# Pole-placement problem

I was following this example of implementing a pole-placement with an Integral controller. Matrices A_star=[-5 -3 0;1 0 0;0 -1 0] B_star=[1;0;0] desired poles -100,-8+i*11,-8-i*11 I know the procedure manually is to compare coefficients of det{sI-(A_star-B_star*K)} and charcteristic polynomial of the poles. I have done this and arrived at the answer of K=[111,1782 ,-18500] Even using MATLAB place command, I got the following result, K=[111.000000000005 1782.00000000008 -18500.0000000005]

from videoThe correct answer is K =

1.0e+04 *

0.0111    0.1782    1.8500


My Last coefficient is differing by a sign and I have spent a day on this. I have verified my matrices and poles, but his end result yields a different value. Note:The video's values for K are correct as they give the stable response in simulink as compared to mine.

• Is the entry at (3,2) in A_Star 1 instead of -1. – Suba Thomas Apr 30 '18 at 16:54
• @SubaThomas no. You can verify from the workspace entry of the video's A_Star. – aadil095 Apr 30 '18 at 16:59
• I should have looked at the video. The answer there is -18500 as expected! youtu.be/CsB0TGRYq08?t=972 – Suba Thomas Apr 30 '18 at 17:37
• But he did not put the minus sign for the gain after integrator, why? – aadil095 Apr 30 '18 at 17:57
• The minus sign is in the summing junction before the integrator. – Suba Thomas Apr 30 '18 at 18:10