Vector Classification of Commercial Vehicles Using a High-Fidelity Inductive Loop Detection System

The impacts of the wide array of commercial vehicles on our roads today are significant in many aspects. However, an accurate understanding of the extents of these impacts is constrained by the inadequate fidelity of commercial vehicle surveillance data due to the limitations of existing classification schemes and surveillance technology deployed. This paper describes the Commercial Vehicle Vector Classification System (CVVCS) – a new commercial vehicle vector-based classification model based on a new prototype standalone inductive loop sensor called the Blade, which is connected to an advanced high-speed sampling Inductive Loop Detector (ILD) card. This new sensor-detector combination yields high fidelity inductive signatures that show axle locations as well as detailed undercarriage inductive vehicle signatures. The results from the calibrated CVVCS developed in this study show 99.0 percent Correct Classification Rate (CCR) for axle configuration, 84.9 percent CCR for drive unit body type and 84.1 percent CCR for trailer unit body type on independent test data. The integrated model was able to match all three classification components of commercial vehicles in the overall data set of 1029 vehicles with an accuracy of 80.8 percent. These results indicate remarkable potential for providing enhanced commercial vehicle surveillance through the implementation of the detection technology and model developed in this study.

Language

  • English

Media Info

  • Media Type: DVD
  • Features: Figures; References; Tables;
  • Pagination: 22p
  • Monograph Title: TRB 89th Annual Meeting Compendium of Papers DVD

Subject/Index Terms

Filing Info

  • Accession Number: 01154616
  • Record Type: Publication
  • Report/Paper Numbers: 10-2211
  • Files: BTRIS, TRIS, TRB
  • Created Date: Jan 25 2010 11:01AM