Hammerstein model identification with set membership errors

Abstract

The problem of the Hammerstein dynamic system identification is considered when the measurement error is characterized in a set-membership context. The proposed approach accomplishes parameter identification through the introduction of a linearized augmented Hammerstein model whose parameter bounds allow us to derive overbounds to the Hammerstein model parameter uncertainties. The issue of the parameter overbound tightness and the bound convergence for increasing number of measurements are discussed in the general case.