Short-term Prediction Study Of Accidents On High-frequency Driving Routes Of Big Trucks

In China,the road traffic accidents caused by lorries occur frequently,and such accidents cause grievous effect.If the probability of accidents on the route of lorries can be predicted,the subsequent control would effectively reduce the probability of road traffic accidents of lorries.The predication has significant importance to improve road safety and reduce the loss of accidents.Various collection methods of real-time traffic data facilitate achieving short-term warning of traffic accidents.Based on the real-time traffic data and the historical accidents information,the short-term accident prediction models on the route,using BP neural network algorithm,support vector machine and random forest algorithm respectively.The proposed model achieving short-term warning of potential traffic accidents where lorries route through frequently,which provides the ground for traffic administration to prevent traffic accident.This paper mainly focuses on the following aspects:(1)This paper analyzes the accident characteristics and the characteristics of high-frequency routes where lorries route through;this paper combines the main influencing factors of road traffic accidents to obtain the key indicators for the short-term accident prediction of high-frequency routes where lorries route through;The index of lorries used to measure the occupation of road by lorries is proposed innovatively;(2)In this paper,a set of data preprocessing methods for the characteristics of road traffic accident data and traffic flow data is designed.Combined with the characteristics of actual data,the characteristic variables required for the prediction research are selected.(3)Based on BP Neural Network algorithm,Support Vector Machine and Random Forest algorithm,this paper constructs three short-term accident prediction model for high-frequency driving routes of large trucks,and combines the actual data of a high-frequency driving route of a large truck to predict short-term accidents.The conclusion that the prediction result of the short-term accident prediction model of the high-frequency driving route of large trucks based on the random forest algorithm is more accurate is obtained.(4)In this paper,the short-term accident prediction model of high-frequency driving route of large trucks based on random forest algorithm is used to analyze the relationship between the two characteristic variables with the largest weight of the model and the predicted probability of accident occurrence.The conclusion that the probability of an accident occurring when the large truck index exceeds 50 and the speed standard deviation exceeds 30 is greatly increased.