How to predict a partial state-space model by specifying an initial state-space model.

2 次查看(过去 30 天)
Using this link (https://jp.mathworks.com/help/ident/ref/ssest.html#btpb7ww-5) as a reference, I predicted a state-space model using tt data as gray-box modeling with system identification.
I constructed init_sys and tried to estimate only the "NaN" part of the matrix, leaving the other elements as fixed values.
However, the output matrices were all free coefficients.
How can I estimate only the "NaN" part?
I thought that specifying "NaN" was not a good idea, so I put the nominal values in init_sys, but the output results were exactly the same as the nominal values.
Is it still dependent on the initial value matrix?
Code is following (This is a linear approximation model of cart pole.)
%grayboxmodeling
% State Space
A = [0 1 0 0
0 0 NaN 0
0 0 0 1
0 0 NaN 0];
B = [0
NaN
0
NaN];
C = eye(4);
D = zeros(4,1);
K = zeros(4,4);
x0 = [0;
0;
-0.03;
0];
Ts = 0.02;
init_sys = idss(A,B,C,D,K,x0,Ts);
apar = init_sys.Structure.A;
apar.Free(1,:) = false;
apar.Free(2,1) = false;
apar.Free(2,2) = false;
apar.Free(2,3) = true;
apar.Free(2,4) = false;
apar.Free(3,:) = false;
apar.Free(4,1) = false;
apar.Free(4,2) = false;
apar.Free(4,4) = false;
init_sys.Structure.A = apar;
bpar = init_sys.Structure.B;
bpar.Free(1,1) = false;
bpar.Free(3,1) = false;
init_sys.Structure.B = bpar;
init_sys.Structure.C.Free = false;
sysgray = ssest(tt,init_sys)

采纳的回答

Tianyu
Tianyu 2024-5-1
Hi 健斗,
NaN is not accepted since the estimation requires a valid initial value to start.
The following code works for me.
%grayboxmodeling
load sdata2 tt2
tt = tt2;
% State Space
A = [0 1 0 0
0 0 1 0
0 0 0 1
0 0 1 0];
B = [0
1
0
1];
C = eye(1,4);
D = zeros(1,1);
K = zeros(4,1);
x0 = [0;
0;
-0.03;
0];
Ts = 0.1;
init_sys = idss(A,B,C,D,K,x0,Ts);
apar = init_sys.Structure.A;
apar.Free(1,:) = false;
apar.Free(2,1) = false;
apar.Free(2,2) = false;
apar.Free(2,3) = true;
apar.Free(2,4) = false;
apar.Free(3,:) = false;
apar.Free(4,1) = false;
apar.Free(4,2) = false;
apar.Free(4,4) = false;
init_sys.Structure.A = apar;
bpar = init_sys.Structure.B;
bpar.Free(1,1) = false;
bpar.Free(3,1) = false;
init_sys.Structure.B = bpar;
init_sys.Structure.C.Free = false;
sysgray = ssest(tt,init_sys)
>> sysgray.A
ans =
0 1.0000 0 0
0 0 2.2653 0
0 0 0 1.0000
0 0 0.2710 0

更多回答(0 个)

类别

Help CenterFile Exchange 中查找有关 Linear Model Identification 的更多信息

产品


版本

R2023a

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by