Vehicle Longitudinal Stochastic Control for Connected and Automated Vehicle Platooning in Highway Systems

Abstract

The vehicle platoon control for highway traffic can help to improve traffic flow efficiency, enhance traffic safety, and reduce fuel consumption. In previous platoon control research, most of the driving behaviors are described using deterministic car-following models. Nevertheless, random factors that may come from the vehicle power train and additional stimuli, can have a greater impact on platoon control in highway traffic. In this paper, a novel stochastic control method is proposed based on a stochastic car-following model which considers microscopic driving behavior. Firstly, the stochastic car-following model is designed that fully accounts for the impact of random factors on platoon control. Secondly, an optimal objective function is constructed and the Hamilton-Jacobi-Bellman equation is used to solve this stochastic control problem, thereby completing the upper-level controller design and obtaining the optimal desired acceleration of the vehicle. Thirdly, the stochastic stability method is applied to analyze the proposed model and obtain the stochastic stability conditions satisfied by the model. Finally, tests are conducted for three different scenarios: stable speed, acceleration, and deceleration with additive and multiplicative noise, as well as the case where the lead vehicle’s speed is based on real vehicle trajectory data. These tests validate the stability and effectiveness of the stochastic car-following model predictive control method from the perspective of control strategy and model respectively. The experimental results show that under the stable parameter conditions of the model, the connected and automated vehicle platoon can achieve accurate speed tracking and maintain an appropriate safe distance in highway system.