Research On Traffic Safety Assessment And Early Warning Management Of Highway On Rainy Days Based On Speed Prediction

With the rapid expansion and improvement of the expressway network,traffic safety issues have become more and more prominent,especially in the case of rainy days,the highway traffic accident rate is higher.In order to ensure the safety of highways and reduce the casualties and economic losses caused by traffic accidents on rainy days,the traffic safety assessment of highway on rainy days based on speed prediction and early warning management method is proposed to evaluate and warn the traffic risks in the short-term.Thereby reducing the risk of driving on rainy days,providing new ideas and references for the traffic management department to the safety management of the highway on rainy days.Firstly,the method of mathematical statistics is used to analyze the speed characteristics of highway and the distribution of traffic accidents on rainy days and analyze the impact of rainfall on highway traffic safety from the perspectives of "people,vehicles and roads".It provides reference and theoretical support for the selection of speed prediction model,the selection of safety assessment indicators for the highway on rainy days and management measures for early warning system.Secondly,based on the principle and advantages of long short-term memory(LSTM)and gated recurrent unit(GRU)networks,the LG(LSTM-GRU)deep learning combination prediction model is proposed to realize the short-term speed prediction of highways on rainy days.The L-G model can better adapt to the complexity,uncertainty and abrupt change of the speed on the rainy day,to achieve higher prediction accuracy and stable prediction effect.At the same time,the feature selection method based on the maximum information coefficient is proposed,and the features which have strong correlation with velocity are selected as the input to improve the prediction accuracy of the model and reduce the over-fitting.Finally,this paper establishes a system for safety assessment and early warning management for highway on rainy days,selects speed,weather and road conditions as evaluation indicators to construct a highway traffic safety assessment index system on rainy days,and applies improved matter-element analysis method to assess the safety of highway on rainy days.According to the results of safety assessment,this paper determines the risk level of driving and formulate traffic control measures and strategies for early warning management.Taking the Jinkou-Anshan section of the Beijing-Hong Kong-Macao Expressway in Hubei Province as an example,the rationality and feasibility of the proposed safety assessment and early warning management method for the highway were verified.There are 31 figures,17 tables and 85 references in this paper.