System Identification is the art and science of building mathematical models from measured input-output data.

You will have to examine the measured data, and create new data sets from the original one by various preprocessing.

Then you will estimate models using the data. The models are estimated within certain classes of candidate descriptions (model structures), typically by choosing that model that gives the best output fit to the measured data. Quite a few models are normally estimated, and their properties are scrutinized and analyzed in different model views.

Ident is a graphical user interface to the System Identification Toolbox, that helps you with these tasks. The ident window governs all data handling, model estimation, and model analysis. The ident window also keeps a detailed record of all data sets and models. Data sets and models are represented as icons in the tables of black boxes.

The process of system identification using ident involves the following six steps:
Importing data into ident.
Examining data.
Preprocessing data.
Estimating models based on data.
Analyzing the models.
Exporting resulting models for further use.

Help topics.

View Windows The ident window contains two check boxes for examining Views of the selected data sets

It contains six check boxes that allow you to examine the Data Viewsgive time and frequency domain plots of the input output data.

The Model Viewsanalysis and displays several model properties.

Help topics.

Data View Windows There are two views to examine data sets. These view windows are opened by checking the corresponding check boxes in the ident window. Select and deselect which data sets to examine by clicking on their icon.

Time Plot
This gives the time plots of inputs and outputs of the selected data sets.
Data spectra
This gives the spectra (periodograms or estimated spactra) of inputs and outputs of the selected data sets.

Help topics.

Model View Windows For text information and access to the model parameters, double-click (middle/right mouse button) on the model icon. For visual infomation, there are six views to analyze and examine model properties. These view windows are opened by checking the corresponding check boxes in the ident window. Select and deselect which models to examine by clicking on their icons.

Model Output
This view shows the simulated or predicted output of the selected models together with the actually measured output. The calculations are carried out for the Model Resids
This view shows the result of residual analysis for the each ofselected models. The residuals (prediction errors) are computed using the Transient Response
This view shows the transient response (impulse or step response) of the selected models.
Frequency Response
This view shows the frequency functions of the selected models.
Zeros and Poles
This view shows the Zeros and Poles of the selected models.
Noise Spectrum
This view shows the spectrum of the additive noise of the selected models. In case a time series is modeled, its spectrum is shown.

Help topics.