Driving Performance Analysis of the Advanced Collision Avoidance System (ACAS) Field Operational Test (FOT) Data and Recommendations for a Driving Workload Manager

This report contains analyses of driving performance data from the Advanced Collision Avoidance System (ACAS) Field Operational Test (FOT), with data from nearly 100 drivers and over 100,000 miles of driving. The analyses compared normal and distracted situations and determined thresholds that distinguish between maneuvering and non-maneuvering situations. Four questions were addressed: (1) How are measures of driver input (steering wheel angle, etc.) and vehicle output (heading, speed, etc.) distributed as a function of 4 road types ((a) ramps, (b) interstates and freeways, (c) arterials and minor arterials, and (d) collectors and local roads). (2) What is the effect of the number of tasks on measures of driver performance as a function of road type. (The distributions for 0 and 1 tasks were similar. For 2 tasks, the range was sometimes 50% less.) (3) How well do linear thresholds distinguish between maneuvering and non-maneuvering situations, and what should those values be. (It varies with the threshold; sometimes the odds were 10:1. Other times they were 1:1.) (4) How effectively do steering and throttle entropy predict distracted and normal driving. (Only steering entropy showed any differences.)Two analysts rated the visual, auditory, cognitive, and psychomotor demands of 68 subtasks (e.g., prepare to eat/drink, converse on the cell phone) performed while driving. Ratings were relative to anchors from the U.S. Army IMPRINT modeling tool (0-to-7 scale). Video clips of those subtasks were sampled from the advanced collision avoidance system (ACAS) field operational test (FOT) database, a naturalistic study of driving previously performed by UMTRI. Key findings were: (1) The most demanding tasks were dialing a phone, answering a phone, lighting a cigar or cigarette, dealing with pet and insect distractions, dealing with spilled drinks and food, typing with 2 thumbs, and drinking from a cup, in that order. (2) Demand levels within subtasks were moderately correlated (visual-cognitive=0.68, visual-psychomotor=0.48, cognitive-auditory=0.42, cognitive -psychomotor=0.34) or close to 0. (3) In terms of these ratings, cognitive demands, both per unit time and when weighted by exposure, were consistently double the value of others. (4) Demands varied to a limited degree among road types. (5) There were consistent differences in demand due to driver age and sex. Researchers are encouraged to use the demand ratings provided and extend them to other tasks so tasks can be compared across experiments.

  • Corporate Authors:

    University of Michigan Transportation Research Institute

    2901 Baxter Road
    Ann Arbor, MI  United States  48109-2150

    Delphi Delco Electronics Systems

    One Corporate Center, M/C E110, P.O. Box 9005
    Kokomo, IN  United States  46904-9005
  • Authors:
    • Eoh, H
    • Green, P A
    • Schweitzer, J
    • Hegedus, E
  • Publication Date: 2006-12


  • English

Media Info

  • Media Type: Print
  • Pagination: 126p

Subject/Index Terms

Filing Info

  • Accession Number: 01155651
  • Record Type: Publication
  • Report/Paper Numbers: UMTRI-2006-18
  • Files: TRIS
  • Created Date: Apr 27 2010 4:23PM