Predictive control of vehicle roll dynamics with rear wheel steering

A pair of model predictive controllers (MPCs) capable of modifying the nominal roll dynamics of a vehicle through control of the planar vehicle dynamics are presented. Each of these controllers is based on a linear model of the vehicle. One controller utilizes differential drive of the rear axle to modify the planar motion, while the other uses rear-wheel-steering (RWS). Results from simulations of a nonlinear vehicle model executing maneuvers with each of the controllers demonstrate that the differential-drive technique results in significant lateral-longitudinal tire force coupling and saturation that degrades the validity of the internal model used for MPC. The RWS controller commands very small steering angles, retaining the validity of the model and showing better performance for the desired roll control task.