Model predictive torque vectoring control with active trail-braking for electric vehicles

Abstract

In this work we present the development of a torque vectoring controller for electric vehicles. The proposed controller distributes drive/brake torque between the four wheels to achieve the desired handling response and, in addition, intervenes in the longitudinal dynamics in cases where the turning radius demand is infeasible at the speed at which the vehicle is traveling. The proposed controller is designed in both the Linear and Nonlinear Model Predictive Control framework, which have shown great promise for real time implementation the last decades. Hence, we compare both controllers and observe their ability to behave under critical nonlinearities of the vehicle dynamics in limit handling conditions and constraints from the actuators and tyre-road interaction. We implement the controllers in a realistic, high fidelity simulation environment to demonstrate their performance using CarMaker and Simulink.