Research On The Test Method Of Collision Prevention Capability Of Intelligent Connected Vehicles Based On Countermeasures

With the vigorous development of the intelligent connected vehicle industry and the popularization of advanced auxiliary driving functions of vehicles,testing the safety and reliability of intelligent connected vehicles has become an indispensable link in the process of industrial landing.Especially when the intelligent connected vehicle encounters the surrounding background vehicles to carry out "confrontational" and "aggressive" behaviors against the vehicle,such as malicious congestion,deliberate anger,etc.At this time,the response strategy of the intelligent connected vehicle should be to avoid collisions as much as possible,so the most important basic ability that intelligent connected vehicle should have is the anti-collision ability in the face of danger.Therefore,how to conduct special tests and improve the anti-collision ability of intelligent connected vehicles has become an important issue that needs to be solved urgently in the test and evaluation of intelligent connected vehicles.This paper mainly elaborates the anti-collision capability test simulation platform of intelligent connected vehicles based on the countermeasures of background vehicles.This paper then gives a detailed introduction to the design of the main vehicle’s DQN decision-making algorithm,the simplified modeling and construction of the simulation environment,the background vehicle control strategy,the anti-collision capability test experimental setting,and the influence of the background vehicle control strategy on the main vehicle training.This paper also selected specific scenarios in the vehicle anti-collision ability simulation test as the entry point,designed a system implementation that can test the basic anti-collision ability of intelligent connected vehicles in a limited space,and implemented simple experiments to verify verify the practicality and implementability of the scheme.The main work of the paper is as follows:First,this research constructed a simulation test platform for highway scenarios.This project selects the highway straight scene as a typical case,uses the Pygame module to build a simulation environment and uses the brief representation of the traffic elements as the simulation of the basic environmental elements.This paper establishes a simplified human driver decision-making model,and designs the DQN algorithm as the main vehicle decision-making module to realize the function of the simplified driver model.Secondly,this paper uses different control strategies of the background vehicle as variables to achieve simplified simulation of the real scene,and then designs the simulation verification experiment according to the control strategy of the background vehicle of different difficulties.The simulation experiment is to load the background vehicle control logic of different difficulties into the environment one by one,record the main vehicle decision model parameters trained in the environment,and then manually load the obtained main vehicle decision model parameters to the untrained blank environment and record experiment results.Through the analysis of the experimental results,it is proved that the training effect of the main vehicle’s anti-collision ability increases with the increase of the running difficulty of the background vehicle,and the higher the difficulty,the more obvious the improvement effect.Finally,as an extension of the virtual simulation test,this project also designed the realization of an intelligent connected vehicle anti-collision capability system in a limited space.The implementation of the system verifies the feasibility of the key basic scenarios in the interaction between the main car and the background car in the virtual simulation experiment in the real car test.