Research And Implementation Of Regional Intelligent Vehicle Risk Analysis System

With the continuous maturity of communication technology and software technology,the development of Internet of vehicles is also more and more rapid.Similarly,the risk of Internet of vehicles platform has attracted more and more people’s attention.Based on the platform of Internet of vehicles,this paper uses big data collection and analysis technology and the corresponding risk analysis model,and combines with the actual background of Internet of vehicles,based on the analysis criteria in the white paper of intelligent Internet of vehicles product test and evaluation,establishes a set of system for analyzing and evaluating the risk information of different objects in the platform of Internet of vehicles.The whole system includes vehicle terminal and service terminal.The vehicle terminal includes mirror network,data acquisition module,message middleware and data transmission module.The collected data includes driver’s unsafe driving behavior,road traffic signs and vehicle driving status collected through mirror network.The service terminal is responsible for storing and analyzing the data transmitted from vehicle terminal,It mainly includes five levels of risk division based on LEC Method,three levels of risk division based on AHP method and big data portrait based on k-means algorithm.The system includes the following aspects:1.The transmission subsystem of vehicle terminal.The vehicle terminal collects driver’s unsafe driving behavior,traffic signs,vehicle driving status and other information through the mirror network,and sends the data to the server through the message middleware and data transmission module.2.Server data analysis subsystem.The subsystem carries out data analysis on the data sent by the vehicle terminal.The analysis contents include: evaluating the driver’s percentage risk score,driver’s 5-level risk behavior division,regional 3-level risk division,evaluating the vehicle safety risk based on the analysis criteria of "white paper on test and evaluation of intelligent connected vehicle products",and forming risk descriptions for different objects according to the above information.3.K-means is used to classify the labels and big data portraits of drivers and regional data objects,and to generate the risk characteristic labels of different objects in the Internet of vehicles,such as the label description of drivers: easily distracted,driving with mobile phones and other risk labels;Label description of the area: risk label of more accidents,poor network,road congestion,etc.4.Distributed architecture based on containerization.The different functions of the server are abstracted into independent business modules,which are deployed and operated through containers to ensure the high availability and high stability of the system,as well as the scalability of the system.