V2V-based Cooperative Collision Avoidance Considering Positioning Accuracy And Communication Quality

The vigorous development and application of intelligent network traffic provide favorable conditions for vehicles to avoid collision.In the intelligent network environment,it can expand drivers' blind areas and enhance drivers' vision through V2 V,V2R and other information interaction,and providing basic information of warning of collision avoidance for vehicles.Therefore,the research on V2 V cooperative warning of collision avoidance in intelligent connected environment has become the focus of current research.Based on the vehicle collision warning model of collision probability,this thesis fully considers the influence of vehicle position accuracy and communication quality on V2 V collision prediction model under the intelligent network environment,that is,it constructs a V2 V cooperative collision avoidance model considering position accuracy and communication quality.The influence of vehicle positioning and communication quality on warning model of collision avoidance can be reduced by vehicle position combined filtering and vehicle position estimation when communication is lost.The effects of position accuracy and communication delay are fully considered in the vehicle warning model of collision avoidance,and the model has an effective warning rate of 97.01%,so as to improve the performance of the warning model and provide theoretical basis for the research of the vehicle warning model of collision avoidance in the intelligent network environment.The main research contents and conclusions of this thesis are as follows:Firstly,the Kalman-Gaussian joint filter model is constructed to filter.The dynamic threshold is used to identify vehicle position drift points,and GPR(Gaussian Process Regression)is used to predict vehicle position,the predicted value and the real observation value are used to construct observation compensation,and the vehicle dynamic filtering is realized by adding compensation value into Kalman observation equation.The vehicle positioning error after combined filtering can be stabilized within 1m,providing a more reliable and stable positioning for the vehicle.Secondly,according to analysis of the vehicle communication quality under the intelligent network environment,and the communication delay and packet loss rate are taken as evaluation indexes to analyze the influence of different driving environments on the communication quality.It is concluded that the communication delay is about 10 ms and the packet loss will change with the change of the environment.According to the situation of packet loss in the communication process,the target vehicle position estimation based on dead reckoning is used to make up for the loss of target vehicle position and reduce the impact on the warning model.Finally,the warning model considering position accuracy and communication quality is established.The vehicle collision probability is obtained by integrating the probability of the vehicle's predicted position in the collision zone at the intersection,the correct alarm rate is 97.01%,the false alarm rate is 1.49%,and the missing alarm rate is 2.99%.The performance of the warning model is verified by hardware in-loop simulation experiment.In this thesis,the establishment of vehicle-vehicle cooperative collision avoidance prediction model considering position accuracy and communication quality under the intelligent network environment,and the establishment of hardware in the loop simulation experiment platform to simulate and verify the prediction model,this prediction model can better ensure the driving safety,and also provide a theoretical basis for the research of vehicle warning of collision avoidance under the intelligent network environment.