3. Review methodology

This research leverages the content analysis-based review method

[28]. This method has been a well-recognized method for reviewing

and synthesizing literature and rationalizing outcomes, and been

widely applied in the researchfifield of engineering/construction man

agement[29,30]. Scopus and ISI Web of Science, the abstract and ci

tation database of peer-reviewed literature, were used to select a

number offifirst-tier articles that are related to VR/AR-CS, from the

authoritative and well-acknowledged scholarly journals in selected

areas such as engineering/construction management, science, and

technology, safety, human factors, etc. The selection criteria are as

follows:

(1) The selected journals should be retrieved from the Science Citation

Index (SCI)-Expanded database or Engineering Index (EI)

Compendex database because they have indexed more than 6500

notable and signifificant journals from more than 150 disciplines.

(2) The selected journals should have signifificant impacts on the tar

geted areas, which could be reflflected from the indexes of citations

or H-index. Book reviews, editorials, and paper in conference pro

ceedings were eliminated. This is to ensure that all retrieved papers

could be investigated using an identical analytical construct in

terms of research aims and methodologies.

To achieve a target and structured review towards the literature

(from 2000 to 2017), a two-stage review method is presented. In stage

1, an exhaustive search was carried out using the Scopus, Google

Fig. 3.Offff-the-shelf VR/AR Systems and Technology Providers.

Fig. 4.VR/AR Hardware and Software Diagram.

X. Li et al.

Automation in Construction 86 (2018) 150–162

152Scholar and Web of Science search engines. Several keywords such as

VR, AR, mixed reality, virtual prototyping, gaming technology, im

mersive, interactive, construction, safety, etc. were attempted so that

the search could cover a broad range of related disciplines. This round

of search identifified 246 most related articles from the below-listed

journals:

•

Automation in Construction (AIC),

•

Journal of Computing in Civil Engineering (JCCE),

•

Advanced Engineering Informatics (AEI),

•

Construction Management and Economics (CME),

•

Journal of Construction Engineering and Management (JCEM),

•

Applied Ergonomics (AE),

•

Ergonomics (EG),

•

Safety Science (SS),

•

Accident Analysis and Prevention (AA&P)

•

Journal of Information Technology in Construction (ITCon)

Publications which do not comprise the above-mentioned keywords

in their titles or abstracts were screened out in stage 2, and the less

relevant and irrelevant papers after a brief visual examination of the

content of the article, leaving a total of 90 publications for further

analysis. The two-stage search may not guarantee a comprehensive

coverage of the papers that are worth reviewing. However, such an

approach suffiffiffices for providing with a considerable amount of sig

nifificant state-of-the-art works, from which the study could generalize

fifindings and recommend future works.

4. Justifification of publication quantity and contribution

Since the emergence of thefifirst VR/AR-CS publication in 2000, the

number of yearly publications remained at around 1.5 and was not

observed a noticeable increase until 2008. From 2008, a signifificant

increase of publications was witnessed, featuring an average amount of

8 papers per year till 2017 (Fig. 5). Among these journals, three con

struction-technology-themed journals (namely, AIC, JCCE and ITCon)

are observed the highest amounts of publications, i.e., 30 for AIC, 12 for

JCCE, 8 for ITCon, 7 for JCEM, 7 for AEI, 5 for ITCon, 3 for SS and 3 for

EG. The number of VR/AR-CS papers published in AIC was thereto

considerably greater than any other selected journals, resulting in the

most contributions of this specifific journal to VR/AR-CS studies.

The difffference of publication quantities among various countries

may imply the extent to which research commitment and value are

observed in these countries. The countries of origin of VR/AR-CS

publication as shown inFig.6are listed alongside the factors of pub

lication quantity. American authors and research institutions leading

the papers contributed 30 papers and thereby scored the highest. A

reasonable justifification would be U.S. has been leading the research

and technology advancement in visualization and informatics for years.

Behind the US, China takes up the second position, leading the rest of

the second-most productive countries/regions such as Australia, Re

public of Korea, and Taiwan. As the total amount of VR/AR-CS papers

from thesefifive key contributors account for 80% of the total publica

tion scale (72 of 90), the contributions to VR/AR-CS research are

thereby perceived as more signifificant than the remaining contributors.

The development of VR/AR-CS research also indicates that the degree

of attention to construction productivity, safety, technological innova

tions and reforms within these countries.

5. Implementation of critical review

As the review deals with numerous papers, it is important for the

methodology to be predicated upon a rational taxonomy that can help

direct the review focus and elicit the valuablefifindings into the area of

research. This paper presents a review taxonomy consolidated by four

specififications, namely, VR/AR technology characteristics incorporating

input and output; VR/AR application domains in safety management;

safety enhancement mechanisms; and safety assessment and evaluation

(Fig.7). Designing proper VR/AR technologies for safety enhancement

is not easy, and always involves a number of technology trade-offffs. It's a

fifine balancing act. A more immersive user experience requires more

interactive human-machine interfacing, necessitating the use of more

advanced hard and software with more powerful algorithms working at

a higher load. A glimpse of VR/AR technology characteristics could

understand the opportunities and challenges that the offff-the-shelf VR/

AR applications still need to catch up or stand a chance of overcoming.

By looking into the VR/AR application aspects and safety enhancement

mechanisms, we could understand how effffectively the reviewed ap

plications could improve the safety outcomes of real construction pro

jects, and how effffectively the VR/AR experiences could enable users to

visualize and recognize complex workplace situations, from which the

knowledge of procedures and skills could be built up, and tasks in a safe

and forgiving environment be carried out. Safety enhancement me

chanisms, when addressed from the perspectives of ergonomics, psy

chology, educational pedagogy, human perception, and cognition, may

elicit knowledge about how VR/AR-based safety training programs can

offffer a more pertinent working environment where users can effffectively

rehearse hazardous factors and ultimately promote the abilities for

hazard cognition and prevention. The analysis of a broad spectrum of

evaluation methods allows us to extract generalfifindings from the re

view work, which in turn could be applied to refifine the existing sys

tems, design new systems, and generate guidelines for formulating the

best safety enhancement paradigms.