Identification toolbox - Parameter identification of a given state space structure

4 views (last 30 days)
Hi together,
is there an opportunity to get parameter values from a known state space structure with the n4sid - function? My problem is as follows - I have given SISO-Signals and a known state space structure - now I want to compute the Parameter values (a_M, omega, k_V) and dynamics for the given state space with the input-output characteristic.
I tried
modelorder = 3;
model_est = n4sid([simout,simin(:,2)],modelorder);
but I have only A, B and C - no throughput D, K, e(t).
A = [a_M 0 0; 0 0 1; (omega^2*k_V) (omega^2) (2*d*omega)]
Or is there any other possibility to get the parameters? Thanks for any help.
Question 2:
What about non-linearity(NL) in the the state-space Model. Is it possible to identify these NLs.? E.g.
A = [a_M 0 0 ... 0;
0 0 1 0 ... 0;
(omega^2*k_V) (omega^2) (2*d*omega) 0 ... 0;
... ;
0 NL 0 NL ... 0;
...;
0 0 1 A_k ... d_K];
Thanks in advance for supporting me. Best Regards, Joh.

Answers (1)

Arkadiy Turevskiy
Arkadiy Turevskiy on 8 May 2014
1. You can enable feedthrough and disturbance estimation for n4sid by setting name-value pair function arguments properly, as described in the doc. However, you won't be able to use n4sid to estimate parameters in a known model structure. For that you need to use idgrey. Check the examples here for how to do it.
Specifically, check out this example , the section called "Specification of Models with Coupled Parameters Using IDGREY Objects".
2. What kind of nonlinearities? The tool cannot give you the universl solution, ie "I automatically determined you have deadzone here". On the other hand, if you can describe nonlinearity with a parameterized equation, the tool can estimate the parameters of the nonlinearity. For that use idnlgrey .
Examples are here .

Community Treasure Hunt

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

Start Hunting!