Abstract:Conflict prediction is prevalent in the field of expressway safety management. As the precursor of crashes, severe conflict draws great concentration in previous studies. However, the conflict mechanism heterogeneity under different traffic states is always ignored. The different traffic environment affects driving behavior, resulting in various casual factors. In order to better predict and understand the influencing factors of severe conflicts, this study proposes severe conflict prediction models under three traffic states based on a 3-hour vehicle trajectory data on the Shanghai Inner Ring Expressway. Firstly, the data is labeled with three traffic states based on the Traffic State Index and road average speed. Modified Time to Collision (MTTC) of vehicle pairs is calculated to identify severe conflicts. Second, the spatial and temporal characteristics of conflicts are analyzed to explore the conflict heterogeneity under different traffic states and the impact area of different ramps. Finally, logistic regression models are established to predict the likelihood of severe conflicts. Compared with the model without division of traffic states, congestion state model and transition state model have an increase of 1.7% and 8.9% in prediction accuracy, respectively. Main casual factors of severe conflicts differ under three traffic states. The occurrence of severe conflicts is more sensitive to the change in traffic volume and ramps when the traffic state becomes more congested. Vehicles in 300-meter area upstream exit ramps/downstream entrance ramps suffers from higher risk of collisions. However, when the traffic state approaches free flow, severe conflicts are more caused by drivers’ unsafe behavior such as high speed and short spacing between vehicles. The findings can help transportation managers figure out the main casual factors of expressway crashes under different traffic states, and thus develop more targeted safety management strategies.

Study on Traffic Accident Prediction Model for Long Sections of Expressway

Crash Prediction Model for Basic Freeway Segments Incorporating Influence of Road Geometrics and Traffic Signs

An Improved Model of Accident Prediction on Freeways Based on Fuzzy Logic

An Accident Prediction Model for Expressway Based on BP Neural Network

Accident Prediction Models for Urban Arterial System

Expressway Crash Prediction Based on Traffic Big Data

Prediction of urban expressway total traffic accident duration based on multiple linear regression and artificial neural network

The analysis of causes and a study of countermeasures of road traffic accidents

Optimizing crash risk models for freeway segments: A focus on the heterogeneous effects of road geometric design features, traffic operation status, and crash units

Short Term Prediction for Occurrence Probability of Expressway Traffic Incidents

Probabilistic Prediction Model for Expressway Traffic Noise Based on Short-Term Monitoring Data

Urban Road Traffic Conflict and Accident Prediction during the Period of Ice and Snow in Cold Region

A Deep Learning Approach to the Prediction of Short-term Traffic Accident Risk

Long-Term Forecasting Oriented to Urban Expressway Traffic Situation

Trajectory Data-Based Severe Conflict Prediction for Expressways under Different Traffic States

Study On Traffic Safety Risk Assessment Of Expressway In Cold Areas

Construction of crash prediction model of freeway basic segment based on interactive influence of explanatory variables

An urban traffic accident prediction model based on the full range of human, vehicle, road and environmental factors

Real-Time Traffic State Prediction and Congestion Mechanism Analysis for Expressways

Real-time safety analysis for expressways considering the heterogeneity of different segment types

Peak Traffic Flow Predictions: Exploiting Toll Data from Large Expressway Networks