Title: A nonlinear correlation-based index to measure independence of sub models in multi-objective identification of MIMO LPV systems

Authors: Daniel Chuk; Gustavo Rodríguez Medina; Enrique Antonio Núñez; Luis V. Gutiérrez; Juan Pedro Gil

Addresses: Instituto de Investigaciones Mineras, Facultad de Ingeniería, Universidad Nacional de San Juan, Argentina ' Instituto de Investigaciones Mineras, Facultad de Ingeniería, Universidad Nacional de San Juan, Argentina ' Instituto de Investigaciones Mineras, Facultad de Ingeniería, Universidad Nacional de San Juan, Argentina ' Instituto de Investigaciones Mineras, Facultad de Ingeniería, Universidad Nacional de San Juan, Argentina ' Instituto de Investigaciones Mineras, Facultad de Ingeniería, Universidad Nacional de San Juan, Argentina

Abstract: A quality index for linear parameter varying (LPV) systems is presented. The purpose of the index is to increase the degree of independence of submodels that contribute to the same output in multiple input multiple output (MIMO) systems and in the usual case where the model parameters depend on filtered versions of the inputs. The importance of the designed index is practically verified in the identification of a nonlinear process with two inputs and two outputs by means of multi-objective optimisation.

Keywords: modelling; nonlinear correlation; multivariable systems; linear parameter varying systems; LPV systems; multi-objective optimisation; submodels; multi-objective identification; MIMO systems; quality index; multiple input multiple output.

DOI: 10.1504/IJMIC.2016.077746

International Journal of Modelling, Identification and Control, 2016 Vol.26 No.1, pp.52 - 58

Received: 15 Apr 2015
Accepted: 27 Aug 2015
Published online: 14 Jul 2016 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article