User`s guide

2 Choosing Your System Identification Strategy
Modeling Multip
le Outputs Directly
You can estimate
the following types of models for multiple-output data:
• Impulse- and ste
p-response m odels
• Frequency-resp
onse models
• Linear ARX model
s
• State-space mod
els
• Nonlinear ARX and
Hammerstein-Wiener models
• Linear and nonli
near ODEs
Tip Estimating m
ultiple-output state-space models directly generally
produces better
results than estimating other types of multiple-output mode ls
directly.
Modeling Multip
leOutputsasaCombinationof
Single-Output
Models
You may find that i
t is harder for a single model to explain the behavior o f
several output
s. If you get a poor fi t estimating a multiple-output model
directly, you c
an try building models for subsets of the most important input
and output cha
nnels.
Use this appro
ach when no feedback is present in the dynamic system and
there a re no co
uplings between the outputs. If you are unsure about the
presence of fe
edback, see “Getting Advice About Your Data” on page 1-84.
To construct
partial models, use subreferencing to create partial data sets,
such that eac
h data set contains all inputs and one output. For more
information
about creating partial data sets, see the following sectio ns in the
System Ident
ification Toolbox User’s Guide:
• For working i
n the System Identification Tool GUI, see “Creating Data Sets
from a Subse
t of Signal Channels” on page 1-31.
• For working
at the command line, see the “Subreferencing iddata Objects”
on page 1-55.
2-22