The Evaluation Method For Surrounding Situation Of Intelligent Vehicle

In recent years,with the increasing number of cars,the traffic safety problems are becoming more and more serious.In order to solve this problem,the vehicle industry is developing towards the direction of more intelligent.If the vehicle can make a prejudgment of the surrounding safety situation of the vehicle ahead of time,so that it can make reasonable decisions and actions,so as to avoid the occurrence of traffic accidents.Therefore,this paper use recurrent neural network to predict the surrounding vehicle's trajectory and build the safety data field around the vehicle.According to the trajectory and safety data field,we propose a method to evaluate the safety situation of the surrounding vehicle.First of all,the VISSIM simulation software is used to build the traffic simulation model of the three lane highway and obtain a lot of vehicle simulation data.Secondly,the sample data that can be used for surrounding vehicle trajectory prediction is extracted from the original data which obtained from the simulation by MATLAB programming.Then,according to the position of the vehicle has the characteristics of context,we build the recurrent neural network prediction system taking vehicle historical relative location sequence point as input and vehicle future relative location sequence point as output using PyTorch deep learning framework,and use the acquired sample data to train and verify recurrent neural network.Finally,the Hadamard product operation is used to fuse the vehicle lateral safety data field and the longitudinal safe data field to form the security data field around the vehicle.According to the prediction trajectory of the surrounding vehicle,the safety data fields of the surrounding vehicles superposition with the safety data field of the intelligent vehicle,respectively.The trend of the safety data field around the intelligent vehicle is obtained,and then use the evaluating indicator to evaluate the surrounding safety situation of the vehicle.The verification shows that the intelligent vehicle peripheral security situation evaluation method proposed in this paper can well determine the security situation of the intelligent vehicle and provide an important basis for the path planning of intelligent vehicle.