Torque Vectoring for Lane-Changing Control during Steering Failures in Autonomous Commercial Vehicles

Abstract

<div class="section abstract"><div class="htmlview paragraph">Lane changing is an essential action in commercial vehicles to prevent collisions. However, steering system malfunctions significantly escalate the risk of head-on collisions. With the advancement of intelligent chassis control technologies, some autonomous commercial vehicles are now equipped with a four-wheel independent braking system. This article develops a lane-changing control strategy during steering failures using torque vectoring through brake allocation. The boundaries of lane-changing capabilities under different speeds via brake allocation are also investigated, offering valuable insights for driving safety during emergency evasions when the steering system fails. Firstly, a dual-track vehicle dynamics model is established, considering the non-linearity of the tires. A quintic polynomial approach is employed for lane-changing trajectory planning. Secondly, a hierarchical controller is designed. The upper layer employs a three-stage cascaded proportional integral controller to determine the total yaw moment required for lane changing, considering the influence of lateral tire forces on brake allocation. The middle layer uses constraint optimization to manage braking force distribution among the four wheels. The lower layer's actuator generates brake torque through brake cylinder pressurization. Finally, the effectiveness and feasibility of the control strategy are validated using joint simulations on Matlab/Simulink and Trucksim over diverse longitudinal distances. Simulation results indicate that autonomous commercial vehicles can execute swift and safe lane changes at varying speeds during steering failures.</div></div>