Multiple Models Book: Local Laguerre models
This chapter presents the use of Laguerre local models to solve some
nonlinear identification and control problems. The proposed architecture
makes use of Laguerre filters as a state representation of the system input
to model stable nonlinear systems.
This architecture does not require direct estimation of the local
model orders, providing an alternative representation
to the traditional ARMAX local model. Its characteristics make it suitable
for modelling nonlinear systems having time delays or dynamic orders which
depend on operational conditions. A design of a non-linear predictive
controller based on Laguerre local representations is presented.
This approach
does not require sophisticated algorithms for performing the calculation
of the control signals. In addition, the use of Laguerre networks provides
more robust operation, because it provides a smooth transition among the
different operating regimes characterised by local models with different
orders.
The identification and control of a simulated continuous stirred tank reactor
illustrates the application of the proposed models.
R Murray-Smith
Last modified: Tue Mar 25 16:35:22 GMT