Lesson 18 of 19 · Module 4: State space analysis

Controllability and observability

Control System Engineering · EEE2510 · Sessions 52–56 · CO6

By the end of this lesson you can

  • Define controllability and observability
  • Apply Kalman's rank tests
  • Explain why they matter for design

Notes

A system is controllable if some input can move the state from any initial value to any final value in finite time. It is observable if the initial state can be found from the output measured over a finite time.

Kalman's tests (for an $n$-th order system)

$$Q_c = \begin{bmatrix}B & AB & A^2B & \cdots & A^{n-1}B\end{bmatrix},\qquad \text{rank}(Q_c) = n$$

$$Q_o = \begin{bmatrix}C \\ CA \\ \vdots \\ CA^{n-1}\end{bmatrix},\qquad \text{rank}(Q_o) = n$$

For a single-input or single-output system, check that the determinant of the square test matrix is not zero.

Why they matter

  • State feedback can place every pole only if the system is controllable.
  • An observer (state estimator) can rebuild every state only if the system is observable.
  • A pole–zero cancellation in the transfer function hides an uncontrollable or unobservable mode.

Lecture

🎬
Lecture video
Coming soon. The recorded lecture for this lesson will appear here.

Exercises

  1. Is $A = \begin{bmatrix}0&1\\-2&-3\end{bmatrix}$, $B = \begin{bmatrix}0\\1\end{bmatrix}$ controllable?
    Show solution
    $Q_c = \begin{bmatrix}0&1\\1&-3\end{bmatrix}$, determinant $-1 \ne 0$: controllable.
  2. Same $A$, $C = [1\ 1]$. Is it observable?
    Show solution
    $CA = [-2\ -2]$. $Q_o = \begin{bmatrix}1&1\\-2&-2\end{bmatrix}$, determinant 0: not observable. (The zero at $s = -1$ cancels the pole at $-1$.)
  3. Why can't you place all poles by state feedback if a system is uncontrollable?
    Show solution
    The uncontrollable mode is not affected by the input at all, so no feedback through the input can move its pole.