MODELING AND IDENTIFICATION OF NON-LINEAR SYSTEMS BY A MULTIMODEL APPROACH : APPLICATION TO A THROTTLE VALVE
Abstract
The control and the diagnosis of non-linear systems require modelling and identification techniques, adapted to numerical computation. This paper deals with such issues. Nonlinearities that are considered are dead zones, hysteresis and saturations. These nonlinearities are simplified according to a multimodel approach that has additive unknown inputs and that divides nonlinearities into several discrete-time linear models. These unknown inputs are used to define commutation conditions from one local model to the other. An automaton is proposed as global model. For identification purpose, we propose an algorithm that uses both classification and linear identification techniques. Finally, this algorithm is applied to the throttle valve of combustion engine.