A Framework for Augmented Reality Assisted Structural Embedment Inspection

Embedments (embeds) placed in concrete or masonry structures are used extensively in construction to connect the final product of one trade-contractor's work to another and are therefore, a critical coordination facet for most construction projects. A failure in this coordination usually leads to lost productivity. Therefore, the need to productively improve the installation quality is paramount. A between-groups experimental study was designed to measure embed placement accuracy within an experimental space. One group inspected the work with a 2-dimensional set of construction plans while another group carried out the work with the assistance of an augmented reality (AR) inspection tool. An AR headset was used that presented a parametric model as a visual overlay on the walls of the experimental space. In this way, the embed placement accuracy could be inspected. The results indicated that accuracy was weakly significant between the two methods of embed inspection. However, a shortcoming discovered during the research required the precision of the AR tool to be tuned because of an image drift within the AR visualization. This paper analyzes the AR shortcoming, differences in accuracy, proposes reasons for the differences, and addresses the accuracy trade-off in a broader context of the framework.