Collaborative bricklaying with a robotic arm using a head-mounted camera in construction

Construction automation is a promising solution to the shortage of skilled workers and the high risk of accidents on construction job sites. However, the automation of construction work still needs improvements due to the dynamic nature of construction environments. Specifically, many construction tasks are performed manually; thus, construction robots need to be designed to collaborate with construction workers. In this context, this study proposes an approach using a head-mounted camera for collaborative bricklaying with a robotic arm in construction environments. The head-mounted camera is capable of capturing first-person view images, thereby avoiding potential visual obstructions caused by carrying or nearby objects during collaboration in construction environments. The approach defined the collaboration steps in performing bricklaying tasks with a robot and identified these steps using a real-time object detection algorithm (i.e., YOLO). A laboratory experiment simulating bricklaying tasks was conducted for data collection, and the proposed approach achieved a mean average precision of 0.968 in classifying five collaboration steps. The results demonstrate that the head-mounted camera-based approach is a reliable solution for operating human-robot collaboration systems. The findings of this study contribute to encouraging the use of collaborative robots to perform repetitive and high-risk construction tasks that can help reduce the physical stress of construction workers.