Data Fusion Solutions to Compute Performance Measures for Urban Arterials

One of the key problems faced by traffic management operators of large urban traffic networks is the lack of sufficient data to compute performance indicators. These indicators, such as travel time,queue length, loss hours, and total time spent, are useful for offline evaluation purposes as well as online traffic control applications. In the latter case, such data are particularly of use in coordination algorithms that require information on the number of vehicles present or queuing in certain areas. This information in turn is used for example to assess the amount of buffer space available to temporarily store or reroute vehicles from more densely used parts of the network. Computing the amount of vehicles present or queuing in a certain area requires counting the number of vehicles that enter or exit that area. In this extended abstract the authors show how through fusing vehicle counts and travel times (measured by any means available), the well-known drift-error can be reduced to virtually zero. In the complete paper the authors show how this algorithm fits in a wider suite of data fusion tools to compute urban traffic performance indicators on the basis of multiple sources of data.


  • English

Media Info

  • Media Type: Digital/other
  • Features: Figures;
  • Pagination: pp 82-86
  • Monograph Title: Celebrating 50 Years of Traffic Flow Theory: A Symposium. August 11-13, 2014, Portland, Oregon
  • Serial:

Subject/Index Terms

Filing Info

  • Accession Number: 01604015
  • Record Type: Publication
  • Files: TRIS, TRB, ATRI
  • Created Date: Jun 23 2016 2:59PM