How can I create an uncertain idpoly model if I know FIR coeffiecients and its uncertainties?

Hello, I need to build an uncertain idpoly model. I have the FIR coefficients (e.g. B(z)=[0 1 2 3 2 1]) and the sampling time (e.g. Ts=1 s), then I build a idpoly model according to the MATLAB help:

    sys=idpoly([],[0 1 2 3 2 1],[],[],[],[],1)

Now the question: I also have an uncertainty in each FIR coefficient which is expressed in standard deviation: std=[0 1e-3 2e-3 3e-3 2e-3 1e-3]. How can I incorporate this knowledge in the idpoly model?

Answer 1

[Michelle Wu]

You may want to use function ' setcov ' to set covariance data in identified model. First, use function 'idpoly' to obtain the identified model (sys in your case). Then, use the following syntax:

 sys1 = setcov(sys,cov)

where cov is the parameter covariance matrix. cov could be represented by an np-by-np semi-positive definite symmetric matrix, where np is equal to the number of parameters of sys (5 in your case). Thus, before using 'setcov', you also need to convert the standard deviation into a covariance matrix. To do so, you may consider using function ' corr2cov ' if you have access to the Financial Toolbox.