Real-time Traffic Safety Evaluation Method For Freeway Under Foggy Environment

Fog can cause a significant drop in visibility,which can affect highway traffic safety.Therefore,it is of great significance to ensure the safety of driving under foggy environment that researching on the traffic safety of highway under foggy environment.In this paper,the factors that affect the real-time traffic safety condition of highway under foggy environment were analyzed,and the method and model of real-time traffic safety assessment under foggy environment were studied,then the real-time traffic safety classification and safety warning under foggy environment has been established,which provides the foundation for traffic management and variable speed limit control.In order to evaluate the real-time safety status of highway and achieve safety warning under foggy environment,the VISSIM software was used to establish the highway network under foggy environment,and simulate the highway traffic flow under different visibility conditions based on the actual traffic data under foggy environment.Traffic conflicts were selected as indicator of highway real-time safety status,and the main factors influencing real-time traffic safety status were identified using random forest and Pearson coefficient.The calculation model of real-time traffic conflicts was established by regression analysis.The traffic conflicts are used as an index to express the real-time safety status of highway.The traffic conflicts are divided into 3 categories by using the K means clustering method,and thereby the traffic safety states are classified into 3 categories: high risk,medium risk and safety state.According to the classification of traffic safety status,an early warning method is established.After the judgment of the current traffic risk level,the high risk state is warned,and the driver is warned in time to reduce the accident risk.The traffic flow data were calculated by the traffic conflicts regression model under accident and non accident conditions,to sums up the fluctuation of traffic conflicts in the case of and without traffic accidents,and concludes the traffic conflict status which is easy to cause traffic accidents.By setting different traffic accident threshold,the accident identification method can get different accident prediction accuracy and false positive rate.Using traffic flow,accidents and meteorological data of I-43 highway in USA,this paper validates real-time traffic safety state evaluation model and warning method under foggy environment.The results show that the traffic safety classification and early warning method can be used to identify the dangerous traffic flow and publish corresponding warning under foggy environment in this paper.According to different of the traffic accident thresholds,we can get different accident prediction accuracy and false alarm rate.When the traffic conflicts 18 was chosen as the accident discrimination threshold,the prediction accuracy of the accident model is 73.33%,the false alarm rate is less than 40%,shows that accident discrimination method has high accuracy,can be used to evaluate the real-time traffic safety status in real time under foggy environment.The fog state classification of traffic safety and early warning method established can be used to identify the fog freeway traffic state and real-time warning,provide a theoretical basis for the foggy environment of highway traffic safety management in real time,variable speed limit control.The model requires higher sampling cycle,sampling distance of traffic flow data and completeness of accident data,however,the actual highway traffic flow data and meteorological data can not be obtained enough,because the traffic flow data acquisition system and meteorological monitoring facilities are not perfected in China.Therefore,the traffic safety state evaluation model established in this paper lacks sufficient data validation.To obtain the accident,traffic flow and meteorological data of the actual highway in China,the highway equipment will be installed on the highway in further study,to modified the traffic conflict model,and carry out practical application.