Subspace identification of Hammerstein systems with nonparametric input backlash and switch nonlinearities

Mathieu Pouliquen (1, 2) , Fouad Giri (2) , Olivier Gehan (2) , Eric Pigeon (2) , Miloud Frikel (2) , Boubekeur Targui (2)

1 LAC - Laboratoire d'automatique de Caen
2 Equipe Automatique - Laboratoire GREYC - UMR6072

In this paper, a subspace identification algorithm for a class of Hammerstein systems is developed. We consider dynamical systems subject to input backlash or switch nonlinearities. The idea is to use a specific input signal allowing the estimation of the nonlinear part and the estimation of a state space model for the linear part. The identification algorithm is a subspace type algorithm. A simulation example is given to illustrate the performances of the present method.

Mots clés: Switches; Estimation; Approximation methods; Observability; Heuristic algorithms; Mathematical model; Robustness; Control nonlinearities; Identification; Nonlinear estimation; Nonlinear systems; State-space methods

Domaines: Automatique / Robotique

Mathieu Pouliquen, Fouad Giri, Olivier Gehan, Eric Pigeon, Miloud Frikel, et al.. Subspace identification of Hammerstein systems with nonparametric input backlash and switch nonlinearities. 52nd IEEE Conference on Decision and Control, Dec 2013, Firenze, Italy. pp.4302-4307, ⟨10.1109/CDC.2013.6760551⟩. ⟨hal-00965073⟩