【Authors】Nan Li1,2,*, Jing Du3, Vicente A. González4, and Jieyu Chen1
1 Dept. of Construction Management, Tsinghua Univ., Beijing, China
2 Hang Lung Center for Real Estate, Tsinghua Univ., Beijing, China
3 Dept. of Civil and Coastal Engineering, Univ. of Florida, Gainesville, U.S.
4 Dept. of Civil and Environmental Engineering, Univ. of Alberta, Edmonton, Canada
【Corresponding e-mail】nanli@tsinghua.edu.cn
【Journal information】Journal of Construction Engineering and Management, Volume 148, Issue 10
【DOI】https://doi.org/10.1061/(ASCE)CO.1943-7862.0002367
【Abstract】
Extended reality (XR) technologies are increasingly being used as a novel research instrument to facilitate scientific inquires in the construction engineering and management (CEM) domain. By allowing humans to interact with immersive environments in controlled and monitored experimental settings, XR technologies have opened new opportunities for researchers to conduct CEM research involving human participants or concerning human behavior. Yet, XR-enabled research, as an independent, rigorous methodology for the CEM domain, is still underexplored. This paper serves as an effort to build an organized knowledge base and workflow for using XR technologies in various CEM research areas and methodological contexts. The paper first investigates the status quo of XR-enabled CEM research, by identifying current research areas in the CEM domain where XR technologies are considered the preferred or recommended methodological solutions. A process model for XR-enabled research is then proposed, with actionable recommendations about how XR-enabled research should be planned, designed, implemented, analyzed, verified, and validated. This process model is demonstrated with two illustrative case studies. Last, the paper discusses the philosophical, methodological, and technological roots of the evolution of XR-enabled CEM research and describes our vision of more enabling, adoptable, and value-adding XR-enabled research in CEM in the near future.
Keywords: Extended reality (XR); Experimental research; Methodology; Process model; Construction engineering and management (CEM)
【Process model】
The goal of a scientific experiment is to establish empirical evidence of a relationship between an independent or experimental variable and a dependent variable that is affected by it. To illustrate the input-output nature of a scientific experiment and the flow of phases that need to be followed to achieve this goal, we developed a process model for XR-enabled experiment in CEM research, as illustrated in Fig. 1. The developed process model consists of four sequential phases, including needs identification, XR environment design, experimental design, and data collection and analysis, and an additional phase of verification and validation that applies to the entire process. The activities that need to be undertaken in each phase, and the particular concerns and considerations associated with each activity, are discussed in detail. Table 1 is proposed to guide CEM researchers in the validation process, suggesting the key question to be tested and the approach to test it in each of the validity and fidelity categories.
Fig.1 The process model for XR-enabled experiment.
Table 1. Validity and fidelity categories, and overall driving questions and testing approaches.
【Cases】
Two illustrative examples of XR-enabled CEM research were presented. We use these examples, each of which addresses a different problem in the CEM domain and uses XR technologies in a different way, to demonstrate how every phase and activity outlined in the last section can be implemented in the context of specific use cases. The first one is VR for exploring cognition load in altered and stressful construction tasks and the second one is VR serious game for investigating earthquake behavioral responses and preparedness in buildings.
Fig.2. The VR environment in case one.
Fig. 3. An example of two alternative actions for participants to choose how to exit the building.
【Discussions and conclusions】
There is an increasing trend of adopting XR technologies in CEM research as an enabling tool to conduct lab or field experiments that involve human participants or concern human behaviors. This trend, according to our observation, can be largely attributed to its philosophical, methodological, and technological roots. From a philosophical perspective, many problems concerned in CEM research emerge at the interface between engineering and the social sciences. Given its interdisciplinary nature, the CEM domain has always faced a pressing need to investigate human and organizational behaviors within complex engineering contexts, and to understand engineering processes under the influence of human and organizational influences. However, it is not uncommon that CEM researchers find it highly difficult, and even prohibitive, to develop the necessary engineering contexts or reproduce the accurate human and organizational influences required for their scientific inquiries, for a variety of logistic, technical, and ethnic reasons. Fortunately, the introduction of XR-enabled methodology provides at least a partial solution to the above challenge. Based on a combinatory use of quantitative (deductive) and qualitative (inductive) methods, XR-enabled methodology offers a uniquely enabling approach to flexible, controllable, reproducible, and explainable experiments in CEM research.
We also found that recent trends in XR technology have significantly lowered the barrier of XR adoption in CEM research and improved the access to high-resolution human assessment data. First, XR devices are becoming more readily available for a broader population, making scalable participation in human-subject experiments possible. As such, scholars may consider nontraditional approaches to access the broadening pool of human subjects, such as via crowdsourcing. Second, with the fast development of computer vision technologies, the quality of XR rendering has been substantially improved. The introduction of the new PhysX physics engine has also enabled a more realistic rendering of physical processes in the XR environment. All of these advances mean that XR is capable of reproducing real-world physics in an unprecedented manner.
Third, we also found that recent XR literature is proposing and testing a multisensory integration approach. For example, haptics as new sensory feedback is being widely tested to enable sensory augmentation in XR. All these trends would expand the horizon of XR-based human subject experiments in various construction tasks. Researchers are empowered by more advanced tools to produce complex, immersive, and interactive scenarios that can induce more realistic individual or collective behaviors by their human participants, and to monitor these behaviors in a more seamless, in-depth, real-time, and non-intrusive manner. This is a main reason behind the rapid growth of the volume of published XR-enabled research in recent years.
