(R(\omega)\)是奇函数,\(X(\omega)\)是偶函数,\(|F(\omega)|\)仍然为偶函数,\(\varphi(\omega)\)仍然为奇函数。

4. 尺度变换特性

\(\mathscr{F}[f(t)]=F(\omega)\),则

\[\mathscr{F}[f(at)]=\frac{1}{|a|}F(\frac{\omega}{a}) \tag{4} \]

其中\(a\)为非零实常数。
证明:

\[\begin{aligned} \mathscr{F}[f(at)]=\int_{-\infty}^{\infty}f(at)e^{-j\omega t}dt \end{aligned} \]

\(x=at\)

\(a>0\)

\[\begin{aligned} \mathscr{F}[f(at)]&=\frac{1}{a}\int_{-\infty}^{\infty}f(at)e^{-j\omega\frac{x}{a}}dx \\ &=\frac{1}{a}F(\frac{\omega}{a}) \end{aligned} \]

\(a<0\)

\[\begin{aligned} \mathscr{F}[f(at)]&=\frac{1}{a}\int_{+\infty}^{-\infty}f(at)e^{-j\omega\frac{x}{a}}dx \\ &=\frac{-1}{a}\int_{-\infty}^{\infty}f(x)e^{-j\omega\frac{x}{a}}dx \\ &=\frac{-1}{a}F(\frac{\omega}{a}) \end{aligned} \]

综合上述两种情况,便可以证明式\((4)\)

对于\(a=-1\)这种特殊情况,\(\mathscr{F}[f(-t)]=F(-\omega)\)

由上可见,信号在时域中压缩(a>1)等效于在频域中扩展;反之,信号在时域中扩展(a<1)则等效于在频域中压缩。对于\(a=-1\)的情况,它说明信号在时域中反褶等效于在频域中也反褶。上述结论是不难理解的,因为信号的波形压缩\(a\)倍,信号随时间变化加快\(a\)倍,所以它所包含的频率分量增加\(a\)倍,也就是说频谱展宽\(a\)倍,根据能量守恒原理,各频率分量的大小必然减小\(a\)倍。

因为\(F(\omega)=\int_{-\infty}^{\infty}f(t)e^{-j\omega t}dt\),所以

\[F(0)=\int_{-\infty}^{\infty}f(t)dt \tag{4a} \]

同样因为:\(f(t)=\frac{1}{2\pi}\int_{-\infty}^{\infty}F(\omega)e^{-j\omega t}d\omega\),所以

\[f(0)=\frac{1}{2\pi}\int_{-\infty}^{\infty}F(\omega)d\omega \tag{4b} \]

上面两式分别说明\(f(t)\)\(F(\omega)\)所覆盖的面积等于\(F(\omega)\)\(2\pi f(t)\)在零点的数值\(F(0)\)\(2\pi f(0)\)

如果\(f(0)\)与F(0)各自等于\(f(t)\)\(F(\omega)\)曲线的最大值,这时定义\(\tau\)\(B\)分别为\(f(t)\)\(F(\omega)\)的等效宽度,可写出如下关系式:

\[\begin{aligned} f(0)\tau &= F(0) \\ F(0)B &= 2\pi f(0) \end{aligned} \]

从而可以得到

\[B=\frac{2\pi}{\tau} \tag{4c} \]

可以看出,信号的等效脉冲宽度与占有的等效带宽成反比,若要压缩信号的持续时间,则不得不以展宽频带作为代价。所以通信系统中,通信速度和占用频带宽度是一对矛盾。

5. 时移特性

\(\mathscr{F}[f(t)]=F(\omega)\),则

\[\mathscr{F}[f(t-t_{0})]=F(\omega)e^{-j\omega t_{0}} \tag{5} \]

证明:

因为

\[\mathscr{F}[f(t-t_{0})]=\int_{-\infty}^{\infty}f(t-t_{0})e^{-j\omega t}dt \]

\[x = t-t_{0} \]

那么

\[\begin{aligned} \mathscr{F}[f(t-t_{0})]&=\mathscr{F}[f(x)]=\int_{-\infty}^{\infty}f(x)e^{-j\omega(x+t_{0})}dx \\ &= e^{-j\omega t_{0}}\int_{-\infty}^{\infty}f(x)e^{-j\omega x}dx \end{aligned} \]

同理可得:

\[\mathscr{F}[f(t+t_{0})]=e^{j\omega t_{0}}\cdot F(\omega) \]

也就是说,信号移动后,其幅度谱不变,而相位谱产生附件变换\((-\omega t_{0}/\omega t_{0})\)

6. 频移特性

\(\mathscr{F}[f(t)]=F(\omega)\),则,

\[\mathscr{F}[f(t)e^{j\omega_{0}t}]=F(\omega-\omega_{0}) \tag{6} \]

证明:

因为:

\[\begin{aligned} \mathscr{F}[f(t)e^{j\omega_{0}t}]&=\int_{-\infty}^{\infty}f(t)e^{j\omega_{0}t}\cdot e^{-j\omega t}dt \\ &= \int_{-\infty}^{\infty}f(t)e^{-j(\omega-\omega_{0})t}dt\end{aligned} \]

所以

\[\mathscr{F}[f(t)e^{j\omega_{0}t}]=F(\omega-\omega_{0}) \]

同理

\[\mathscr{F}[f(t)e^{-j\omega_{0}t}]=F(\omega+\omega_{0}) \]

可见,若时间信号\(f(t)\)乘以\(e^{j\omega_{0}t}\),等效于\(f(t)\)的频谱\(F(\omega)\)沿频率轴右移\(\omega_{0}\),或者说在频域中将频谱沿频率轴右移\(\omega_{0}\)等效于在时域中信号乘以因子\(e^{j\omega_{0}t}\)

频谱搬移技术在通信系统中得到广泛应用,诸如调幅、同步解调、变频等过程都是在频谱搬移的基础上完成的。频谱搬移的实现原理是将信号\(f(t)\)乘以所谓载频信号\(cos(\omega_{0}t)/sin(\omega_{0}t)\)

因为:

\[\begin{aligned} cos(\omega_{0}t) &= \frac{1}{2}(e^{j\omega_{0}t}+e^{-j\omega_{0}t}) \\ sin(\omega_{0}t) &= \frac{1}{2j}(e^{j\omega_{0}t}-e^{-j\omega_{0}t}) \end{aligned} \]

那么,可以导出:

\[\begin{aligned} \mathscr{F}[f(t)cos(\omega_{0}t)] &= \frac{1}{2}[F(\omega+\omega_{0})+F(\omega-\omega_{0})] \\ \mathscr{F}[f(t)sin(\omega_{0}t)] &= \frac{1}{2}[F(\omega+\omega_{0})-F(\omega-\omega_{0})] \end{aligned} \tag{6a} \]

所以,若时间信号\(f(t)\)乘以\(cos(\omega_{0}t)\)\(sin(\omega_{0}t)\),等效于\(f(t)\)的频谱\(F(\omega)\)一分为二,沿频率轴向左向右各平移\(\omega_{0}\)

已知直流信号的频谱是位于\(\omega=0\)点的冲激函数,也即:

\[\mathscr{F}[1]=2\pi\delta(\omega) \]

利用频移定理,可以得到:

\[\mathscr{F}[cos(\omega_{0}t)]=\pi[\delta(\omega+\omega_{0})+\delta(\omega-\omega_{0})] \]

\[\mathscr{F}[sin(\omega_{0}t)]=j\pi[\delta(\omega+\omega_{0})-\delta(\omega-\omega_{0})] \]

7. 微分特性

\(\mathscr{F}[f(t)]=F(\omega)\),则

\[\begin{aligned} \mathscr{F}[\frac{df(t)}{dt}] &= j\omega F(\omega) \\ \mathscr{F}[\frac{d^{n}f(t)}{dt^{n}}] &= (j\omega)^{n} F(\omega) \\ \end{aligned} \tag{7} \]

证明:

因为:

\[f(t)=\frac{1}{2\pi}\int_{-\infty}^{\infty}F(\omega)e^{j\omega t}d\omega \]

两边对t求导,得:

\[\frac{df(t)}{dt}=\frac{1}{2\pi}\int_{-\infty}^{\infty}[j\omega F(\omega)]e^{j\omega t}dt \]

所以:

\[\mathscr{F}[\frac{df(t)}{dt}] = j\omega F(\omega) \]

同理,可以推出:

\[\mathscr{F}[\frac{d^{n}f(t)}{dt^{n}}] = (j\omega)^{n} F(\omega) \]

同理可以推出,频域的微分特性如下:

\[\begin{aligned} \mathscr{F}^{-1}[\frac{dF(\omega)}{d\omega}] &= (-jt) f(t) \\ \mathscr{F}^{-1}[\frac{d^{n}F(\omega)}{d\omega^{n}}] &= (-jt)^{n} f(t) \\ \end{aligned} \tag{7a} \]

8. 积分特性

\(\mathscr{F}[f(t)]=F(\omega)\),则

\[\mathscr{F}[\int_{-\infty}^{t}f(\tau)d\tau]=\frac{F(\omega)}{j\omega}+\pi F(0)\delta(\omega) \tag{8} \]

证明:

\[\begin{aligned} \mathscr{F}[\int_{-\infty}^{t}f(\tau)d\tau]&=\int_{-\infty}^{\infty}[\int_{-\infty}^{t}f(\tau)d\tau]e^{-j\omega t}dt \\ &=\int_{-\infty}^{\infty}[\int_{-\infty}^{\infty}f(\tau)u(t-\tau)d\tau]e^{-j\omega t}dt \end{aligned} \tag{8a} \]

上式将被积函数\(f(t)\)乘以\(u(t-\tau)\),同时将积分上限\(t\)改为\(\infty\),结果不变。交换积分次序,并引用延时阶跃信号的傅里叶变换关系式,式\((8a)\)成为:

\[\begin{aligned} &\quad \int_{-\infty}^{\infty}f(\tau)[\int_{-\infty}^{\infty}u(t-\tau)e^{-j\omega t}dt]d\tau \\ &=\int_{-\infty}^{\infty}f(\tau)\pi \delta(\omega)e^{-j\omega \tau}d\tau + \int_{-\infty}^{\infty}f(\tau)\frac{e^{-j\omega\tau}}{j\omega}d\tau\\ &=\pi F(0)\delta(\omega)+\frac{F(\omega)}{j\omega} \end{aligned} \tag{8a} \]

9. 卷积特性

(1)时域卷积定理

给定两个时间函数\(f_{1}(t), f_{2}(t)\),已知\(\mathscr{F}[f_{1}(t)]=F_{1}(\omega),\mathscr{F}[f_{2}(t)]=F_{2}(\omega)\),那么

\[\mathscr{F}[f_{1}(t)*f_{2}(t)]=F_{1}(\omega)F_{2}(\omega) \tag{9a} \]

证明:

根据卷积的定义

\[f_{1}(t)*f_{2}(t)=\int_{-\infty}^{\infty}f_{1}(\tau)f_{2}(t-\tau)d\tau \]

因此:

\[\begin{aligned} \mathscr{F}[f_{1}(t)*f_{2}(t)]&=\int_{-\infty}^{\infty}[\int_{-\infty}^{\infty}f_{1}(\tau)f_{2}(t-\tau)d\tau]e^{-j\omega t}dt \\ &= \int_{-\infty}^{\infty}f_{1}(\tau)[\int_{-\infty}^{\infty}f_{2}(t-\tau)e^{-j\omega t}dt]d\tau \\ &= \int_{-\infty}^{\infty}f_{1}(\tau)F_{2}(\omega)e^{-j\omega\tau}d\tau \\ &= F_{2}(\omega)\int_{-\infty}^{\infty}f_{1}(\tau)e^{-j\omega\tau}d\tau \\ &= F_{1}(\omega)F_{2}(\omega) \end{aligned}\]

(2)频域卷积定理

\[\begin{aligned} \mathscr{F}[f_{1}(t)]=F_{1}(\omega) \\ \mathscr{F}[f_{2}(t)]=F_{2}(\omega) \\ \end{aligned}\]

那么

\[\mathscr{F}[f_{1}(t)\cdot f_{2}(t)]=\frac{1}{2\pi}F_{1}(\omega)*F_{2}(\omega) \tag{9b} \]

posted @ 2022-02-09 13:24  Vinson88  阅读(13595)  评论(0编辑  收藏  举报