Cooperative lateral vehicle guidance control for automated vehicles with Steer-by-Wire systems

Abstract

With the global trend towards automated driving, fault-tolerant onboard power supply systems are introduced into modern vehicles and the level of driving automation is continuously increasing. These advancements contribute to the applicability of Steer-by-Wire systems and the development of automated lateral vehicle guidance control functions. For the market acceptance of automated driving, the lateral vehicle guidance control function must hereby be cooperative, that is it must accept driver interventions. Existing approaches for automated lateral vehicle guidance commonly do not consider driver interventions. If unconsidered in the control loop, the driver intervention is interpreted as an external disturbance that is actively compensated by feedback. This thesis addresses the development of a cooperative lateral vehicle guidance control concept, which enables a true coexistence between manual steering control by the driver and automated steering control. To this end, the subordinate controls of the Steer-by-Wire system for the manual and automated driving mode are initially presented. These include the steering feel generation and steering torque control of the Steer-by-Wire Handwheel Actuator for the manual driving mode, which is structurally extended to a cascade steering position control for the automated driving mode. Subsequently, a superposition control is introduced, which fuses steering torque and position control. The resulting cooperative Handwheel Actuator control achieves precise tracking of the reference steering position in automated driving mode but accepts driver interventions. Thus, the driver can override the active control and experiences a natural steering feel. The transitions hereby are seamless as no blending, gain scheduling or controller output saturation is required. Subsequently, the superimposed lateral vehicle guidance controller for the automated driving mode is described, which computes the reference steering position for the respective Steer-by-Wire controls. In contrast to existing approaches, the plant model equations are rearranged to isolate the vehicle speed dependent dynamics. Thereafter, the concept of inverse nonlinearity control is employed, using a virtual control loop and feedback linearization for an online inversion of the nonlinear plant dynamics. The remaining plant is fully linear and independent of vehicle speed. Consequently, one controller can be synthesized that is valid for all vehicle speeds. The closed and open loop system thereby have the same dynamics independent of vehicle speed, which significantly simplifies control synthesis, analysis, and performance tuning in the vehicle. For considering the future reference path information and constraints on the maximum steering position within the control law, a linear Model Predictive Controller synthesis is selected. The combination of inverse nonlinearity control and linear Model Predictive Controller thus results in a Nonlinear Adaptive Model Predictive Control concept, which makes commonly applied gain scheduling fully obsolete. The controller is structurally extended by a cooperative dynamic feedforward control for considering driver interventions within the control loop. Consequently, the driver can override the active control and seamlessly modify the lateral vehicle motion. A variety of nonlinear simulation analyses and real vehicle tests demonstrate the effectiveness of the proposed control concept.