Human-centered computing and visual analytics for future of work in construction

Artificial intelligence (AI) is revolutionizing various systems within the Architecture, Engineering, Construction, and Facilities Management (AEC/FM) domains. The rapid advancements in computational methods, engineering knowledge, and sensor technologies is transforming the current construction practices that are heavily reliant on human intervention. Therefore, visionaries are foreseeing that future construction works will be collaborated by humans and machines which will lead to unprecedented socio-economic outcomes in the safety, health, and productivity of construction workers. This Dissertation aims to advance the fundamental knowledge for effectively implementing human-machine collaboration in the construction site. Particularly, the ultimate objective of this Dissertation is to design AI-based autonomous systems for continuously monitoring workplace safety and productivity. Toward this goal, firstly, a content retrieval scheme is designed to analyze a large volume of construction imagery at a rapid speed. Next, an object recognition framework is developed to detect construction-related objects from digital images or videos in real-time. By further extending this framework, an automated safety monitoring system is subsequently designed to verify workers’ compliance with the requirements related to personal protective equipment (PPE). Next, an AI-enabled image enhancement technique is developed to improve the quality of visual data to achieve better performance from the detection models in the preceding steps. Finally, an active vision system is proposed that enables an autonomous camera to intelligently navigates through a jobsite to monitor objects of interest for their safety and productivity.