Abstract
Transit agencies around the world are increasingly mounting video cameras inside and outside their buses for liability, safety, and security reasons. Some of the cameras provide fields of view that allow observation of vehicles traveling on the surrounding roadways. Such video imagery could conceivably be used to estimate traffic volumes on roadway segments traversed by the transit buses. Transit buses are attractive platforms for acquiring the information that leads to traffic volume estimates, since a fleet of transit buses collectively covers most major surface streets in an urban area and the buses regularly and repeatedly cover the same roadway segments, which would allow for multiple, independent estimates of roadway segment flows across days and by time of day. Since the video cameras are already installed for other purposes, the costs of estimating traffic flows from video obtained from transit buses in regular service would be minimal. Therefore, traffic flows could be estimated with much greater geographic coverage, with much greater frequency, and with much lower cost than is presently available from existing traffic volume observation methods.
Description
Background and Motivation: Transit agencies around the world are increasingly mounting video cameras inside and outside their buses for liability, safety, and security reasons. Some of the cameras provide fields of view that allow observation of vehicles traveling on the surrounding roadways. Such video imagery could conceivably be used to estimate traffic volumes on roadway segments traversed by the transit buses. Transit buses are attractive platforms for acquiring the information that leads to traffic volume estimates, since a fleet of transit buses collectively covers most major surface streets in an urban area and the buses regularly and repeatedly cover the same roadway segments, which would allow for multiple, independent estimates of roadway segment flows across days and by time of day. Since the video cameras are already installed for other purposes, the costs of estimating traffic flows from video obtained from transit buses in regular service would be minimal. Therefore, traffic flows could be estimated with much greater geographic coverage, with much greater frequency, and with much lower cost than is presently available from existing traffic volume observation methods.
The team of investigators previously revised the traditional moving observer method to estimate traffic segment volumes from bus based video imagery and demonstrated the potential of using imagery obtained from The Ohio State University (OSU) Campus Area Bus Service (CABS) buses in regular operation. Roadway segment volumes estimated from CABS video imagery were compared to volumes obtained from deployed road tubes. As would be expected from a single bus pass that represents a very short observation period, noisy estimates resulted. However, statistical analysis indicates that multiple estimates from multiple bus passes across the same roadway segment over many days can provide good estimates of typical segment volumes.
Objectives and Tasks: This project will develop and evaluate approaches to refine the method to estimate a segment traffic volume from a single bus pass and to aggregate the segment volume estimates obtained from multiple bus passes to provide good estimates of “typical” (e.g., average) volumes by homogenous periods (e.g., weekday by hour). In addition, the potential to improve the accuracy of aggregate network volume estimates (e.g., vehicle miles traveled) will be investigated, and the potential to replace or complement existing traffic volume observation methods will be explored and evaluated.
The investigators will again work with OSU’s Transportation and Traffic Management (TTM) to obtain video imagery from OSU CABS buses in regular service operation, which will be used to determine “video-based” roadway segment volumes. In parallel, segment traffic volumes will be estimated using traditional data collection methods, namely, road tubes and manual counts to compare to the video-based estimates for evaluation purposes.
The methodological research will consist of the following main tasks and expected outcomes:
Year 1
• Investigate improvements to the method used to estimate a traffic volume from a single bus pass
• Investigate ways to aggregate observations from individual passes across passes (e.g., for times of day and days of the week)
Expected outcomes: An approach to provide good quality traffic volume estimates from video imagery recorded during multiple bus passes over time and determination of traffic and infrastructure conditions that lead to better or worse accuracy.
Year 2
• Investigate the potential to improve the accuracy of aggregate network volume estimates, such as vehicle miles traveled
• Investigate the ability to use the estimates obtained from bus video imagery to complement or replace traditional traffic volume observations
• Develop approaches for ongoing volume estimation and monitoring
Expected outcomes: Quantification of improvements to urban surface street volume estimation using transit bus video imagery instead of or in combination with traditional data collection methods in terms of estimation accuracy and resources required. Proposed approach for urban transportation agencies to use available bus-based video on an ongoing basis to update traffic volume estimates across the urban roadway network.
Years 1 & 2
In addition to developing and evaluating these methodological approaches, traffic volume estimates developed using the OSU campus roadway network as a living lab would be provided to OSU campus transportation and infrastructure planners each year. The research team has worked with these groups over many years, and lead members of these groups have confirmed that no systematic traffic volume database exists and indicated a desire to populate such a database and update it on an ongoing basis.
Expected outcomes: In addition to providing otherwise unavailable information in an outreach effort, the provision of such information would be valuable in motivating expanded use of the approaches developed, both in terms of larger geographic coverage within an urban areas and across urban areas.
Timeline
Year 1
• Investigate improvements to the method used to estimate a traffic volume from a single bus pass
• Investigate ways to aggregate observations from individual passes across passes (e.g., for times of day and days of the week)
Year 2
• Investigate the potential to improve the accuracy of aggregate network volume estimates, such as vehicle miles traveled
• Investigate the ability to use the estimates obtained from bus video imagery to complement or replace traditional traffic volume observations
• Develop approaches for ongoing volume estimation and monitoring
Strategic Description / RD&T
Deployment Plan
Years 1 & 2
In addition to developing and evaluating these methodological approaches, traffic volume estimates developed using the OSU campus roadway network as a living lab would be provided to OSU campus transportation and infrastructure planners each year. The research team has worked with these groups over many years, and lead members of these groups have confirmed that no systematic traffic volume database exists and indicated a desire to populate such a database and update it on an ongoing basis.
Expected Outcomes/Impacts
Year 1 Expected outcomes: An approach to provide good quality traffic volume estimates from video imagery recorded during multiple bus passes over time and determination of traffic and infrastructure conditions that lead to better or worse accuracy.
Year 2 Expected outcomes: Quantification of improvements to urban surface street volume estimation using transit bus video imagery instead of or in combination with traditional data collection methods in terms of estimation accuracy and resources required. Proposed approach for urban transportation agencies to use available bus-based video on an ongoing basis to update traffic volume estimates across the urban roadway network.
Years 1 & 2 Deployment Expected outcomes: In addition to providing otherwise unavailable information in an outreach effort, the provision of such information would be valuable in motivating expanded use of the approaches developed, both in terms of larger geographic coverage within an urban areas and across urban areas.
Expected Outputs
TRID
Individuals Involved
| Email |
Name |
Affiliation |
Role |
Position |
| coifman.1@osu.edu |
Coifman, Benjamin |
The Ohio State University |
Co-PI |
Faculty - Tenured |
| mccord.2@osu.edu |
McCord, Mark |
The Ohio State University |
PI |
Faculty - Tenured |
| mishalani.1@osu.edu |
Mishalani, Rabi |
The Ohio State University |
Co-PI |
Faculty - Tenured |
| redmill.1@osu.edu |
Redmill, Keith |
The Ohio State University |
Other |
Faculty - Research/Systems |
Budget
Amount of UTC Funds Awarded
$191112.00
Total Project Budget (from all funding sources)
$498864.00
Documents
| Type |
Name |
Uploaded |
| Data Management Plan |
dmp-McCord-2020.docx |
Jan. 6, 2020, 1:34 p.m. |
| Progress Report |
318_Progress_Report_2020-03-31 |
March 27, 2020, 2:16 p.m. |
| Progress Report |
318_Progress_Report_2020-09-30 |
Sept. 26, 2020, 8:45 a.m. |
| Publication |
Municipal Vehicles as Sensor Platforms to Monitor Roadway Traffic |
Feb. 24, 2021, 7:48 a.m. |
| Presentation |
Time-of-Day_Traffic_Volumes_Using_Video_Imagery_Obtained_from_Transit_Buses_in_Regular_Operation_OSU_McCord_et_al_2020-10-30.pdf |
March 30, 2021, 6:27 a.m. |
| Presentation |
Estimating_Traffic_Volumes_from_Bus-based_Video_OSU_McCord_et_al_2020-11-20.pdf |
March 30, 2021, 6:27 a.m. |
| Progress Report |
318_Progress_Report_2021-03-31 |
March 30, 2021, 6:28 a.m. |
| Publication |
MS_Thesis_ShahrzadCharmchiToosi_OSU_2021.pdf |
Oct. 6, 2021, 2 p.m. |
| Progress Report |
318_Progress_Report_2021-09-30 |
Oct. 6, 2021, 2:03 p.m. |
| Presentation |
Smart_Mobility_Connection__McCord_et_al_2022-02-11_revised.pdf |
March 30, 2022, 10:17 a.m. |
| Presentation |
McCord_63-2.pdf |
March 30, 2022, 10:17 a.m. |
| Presentation |
Smart_Mobility_Connection__McCord_et_al_2022-02-11_revised_vthdOf8.pdf |
March 30, 2022, 10:17 a.m. |
| Presentation |
McCord_63-2_w1pPs6b.pdf |
March 30, 2022, 10:17 a.m. |
| Progress Report |
318_Progress_Report_2022-03-30 |
March 30, 2022, 10:19 a.m. |
| Progress Report |
318_Progress_Report_2022-09-30 |
Sept. 30, 2022, 8:44 a.m. |
| Progress Report |
318_Progress_Report_2023-03-31 |
April 5, 2023, 11:51 a.m. |
| Presentation |
McCord_Mishalani_Shah_TRB_Poster_2023_01_09.pdf |
April 7, 2023, 8:38 a.m. |
| Presentation |
McCord_Mishalani_Coifman_Shah_Galdino_CMU_Poster_2022_11_3.pdf |
April 7, 2023, 8:38 a.m. |
| Final Report |
Final_Report_-_McCord_318.pdf |
Aug. 3, 2023, 9:31 a.m. |
| Publication |
Investigating_Temporal_Patterns_in_Traffic_Volumes_Obtained_using_Video_Imagery_from_Buses_Operating_in_Regular_Service.pdf |
Oct. 15, 2023, 6 p.m. |
| Presentation |
McCord_Mishalani_Shah_WCTR_2023_Presentation__3459.pdf |
Oct. 15, 2023, 6 p.m. |
| Progress Report |
318_Progress_Report_2023-09-30 |
Oct. 15, 2023, 6 p.m. |
Match Sources
No match sources!
Partners
| Name |
Type |
| The Ohio State University Transportation and Traffic Services |
Deployment Partner Deployment Partner |
| Mid-Ohio Regional Planning Commission |
Deployment Partner Deployment Partner |
