A Computational Framework For Robotic Quality Assessment and Management In Construction

As an integrated process in construction projects,quality assessment and management (QA&M) can be importantto prevent failures during construction. The existing QA&Mpractice such as the evaluation of the geometric tolerance andsurface qualities is mostly performed manually which can belabor-intensive and tedious. This study proposes a computationalframework for a robot to perform automatic QA&M in unknownenvironments. The framework is composed of three parts: (1)motion planning; (2) defect detection; and (3) defect registration.The motion planning component generates efficient robotic pathfor autonomous exploration and surface inspection. The defectdetection component quantifies surface anomalies within a userdefinedarea of interests through multiple sensor measurements.The defect registration component localizes the detected defectsand registers the defects to a site model. To demonstrate thefeasibility of the proposed framework, we present a user case forassessing geometric tolerance and surface quality of a 1500 mm(L) x 745 mm (W) x 1980 mm interior wall mockup. The resultof the case study shows that the proposed framework has thepotential to provide reliable geometric measurement and defectdetection for gypsum wall panels in a lab environment.