Gaussian Process for Internal Model Control
Gregor Gregor, Gordon Lightbody · 2002
To improve transparency and reduce the curse of di- mensionality of non{linear black{box models, the local modelling ap- proach was proposed. Poor transient response of Local Model net- works led to the use of non{parametrical probabilistic models such as the Gaussian Process prior approach. Recently, Gaussian Process models were applied in the Minimum Variance Control. This pa- per introduces the use of the Gaussian Process model for Non{linear Internal Model control. The invertibility of the Gaussian Process model is discussed and the use of predicted variance is illustrated on a simulated example. Keywords | Non{linear Internal Model Control, Non{linear modelling, Gaussian processes, Invertibility, Covariance function