Successive Linearization Based Predictive Vehicle Torque Vectoring

Abstract

In this paper, a linear time-varying model predictive torque vectoring controller is proposed. The controller is based on the non-linear two-track model of the vehicle and TMEasy tire model with neglected load transfer. The model is linearized along the trajectory over the prediction horizon. The steering angle is provided by the driver while the model predictive controller (MPC) calculates the optimal torques on each wheel. In order to test if the controller improves reference tracking while fulfilling given constraints, tests using the sine with dwell maneuver are performed and the results are compared to an MPC with linear tire model and to the passive vehicle behavior.