Project: #318 Determining segment and network traffic volumes from video data obtained from transit buses in regular service: Developments and evaluation of approaches for ongoing use across urban networks Progress Report - Reporting Period Ending: March 31, 2021 Principal Investigator: Mark McCord Status: Active Start Date: June 1, 2020 End Date: June 30, 2022 Research Type: Advanced Grant Type: Research Grant Program: FAST Act - Mobility National (2016 - 2022) Grant Cycle: Mobility21 - The Ohio State University Progress Report (Last Updated: March 30, 2021, 6:53 a.m.) % Project Completed to Date: 30 % Grant Award Expended: 15 % Match Expended & Document: 75 USDOT Requirements Accomplishments Goals and Objectives • Refine approaches to determine traffic volumes and related network measures in urban areas from video sensors mounted on transit buses in operational service • Demonstrate performance of the method • Motivate practical use of the method and begin applications • Expand method to an approach for ongoing monitoring of traffic measures Accomplishments • A large empirical data collection of concurrent road tube, manually observed, and video data on the campus of The Ohio State University (OSU) was undertaken on November 5, 2020. The data were subsequently processed and analyzed for research, education and outreach purposes. The concurrent video and traditionally collected (i.e., road tube and manually observed) data support validation studies and improvements of approaches as presented in many items below. The data collection, processing, and analysis were used in a term project of a Civil Engineering class consisting of a mix of 27 undergraduate and graduate students. The vehicle miles travelled estimated (see below) are presented to OSU transportation planning and operations managers. o Regarding the data collection and processing: ? Twenty-four (24) hours of road tube data on 10 roadway segment-directions were obtained ? Approximately 63 hours of bus-based video imagery extracted data were obtained on 4 different bus routes for empirical volume estimation ? Approximately 35 hours of the video imagery were preprocessed into time- and location-specific vehicle observations, and these vehicle observations were then processed into 336 segment-direction-hours of traffic volumes ? Fifty-two (52) segment-direction-hours of manually collected traffic volumes were obtained o Regarding analysis of the data collected on 11/5/2020: ? Vehicle miles traveled (VMT) values during a 6-hour period over a roadway network of 8.0 direction-miles of roadway were determined separately from the hourly volumes derived from the bus video imagery extracted data and from the hourly volumes derived from the combination of road tube and manually collected ? VMT values over a roadway network and time-of-day period that is common among the 2020 study conducted during this reporting period and the previously conducted 2019 and 2018 studies were determined and contrasted to indicate reduced travel during the COVID (2020) period and the stability of the estimates in the 2018 and 2019 periods • In an attempt to reduce the effort associated with semi-manually processing video data into location- and time-specific vehicle observations, software that can automatically determine roadway segments of interest and indicate which segments should be subsequently submitted to additional processing was developed and refined on another project; this segmenting software was used to determine roadway segments on 643 video clip; approximately 40% of these video clips were segmented successfully in fully automatic mode; the other clips were then triaged for segmentation with the previously used, semi-manual approach • Empirical analyses were conducted which indicate that traffic volumes estimated from commonly used, manual traffic count procedures and volumes obtained from road-tube tube are not different from each other when used for comparisons with video-based estimates, indicating that the “manual” and “tube” data can be pooled into what we refer to as estimates from “traditional” data, thereby allowing larger sample sizes • Empirical analyses indicated statistically significant differences in relative differences between video-based and traditionally estimated hourly vehicle volumes, depending on whether the segment-direction-hour data were obtained in 2019 or 2018; however o The differences in the relative errors, depending on the year, were seen to be consistent with results from an estimated regression model of the relative difference between video-based and traditionally estimated hourly volumes as a function of traffic volume and equivalent duration of the bus-based video data; this result supports the validity and usefulness of the regression model developed o When only using data from the same segment-direction-hours in the two years, the statistical significance of the differences was eliminated; indeed, the mean relative error was identical for the two years; this result supports the reliability of our estimation approach • Alternative approaches to determine point estimates of traffic volumes over a time interval (e.g., 15-minute interval, 60- minute interval) from bus-based video data were developed; preliminary validation studies (comparing video-based estimates to concurrent road tube data) indicate that the alternative approaches perform better than methods used to date • Efforts were begun to develop approaches to determine estimates of uncertainty distributions of traffic volumes over a time interval from bus-based video data; preliminary empirical results were conducted that are spurring comparisons and refinements • Methods were developed and compared to estimate the hourly vehicle volume on an average homogeneous day from bus-based video imagery obtained on different days during the same hour of the day o One of the methods was determined to be preferred o Empirical hourly volume estimates for specified hours on an average homogeneous day were determined from video imagery and from traditionally collected data; comparisons indicate that the relative differences between the video- and traditionally-determined estimates for “an average day” are less than the relative differences obtained for a single day; this result supports the overall motivation for this project, namely, the hypothesized ability of the repeated coverage of transit buses to allow large enough samples to yield good estimates of traffic volumes for homogeneous time periods Training and professional development • Three graduate students and five undergraduate students continued to be involved with various efforts of this project; three additional undergraduate students have expressed interest in being involved with this project • One undergraduate student has taken responsibility for training students on various data processing tasks. Dissemination • An online presentation was made to the Mobility 21, Smart Mobility Connections Seminar Series • An online presentation was made to the Traffic21/Mobility 21 Deployment Partner Consortium Symposium • Investigators meet regularly with administrators from The Ohio State University’s Transportation and Traffic Management (TTM) to update them on empirical results obtained; TTM provides the video imagery upon request of the investigators as a result of these meetings Plans for upcoming period • Train new undergraduates in video processing • Process additional video imagery into digital location- and time-specific vehicle observations • Develop volume estimates from processed video imagery • Refine and validate approaches that produce point-estimates of video-based traffic volume estimates • Refine and validate approaches that produce estimates of uncertainty distributions of video-based traffic volumes • Validate benefits of estimating average day traffic volumes in a time-of-day period relative to estimating single day traffic volumes in the time-of-day period • Investigate the accuracy in estimating temporal and spatial patterns in traffic volumes from video imagery • Continue interactions with OSU TTM Impacts • Regular meetings with The Ohio State University (OSU) Transportation and Traffic Management (TTM) office continue to generate interest in the empirical traffic flow estimates being produced and motivate ongoing collaboration in project research and outreach efforts Other Physical collections: During the period, manually collected traffic counts, road tube counts, and concurrent video images were obtained and processed into traffic volume estimates Curricula: During the period, the approach being refined and evaluated in this study was presented as a module and served as the basis of a term project in one undergraduate/graduate Civil Engineering class consisting of 27 students. The project was also presented in an undergraduate Civil Engineering class consisting of 52 students Outcomes New Partners We continued to work closely with Transportation and Traffic Management (TTM) at The Ohio State University (OSU). TTM oversees all OSU transportation operations, other than parking, and operates the Campus Area Bus Service (CABS), with a fleet of approximately fifty 40-foot transit buses serving close to 5 million passengers per year on fixed route scheduled services. We also coordinated field validation studies with the transportation division of the Mid-Ohio Regional Planning Commission (MORPC), the greater Columbus area MPO. Issues It appears that some charges made to matching funds should have been made to the award. We are working on correcting these errors.