LQR-UKF active comfort control of passenger vehicles with uncertain dynamics

Abstract

This study considers the passenger comfort problem in modern sedan vehicles and presents a novel LQR framework that treats the induced vibration attenuation of the chassis frame, under the assumption of uncertain vehicle dynamics. In contrast to the conventional LQR controller, where the gain matrix is precalculated offline on the basis of an associated cost function, the presence of uncertainty calls for a formulation that is able to handle its propagation into the solution of the optimal control problem. To this purpose, a novel scheme is presented, which combines the use of the unscented Kalman filter with a modified recursive solution of the Riccati equation that allows calculation of the control gain in hard real–time. The numerical results confirm the efficacy of the proposed approach and call for further investigation, especially in terms of the technological aspects of its actual implementation.