How to identify a sytem by the System Identification Toolbox that is invertable ?

Hallo,

I want to design an Internal Model Controller, but I have a non-linear system. Therefore I want to generate data from the non-linear model, to identify a linear model by the System Identification Toolbox from that data. (Because I want to set the model for the IMC as M=L+(N-L) so I only have to invert the linear model for control -> Q=inv(L))

Now my problem is that for the identified system D=0 (singular) so no inv() exist. So my question is if there is a possibility to identify a system with a nonsingular D ?

Greetings

Lucas

Answer 1

[Rajiv Singh]

Tyically yes.

• If you are estimating state-space model, use "feedthough" name-value pair, as in ssest(Data, order, 'Feedthrough', 'on')
• If you are estimating continuous-time transfer function, specify number of poles = number of zeros, as in tfest(data, N, N)
• If you are estimating a discrete-time transfer function, use "feedthough" name-value pair, as in tfest(data, np, nz, 'Ts', Ts, 'Feedthrough', true)