FW-HTF-P: Reshaping Construction Work Conventions: Endowing Collaborative Construction Robots with Social Intelligence for Contextually-Appropriate Robot Behaviors

Georgia Tech Research Corporation, Atlanta GA

Investigators

Abstract

Construction robots have demonstrated potential benefits to alleviate persistent issues such as stagnant productivity, low profitability, high accident rates, and labor scarcity, and the construction robot market continuously grows at a rate of over 10%. As more and more onsite construction robots are expected to be introduced to jobsite to work with humans, a new imperative objective for a new workplace is to assure the safety of human co-workers. At the construction jobsites, construction workers are constantly in contact with other workers and equipment. To avoid safety risks arising from this dynamic environment, workers have established a shared social norm to respect other workers' work conventions, and every individual is responsible for other workers' safety. However, most of the current construction robots do not understand such work-related norms and are incapable of performing human-like actions or adapting to dynamic human behaviors. To address this issue, this project development grant will support the development of robot social intelligence to understand and abide by work-related social norms in the construction work context. Such intelligence is required to enhance robots' capacities to understand human behaviors, adaptively act upon human behaviors properly, and in turn, acquire trust and acceptance from human co-partners. This research can establish a fundamental understanding of the complex human-robot interactions and trust dynamics in the construction work context. The methods and findings resulting from this research can be extended to human-robot collaboration in various fields outside the construction industry to build a safe and effective human-robot collaborative workplace policy that assures improved safety and productivity. This project development grant aims to explore socially-appropriate robot motion planners and examine whether these behaviors can affect the human-robot trust dynamics. The proposed activities for this project include 1) understanding work contexts that are essential for understanding and predicting human state and intent, 2) exploring social norm-appropriate robot motion controllers for social navigation and collaborative manipulation tasks, and 3) analyzing human trust via trust calibration and negotiation. These activities will facilitate and expedite the development of socially compliant robots to function adaptively and efficiently through improved recognition of human workers' intentions and social norms. The team has formed a unique interdisciplinary group of researchers across civil engineering, computer science, and robotics to achieve this convergent research objective. This research team will bring together diverse disciplinary perspectives to support convergence research while contributing to their domains. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.