A Modified Vehicle Following Control System on the Curved Road Based on Model Predictive Control

In this paper, a decoupled control framework is developed to fulfill the automated vehicle following. Human driver's habits on curved road are incorporated into controller. Firstly, the overall structure of vehicle following control system is given. Secondly, in the upper-level controller the motion relation of vehicle longitudinal is set up by Preview G-Vectoring Control (PGVC). Hence, the acceleration of host vehicle is adjusted according to the variety of road contour during vehicle-following. Then, an acceleration-predictor of preceding vehicle in prediction horizon is also incorporated into MPC structure. Moreover, a practical lower-level controller for acceleration follow is proposed respectively. Finally, simulation results demonstrate the control framework has good performance of vehicle-following ability on the curved road, and its velocity characteristic is more similar with human's driving behavior.