Different results in Matlab and Python for StateSystem #1077
|
While migrating to Python from Matlab, I get different result for StateSystem transform. tf1 = Numerator: {[0 0 0 -4]} tf2 = Numerator: {[0 0 0]} tf3 = Numerator: {[0 0 0.7500]} MatlabCode: sys=[tf1,tf2;tf3,0];
sys_ss=ss(sys);the output is: In Python，I use control.ss: import control
num11 = [0,0,0,-4]
num12 = [0,0,0]
num21 = [0,0,0.7500]
num22 = [0]
row1 = [num11, num12]
row2 = [num21, num22]
numerator = [row1, row2]
denominator = [
[[1000000, 30000, 300,1], [5600,150,1]],
[[5600,150,1], [1]]
]
G = control.tf(numerator,
denominator)
control.ss(G )Then I got different StateSpace， What am I missing here? How can I fix to get the exact same results? |
Replies: 4 comments 2 replies
|
Different results are normal, this is not an error! Try to cast again to tf to see the result. (Sorry, I can't check it because I don't have Matlab anymore!) For the same tf (or tf's) there are infinity possible state-space representations, depending on the chosen state variables! |
|
method balance() in Matlab has different output compared to Python，so they have different results. |
|
Can you post the results you are getting from MATLAB (I don't have it installed on my machine since I only use Python). |
|
A little problem is given by the very little values in the state matrix (...*e-17): I also found that python control doesn't approximate these values to "0". |
This comment was marked as spam.
This comment was marked as spam.
|
Yes, I had the same experience. See this related issue: |
A little problem is given by the very little values in the state matrix (...*e-17): I also found that python control doesn't approximate these values to "0".