Find $K$ equal to the eigenvalues of $A - B K$

323 Views Asked by At

I've a little trouble with this old exam question. It is a multiple choice exam question!


A discrete state-space is represented as $$x(t-1) = Ax(t) + Bu(t)$$ where $$A = \begin{bmatrix}-3&1&-1\\0&-2&-1\\1&0&-2\end{bmatrix} \qquad B = \begin{bmatrix}0.5\\0.5\\-0.5\end{bmatrix}$$ Find a $K \in \mathbb{R}^{1x3}$ equal to the eigenvalues of $A-BK$.


This problem looks pretty hard because it is a 3th order system. It isn't possible to use Matlab or a calculator. The only way I can think of when solving this problem would be to take for K:

$K = \begin{bmatrix}k1\\k2\\k3\end{bmatrix}$ and work out the $det((A-BK)-I\lambda)=0$ equation where the solution got a form of: $(\lambda -k_1)(\lambda -k_2)(\lambda -k_3)$ and thus the eigenvalues would be equal to $\lambda_1 = k_1,\lambda_2 = k_2,\lambda_3 = k_3$. The equation $(A-BK)-I\lambda$ is: $(A-BK)-I\lambda =\begin{bmatrix}-3-\lambda&1&-1\\0&-2-\lambda&-1\\1&0&-2-\lambda\end{bmatrix} + \begin{bmatrix}-0.5k_1&-0.5k_2&-0.5k_3\\-0.5k_1&-0.5k_2&-0.5k_3\\0.5k_1&0.5k_2&0.5k_3\end{bmatrix}$

I calculated the controllability matrix, which is $M_c = \begin{bmatrix}0.5&-0.5&-0.5\\0.5&-0.5&-0.5\\-0.5&1.5&-3.5\end{bmatrix}$ which is rank 2 (< rank n = 3) and thus uncontrollable.

To get to a solution for $det((A-BK)-I\lambda)=0$ it will take ages. There is probably a trick which needs to be used or something I missed. Thanks in advance!


This question is solved. The system (A,B) has one uncontrollable eigenvalue, which is at -3. This eigenvalue can't be relocated. There was only one answer of the multiple choice options which had this eigenvalue. A bit silly, because the equation did not have to be worked out completely. The possible answers were needed to get to the final answer.

1

There are 1 best solutions below

0
On BEST ANSWER

This question is solved. The system (A,B) has one uncontrollable eigenvalue, which is at -3. This eigenvalue can't be relocated. There was only one answer of the multiple choice options which had this eigenvalue. A bit silly, because the equation did not have to be worked out completely. The possible multiple choice answers were needed to get to the final answer.