Collaborative Research: Measuring, Predicting, and Improving Construction Safety by Improving Hazard Signal Detection with Augmented Virtual Environments

与其他行业相比，建筑行业每年的受伤率不成比例。造成这些伤害的主要原因之一是人为失误，而与信号检测相关的技能不足（定义为个人检测表明存在安全隐患的视觉刺激的能力）可能会加剧这种失误。从心理学和工程学的角度来看，人们对信号检测知之甚少，也不知道这项技能能否被准确地测量、预测或利用新兴技术加以改进。该奖项支持跨学科研究，调查信号检测是否可以通过增强虚拟建筑环境的暴露和反馈来可靠地改进。此外，该研究还探讨了个人在虚拟环境中的信号检测技能是否可以预测他们在真实建筑环境中的技能。随着新知识的获得，将通过向国家实践社区介绍与行业从业人员分享，将其纳入土木工程和心理学课程的多个层次，并嵌入将广泛传播用于教育和安全培训的虚拟系统的最终版本。本研究的主要目的是验证以下假设：增强虚拟环境可用于将信号检测技能提高30%以上，并且现场环境中信号检测技能的一半以上可变性可以通过增强虚拟系统中的表现来预测。这些假设将使用从制药研究中改编的多基线设计来测试，以解决建筑行业的动态特性。实验结果将使用中断时间序列回归模型进行统计验证。如果成功，多基线设计的适应和统计验证将为建筑实验研究提供新的途径。此外，在虚拟环境和现场环境中的表现之间的密切关系表明，高保真增强虚拟环境为测量信号检测技能提供了无风险的代理，可能会改变施工安全研究的方式。

Each year the construction industry has a disproportionate injury rate compared to other industries. One of the main causes of these injuries is human error, which could be exacerbated by skill deficiency associated with signal detection (defined as the ability of an individual to detect visual stimuli that indicate the existence of a safety hazard). From psychological and engineering standpoints little is known about signal detection or if this skill can be accurately measured, predicted, or improved with emerging technologies. This award supports interdisciplinary research that investigates if signal detection can be reliably improved through exposure to and feedback from an augmented virtual construction environment. Additionally, the research explores if an individual's signal detection skill in a virtual environment can predict their skill in real construction environments. As new knowledge is gained, it will be shared with industry practitioners through presentations to national communities of practice, integrated into multiple levels of civil engineering and psychology curricula, and embedded in the final version of the virtual system that will disseminated broadly for education and safety training.The main objective of this research is to test hypotheses that augmented virtual environments can be used to improve signal detection skill by over 30% and that over half of the variability in signal detection skill in field environments can be predicted by one's performance in an augmented virtual system. These hypotheses will be tested using a multiple baseline design that has been adapted from pharmaceutical research to address the dynamic nature of the construction industry. The experimental results will be statistically verified using interrupted time-series regression models. If successful, the adaptation of multiple baseline design and statistical verification would serve as a new approach of experimental research for construction. Additionally, a strong relationship between performance in virtual and field environments would indicate that high-fidelity augmented virtual environments provide a risk-free proxy for measuring signal detection skill, potentially transforming the way that construction safety is studied.