Automatic Control Linear Systems PART VII
Should be the last one
Table of Contents
Lec 8 Linear feedback controllers, s.s. observers
I. Lyapunov method
a. Properties
Let us consider Lyapunov functions for investigation stability of linear systems. Lyapunov functions have the next properties:
- Lyapunov function V (π₯) must be positive definite: for any π₯ β π π
- Lyapunov function V (π₯) is positive definite and V (π₯) = 0 in case π₯ is null-vector (zero).
- Lyapunov functions must increases (decreases) uniformly with uniform increasing (decreasing) of π₯-vector norm.
That is, $\lim_{x\to 0}V(x)=0$ and $\lim_{x\to \infty}V(x)=\infty$
b. Lyapunov Theorem
The equilibrium π₯ = 0 is asymptotic stable if exists Lyapunov function V (π₯) such that for any motion trajectories π₯(π‘) starting from the arbitrary initial conditions for any time βπ‘ β₯ 0 the derivative of the function is negative: $\frac{dV(x(t))}{dt} < 0$.
c. Example
- P is symmetrical positive definite matrix. Its dimension is the same with A.
- In linear algebra, a symmetric nxn real matrix) Mis said to be positive-definite
II. State Feedback design
a. Scenary
- note that the original inequality doesnβt have LMI.
b. Example
- In u(t) there should be a minus sign.
c. MATLAB realization
d. Modal Control
- ita and gamma in capital.
e. Example
- -BKM=-BH
f. Observer Design
- In former lectures, we assumed that we obtain all states from output, but in practice, all states requries more moeny for sensors.
- Observer helps to see state that canβt be obtained directly.