Link to the University of Pittsburgh Homepage
Link to the University Library System Homepage Link to the Contact Us Form

Evaluating the Quality of Automatic Pedestrian Detection at Major Arterial Intersection

Rodriguez, Daniel (2022) Evaluating the Quality of Automatic Pedestrian Detection at Major Arterial Intersection. Master's Thesis, University of Pittsburgh. (Unpublished)

Download (4MB) | Preview


Improvements in detection technology can provide full actuation for vehicles and pedestrians at a
signalized intersection. The goal of this study was to assess the accuracy of an automatic camera detection
system from GRIDSMART to detect pedestrians and make calls to the controller to provide pedestrian
phases only when needed. Manual actuation has been proven ineffective due to individual behavior, so
automatic detection is of interest as a solution in lower pedestrian areas to improve signal operations for all
users. Video from the GRIDSMART system was analyzed to determine the accuracy of pedestrian
detection. The detection system was not capable of accurately counting pedestrians due to vehicle
interference and configuration issues. Pedestrian calls were made only for those who came to a stop in the
detection zones. Weather and time of day also impacted accuracy with errors increasing during night and
rainy conditions. The detection system averaged 2.4 calls/pedestrian which resulted in unnecessary
pedestrian phases. A VISSIM model was created to determine the effects on vehicle delay from the current
operations, and to determine the most appropriate pedestrian treatment for the subject intersection. Vehicle
volumes on the critical approaches are significant enough that pedestrian green times do not extend the
vehicle phases when no demand is present. The detection errors did not lead to statistically significant delay
increases for vehicles or pedestrians, but there was no benefit identified. For this specific case, the
intersection can be operated under pedestrian recall without increasing vehicle delay, even for low
pedestrian volumes. For higher pedestrian volumes that can be expected in peak conditions, pedestrian
recall would also be an appropriate signal operation. The detection software does not appear suited for
controlling pedestrian crossings until improvements are made. Other intersections where pedestrian recall
would increase vehicle delays may prove to be better candidates.


Social Networking:
Share |


Item Type: University of Pittsburgh ETD
Status: Unpublished
CreatorsEmailPitt UsernameORCID
Rodriguez, Danieldar212@pitt.eduDAR2120000-0001-8152-9747
ETD Committee:
TitleMemberEmail AddressPitt UsernameORCID
Thesis AdvisorStevanovic,
Committee MemberAlavi,
Committee MemberFascetti,
Date: 10 June 2022
Date Type: Publication
Defense Date: 29 March 2022
Approval Date: 10 June 2022
Submission Date: 16 March 2022
Access Restriction: No restriction; Release the ETD for access worldwide immediately.
Number of Pages: 107
Institution: University of Pittsburgh
Schools and Programs: Swanson School of Engineering > Civil and Environmental Engineering
Degree: MSCE - Master of Science in Civil Engineering
Thesis Type: Master's Thesis
Refereed: Yes
Uncontrolled Keywords: gridsmart, pedestrian treatments, pedestrian recall, actuated signal operations, signalized intersection, automatic detection, manual actuation, automatic actuation,
Date Deposited: 10 Jun 2022 18:46
Last Modified: 10 Jun 2022 18:46


Monthly Views for the past 3 years

Plum Analytics

Actions (login required)

View Item View Item