Signal Processing and Alert Logic Evaluation for IoT-Based Work Zone Proximity Safety System

Construction projects are dynamic by nature because of continuously moving resources such as heavy equipment and workers. This nature necessarily results in proximity hazards, especially on a large-scale construction site. Especially, struck-by accidents still account for about 38% of the total injuries in the US construction industry. Although there have been several efforts to mitigate the hazards, the statistics show that the hazards still persist. To provide a practical solution to this problem, this study proposes an Internet of Things (IoT)-based proximity warning system that provides an alert to workers whenever they are close to heavy equipment. The system includes equipment protection units (EPUs), personal protection units (PPUs), and Bluetooth low energy (BLE) beacons. A framework of signal processing and alert logic was developed to estimate the distance between PPUs and EPUs and to activate alerting modules timely. By calculating the distance based on the signal strength using a particle filtering method, EPUs and PPUs provide auditory and vibration alerts to equipment operators and workers when they are in an alert range. Also, practical alert logic was developed based on the site worker's feedback. This study validates the system performance with different signal processing methods and alert logic through five real-world field tests. The system achieved a precision, recall, and F1-score of 89.21%, 97.45%, and 0.931 from the field tests, respectively. Also, positive feedback was obtained from the participating workers. The proposed IoT-based proximity warning system has a high potential for a practical solution to proximity hazards in actual construction sites.