Model-based integrated control with vehicle states and parameters estimation

Abstract

The focus of this paper is the identification of a nonlinear full vehicle model. The identification task is realized as a weighted nonlinear least squares optimization problem, which minimizes the 2-norm of the errors between the power spectral density (PSD) of the measured sensor signals and that of the simulated sensor signals. Subsequently the parameterized vehicle model is reduced to a quarter vehicle model. This model is then utilized during the synthesis of an LPV controller employing semi-active dampers. Results from a four-post-test-rig showing the correlation of the identified full-vehicle model to the real vehicle are given at the end of the paper, followed by those from simulations and a road test demonstrating the performance of the LPV controlled suspension compared to passive suspension set-ups.