Research On Sensitivity And Detection Probability Model Of Traffic Incident Based On Freeway Travel Time

With the rapid development of Chinese economy and the active transportation,the freeway mileage and traffic volume also increased rapidly,the number of traffic incidents and the severity of the incidents also increased significantly.How to detect incidents accurately and efficiently and grasp the information of incidents roundly,improve the safety,efficiency and comfort of the daily travel of the public,and put forward new challenges to the operation and management of the freeway.Therefore,based on the accurate road network construction data and rich data resources of freeway in Shaanxi Province,the paper analyzes the laws of vehicle driving through theoretical modeling and real data,from which to dig operation of road network.Based on the analysis and data validation of the influencing factors of freeway travel time,the paper puts forward that the frequency of travel time obey the law of logarithmic normal distribution on the different work days,rest days,hours and sections,and gives the mainstream travel time algorithm according to the parameter estimation algorithm.Analyze the vehicle driving time delay and the spatial queuing characteristic under the traffic incidents.Judge the delay OD and detect the delay vehicle through the travel time comparison,and calculate the delay scale and other indicators caused by the same incidents.For vehicles affected by traffic incidents,think about the influencing factors of freeway travel time such as travel path,time period,vehicle type and the impact time of the traffic incidents,analyze the degree of travel time fluctuation under the incidents influences of different path length,and establish the sensitivity model on the basis of the proportion of vehicle type and the proportion of OD combination.After this,paper also puts forward that vehicle which affected by the incidents will improve their speed to make up the delay time of the incidents.Through making curve fitting to the vehicle travel time of the same entrance station,establish the distance-travel time compensation model.On the basis of this,propose two kinds of methods to extrapolate the generation position and time information of incidents based on path relation and vehicle driving characteristics.Based on the decision of delay path,select statistical eigenvalues such as sensitivity of travel time?time delay degree as indicators,establish the probability model of incidents detection based on logistic algorithm.Finally,relying on the toll data and traffic incidents report of a freeway management center in Shaanxi province as experimental samples.The Oracle database is accessed through Visual C++ programming,and the detection algorithm of freeway traffic incidents information and incidents detection probability model are verified.The data verification results show that algorithms and detection models have higher accuracy and lower false positives.