Note: This content is accessible to all versions of every browser. However, this browser does not seem to support current Web standards, preventing the display of our site's design details.


Nonlinear Dynamic Model Structures: Options and Issues


R. Pearson

Purdue University,, School of Electrical Engineering and Computer Science.

Various authors have noted that a significant practical obstacle to the wider use of nonlinear model predictive control (MPC) and other similar model-based control strategies is the lack of suitable nonlinear models. Fundamental models, when available, are generally too complex for direct use in controller design, motivating interest in model reduction procedures, empirical modelling procedures, or gray-box modelling procedures. Generally, these procedures permit some degree of control over the final structure, and this is both an advantage and a disadvantage: it permits us to specify convenient model structures, but it also permits us to specify model structures that cannot capture the dominant process dynamics we wish to model. This paper considers the problem of nonlinear model structure selection from the perspective of structure/behavior relations, permitting us {em a priori} to either select model structures that exhibit certain desirable forms of behavior (e.g., BIBO stability) or reject model structures that exhibit certain undesirable forms of behavior (e.g., chaotic step responses). Specific model structures considered here include nonlinear FIR models (both Volterra models and nonpolynomial FIR models), the output-affine class (which includes Hammerstein, AR-Volterra, and bilinear models as special cases), Lur'e models, and artificial neural networks.


Type of Publication:

(05)Plenary/Invited/Honorary Lecture

File Download:

Request a copy of this publication.
(Uses JavaScript)
% No recipe for automatically generating a BibTex entry for (05)Plenary/Invited/Honorary Lecture
Permanent link