Intelligent Inspection and Warning Robotic System for Onsite Construction Safety Monitoring Using Computer Vision and Unmanned Ground Vehicle

Worker safety is a critical factor to construction success and should be properly monitored and managed at jobsites. While many vision-based worker safety inspection/monitoring systems were developed by previous studies, they commonly suffer from low mobility of stationary cameras and the lack of taking real-time actions. To address these challenges, this paper proposes an intelligent inspection and warning robotic system using an unmanned ground vehicle (UGV). This robotic system (1) automatically and movably detects construction workers and personal protective equipment (PPE) using state-of-the-art YOLOv8 architecture deep learning-based computer vision model, and (2) dynamically warns/reminds workers of wearing the undetected required PPE. The developed system includes (1) a robotic vehicle prototype to provide mobility, (2) a high-resolution camera to collect visual data, (3) a speaker for auditory warning/reminder information, and (4) a single board computer for real-time data processing. The proposed system was tested at a real construction site. Field test results showed that it can reliably detect construction workers and their PPE and then play voice messages to remind them to wear the required PPE when it is not detected. Ultimately, this paper contributes to the body of knowledge by developing an intelligent UGV-based system for improving onsite construction safety management.