Design and Verification of Autonomous Steering Control Based on Driver Modeling

Abstract

The paper proposes the design and verification of an autonomous steering control system. The method is based on the control-oriented modeling of the human driver, which results in the comfort of the passengers and provides predictability for the drivers of the conventional vehicles and the pedestrians. The optimization procedure is based on test measurements, which is able to guarantee the steering characteristics of the driver in the autonomous control system. The designed controller is based on the robust $\mathcal{H}_{\infty}$ method, which is extended with a prediction rule. The proposed autonomous steering functionality is demonstrated through real vehicle tests, which show that the control system approximates the steering behavior of the driver accurately.