Incorporate Probe Data to Expand the Range of Smart Work Zones and Provide Measurements from Locations Beyond the Range of the Sensors.

Proof-of-concept Study Assessed the Feasibility of Using Internet of Things and Computer Vision Technologies To Improve Roadway Work Zone Safety.

Date Posted

Using IoT Technology to Create Smart Work Zones

Summary Information

Work zone crashes pose a significant problem due to their impact on safety and mobility. This study assessed the feasibility of using Internet of Things (IoT), artificial intelligence (AI), and computer vision technologies to improve roadway work zone safety. Two proof-of-concept systems were developed: (i) a work zone intrusion warning system and (ii) a vehicle queue detection system. The work zone intrusion alert system comprised a mobile device attached on a tripod to monitor the restricted area and that run a software application designed to alert workers when an intrusion occurs. The workers received alerts instantly through sounds and vibrations generated by their mobile devices. The findings of the tests indicated its potential to provide a robust technical approach to improving work zone safety. The queue detection system was designed to detect stopped vehicles inside a polygon area selected by the user to denote the work zone area. Both systems utilized AI-based object detection methods and were tested using publicly available highway traffic videos as the simulated test environment. The developed systems demonstrated the feasibility of building low-cost AI and computer vision-based systems that could be used to build smart work zones.

  • Expand the range of work zones by using probe data to fill in detection gaps. Incorporating probe data with sensor data can provide more accurate queue warnings, reduce work zone safety costs, and improve system reliability and efficiency.
  • Integrate mobile apps with smart work zone technologies. Operators should disseminate roadway information through mobile apps to improve the efficacy and expand the range of work zone alerts. Mobile apps can address the limitations posed by dynamic message displays, which require considerable costs per unit.
  • Further explore the feasibility of IoT, AI, and computer vision technologies for future work zone intrusion detection systems. Tests found that the two proof-of-concept systems could be easily deployed at test sites, suggesting that they should be further developed. The developed systems should then be evaluated at a work zone site to identify areas for improvement.
  • Consider the capabilities of radar-based sensors and thermal imaging solutions when improving smart work zone technologies. Work zone intrusion warning systems and queue detection systems could be improved to perform more reliably at nighttime and under unfavorable weather conditions. These technologies may be the most beneficial; they should be assessed for short and long-term deployment.