Study And Implementation Of Vehicle Collision Warning Of Vehicle Infrastructure Cooperative Systems Based On V2X

At present,the number of car ownership is increasing,and the road traffic network is becoming more and more complicated.In complex traffic scenes,traffic accidents are easily caused by the driver's sight being affected or the speed and safety distance controlled improperly.The anti-collision warning technology based on radar and machine vision has obvious warning effect in the line of sight range,but the anti-collision warning effect in the over-the-horizon range is weak and vulnerable to the environment and road conditions.The development of V2X vehicle networking technology effectively realizes the perception and re-cognition of vehicles in the over-the-horizon range,and provides a technical basis for vehicle collision avoidance warning based on V2X communication.Based on the research and analysis of V2X vehicle network anti-collision warning,this paper designs a vehicle-based cooperative anti-collision warning scheme based on V2X communication technology.The scheme design improves the safety distance model and hidden markov model to realize the anti-collision warning function in different scenarios.Aiming at the problem that the traditional safety distance model does not reasonably quantify the driver's reaction time,eliminate the brake gap time and“looming effect”parameters,and make the risk judgment accuracy low,the paper proposes the modified safety distance model to improve the model accuracy.Based on the traditional safety distance model,the modified safety distance model introduces statistical values to quantify the driver's reaction time and eliminate the brake gap time.The parameter_cT quantifies the“looming effect”to optimize the model and simultaneously use the real time of exchanging data through V2X communication to improve the accuracy of the model effectively.The modified safety distance model predicts vehicle road and vehicle reminder scenes with high accuracy and low time delay,which can significantly improve driving safety.Aiming at the problem that the accuracy of the dangerous situation is only determined by the motion model in the vehicle following scene,the paper proposes the modified hidden markov model,analyse the driver's driving behavior data to train model parameters,through the K-means algorithm to process abd discretize the data of vehicle,then train the model parameters by the supervised learning method.At the same time,the paper through the modified safety distance model to mark the state data with dangerous state,which improves the real-time and accuracy of Viterbi prediction.The implementation of anti-collision warning of vehicle infrastructure cooperative systems adopts the modular design idea,according to the system function to divided the system into communication module,data module and anti-collision warning judgment module and adopts concurrently computation in the Data acquisition,data processing and synchronization,anti-collision warning service processing and synchronization.The implementation enhances the real-time performance of the system and the reliability of judgment fully.In order to verify the feasibility of the algorithm,the paper built a simulation verification platform consisting of IPG CarMaker simulation platform,Simulink tool and vehicle terminal equipment for testing and verification,and then verified the field in some scenarios under the condition of ensuring traffic safety and the implementation of the anti-collision warning function.