BASELINE REQUIREMENTS FOR LANE SENSING TO AVOID RUN-OFF-ROAD ACCIDENTS. FINAL REPORT

Over a third of U.S. highway fatalities occur in single vehicle roadway departure accidents. Several research activities are described that address the design of a sensing and processing system necessary to enable prediction and warning of run-off-road accidents, and to eventually support an intervention type of path assist function. These activities are: the development of computer tools for design and analysis of road-departure prevention systems; the design and analysis of a warning and intervention system; and the development of a trailer platform to collect video of Michigan public highway lane markings for use in statistically characterizing their visual characteristics. Two computer tools are developed. The first is a basic driving simulator usable in either desktop or laboratory settings, allowing testing and evaluation of road-departure prevention system designs in their early stage. The second tool utilizes both this simulator mode and a simulation environment driven by a graphical user interface. This tool allows detailed evaluation of sensing, algorithmic, and vehicle dynamics issues. A proposed warning and intervention system is developed in detail for a computer vision-based roadway sensing system. Decisions to trigger warnings or initiate interventions are made using an online driver state assessment and a computation of the time to lane-crossing. Extensive development of algorithms for accurate and robust computation of time to lane-crossing and its associated uncertainty is reported. An intervention scheme using differential braking is described. The third activity is the design of a trailer-based vision system for measuring and analyzing the optical characteristics of paint stripe markings over hundreds of miles of Michigan public highways. The platform design is described; software to process the video is reported. Plans for testing are presented. Finally, a follow-on effort to implement the road-departure prevention system is briefly described.

  • Supplemental Notes:
    • Sponsored by the ITS Research Center of Excellence, University of Michigan, 2609 Draper Drive, Ann Arbor, MI 48109.
  • Corporate Authors:

    University of Michigan, Ann Arbor

    Mechanical Engineering and Applied Mechanics
    Ann Arbor, MI  United States  48109

    University of Michigan Transportation Research Institute

    2901 Baxter Road
    Ann Arbor, MI  United States  48109-2150
  • Authors:
    • Ulsoy, A G
    • Ervin, R D
    • Peng, H
  • Publication Date: 1995-1

Language

  • English

Media Info

  • Features: Appendices; Figures; References;
  • Pagination: 124 p.

Subject/Index Terms

Filing Info

  • Accession Number: 00723831
  • Record Type: Publication
  • Report/Paper Numbers: RCE-939404
  • Files: TRIS
  • Created Date: Jul 22 1996 12:00AM