Estimating Traffic Parameters Of Expressway Section Based On Mass OD Information Of Network

The traffic parameters of expressway network section has become an importantelement of the traffic management,traffic prediction and traffic planning for trafficadministrative department.How to get more accurate and gain a wider range of trafficparameters data of expressway section has become the key research points of trafficinformation collection techonology.Tradditional information acquisition for trafficparameters of expressway section is mainly composed of stationary car inspectiondevice,which were imbedded in the pavement of main roads,can not provide enoughinformation of expressway section in the road network.Estimating traffic parameters ofexpressway section based on magnanimity OD(Origin－Destination) data, is a kind of newway of traffic information collection.This method don’t need to configure a lot of expensiveequipments,can use the existing road network data which covering the entire road network,access to accurate,comprehensive and authenticated traffic parameters information ofexpressway network section.This paper introduced the method of computing OD traffic information of the vehiclebased on charging data.And with the Oracle database as a platform,Visual C++programming to access Oracle databases to compute OD traffic information of the vehicleon the expressway.The vehicle type classification presented in the paper has highprecision,small partition granularity.This paper analyzed the vehicle models,driving pathand speed information based on the OD information table of the vehicle.Then worked outthe traffic flow, vehicle speed and occupancy between all adjacent toll station of theexpressway network in Shaanxi Province.Thesis adopted the method of gradual approximation,started with a month’s timeperiod,continued to narrow the time period of observing traffic parameters, with the VisualC++as a platform, programmed to compute traffic flow and occupancy in the different timeperiods based on the charging data.This paper used the traffic flow detected by vehicle detector as the benchmark data,evaluated the traffic flow error in the different time periods,then got the time precision of traffic flow based on the charging data.Analyzed the trafficflow and occupancy of the section,results showed that the traffic parameters based oncharging data accord to facts, has good adaptability and accuracy.Combined with charging data and vehicle detector data to estimate the section trafficvolume is one of the features of this paper. The paper established the BP neural networkmodel, with traffic volume of one hour based on OD charging data as network input, trafficvolume detected by vehicle detector as network output,programmed on Matlab platform toimplement simulation experiment of correcting algorithm of traffic parameters based on BPneural network. Then compared the average relative error of section traffic volume beforeand after correction, the results showed that the correction algorithm of traffic volume basedon BP neural network is effective.