Accurate Modelling and Identification of Vehicle's Nonlinear Lateral Dynamics
Abstract
The use of model-based controller or fault diagnosis systems inautomotive control requires reliable dynamics models. Thereliability consists in accurate prediction of the vehicle'sbehavior, in order to evaluate current driving situations. Modelparameters have to be precisely identified. This depends on themodel's capability to reproduce the complex dynamics of the realsystem, by remaining as simple as possible, for onlinecomputation. In this article, a practical approach of modellingvehicle dynamics is presented. Conventional simplifications areremoved by coupling lateral and roll motion. Besides,nonlineartire characteristic is presented. The accuracy and performance ofthe identified model is demonstrated by tests on real automotivesystems.
