Highway Travel Time Prediction Research Based On Adaptive Interpolation Kalman Filter And Toll Collection Data

ABSTRACT:Travel time between highway stations can measure the road traffic efficiency and traffic condition, which is the important basis for Traffic Control Department to achieve traffic control and guidance. Travel time is highly concerned by travelers, which has become the key factors of advanced traveler information system and route guidance system. Toll collection data was used to predict travel time between highway stations, specific studies are as follows:(1) Compared with electronic toll collection data, manual toll collection data contains vehicles queued payment time. A new processing standard of data fusion based on real-time data in MTC and ETC database is put forward, which contains extremely abnormal data processing method and data fusion method, as well as improves cycle vehicle data number.(2) Considering the abnormal toll collection data was difficult to eliminate, a calculation model of improved average travel time is researched, which contains Four-way split data exclude method ideas. The model improves the quality of toll collection data and the calculation accuracy of average vehicle travel time.(3) Considering the weak nonlinear and adaptive of Kalman filtering algorithm, a Kalman filtering algorithm fused with adaptive interpolation is researched. Equidistant interpolation method is introduced to reconstruct the time series between real-time and historical cycle travel time, then, Kalman filtering prediction model is built based on least square method, and travel time prediction principle is elaborated, which is based on Sage-Husa adaptive interpolation algorithm.(4) The actual case is used to verify the effectiveness of the algorithm, and the case results show that the relative error of the travel time obtained by interpolation adaptive algorithm is less than7.5%under the all period of normal, accident and holiday traffic flow, and the relative error under the accident period is less than10%.(5) Travel time prediction system architecture and prediction system logic is built. Toll collection data database, travel time algorithm and publishing interface design is elaborated. Then the travel time prediction system based on C#and SQL Server2008is developed.(6) The stability test environment of offline travel time system is built. After stability test, the system is distributed in the highway information center to predict highway travel time in real-time. The practical application shows that the travel time prediction system performs well, and can provide an effective time reference for public in highway.