Pilot Application of Biometric-Based Vehicle Occupancy Detection on Managed Lanes for Congestion Reduction [Policy Brief]

Details

• Alternative Title:
  Pilot of Vehicle Occupancy Detection Technology [Policy Brief]
• Creators:
  Hendricks, Sara J. ; Winters, Philip L.
• Corporate Creators:
  University of South Florida. Center for Urban Transportation Research
• Corporate Contributors:
  National Institute for Congestion Reduction ; United States. Department of Transportation. University Transportation Centers (UTC) Program ; United States. Department of Transportation. Office of the Assistant Secretary for Research and Technology
• Subject/TRT Terms:
  Detection And Identification Systems High Occupancy Vehicle Lanes Incentives Managed Lanes Traffic Lanes Vehicle Occupancy
• DOI:
  https://doi.org/10.5038/CUTR-NICR-Y2-2-6
• Resource Type:
  Brief
• Geographical Coverage:
  United States ; Florida
• Edition:
  Policy Brief
• Corporate Publisher:
  National Institute for Congestion Reduction
• Abstract:
  This study tested the accuracy and reliability of a biometric vehicle occupancy detection technology. Downloaded onto a carpooler’s mobile device, the app enables measuring the number of passengers in a vehicle for purposes of high occupancy vehicle (HOV) and high occupancy toll (HOT) lane management and incentivization. Volunteers recruited from South Florida, the Tampa Bay area, and Utah used the app to log carpool trips that were recorded and validated by the app. The data were compared against digital photos taken at the same time as the trip to verify that the carpool validation was correct. A survey was conducted to gather information about the user-friendliness of the app. Evaluators reviewed all 837 logged trips. Evaluators determined one False Positive among trips validated by the app, with a positive predictive value of 0.9803. Evaluators determined 13 False Negatives among trips unvalidated by the app, with a negative predictive value of 0.9030. Most of the False Negatives were due to occurrences of an overly high similarity threshold that can be adjusted. Evaluators also tested the app using single occupancy vehicles (SOV) for staged True Negative scenarios. After more than 30 attempts, evaluators failed to ‘fool’ the app into verifying an SOV as a carpool. During the pilot, the app continually improved performance, both in accuracy and in user experience. For the app to work as designed, carpoolers must remember to capture an initial snapshot with their mobile device, before the start of the carpool trip and revalidate at the end of the trip, if prompted. Carpoolers must look directly at their phone camera and wait for the verification signal to ensure sufficient biometric facial data has successfully been recorded. The app is configurable, thus requiring a policy decision regarding the balance between reducing the seconds required for trip validation, increasing convenience, versus heightened accuracy, by selected threshold for similarity. A soft launch of the app should provide an initial period for determining the acceptable balance of False Negatives and False Positives prior to a full deployment. There will be a need for customer support when deployed. Future research could focus on back-office integration with app functions, automated reward delivery, and testing the accuracy and reliability of counts for three or more vehicle occupants. Finally, social marketing research could provide insight into what would make the app more attractive than the alternatives for specific market segments, and what benefits would outweigh the barriers to using the app.
• Format:
  PDF
• Funding:
  69A3551947136 ; 79075-10
• Collection(s):
  University Transportation Centers
• Main Document Checksum:
  urn:sha-512:7f9bd9fa1e922fd95e454d0e0fc3022df54dd3db6c63cd40bc82f0663c08d4f6858dc96a09d8a784920167bb42fe5b85163eeeb35cecabc24725874e91a344d3
• Download URL:
  https://rosap.ntl.bts.gov/view/dot/85520/dot_85520_DS1.pdf
• File Type:
  [PDF - 286.80 KB ]