HOWTO.txt

systemID7States.m Documentation:

Here's how to use the fitting code.

1) Get some data. I'm not sure that the data has to be uniformly sampled in
time, but it is probably better if it is.

2) Save the data as an iddata object. If the sampling rate is nonuniform, 
see this entry from the iddata help page:
Non-equal sampling:
    The property 'SamplingInstants' gives the sampling instants of the
    data points. It can always be retrieved by 
    GET(DAT,'SamplingInstants') and is then computed from
    DAT.Ts and DAT.Tstart. 'SamplingInstants' can also be set to an
    arbitrary vector of the same length as the data, so that non-equal
    sampling can be handled. Ts is then automatically set to [].
    
3) On line 58 of systemID7States, replace simdata.mat with the name of the 
.mat file with your iddata object. By default, the code assumes that the
initial state of the vehicle at the beginning of the experiment is all 
zeros. This will almost certainly not be the case, so if you need to
replace the default value with another guess, do it in this cell by 
uncommenting the appropriate lines.

4) Compare the behavior of the model with default parameter values to the 
experimental data. On lines 78 and 80, replace simData with the name of 
your iddata object. If the behavior is really different, you may have to
modify the initial guess of the parameter values in the next step so that
the optimization code converges.

5) If necessary, modify the default initial guess of the parameter values 
by modifying the vector parameterValues defined on line 99. You can then 
re-run the previous cell to see the effects of the change.

6) Run the optimization code to estimate the values of the parameters. This
code takes a while to run but outputs some information about its progress.