A 4WD drift assist control framework on general rally track

Abstract

In this paper, we present a novel state machine-based 4-Wheel-Drive (4WD) drift assist controller with four drive modes, to allow the driver transition into and out of drifting in corners intuitively. To interpret the driver’s intention to drift, we rely on the driver’s inputs and vehicle motion signals. During drifting, without prior knowledge of path geometry, we introduce a ‘Trajectory Radius Predictor’ to assist the driver stay on the desired course and provide the necessary reference yaw rate for the subsequent torque vectoring controller. Additionally, we restrict the sideslip rate to allow the driver sufficient reaction time to counter steer and follow the desired track when the sideslip changes during drifting. To validate our controller, we consider different driving behaviours from both the CarMaker driver model and a human driver using the driving simulator of Cranfield University. The Pikes Peak rally track in IPG CarMaker is utilised for evaluation, and the results have demonstrated the effectiveness of the proposed framework.