6533b7d5fe1ef96bd1263d16

RESEARCH PRODUCT

Identification for a general class of LPV Models

subject

description

Abstract In this paper we consider the problem of identifying discrete-time Linear Parameter Varying (LPV) models of non-linear or time-varying systems. LPV models are considered for their connection with the industrial practice of gain-scheduling. We assume that inputs, outputs and the scheduling parameters are measured, and a form of the functional dependence of the coefficients on the parameters is known. We show how the identification problem can be reduced to a linear regression so that a Least Mean Square identification algorithm can be reformulated. Conditions on the persistency of excitation in terms of the inputs and parameter trajectories are given to ensure the consistency of the algorithms when the functional dependence is of polynomial type. Simulation example are given for both periodic and non periodic trajectory of the scheduling parameters.