Autonomous vehicle longitudinal following control based on model predictive control

摘要

In this paper, an approach to control the longitudinal motion as well as reject disturbance of a platoon of autonomous vehicles is presented. Firstly, we consider the longitudinal acceleration and velocity of the leading vehicle as disturbance, and establish a model which describes the longitudinal dynamics of inter-vehicle. Secondly, constant time headway (CTH) strategy is adopted to design a longitudinal upper controller based on model predictive control (MPC) to obtain the desired longitudinal acceleration. For rejecting unknown disturbance, a major contribution that differs this paper from previous research is that the upper level controller has two components: a nominal control action and an ancillary control law. The lower level controller adopts simple PID algorithm to determine either a throttle or brake input. Finally, simulations are carried out using vehicle dynamics software veDYNA to verify the effectiveness of the proposed method. Simulation results show that the controller is able to follow the desired spacing as well as reduce the fluctuations of the leading vehicle.