This study explores the feasibility of improving road work zone safety by using state-of-the-art Internet of things (IoT), artificial intelligence (AI), and computer vision technologies. This project included an in-depth analysis of the key technologies and methods that have the potential to improve work zone safety. The report also explains and illustrates how existing and emerging work zone safety systems and methods are typically implemented through a number of cases. It also reports the development of two proof-of-concept systems geared toward the needs of today's smart work zones and the evaluation of their effectiveness and reliability through lab experiments.
In order to gain an understanding of the major triggers of the most harmful crashes in work zones, the project team analyzed crash data from North Carolina. The team also conducted a thorough literature review to determine the current state of practice in smart work zone implementations in the United States together with the technical capabilities of the most prominent products on the market. The findings of those research activities led the team to focus on two core smart work zone elements: queue detection and work zone intrusion detection. Queue detection is a key technology in many smart work zone applications, such as dynamic lane merge systems and queue warning systems. Intrusion detection is a key element of systems that protect workers from vehicles entering into restricted work areas. This report identifies three commercially available devices that can have the greatest potential to be used in the field as part of a smart work zone to improve the work zone safety.
Driven by the insights gained through the literature review and from the analysis of the North Carolina work zone crash data, two proof-of-concept systems were developed using IoT, AI, and computer vision technologies for work zone safety. The developed systems provide capabilities for two functions: (1) work zone intrusion warning and (2) vehicle queue detection. The first of these systems is a proof-of-concept intrusion alert system, comprising a mobile device attached on a tripod to monitor the restricted area and that runs a software application designed to alert workers when an intrusion occurs. The workers receive alerts instantly through sounds and vibrations generated by their mobile devices. The system was tested in a simulated test environment and the findings of the tests indicated its good potential to provide a robust technical approach to improving work zone safety. The second system is a proof-of-concept queue warning system, which was also developed and tested as part of this project. The results indicated it had significant potential to be used in smart work zones as a low-cost and easy-to-deploy system. Both systems were implemented to run on Android smartphones. However, the software is extremely portable, and therefore the efficient technical design means it can be embedded in any type of hardware.
