State space LPV model identification using LS-SVM: A case-study with dynamic dependence

Abstract

In this paper the nonparametric identification of state-space linear parameter-varying models with dynamic mapping between the scheduling signal and the model matrices is considered. Indeed, we are particularly interested on the problem of estimating a model using data generated from an LPV system with static dependence, which is however represented on a different state-basis from the one considered by the estimator. © 2016 IEEE.