Evaluating Vulnerability to Uncertainty in Human-Robot Collaboration: A Case Study in Drywall Finishing

Abstract

Construction projects are exposed to high levels of logistical uncertainty like material shortages, power outages, and unpredictable weather events. Proactive planning around these uncertainties is complicated by the many resource interactions and interdependencies needed for the completion of most construction tasks. Exposure to uncertainty and a high sensitivity to that uncertainty makes project schedules more vulnerable to delay. The application of human-robot collaboration (HRC) in construction tasks has the potential to reduce uncertainty related to labor but may increase vulnerability in other unexpected ways. This paper explores how HRC in a subset of construction tasks, specifically in drywall finishing, affects project vulnerability to simulated disruptions. Data from jobsite observations and worker interviews are used to develop a meta-network model of the drywall finishing process, which is integrated a semi-automated robot named “Canvas” in an HRC application. The results identify the circumstances under which HRC in drywall finishing makes the project more or less vulnerable to uncertainty. The findings of the research will aid project managers by enabling more resilient planning of HRC applications and provide guidance to robotic manufacturers improving the integration of their systems on construction projects.