Incident Identification System for Missouri’s I-270 Corridor Reduced Incident Response Times by an Average of 10 Minutes.
Three Advanced Technologies Deployed During Construction on I-270 in Missouri To Predict and Respond to Incidents.
St. Louis, Missouri, United States
I-270 Predictive Layered Operations Initiative
Summary Information
The I-270 Predictive Layered Operations Initiative (PLOI) was an effort by the Missouri Department of Transportation (MoDOT) to improve traffic management and enhance roadway safety along the I-270 corridor, implemented as part of the broader I-270 North Design-Build Project. This major roadway reconstruction effort incorporated strategies intended to maintain safety and traffic operations during construction. The project covered an 8.6-mile segment of I-270 and was carried out between 2019 and 2023. Funded in part by an Advanced Transportation and Congestion Management Technologies Deployment (ATCMTD) grant, the PLOI included three advanced technologies: (1) Crash Risk Location Prediction and Incident Identification, to predict the location and timing of incidents using machine learning and data from connected vehicles, weather forecasts, and special events, among other sources, (2) Advanced Video Analytics, to identify and respond to incidents more quickly using analyzed CCTV camera feeds, and (3) Integrated Modeling for Road Condition Prediction (IMRCP), to refine weather and traffic forecasts to decrease weather related crashes.
METHODOLOGY
MoDOT set performance metrics for each technology and identified the data sources linked to each performance measure, as well as any data limitations or constraints. The agency also undertook a verification process for the tools, to determine whether they were providing high-quality information. Data used for the verification process were collected quarterly in 2022 (February, April, July, and October) and in 2023 (January, April, July, and October).
FINDINGS
- The incident identification system notifications preceded verified Advanced Traffic Management System (ATMS) events by 45 percent of the time, with an average of 10 minutes saved overall.
- The total percentage of verified true alerts using advanced video analytics ranged from 70 percent to 100 percent in 2022, and 66 percent to 98 percent in 2023.
- A benefit-cost (B/C) analysis of the crash risk location prediction tool found a B/C ratio of up to 12.33, assuming 20 percent accuracy and 30 percent police deployment.
- Combined with the benefits from the incident identification system and advanced video analytics, the crash risk location prediction tool’s B/C ratio was as much as 16.81, assuming 20 percent accuracy and 30 percent police deployment.
