Environment-on-Board Predictive Braking Control Functions for Autonomous Driving During Sudden Changes in the Road Friction Coefficient on Sharp Curves

Abstract

Research Questions/Objective In the context of autonomous and connected vehicle technology, this research aims to develop an environment-on-board predictive braking control system, regulating the vehicle velocity to a desired value, in order to maintain and enhance path-following performance and vehicle stability when a vehicle is driven on a path where the road friction coefficient changes suddenly on a curve. If the vehicle enters a wet road from a dry road and moves on to a dry road at a high velocity, the vehicle suffers unexpected lateral disturbances, thereby causing it to destabilize and leading to deviation from the lane. Therefore, in this research, it is important to find a velocity, which the vehicle can turns on a curve that has an arbitrary curvature radius and wet road friction coefficient without deviating from the lane. Methodology In order to design the predictive braking control, the desired velocity needs to be calculated. By transforming the equation of simplified linear two-degree-of-freedom bicycle model the velocity can be expressed in terms of the curvature radius, road friction coefficient, lateral deviation and each vehicle parameter. Hence, if the definition of lane deviation is that lateral deviation is equal to 0.85 m, substituting its value gives the desired velocity that should be decelerated before entering a wet road. In addition, PID control is used to follow that velocity. The effectiveness of this control system is verified using a nonlinear four-degree-of-freedom four-wheel model and a full vehicle model of the IPG CarMaker under various conditions. Results To verify the effectiveness of the proposed control system, simulations of a vehicle driving through a curve which has a wet road surface are performed at various conditions that the wet road position, wet road friction coefficient and initial velocity differ from each other. Furthermore, the vehicle behavior such as side slip angle and lateral deviation, and the tire workloads are analyzed to evaluate vehicle stability and path-following performance. The results of these analysis show that the vehicle can be stabilize and it can be prevented from deviating from the lane by applying the control which decelerates the vehicle to the desired velocity before entering a wet road. Conclusions Based on the long-horizon road friction and curvature information from infrastructure together with connectivity between intelligent vehicles, the environment-on-board predictive braking control is developed to prevent vehicles from deviating from the lane even when the road friction coefficient changes suddenly on a sharp curve, and its effectiveness is verified through full-vehicle simulation. This system improves the vehicle stability and path-following performance under various conditions, including the position of a wet road.