Robotics in Construction: A Critical Review of Reinforcement Learning, Imitation Learning, and Industry-specific Challenges for Adoption

Abstract

This paper provides a comprehensive review of the role of robotics in addressing critical construction industry challenges, including labor shortages, productivity inefficiencies, and safety concerns. The review focuses particularly on reinforcement and imitation learning as promising yet underexplored paradigms within construction robotics. While these approaches have shown transformative potential in other fields, their application in construction is constrained by the sector’s unstructured, dynamic environments, which demand specialized and adaptable robotic systems. The review further examines the role of deep learning in advancing construction machinery, emphasizing applications in perception, navigation, control, and human-robot interaction. However, challenges persist in the form of limited datasets, interpretability issues, and the need for higher levels of autonomous intelligence. Additionally, the paper identifies industry-specific adoption barriers, such as economic, technical, and cultural factors, which continue to hinder robotics integration in construction. By presenting these insights, this review establishes a foundation for understanding the current state of construction robotics and the opportunities for advancing its application across the industry.