Exploring the Impact of Augmented Reality on Work Performance in a Full-Scale MEP Assembly Task: Study of Industry and Novice Populations

Building information modeling (BIM) has revolutionized the construction industry; however, field personnel still rely on traditional methods for design interpretation. Augmented reality (AR) head-mounted display devices (HMDDs) offer a promising alternative for delivering three-dimensional design information, yet their effectiveness must be validated across diverse populations with a significant sample size. This study examines the impact of AR HMDDs on mechanical, electrical, and plumbing (MEP) assembly tasks, evaluating key performance metrics: task completion time, rework rate, and error rate. Both novice and experienced participants completed assembly tasks in industrial and laboratory settings, marking the first comparative analysis of AR models with different levels of detail (LOD 300 and LOD 400) against traditional isometric paper plans. Findings indicate that the AR LOD 400 model significantly improved all performance metrics across both populations, except for the error rate among industry professionals. The AR LOD 300 model notably reduced rework rates, while traditional paper plans were the least effective. Age did not significantly impact performance, whereas higher spatial cognition enhanced novices' efficiency. Participants acknowledged AR HMDDs, particularly the LOD 400 model, as beneficial for design comprehension; however, some also reported distractions. These insights highlight the need for user-centered AR interface design and tailored training strategies to enhance usability and efficiency. As the first study to investigate the influence of AR model LODs on construction task performance with a robust sample size, the results provide valuable guidance for optimizing AR HMD technology, training protocols, and design information delivery in the construction industry.