SENSING OF VEHICLES RUNNING ON PLURAL LANES WITH AN ARTIFICIAL RETINA VEHICLE SENSOR

This paper describes how, in cities with heavy snowfalls, snow removal operations often form large walls of snow close to the walkways in both directions. These walls narrow the traffic lane and cause an increasing number of vehicles to operate on the central portion of the road. As a traffic information collection system, a supersonic type vehicle sensor is most popular. However, this sensor is a spot sensing type and often cannot sense these vehicles running on the center of the two lanes and this causes measurement errors. This paper investigates the improvement of these measurement errors using a built in artificial retina tip in the vehicle sensor. This artificial retina tips is a CMOS image sensor and has an arithmetic processing function and an analogue adjustment function so that in can work as a small-scale image recognition device. An image vehicle sensor equipped with an artificial retina tip can monitor two lanes with a single camera. This image type vehicle sensor was tested in Sapporo City during the winter season and the results are presented in this paper.

• Supplemental Notes:
  • Conference Proceedings available on CD-ROM.
• Corporate Authors:

  World Road Association (PIARC)

  La Grande Arche, Paroi Nord, Niveau 5
  F-92055 La Defense Cedex,   France 
• Authors:
  • Mashima, M
  • HAYAKAWA, Y
  • Munakata, I
  • Hara, H
• Conference:
  • New Challenges for Winter Road Service. XIth International Winter Road Congress
  • Location: Sapporo, Japan
  • Date: 2002-1-28 to 2002-1-31
• Publication Date: 2002

Language

• English

Media Info

• Pagination: 7p

Subject/Index Terms

• TRT Terms: Monitoring; Sensors; Traffic control; Traffic flow; Traffic safety; Vehicle detectors; Vehicle safety; Winter maintenance
• Subject Areas: Highways; Maintenance and Preservation; Operations and Traffic Management; Planning and Forecasting; I62: Winter Maintenance; I72: Traffic and Transport Planning;

Filing Info

• Accession Number: 00925090
• Record Type: Publication
• Files: TRIS
• Created Date: May 3 2002 12:00AM