How to estimate time-varying state space in Matlab?

Actually, I have input and output data which by using iddata commend I made a data include input, output and sample time but when I apply them for estimating state space in continuse time domain it gives an error like this:There are too many parameters to estimate for chosen estimation data size. Reduce model order or use a larger data set. and when I reduce model order to zero but it gives an error : NaN and Inf not allowed.

I would appreicate it if someone could guide me how to estimate state space ?

Answer 1

[Sanjana]

Hi,

I understand that you need guidance with estimating state space. I see that you have created "iddata"object from input, output and sample time details of the data, and in order to estimate State-Space Models at the Command Line, you can use

• ‘n4sid’ function – it supports both time-domain and frequency-domain data, uses noniterative subspace method for estimating linear space models
• ‘ssest’ function- supports both time and frequency domain data, uses Iterative method that uses prediction error minimization algorithm.

For more information on the above functions, refer to the below documentation link

https://www.mathworks.com/help/releases/R2021b/ident/ug/estimate-state-space-time-series-models.html?searchHighlight=estimate%20state%20space%20time%20series%20model&s_tid=doc_srchtitle

Hope this helps!