Inference of disjoint linear and nonlinear sub-domains of a nonlinear mapping

This paper investigates new ways of inferring nonlinear dependence from measured data. The existence of unique linear and nonlinear sub-spaces which are structural invariants of general nonlinear mappings is established and necessary and sufficient conditions determining these sub-spaces are derived. The importance of these invariants in an identification context is that they provide a tractable framework for minimising the dimensionality of the nonlinear modelling task. Specifically, once the linear/nonlinear sub-spaces are known, by definition the explanatory variables may be transformed to form two disjoint sub-sets spanning, respectively, the linear and nonlinear sub-spaces. The nonlinear modelling task is confined to the latter sub-set, which will typically have a smaller number of elements than the original set of explanatory variables. Constructive algorithms are proposed for inferring the linear and nonlinear sub-spaces from noisy data.

[1]  T. Johansen,et al.  On transient dynamics, off-equilibrium behaviour and identification in blended multiple model structures , 1999, 1999 European Control Conference (ECC).

[2]  Roderick Murray-Smith,et al.  The operating regime approach to nonlinear modelling and control , 1997 .

[3]  Christopher K. I. Williams.  Prediction with Gaussian Processes: From Linear Regression to Linear Prediction and Beyond , 1999, Learning in Graphical Models.

[4]  W. Leithead,et al.  Gain-scheduled and nonlinear systems : dynamic analysis by velocity-based linearization families , 1998 .

[5]  B. Silverman,et al.  Nonparametric regression and generalized linear models , 1994 .

[6]  Lennart Ljung,et al.  Nonlinear black-box modeling in system identification: a unified overview , 1995, Autom..

[7]  Tor Arne Johansen,et al.  Identification of non-linear system structure and parameters using regime decomposition , 1995, Autom..

[8]  L. Ljung.  Nonlinear Black Box Models in Systems Identification , 1997 .

[9]  Bernard Delyon,et al.  Nonlinear black-box models in system identification: Mathematical foundations , 1995, Autom..

[10]  William Leithead,et al.  Gain-scheduled controller design: An analytic framework directly incorporating non-equilibrium plant dynamics , 1998 .

[11]  William Leithead,et al.  Survey of gain-scheduling analysis and design , 2000 .

[12]  A. O'Hagan,et al.  Curve Fitting and Optimal Design for Prediction , 1978 .

[13]  W. Leithead,et al.  Gain-Scheduled & Nonlinear Systems : Dynamic Analysis by Velocity-Based Linearisation Families , 1998 .

[14]  Michael Athans,et al.  Analysis of gain scheduled control for nonlinear plants , 1990 .

[15]  Heekuck Oh,et al.  Neural Networks for Pattern Recognition , 1993, Adv. Comput..

[16]  Geoffrey E. Hinton,et al.  Bayesian Learning for Neural Networks , 1995 .

[17]  Lennart Ljung,et al.  Nonlinear Black Box Modeling in System Identification , 1995 .

[18]  William Leithead,et al.  Analysis of gain-scheduled and nonlinear systems by velocity-based linearisation families , 1997, Proceedings of the 36th IEEE Conference on Decision and Control.

[19]  S. Billings,et al.  Orthogonal parameter estimation algorithm for non-linear stochastic systems , 1988 .

[20]  Tor Arne Johansen,et al.  Design and analysis of gain-scheduled control using local controller networks , 1997 .