FW-HTF-RM: Improving Construction Work Performance through Human-Centered Augmented Reality

美国建筑业占国内生产总值的近15%，占总劳动力的6%以上。建筑业负责更新美国的基础设施，以满足日常对清洁水、交通和生活空间的需求。最近，建筑业经历了严重的劳动力短缺，这促使人们需要向工艺工人提供最优信息，以高效、安全的方式建设基础设施，并促进未来工人的社会和经济福祉。越来越多的文献证明了增强现实（AR）和人工智能（AI）在改变工作场所方面的令人兴奋的潜力，但大多数现有研究都集中在办公室、工厂和医疗工作者身上。本项目旨在探索这些技术是否以及如何具体地改善施工中的工作绩效。在目前的建筑行业实践中，设计信息通过二维平面图和书面规范提供给建筑工人。新技术开始使设计信息能够在多维模型中表示，这些模型可以模拟施工过程并有助于可视化。随着AR和AI技术的成熟，人们认为它们可以通过增强信息获取来促进更有效的建设。然而，目前尚不清楚如何将创新信息与AR一起呈现，以最好地支持未来的建筑工作。此外，由于人工智能被考虑用于创建项目信息，我们必须了解来自人类或人工智能的设计感知来源如何影响用户对信息的信任和对后续决策的自信。在这个项目中，研究人员将测试两个主要假设：(1)AR中提供的设计信息的细节水平如何以及在多大程度上影响工作绩效的维度；(2)设计信息的感知来源如何以及在多大程度上影响对信息的信任和对决策的自信。次要假设将测试细节水平和设计的感知来源如何影响空间推理；计划的方法和时间安排；以及对错误和安全隐患的预测。这些假设将通过一个实验进行测试，在这个实验中，设计信息的详细程度被操纵，错误、安全隐患和不确定性被有意地嵌入到测试任务中。实验将包括一个实际建筑项目的比例模型。在研究条件下，将有200名工艺工人和200名本科生参与研究实验，分别代表专家和新手的技能水平。除了定量分析外，在每次实验试验结束后，将对研究参与者进行访谈，以解释结果。这项研究将有可能建立新的知识，关于什么以及如何使用新兴技术将设计信息最好地传递给未来的工艺工人，以实现高效、安全和高质量的施工。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

The construction industry in the United States accounts for nearly 15% of Gross Domestic Product and over 6% of the total workforce. The construction industry is responsible for renewing US infrastructure to meet daily needs for clean water, transportation, and living space. Recently, construction has experienced significant workforce shortages that have precipitated the need to deliver craft workers with optimal information to construct infrastructure efficiently, safely, and in a manner that promotes future workers' social and economic well-being. A growing body of literature has demonstrated the exciting potential of augmented reality (AR) and artificial intelligence (AI) to transform workplaces, but most existing studies focus on office, factory, and medical workers. This project aims to explore if and how these technologies specifically improve work performance in construction. In current construction industry practice, design information is provided to construction workers through two-dimensional plans and written specifications. New technology is beginning to enable design information to be represented in multi-dimensional models that simulate the construction process and aid with visualization. As AR and AI technologies mature, it is posited that they can promote more effective construction via enhanced access to information. However, it is unclear how innovative information should be presented with AR to best support future construction work. Additionally, as AI is considered for the creation of project information, we must understand how the perceived origin of a design from either humans or AI impacts the user's trust of the information and self-confidence in subsequent decisions.In this project, researchers will test two main hypotheses: (1) How and to what extent the level of detail of design information delivered in AR impacts dimensions of work performance and (2) How and to what extent the perceived origin of design information impacts trust in the information and self-confidence in decisions. Secondary hypotheses will test how level of detail and perceived origin of design impact spatial reasoning; methods and timing of planning; and anticipation of errors and safety hazards. These hypotheses will be tested via an experiment where the level of detail of design information is manipulated and errors, safety hazards, and uncertainty are purposely embedded in the test task. The experiment will involve a scale model of an actual construction project. Two hundred craft workers and two hundred undergraduate students will participate in the research experiment to represent expert and novice skill levels in the study conditions. In addition to quantitative analyses, interviews will be conducted with the study participants after each experimental trial to explain the results. This research will have the potential to build new knowledge on what and how design information can be best delivered to future craft workers using emerging technologies to enable efficient, safe, and high-quality construction.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.