Real-Time Accuracy Degree Forecast of Estimated Link Average Travel Time Based on Data Fusion Method

The paper brings forward a BP neural network model to forecast the real-time accuracy degree of estimated link average travel time based on data fusion method, and four variables (link average density, traffic volume, link average travel time based on floating car data, and floating car sampling size) are taken as input variables. Among these four variables, link average density and traffic volume can be obtained by loop detectors from SCATS, while link average travel time and floating car sampling size can be acquired with FCD. Then the reasons why those four variables are chosen are given with the support of statistical analysis. The model consists of three parts, the initial data generated module, data fusion module based on BP network and results analysis module. At last, an arterial road in Hangzhou is chosen as object link, 400 groups of data is being utilized to verify the model, and the results prove to be very satisfactory.


  • English

Media Info

  • Media Type: Web
  • Features: Figures; References; Tables;
  • Pagination: pp 1077-1086
  • Monograph Title: CICTP 2012: Multimodal Transportation Systems—Convenient, Safe, Cost-Effective, Efficient

Subject/Index Terms

Filing Info

  • Accession Number: 01501013
  • Record Type: Publication
  • ISBN: 9780784412442
  • Files: TRIS, ASCE
  • Created Date: Dec 10 2013 8:10AM