How to extract idpoly property to workspace

Hi, how can I extract and put in my workspace variable A and NoiseVariance from idpoly to do further processing?

Answer 1

[Star Strider]

There are several properties that can be retrieved from the System Identification Toolbox ‘data’ objects. See the documentation on idpoly and others for details.

Try something like this —

LD = load(websave('idpoly','https://www.mathworks.com/matlabcentral/answers/uploaded_files/1248207/idpoly.mat'))

LD = struct with fields:
    a: [1×0 idpoly]

a = LD.a

a =
Discrete-time AR model: A(z)y(t) = e(t)                                                                                                                                  
                                                                                                                                                                         
  A(z) = 1 + (-7.954 - 0.3i) z^-1 + (26.36 + 2.3i) z^-2 + (-45.35 - 7.6i) z^-3 + (38.88 + 13i) z^-4 + (-7.968 - 12i) z^-5 + (-8.478 + 6.2i) z^-6 + (-3.663 - 2.7i) z^-7  
          + (11.26 + 2.5i) z^-8 + (0.9163 + 0.051i) z^-9 + (-7.189 - 2.7i) z^-10 + (-4.175 + 0.46i) z^-11 + (12.39 + 3.7i) z^-12 + (-4.697 - 5.7i) z^-13 + (-6.017       
              + 3.4i) z^-14 + (6.773 + 3.6i) z^-15 + (-1.439 - 10i) z^-16 + (-1.529 + 11i) z^-17 + (1.173 - 5.9i) z^-18 + (-0.3192 + 1.7i) z^-19 + (0.0307 - 0.21i) z^-20
                                                                                                                                                                         
Sample time: 1 seconds
  
Parameterization:
   Polynomial orders:   na=20
   Number of free coefficients: 20
   Use "polydata", "getpvec", "getcov" for parameters and their uncertainties.

Status:                                                                
Estimated using AR ('burg/now') on time domain data (complex) "est_x1".
Fit to estimation data: 100%                                           
FPE: 2.355e-25, MSE: 2.008e-25                                         
 

Coefficients = a.A

Coefficients = 
   1.0000 + 0.0000i  -7.9541 - 0.2965i  26.3582 + 2.3178i -45.3509 - 7.5709i  38.8797 +13.0764i  -7.9676 -12.3483i  -8.4781 + 6.1760i  -3.6634 - 2.7037i  11.2582 + 2.5171i   0.9163 + 0.0513i  -7.1889 - 2.6674i  -4.1745 + 0.4615i  12.3895 + 3.7428i  -4.6974 - 5.6726i  -6.0167 + 3.4192i   6.7730 + 3.5819i  -1.4387 -10.4534i  -1.5295 +10.7850i   1.1733 - 5.9423i  -0.3192 + 1.7405i   0.0307 - 0.2143i

Variable = a.Variable

Variable = 'z^-1'

Structure = a.Structure

Structure =
                 A: [1×1 param.Continuous]
           IODelay: [1×0 double]
    IntegrateNoise: 0

AR model structure.

You can use the ‘Coefficients’ vector with polyval to evaluate it.

t = linspace(0, 1.5, 150);
A = polyval(Coefficients, t);

figure
plot(t, real(A), 'DisplayName','Re(A)')
hold on
plot(t, imag(A), 'DisplayName','Im(A)')
plot(t, abs(A), 'DisplayName','|A|')
hold off
grid
legend('Location','best')

.