Shared steering control between a human and an automation designed for low curvature road

Abstract

In order to facilitate driving and prevent accidents due to lane departure, this paper focuses on the development of an assistance device for lane keeping of a passenger electric vehicle. The goal is to develop a shared control mode based on the anticipation of risk and the prediction of the most probable actions of the driver. The paper is mainly articulated around two axes: the cybernetic modelling of the driver in his task of lateral control of the vehicle and the design of a shared steering control using a Driver-Vehicle-Road (DVR) model. Since many vehicle/road interaction factors (such as road adhesion, aerodynamic forces) and actuator dynamics are very poorly known, the proposed work addresses shared control resulting from applying the linear-quadratic control (LQR) synthesis to the global model (DVR). Innovative criteria were used for assessing the time to lane crossing, the level of sharing between the driver and lateral assistance, as well as their cooperative or conflicting behaviour. All simulation tests are carried out on a 5.8 km track that has several pretty tight radius turns (up to 150 m). The implementation of LQR controller has been made for a longitudinal speed set at 75 km/h. The LQR is sufficient to keep the vehicle under control on a road with low radius curvature. Preliminary simulation results in MATLAB/Simulink are presented to explain the concept.