Short-Term Congestion Prediction on Freeways

The paper describes a comparative analysis on several methods (the naïve method, multi-linear regression, time series analysis, multi-layer feedforward neural networks, radial basis function neural networks, Elman neural networks, self-organizing map neural networks, fuzzy logic) that can be used to come to short-term congestion prediction. The field data sets that were used were acquired on two locations during the morning and evening peak periods. The research results show that both the multi-layer feedforward and the Elman neural networks outperform the other methods. The paper concludes with remarks on the importance to have detectors upstream of the prediction point and to keep in mind what kind of congestion has to be predicted.

• Corporate Authors:

  World Conference on Transport Research Society

  Secretariat, 14 Avenue Berthelot
  69363 Lyon cedex 07,   France 
• Authors:
  • Huisken, Giovanni
  • Tillema, Frans
• Conference:
  • 10th World Conference on Transport Research
  • Location: Istanbul , Turkey
  • Date: 2004-7-4 to 2004-7-8
• Publication Date: 2004

Language

• English

Media Info

• Media Type: CD-ROM
• Features: Figures; Photos; References;
• Pagination: 17p
• Monograph Title: 10th World Conference on Transport Research

Subject/Index Terms

• TRT Terms: Congestion management systems; Congestion pricing; Forecasting; Fuzzy logic; Fuzzy systems; Linear regression analysis; Neural networks; Time series analysis; Traffic congestion
• Subject Areas: Highways; Operations and Traffic Management; Planning and Forecasting; I72: Traffic and Transport Planning;

Filing Info

• Accession Number: 01084996
• Record Type: Publication
• Files: TRIS
• Created Date: Jan 28 2008 8:14AM