Road Traffic Risk Evaluation Method Based On Multi-source Information

With the rapid development of China’s social economy,the number of traffic accidents caused by the increase of car ownership and traffic travel demand remains high,resulting in huge casualties and economic losses.Accurate evaluation of road traffic risks can not only provide effective information for residents’ immediate travel to avoid risks,but also provide theoretical support for traffic management departments’ risk warning,prevention and control and governance.The existing road traffic risk evaluation is mostly based on historical accident data to analyze the impact of various risk-causing factors on the accident,however,the actual traffic risk exists in multiple risk sources of human-vehicle-road-environment,and each risk source has numerous influencing factors,these static risk factors have an important impact on the safety of vehicle operation;at the same time,dynamic traffic flow information(such as rapid acceleration behavior,rapid deceleration behavior,rapid lane change At the same time,dynamic traffic flow information(such as rapid acceleration behavior,rapid deceleration behavior,rapid lane change behavior,etc.)also plays a role in road traffic risk,so the comprehensive consideration of multiple sources of road traffic risk information,the construction of road traffic risk evaluation method is to realize the premise of road traffic risk active prevention and control.The paper analyzes the characteristics of traffic accident and traffic violation data from four aspects of humanvehicle-road-environment respectively,and establishes the framework and method of road traffic risk evaluation based on static information;at the same time,it considers the influence of abnormal driving behavior on road traffic risk,and constructs a comprehensive evaluation method of road traffic risk considering multi-source information.The research in this paper will provide technical support for the evaluation of road risk status and risk prevention and management.The paper firstly analyzes the data characteristics of different accidents and illegal behaviors in the data of traffic accidents and traffic violations from four types of risk sources: drivers,vehicles,roads and environment,completes the multidimensional analysis of traffic risk causation,selects various risk source evaluation indexes,and proposes the multidimensional risk source evaluation index framework;then uses the combination of hierarchical analysis and entropy weighting method to determine the weights of each evaluation index factor Then,the combination of hierarchical analysis and entropy weighting method is used to determine the weights of each evaluation index factor,apply the principle of topology,construct a static traffic risk evaluation model based on the topologic element method,and realize the classification of static traffic risk level of the road.This paper combines the data of traffic accidents,traffic violations and abnormal driving behaviors,applies the Delphi method to judge the road traffic risk level,and designs the road traffic risk level based on the Delphi-XGBoost algorithm.Delphi-XGBoost road traffic risk discrimination model is designed and compared with Delphi-RF algorithm and Delphi-SVM algorithm for experiments,and the results show that the Delphi-XGBoost discrimination model proposed in this paper has better performance.