Smart Work Zone Queue Detection System Achieved Up to 85 Percent System Accuracy and Over 91 Percent System Uptime While Aligning with Industry Cost Benchmarks.
Evaluation of a Queue Detection System Deployed as Part of a Smart Work Zone on I-44 in Missouri.
Springfield
TSMO Strategies and Automation in Work Zones
Summary Information
Effective work zone traffic management reduces delays, protects travelers and workers, supports timely construction, and maintains access for travelers. To advance its Transportation Systems Management and Operations (TSMO) and work zone management priorities, the Missouri Department of Transportation (MoDOT) has been identifying practical strategies that address workforce needs and improve safety. The objective of this study was to evaluate the effectiveness of Smart Work Zone (SWZ) strategies, with the goal of improving work zone safety and operations and applying best practices to construction projects statewide. The evaluation focused on a road-widening and bridge/ramp construction project on Interstate 44 (I-44) in Greene County, from Kansas Expressway/Missouri Route 13 to US Route 65 in Springfield. A queue detection system (with radar sensors and portable changeable message signs), and temporary long-term rumble strips were deployed along an approximately 4.5-mile segment of I-44 in April 2025. The field evaluation took place from May to November 2025.
METHODOLOGY
A queue detection system with eight radar sensors and six portable changeable message signs (CMS) and temporary long-term rumble strips were deployed on the selected segment of I-44. The system was programmed to detect slow or stopped conditions with three or more vehicles, and displaying “SLOW TRAFFIC AHEAD” or “STOPPED TRAFFIC AHEAD” on a CMS depending on detected speeds. After traffic returned to free-flow speeds for three minutes, the message deactivated and reverted to the standard message, “ROAD WORK/XX MILES AHEAD.” Collected data included gap, headway, speed, 85th-percentile speed, vehicle count, and classification by lane from upstream radar sensors in one-minute bins from May to November 2025. During the field evaluation, the data were used along with the timestamps for messages displayed to generate performance metrics such as number of warnings issued, system downtime, false positives, false negatives, queue length, travel delay, driver reaction to rumble strips, and driver reaction to queue feedback. System accuracy was defined as the number of true positives (a change in CMS message as designed) divided by the total of true positives and false positive and negatives, expressed as a percent. System uptime was measured by the proportion of one-minute data points that were present over the course of a day, week, month, and the project duration, for each sensor. Message delay was defined as the amount of elapsed time from the radar-detected speed threshold change to the time the CMS changed.
FINDINGS
- Overall system accuracy was 84.9 percent on the eastbound (EB) and 80.8 percent on the westbound (WB) direction. Accuracy measured correct message changes and excluded periods when sensors were not functioning properly. Results showed that false positives were 4.3 percent for both directions. False negatives were 10.9 percent for the EB direction and 14.9 percent for the WB direction.
- System uptime for the sensors ranged from 91.8 to over 99 percent.
- Median message delay was 55 seconds for the EB devices and 59 seconds for the WB devices.
- Queue lengths of 2.5 miles or longer occurred 17 times during the evaluation period, with most events lasting less than 20 minutes.
- Economic evaluation of the SWZ deployment found that the costs aligned with industry benchmarks, accounting for approximately 3-4 percent of total temporary traffic control (TTC) costs.
