Research On Intelligent Vehicle Longitudinal Control Strategy Based On Vehicle-Road Information Interaction

Road traffic is a complex system in which multiple traffic participants participate,influence and restrict each other.Under the condition of limited road resources,with the increase of the number of vehicles,the proportion of front and side collision between vehicles keeps increasing,which has become the most common accident type in road traffic accidents.In order to improve driving safety in complex road scenes,this paper considers the impact of other vehicles’ driving conditions on the host vehicle.It focuses on the longitudinal control strategy of intelligent vehicles,based on vehicle road information interaction,from the recognition of environmental vehicle driving patterns and the longitudinal direction of intelligent vehicles.Research has been carried out on the proposal of the control strategy and the design of the longitudinal controller.Firstly,based on the traffic environment data obtained by CVIS,a method of environmental vehicle road pattern recognition based on vehicle hidden Markov is established.The driving state of vehicles in the multi-lane scenario of highway is analyzed,and the driving modes of other vehicles are divided into three types: lane change on the left,lane change on the straight and lane change on the right.By studying the concept of hidden Markov model and classical algorithms,the lateral offset and lateral offset speed of the center of mass of the vehicle relative to the lane center are selected as observation variables,and the driving mode of the vehicle is used as a hidden variable.Based on GMM-HMM(Gaussian Mixture-Hidden Markov Model),the environmental vehicle driving pattern recognition method.The established model is trained by using the NGSIM data set and the model parameters are obtained,and then the forward-backward algorithm is used to identify the driving mode of the environmental vehicle according to the driving state of the environment.Secondly,the types of traffic conflicts that may occur when vehicles change lanes and their safety are analyzed,and the longitudinal conversion safety distance models are established for different following targets,and the vertical control strategy of intelligent vehicles is proposed.The lane change trajectory prediction model of the environment vehicle is established based on the quintic polynomial for the lane change driving mode of the environment vehicle which may affect the longitudinal driving of the intelligent vehicle.According to the types of traffic conflicts between other vehicles and the host vehicle when other vehicles change lanes,the safety analysis of the two main traffic conflict types is carried out,and the minimum longitudinal safe distance is defined.Aiming at the difference of the target that the intelligent vehicle follows when driving longitudinally,two safety distance models are established.The distance between the vehicles in response to the following behavior of the main vehicle is defined,and the longitudinal driving mode of the intelligent vehicle is divided into two types:cruise mode and car-following mode.And on the basis of recognizing the driving mode of environmental vehicles,a longitudinal control strategy of intelligent vehicles is established that considers the state of surrounding vehicles.Thirdly,in order to follow the desired speed and distance accurately,the intelligent vehicle front controller is designed.Aiming at the longitudinal driving mode of intelligent vehicle,the upper controllers of cruise mode and following mode are established respectively based on PID algorithm and MPC algorithm.And the lower level controller is designed using vehicle inverse longitudinal dynamics.Finally,in order to analyze the quality of the longitudinal control strategy and the controller,MATLAB/Simulink and Car Sim are used for simulation to realize the longitudinal control of intelligent vehicles under different working conditions.The cruising condition and the following condition in the scene of other vehicles going straight and lane changing were simulated respectively,and analyze its quality through the simulation data.