how to fit a Nonlinear model with multiple input and single output by using dataset in matlab

i have input and output measured data for my system and i have ODE model equations which is bilinear and it has 11 states.

i want to fit the model with the measured data and i have to estimate the parmeters of that model, can anyone kindly help me to solve this problem. Thanks in advance.

Answer 1

[Thiago Henrique Gomes Lobato]

Without your data nor any code is difficult to specifically help you. Still, a general thing you may do, if you already have the model, is to simply optimize the parameters. An way to do it is:

Par0 = InitialParameters;
[Par,fval] = fminsearch(@(Par)Error(Par,Input,Output),Par0)

function e = Error(Par,Input,Output)

ModelOuput = YourModel(Par,Input);
e          = rms(ModelOutput-Output);
end

Just substitute the variables/functions accordingly.

[Padmapritha T]

Thank you for your reply sir,

i did this by the following code in matlab.

but the FPE was very high, is there any possibility to reduce the FPE.

data = iddata(Glu,input,Ts);

data.Tstart = 0;

data.TimeUnit = 's';

nx = 11;

in_sys = n4sid(data,nx,'Ts',900);

sys = pem(data,in_sys);

copmare(data, sys);

where Glu is my measured output data and my system has 2 inputs.

[Thiago Henrique Gomes Lobato]

You can try change the model order https://de.mathworks.com/help/ident/ref/n4sid.html#mw_3b314c29-7732-49b5-8f64-0ad3d618a94f