Investigating the Use of Dual Tire Variable for Vehicle Classification

This paper discusses the improvement of the current vehicle classification algorithm which uses the number of axles and axle spacing to collect traffic data from the highway. This algorithm has a limitation of misclassifying passenger vehicles as trucks and single unit trucks towing light trailers as multi unit trucks. The proposed algorithm adds the number of tires per axle as a discriminating variable besides the number of axles and axle spacings. The use of this new variable parallels the maturity of technology that can count the number of tires on each axle. A total of 5,404 vehicle records were downloaded from three permanent traffic classification stations with a video camera being used to determine presence of dual tires and the true class of each vehicle according to Scheme F. Three algorithms utilizing number of axles, axle spacing, and number of dual tires per axle were built and tested on the field data. The best of the three algorithms reduced the overall and truck classification error from 8.5% and 27.2% to 1.7% and 6.9%, respectively. One of the advantages of using the dual tire variable was that trucks were no longer misclassified as passenger vehicles. Field validation of the proposed algorithm will likely show improvement in the accuracy of vehicle class data collected according to the 13 categories of the Scheme F.

Language

  • English

Media Info

  • Media Type: CD-ROM
  • Features: Figures; Photos; References; Tables;
  • Pagination: 16p
  • Monograph Title: TRB 86th Annual Meeting Compendium of Papers CD-ROM

Subject/Index Terms

Filing Info

  • Accession Number: 01049552
  • Record Type: Publication
  • Report/Paper Numbers: 07-2241
  • Files: TRIS, TRB
  • Created Date: Feb 8 2007 6:55PM