The Application And Research Of Forecast Analysis Based On Expressway Networking Toll Data

Expressway networking toll data is the direct result of expressway network toll management system,with abundant fields and content.It contains large amounts of data and updates in time.It can obtain large amounts of information hidden in the original data by depth analysis and mining of the toll data.Expressway route selection,section vehicle flow and section travel time forecast were studied by optimization algorithm,modeling,case analysis and application,which was based on feature analysis of expressway networking toll data.All the research results both better the selections of drivers and improve management level of expressway management departments.Since correlated traffic forecast studies both at home and abroad only pay attention to some certain aspects and it lacks of comprehensive forecast research,specific studies in this paper are as follow:First of all,a pretreatment method for original expressway networking toll data was proposed.To minimize impact to the results of abnormal data,which occupies a quite proportion of original data,abnormal data was divided into redundant data,missing data and noise data,and each kind of data processing methods was described.The feasibility and accuracy of this method was proved by case analysis with MATLAB.Secondly,a vehicle travel path forecast model was built based on Markov forecast method.To improve the precision of prediction,original data was divided into 9 vehicle classifications.To solve state transition probability matrix,statistical method and linear equations method were selected to analyze,and cases were taken to validate that statistical method is more suitable for toll data prediction.Thirdly,expressway section traffic state forecast research was conducted based on travel path prediction.Section vehicle flow and travel time were chosen as section state forecast indexes.A section traffic statistics method based on toll data was proposed for vehicle flow forecast.Cases analysis show that section vehicle flow forecast based on Adaptive Kalman Filtering method can avoid the shortcoming of Kalman.Filtering method and improve forecast accuracy.Then,the relationship of section vehicle flow and travel time was proved,and a section travel time forecast method was proposed and modified.The feasibility and accuracy of this method was proved by case analysis with MATLAB.Finally,the application scenarios for expressway networking toll data forecast was proposed based on the above forecast research results.Forecast of real-time and non-real-time were discussed for comparative analysis.And traffic forecast for different traffic characteristics of weekdays,weekends and vacations were conducted.The research results are useful for some research in related fields.