Integration of a Robust Automated Pedestrian Detection System for Signalized Intersections

Details

• Creators:
  Lin, Pei-Sung ; Kourtellis, Achilleas ; Wang, Zhenyu ; Chen, Cong
• Corporate Creators:
  University of South Florida. Center for Urban Transportation Research
• Corporate Contributors:
  Florida. Department of Transportation. Research Center
• Subject/TRT Terms:
  Automation Flashing Beacons Midblock Crossings Pedestrian Detectors Pedestrian Safety Signalized Intersections Traffic Safety
• Resource Type:
  Tech Report
• Geographical Coverage:
  United States
• Edition:
  Final report, October 2017–December 2019
• Contracting Officer:
  El-Urfali, Alan
• Corporate Publisher:
  Florida. Department of Transportation
• Abstract:
  Improving pedestrian safety and making roadway facilities safer and friendlier for pedestrians are among the top priorities and transportation goals in Florida. The actuation of pedestrian “Walk” signal indications at signalized intersections and the triggering of rectangular rapid flash beacons (RRFBs) or high-intensity activated crosswalks (HAWKs) at midblock crosswalks, require a pedestrian to push a button. However, 40%–50% of pedestrians of pedestrians do not push the button. The Center for Urban Transportation Research (CUTR), in close coordination with the Florida Department of Transportation (FDOT), researched automatic pedestrian detection systems for use at midblock crosswalks to automatically activate the RRFBs, and at signalized intersections to automatically place a pedestrian call to the traffic signal controller. The CUTR research team reviewed various automatic pedestrian detection systems for their functionality and performance. Three systems were selected for this research. They were first tested under controlled conditions, and at midblock and intersection locations. One of the three systems was able to meet all desired performance requirements. This system was further deployed and evaluated at two midblock crosswalks and one signalized intersection. The evaluation results showed that the system produced an overall 92% detection system accuracy at midblock locations, with only 2% false detections. The system was able to detect pedestrians 94% of the time and place a pedestrian service call 90% of the time at a signalized intersection. An important contribution of this research project is the work conducted to connect the automatic pedestrian detection systems to a traffic signal controller to place a pedestrian call when a pedestrian is detected and to remove the call when the pedestrian walks out of the detection zone prematurely before the call is served. Testing showed that the system was able to detect the disappearance of pedestrians 98% of the time and removed the pedestrian call 97% of the time when they left the detection zone early. This result showed the capability of the automatic detection system and the advanced traffic signal controller with a custom script to administer removal of a pedestrian call when it is not needed. This capability is useful for minimizing unnecessary vehicle delay. This research provided a key step to apply automatic pedestrian detection to further enhance pedestrian safety at signalized intersections and midblock crosswalks, and reduce unnecessary vehicle delay.
• Format:
  PDF
• Funding:
  BDV25-977-44
• Collection(s):
  US Transportation Collection
• Main Document Checksum:
  urn:sha256:ded2de394e8b2b23be9001d0ad570c4ce091f25f6fac4e04a614639812064789
• Download URL:
  https://rosap.ntl.bts.gov/view/dot/64442/dot_64442_DS1.pdf
• File Type:
  [PDF - 4.01 MB ]