System Identification Toolbox

Parametric Model Estimation

The System Identification Toolbox contains several functions for parametric model estimation. They all share the same command structure.

• m = function(Data,modstruc)
  m =   ...
  function(Data,modstruc,'Property1',Value1,...'PropertyN',ValueN
  )
  

The argument Data is an iddata object that contains the output and input data sequences, while modstruc specifies the particular structure of the model to be estimated. The resulting estimated model is contained in m. It is a model object that stores various information. The model objects will be described in Defining Model Structures, but for most use of the toolbox, you do not have to consider the details of these objects. Just typing the model name

• m
  

will give a concise display of the model. The command

• present(m)
  

gives some more details, while

• get(m)
  

gives a complete list of the model's properties. The property values can be easily retrieved just by dot-referencing; for example,

• m.par
  

retrieves the estimated parameters.

In the function call (...,'Property1', Value1,...,'PropertyN',ValueN) is a list of properties that can be assigned to affect the model structure, as well as the estimation algorithm. A list of typical properties is given at the end of this section. The model m is also immediately prepared for displaying and analyzing its characteristics as well as for transforming it to other representations, as in

• bode(m)
  compare(Data,m)
  [A,B,C,D, K] = ssdata(m)
  

See Examining Models for a detailed discussion of these possibilities.

In the following, Data denotes an iddata object that contains the input output data as described in the previous section. It may also just contain an output signal, i.e., a time series.

More on the Data Representation in iddata

ARX Models