Study On The Freeway Traffic Accident Duration And Influence Range

Predicting the duration of freeway accidents is of great importance in increasing accidentrescue efficiency, releasing predictive and guiding traffic information in time, and quicklyeliminating accident impacts, etc. Forecasting the accident influencing area provides thefoundation for evaluating the accident influence degree, deciding whether and how to take theemergent traffic organization measures, etc. The research into these two aspects can serve asthe theory and methodology base for increasing the processing efficiency of freewayaccidents and quickly eliminating the negative effects.First, the related fundamental theories the temporal and spatial influence of freewayaccidents are summarized, consisting of the composition and influencing factors of trafficaccident duration, traffic wave theory, queuing theory, decision tree theory and travel pathchoice theory.Second, on the basis of deeply analyzing the corresponding influencing factors of thefour stages of traffic accident duration, that is, accident discovery time, accident responsetime, accident clearance time and traffic recovery time, different models are built for eachstage. Among them, the accident discovery time are classifiedly predicted according to thealarm mode. The accident response time and clearance time are both forecasted using thedecision tree method. For the accident recovery time, traffic wave theory is applied whenthere are no intervention measures used, while the queuing theory is used when interventionmeasures are taken. When durations of each stage are predicted, the total accident durationcan be obtained. This model for predicting the freeway accident duration is called the stagedmodel.Then, the route choice behavior of the upstream vehicles is seen as a course of theredistribution of traffic flows, upon which the accident affected range is analyzed. The originsand destinations of vehicles’ route choices are firstly determined, after which the functions ofroutes’ impedance are analyzed. Then, on the basis of the two improved models of the traditional Logit route choice model, a new improved model is built to calculate the selectiveprobability of each route. The traffic accident duration is divided into several short periods fortraffic assignment use. Then, the capacity restrained-multipath traffic assignment method isused several times and the traffic volumes of related roads corresponding to each trafficassignment period can be obtained. With the level of service of these roads under normalconditions and traffic accident conditions compared, the significantly influenced roads in thegiven period can be determined. Thus, the traffic accident influence range in each period canbe specified. The largest one of these ranges is regarded as the freeway traffic accidentinfluence range.