Feedforward control approach for all-wheel drive super cars with torque vectoring capabilities

Abstract

To improve their performances, modern sport cars can use Torque Vectoring managed by a control allocation system to generate a desired yaw moment. The goal here is to fit a reference yaw moment calculated through a model and find the torques to apply on each wheel to achieve it. However, drivers will not accept a synthetic driving feeling that can occurs with a feedback control. This is why this project investigates a model-based and feed-forward controller that enables to manage the whole vehicle dynamics, to reach a desired behaviour. The work presented here is based on a control TV management strategy first developed for Porsche. It aims at adapting this strategy to a hybrid AWD vehicle with the rear axle powered by an ICE and two electric motors powering the front wheels. A double track vehicle model is used to emulate the car’s dynamics, associated to a Pacejka’s Tyre model. Starting with the algorithm’s architecture, all the equations behind each step have been found. Therefore, a generical algorithm has been created to identify the process and the requirements for a future implementation in a Model-In-the-Loop developed by Mclaren. An optimizing function has been created to find the optimal longitudinal distribution while the whole process adjusts the TV moment to reach a reference yaw rate. Then, the complete algorithm has been implemented on Matlab/Simulink as a first try. However, a misunderstanding of one of the algorithm’s step has not enabled to obtain satisfying results.