Implementation of a Real Time Data Driven Smart Work Zone Management System to Provide Queue Alerts to Stakeholders

Queueing frequently occurs in advance of interstate work zones, and back-of-queue crashes are of concern to all agencies. Over 30,000 crashes on 2,280 directional miles of Indiana interstates in calendar years 2014 and 2015 were reviewed to determine if they were associated with interstate queueing. Crashes on interstates were found to be 20-24 times more likely to occur when a queue is present. For almost 75% of the identified back-of-queue crashes, queues existed for 15 minutes or more, suggesting that improved alerting to drivers has significant potential to reduce back-of-queue crashes. This increased crash rate is cited as a strong motivator for agencies to invest in queue warning systems. This paper reports on a real-time probe vehicle data-based system for generating texts/emails to alert the Indiana Department of Transportation to interstate work zone queueing. Six work zones were identified for this study, with 10-mile upstream approach boundaries. When average speeds drop below 45 MPH, queue monitoring algorithms are triggered. When this queue grows to be at least 1 mile long or has a speed drop of at least 15 MPH at the back of the queue, an alert is sent to selected individuals. Still camera images, work schedules, and crash reports were used to ground-truth the alert system. The paper concludes by recommending the use of probe vehicle data for queue monitoring due to the scalable, cost effective nature of the technology and the ability to rapidly deploy monitored areas.