Study On Travel Time Estimation Of Urban Expressway Under Incident Conditions

Travel time as one of the most important traffic parameters,which has great significance for people to assign daily's work and life plan.However,because the traffic accident is uncertainty and unpredictability,especially with the growing traffic flow,a slightly incident may lead to a large scale of travel time delay in the urban expressway,which has been bringing inconvenience to the residents.Therefore,it is very important to reduce the travel time delay by studying the characteristics of traffic flow under accident conditions,establishing travel time estimation model in the incident scene,and constructing the subsystem of intelligent transportation system with the features of travel time estimation,information releasing and traffic inducing.Firstly,this paper studied the basic theory and method of the urban expressway travel time estimation under incident conditions in three aspects: road capacity under incident conditions,the theory of traffic flow queuing and dissipative,travel time estimation.Secondly,by analyzing the advantages and disadvantages of the current traffic travel time estimation theory,established the fluctuationtheory-BP neural network urban expressway travel time estimation model under incident conditions based on statistical the characteristic of the traffic flow.Finally,based on the historical data of the traffic accidents in Beijing's Fourth urban expressway,we calculate actual travel time according to the travel time between every two loop detectors,and calculate prediction travel time using the estimation model with the data of the volume of traffic,the location speed 36 hours before and after the occurrence of traffic incident,the statistical time of each accident stage.Three common criteria including mean absolute error(MAE),root mean square error(RMSE),and mean absolute relative error(MARE)are utilizedto evaluate the estimation performance.The results show that the combination model has higher prediction accuracy than the single prediction model.Applying the travel time estimation model under incident conditionsto the intelligent transportation system,and construct the basic framework of the urban expresswaytravel time estimation subsystem under the intelligent transportation system.