控制系统状态空间表达式的解(2)——计算矩阵指数函数的常用方法

方法一

根据矩阵指数函数的定义直接计算：
$$e^{At}\stackrel{def}{=}I+At+\frac{1}{2!}{A}^2{t}^2+...=\sum _{k=0}^{\infty }\frac{1}{k!}{A}^{k}t{}^k$$

(用定义计算，很少用到8)

---

方法二

将$A$阵化为对角标准型或约当标准型求解

1.$A$的特征值不存在重根
若$A$的特征值${\lambda }_1,{\lambda }_2,...,{\lambda }_n$不存在重根，则在求出使$A$阵实现对角化
$$T^{-1}AT=\left[\begin{array}{cccc}{\lambda }_1& & & \\ & {\lambda }_2& & \\ & & \ddots & \\ & & & {\lambda }_n\end{array}\right]$$
的变换阵$T^{-1}$，$T$后，即有指数函数矩阵：
$$e^{At}=T\left[\begin{array}{cccc}{e}^{{\lambda }_{1}t}& & & \\ & e^{{\lambda }_{2}t}& & \\ & & \ddots & \\ & & & e^{{\lambda }_{n}t}\end{array}\right]T^{-1}$$
证明：
$$由:T^{-1}AT=\left[\begin{array}{cccc}{\lambda }_1& & & \\ & {\lambda }_2& & \\ & & \ddots & \\ & & & {\lambda }_n\end{array}\right]可得:A=T\left[\begin{array}{cccc}{\lambda }_1& & & \\ & {\lambda }_2& & \\ & & \ddots & \\ & & & {\lambda }_n\end{array}\right]T^{-1}$$
而:
$$\begin{array}{rl}{e}^{At}& =\sum _{k=0}^{\infty }\frac{1}{k!}{A}^{k}t{}^k\\ & =\sum _{k=0}^{\infty }\frac{1}{k!}{\left(T\left[\begin{array}{cccc}{\lambda }_1& & & \\ & {\lambda }_2& & \\ & & \ddots & \\ & & & {\lambda }_n\end{array}\right]T^{-1}\right)}^k{t}^k\end{array}$$

接下来，先考虑$k=2$的情况：
$$\begin{array}{rl}{\left(T\left[\begin{array}{cccc}{\lambda }_1& & & \\ & {\lambda }_2& & \\ & & \ddots & \\ & & & {\lambda }_n\end{array}\right]T^{-1}\right)}^2& =T\left[\begin{array}{cccc}{\lambda }_1& & & \\ & {\lambda }_2& & \\ & & \ddots & \\ & & & {\lambda }_n\end{array}\right]T^{-1}T\left[\begin{array}{cccc}{\lambda }_1& & & \\ & {\lambda }_2& & \\ & & \ddots & \\ & & & {\lambda }_n\end{array}\right]T^{-1}\\ & =T{\left[\begin{array}{cccc}{\lambda }_1& & & \\ & {\lambda }_2& & \\ & & \ddots & \\ & & & {\lambda }_n\end{array}\right]}^2{T}^{-1}\end{array}$$
将结果带入上式中有
$$\begin{array}{rl}{e}^{At}& =\sum _{k=0}^{\infty }\frac{1}{k!}{A}^{k}t{}^k\\ & =\sum _{k=0}^{\infty }\frac{1}{k!}{\left(\left[\begin{array}{cccc}{\lambda }_1& & & \\ & {\lambda }_2& & \\ & & \ddots & \\ & & & {\lambda }_n\end{array}\right]T^{-1}\right)}^k{t}^k\\ & =\sum _{k=0}^{\infty }\frac{1}{k!}T{\left[\begin{array}{cccc}{\lambda }_1& & & \\ & {\lambda }_2& & \\ & & \ddots & \\ & & & {\lambda }_n\end{array}\right]}^k{T}^{-1}{t}^k\end{array}$$
而$\frac{1}{k!}$和$t^k$都是常数，将其乘到对角阵中去
$$\begin{array}{rl}& =T\left[\begin{array}{cccc}{\sum }_{k=0}^{\infty }\frac{1}{k!}{\lambda }_1^k{t}^k& & & \\ & {\sum }_{k=0}^{\infty }\frac{1}{k!}{\lambda }_2^k{t}^k& & \\ & & \ddots & \\ & & & {\sum }_{k=0}^{\infty }\frac{1}{k!}{\lambda }_n^k{t}^k\end{array}\right]T^{-1}\\ & =T\left[\begin{array}{cccc}{e}^{{\lambda }_{1}t}& & & \\ & e^{{\lambda }_{2}t}& & \\ & & \ddots & \\ & & & e^{{\lambda }_{k}t}\end{array}\right]T^{-1}\end{array}$$
证毕

2.$A$的特征值存在重根时
暂时略

---

方法三

拉氏变换法:
$$e^{At}=L^{-1}[(sI-A{)}^{-1}]$$
证明:
由矩阵指数函数的定义：
$$e^{At}=I+At+\frac{1}{2!}{A}^2{t}^2+...=\sum _{k=0}^{\infty }\frac{1}{k!}{A}^{k}t{}^k$$
两边取拉氏变换:

根据:
$$L(\frac{t^k}{k!})=\frac{1}{s^{k+1}}$$

$$\begin{array}{rl}L(e^{At})& =\frac{1}{s}I+\frac{1}{s^2}A+\frac{1}{s^3}{A}^2+...\\ & =\sum _{k=0}^{\infty }\frac{1}{s^{k+1}}{A}^k\\ & =s^{-1}\sum _{k=0}^{+\infty }(s^{-1}A{)}^k\end{array}$$

根据：
$$\begin{array}{rl}1+x+x^2+...+x^k+...& =\sum _{k=0}^{+\infty }{x}^k\\ & =\frac{1}{1-x}=(1-x{)}^{-1}\end{array}$$

有：
$$\begin{array}{rl}L(e^{At})& =s^{-1}\sum _{k=0}^{+\infty }(s^{-1}A{)}^k\\ & =s^{-1}(I-s^{-1}A{)}^{-1}\\ & =(sI-A{)}^{-1}\end{array}$$
两边取拉式反变换：
$$e^{At}=L^{-1}[(sI-A{)}^{-1}]$$
证毕

---

凯莱-哈米尔顿定理：

设$A\in R^{x\times x}$，其特征多项式为:
$$\begin{array}{rl}D(\lambda )& =|\lambda I-A|\\ & ={\lambda }^n+a_{n-1}{\lambda }^{n-1}+...+a_1\lambda +a_0\\ & =0\end{array}$$
则矩阵A必须满足其特征多项式，即:
$$A^n+a_{n-1}{A}^{n-1}+...+a_{1}A+a_{0}I=0$$

证明略