How to estimate the inital state with an ssest output model?
I am trying to linearize a model arround a certain operating point, the model is highly non linear so things like small angle approximation or Taylor-Series dont really work. For this is used ssest. The model has a total of 5 states but ssest suggests to linearize with 8 states, so it is creating transitional states. This concept is not completely new to me but i have never worked with it. The compare function shows a very good fit, but I dont know how the inital state is created since in the iddata the original inital state has 5 values and the linearized model demands 8. Compare outputs an Initial state which I tried to recreate but couldnt. Can anybody explain to me how to translate the original nonlinear state to the new states of the estimatete SS-Model?
John Williams answered .
2025-11-20
To estimate the initial state with an ssest output model in MATLAB, you can use the following approach:
Estimate the State-Space Model: Use the ssest function to estimate the state-space model from your data. This function returns the estimated model and, optionally, the initial state.
Return Initial States: Use the syntax [sys, x0] = ssest(data, nx) to get the initial state x0 along with the estimated model sys.
Here's an example:
% Load your data
load iddata1 % Replace with your actual data
% Define the model order
nx = 4; % Example model order
% Estimate the state-space model and initial state
[sys, x0] = ssest(z1, nx);
% Display the initial state
disp('Initial State:');
disp(x0);
In this example, z1 is your data, and nx is the order of the state-space model you want to estimate. The ssest function will return the estimated model sys and the initial state x0