Research On Highway Traffic Feature Portrait And Risk Assessment Model Based On ETC Gantry Data

After the cancellation of the provincial boundary toll station project,the national expressway network has achieved full coverage of the ETC gantry system,which provides more comprehensive data support for the analysis and risk assessment of highway traffic characteristics.How to dig deep into the data collected by the gantry system to improve road traffic efficiency and reduce the risk of highway traffic accidents has become an urgent problem to be solved.This paper uses data mining technology to study the driving characteristics of highway vehicles,the section traffic operation characteristics,and the evaluation of vehicle driving risks and section operation risk levels under different feature portraits.The main work of the paper is as follows:1.Analyze the basic content and main fields of ETC mast monitoring data,meteorological data,and highway traffic accident data.Preprocess the data.On this basis,integrate the complete gantry trajectory data of a single passage of a bicycle.Calculate the traffic flow data under the ETC gantry at different locations with different time granularity.Obtain spatiotemporal matching with meteorological data and traffic accident data to acquire traffic flow data and traffic accident data including meteorological environment.2.Analyze the vehicle driving characteristics index from the two aspects of vehicle risk driving behavior and highway vehicle inspection.Use the "TOPSIS-gray correlation degree analysis" entropy weight method for weight distribution.Use the weighted average method to calculate the index of different risk driving behavior in a single passage of the vehicle.Calculate the abnormal inspection status index of the vehicle using the probability and statistics method to form the vehicle driving feature portrait.This provides a decision-making basis for the highway management department to identify typical vehicles.3.Use the clustering algorithm based on the HMM model,the fusion Granger causal analysis,and the CCF function algorithm to explore the spatiotemporal distribution characteristics of traffic flow.Divide the traffic states of different characteristics based on the three-phase traffic flow theory to form a section traffic operation feature portrait.Use the Poisson regression model combined with Bayesian parameter estimation to analyze the correlation between traffic operation characteristics and traffic accident frequency.Finally,analyze the portrait of the operation characteristics of the West Hanbei Expressway and the correlation between the traffic operation characteristics and traffic accidents.The analysis results are consistent with the actual situation of the road section.4.Use the weight allocation method combining the AHP method and the "TOPSIS-gray correlation degree analysis" entropy weight method to establish a vehicle driving risk assessment model based on fuzzy comprehensive evaluation.Verify the effectiveness of different vehicle driving feature portraits and other risk indicators on the weight allocation of different driving risk levels.Use the K-prototypes clustering algorithm to divide the traffic safety status of the section and match the traffic operation characteristic portrait of the section.Use Bayesian conditional logistic regression to establish a section operation risk assessment model for dividing traffic safety status.Use MCMC to infer and estimate the model parameters to evaluate the risk of traffic feature portraits in different sections of the West Hanbei Expressway.The evaluation results are in line with the actual situation of the road section.The vehicle and segment risk assessment model constructed in this paper can provide a basis for the supervision of high-risk vehicles and sections by highway authorities.