Nonlinear Identification of vehicle's coupled lateral and roll dynamics

verfasst von
H. Abdellatif, Bodo Heimann, Jan Martin Hoffmann
Abstract

The implementation and design of modern car dynamics controller or driver assistance systems require suitable as well as high accurate dynamics models. The parameters of such models are generally unknown and have to be identified and used to design the controller systems. A vehicle's model is presented, by removing conventional simplification and coupling the lateral and roll car dynamics. The resulting model is high nonlinear in its parameters, thus the methods of nonlinear optimization are used for its identification. Nonlinear optimization with nonlinear least square approach and based on the trust region methods has been successfully implemented. The sensitivity analysis in frequency domain is used to improve the convergence of the algorithms. The application of the developed technique on experimental data demonstrates the accuracy of the considered model, as well as the successful implementation of the identification method. The real car dynamics could been reconstructed with high accuracy, for a wide range of steering frequencies and vehicle's speed.

Organisationseinheit(en)
Institut für Mechatronische Systeme
Externe Organisation(en)
Karlsruher Institut für Technologie (KIT)
Typ
Aufsatz in Konferenzband
Publikationsdatum
2003
Publikationsstatus
Veröffentlicht
Peer-reviewed
Ja