LMIs in Control/pages/Optimal Output Feedback Hinf LMI

Optimal Output Feedback $H_{\mathrm{\infty }}$ LMI

Optimal output feedback control is a problem which arises from not knowing all information about the output of the system. It correlates to the state feedback situation where the part of the state is unknown. This issue can arise in decentralized control problems, for example, and requires the use of an "observer-like" solution. One such method is the use of a Kalman Filter, a more classical technique. However, other methods exist that do not implement a Kalman Filter such as the one below which uses an LMI to preform the output feeback. The $H_{\mathrm{\infty }}$ control methods form an optimization problem which attempts to minimize the $H_{\mathrm{\infty }}$ norm of the system.

The System

The system is represented using the 9-matrix notation shown below.

$\left[\begin{array}{c}\dot{x}\\ z\\ y\end{array}\right]=\left[\begin{array}{ccc}A& B_1& B_2\\ C_1& D_{11}& D_{12}\\ C_2& D_{21}& D_{22}\end{array}\right]\left[\begin{array}{c}x\\ w\\ u\end{array}\right]$

where $x(t)\in {ℝ}^n$ is the state, $z(t)\in {ℝ}^p$ is the regulated output, $y(t)\in {ℝ}^q$ is the sensed output, $w(t)\in {ℝ}^r$ is the exogenous input, and $u(t)\in {ℝ}^m$ is the actuator input, at any $t\in ℝ$.

The Data

$A$, $B_1$, $B_2$, $C_1$, $C_2$, $D_{11}$, $D_{12}$, $D_{21}$, $D_{22}$ are known.

The LMI: Optimal Output Feedback $H_{\mathrm{\infty }}$ Control LMI

The following are equivalent.

1) There exists a $\hat{K}=\left[\begin{array}{cc}{A}_K& B_K\\ C_K& D_K\end{array}\right]$ such that $||S(K,P)|{|}_{H_{\mathrm{\infty }}}<\gamma$

2) There exists $X_1$, $Y_1$, $Z$, $A_n$, $B_n$, $C_n$, $D_n$ such that

$\left[\begin{array}{cc}{X}_1& I\\ I& Y_1\end{array}\right]>0$

$\left[\begin{array}{cccc}A{Y}_1+Y_1{A}^{\text{T}}+B_2{C}_n+C_n{B}_2^{\text{T}}& {*}^{\text{T}}& {*}^{\text{T}}& {*}^{\text{T}}\\ A^{\text{T}}+A_n+(B_2{D}_n{C}_2{)}^{\text{T}}& X_{1}A+A^{\text{T}}+B_n{C}_2+C_2^{\text{T}}{B}_n^{\text{T}}& {*}^{\text{T}}& {*}^{\text{T}}\\ (B_1+B_2{D}_n{D}_{21}{)}^{\text{T}}& (X_1{B}_1+B_n{D}_{21}{)}^{\text{T}}& -\gamma I& {*}^{\text{T}}\\ C_1{Y}_1+D_{12}{C}_n& C_1+D_{12}{D}_n{C}_2& D_{11}+D_{12}{D}_n{D}_{21}& -\gamma I\end{array}\right]<0$

Conclusion:

The above LMI determines the the upper bound $\gamma$ on the $H_{\mathrm{\infty }}$ norm. In addition to this the controller $\hat{K}(A_K,B_K,C_K,D_K)$ can also be recovered.

$D_K=(I+D_{K2}{D}_{22}{)}^{-1}{D}_{K2}$

$B_K=B_{K2}(I+D_{22}{D}_K)$

$C_K=(I+D_K{D}_{22})C_{K2}$

$A_K=A_{K2}-B_K(I+D_{22}{D}_K{)}^{-1}{D}_{22}{C}_K$

where,

$\left[\begin{array}{cc}{A}_{K2}& B_{K2}\\ C_{K2}& D_{K2}\end{array}\right]={\left[\begin{array}{cc}{X}_2& X_1{B}_2\\ 0& I\end{array}\right]}^{-1}[\left[\begin{array}{cc}{A}_n& B_n\\ C_n& D_n\end{array}\right]-\left[\begin{array}{cc}{X}_{1}A{Y}_1& 0\\ 0& 0\end{array}\right]]{\left[\begin{array}{cc}{Y}_2^T& 0\\ C_2{Y}_1& I\end{array}\right]}^{-1}$

for any full-rank $X_2$ and $Y_2$ such that

$\left[\begin{array}{cc}{X}_1& X_2\\ X_2^T& X_3\end{array}\right]={\left[\begin{array}{cc}{Y}_1& Y_2{B}_2\\ Y_2^T& Y_3\end{array}\right]}^{-1}$.

Implementation

This implementation requires Yalmip and Sedumi. https://github.com/eoskowro/LMI/blob/master/OF_Hinf.m

Related LMIs

Optimal Output Feedback H2

External Links

• LMI Methods in Optimal and Robust Control - A course on LMIs in Control by Matthew Peet.
• LMI Properties and Applications in Systems, Stability, and Control Theory - A List of LMIs by Ryan Caverly and James Forbes.
• LMIs in Systems and Control Theory - A downloadable book on LMIs by Stephen Boyd.
• LMIs in Control Systems: Analysis, Design and Applications - by Guang-Ren Duan and Hai-Hua Yu, CRC Press, Taylor & amp; Francis Group, 2013.