Research On The Control Technology Of Multi-mode CACC System Of Vehicle Based On V2X Environment

With the development of Internet of vehicles communication technology and intelligent vehicle technology,vehicle cooperative adaptive cruise technology is gradually developed.This paper studies the adaptive cruise control technology of vehicles in the vehicle network environment based on the hardware-in-the-loop test bench as the research platform.Traditional adaptive cruise(ACC)technology relies on millimeter wave radar and other vehicle sensors to collect the information of obstacles or vehicles in front.In this paper,the vehicle is in the V2 X environment,and it directly relies on the V2 X network to achieve information transmission,which is a cooperative adaptive cruise control(CACC).Compared with the traditional mode,cooperative adaptive cruise has the advantages of strong reliability,fast data receiving rate,and so on.Therefore,the vehicle has the advantages of better safety and fuel economy.In terms of control system architecture,cooperative adaptive cruise control system is similar to adaptive cruise control system.At present,the mainstream solution is to use hierarchical structure,that is,there are both upper controller and lower controller.The upper controller outputs the target acceleration value according to the distance between vehicles,the relative speed of two vehicles,the relative acceleration of two vehicles and other information,and the lower controller controls the actuator of the vehicle to accelerate or brake according to the target acceleration value.Because the lower controller algorithm has strong applicability,fuzzy-PID algorithm is widely used,so the core of the research is the upper control algorithm.In this paper,the upper controller adopts deep reinforcement learning algorithm,which outputs the target acceleration according to the set reward and punishment function and the result of vehicle environment interaction.After consulting the domestic and foreign literature on Cooperative adaptive cruise and analyzing the key technologies and methods of cooperative adaptive cruise control system,this paper divides the car following mode into seven modes: constant speed cruise,steady-state follow,approaching the car in front,acceleration,deceleration,collision avoidance and signal lights mode.The division basis of the seven modes and the weight coefficient of the evaluation index under each mode are clarified,and the optimization is mainly carried out for the safety,fuel economy and comfort of the vehicle.In order to verify the feasibility and effectiveness of the algorithm,the vehicle dynamics simulation software Car Sim and the deep reinforcement learning software tensorflow and python are used for CO simulation,and the experiment is carried out on the hardware platform.