Highlights

4.1.1. Immersive VR

IVR utilizes specialized hardware to provide immersive sensory experiences to users. Such hardware includes head-mounted devices (HMDs) and sensor gloves (Wang et al., 2018). Waly and Thabet (2003) demonstrated IVR by developing the CAVE. Based on the user’s location, the user is immersed in a virtual world. The user’s position in the virtual environment varies by his/her position in the real world (Wang et al., 2018). The degree of immersion is the primary factor distinguishing an HMD and CAVE VR training session from a desktop computer VR training session (Makransky and Petersen, 2021). According to Cummings and Bailenson (2016), immersion describes the degree to which a system can shut out the outside world and offer vivid experiences. The extent of immersion is determined by hardware quality and the amount of senses triggered by the technology. HMD and CAVE VR experiences are commonly considered high immersion (Makransky and Petersen, 2021).

According to Li et al. (2018), IVR is a full-immersive VR technology that gives users the maximum level of immersion (the deep sense of being in the virtual world). It is the most expensive system, which takes only 6.7% of the VR applications in construction safety. One of its key applied areas is in the construction sites’ VR walk-throughs to detect hazards (Sacks et al., 2013). A virtual structural analysis program (VSAP) represents a typical IVR technology designed by the Virginia Polytechnic Institute and State University (Setareh et al., 2005). VSAP was developed with the intention of investigating building structural behavior within a virtual environment (VE). Notably, VSAP made a significant contribution by introducing an immersive portal interface. This was in response to the high cost of traditional immersive interfaces and the compromise in quality with desktop interfaces, which were more cost-effective. Subsequently, a customized Virginia Tech CAVE (VT CAVE) was designed, featuring a cubic room measuring 3 m × 3 m × 2.75 m. According to Wang et al. (2018), it was verified that the VT CAVE had superior effectiveness in terms of usage.

Sacks and Pikas (2013) utilized a 3D IVR power wall for the purpose of construction safety training. This power wall setup consists of three rear-projection screens within an open structure resembling a three-area CAVE, employing 3D projection stereo with dynamic glass. Participants engaged in the training using an XBOX controller and a head tracking system, monitored by eight cameras positioned atop the screens. The training employed three distinct software applications: Autodesk Revit was used to model the building displayed in the system, 3D Studio MAX was adopted to model other 3D geometry, and EON Studio v6 was used to generate the VR scenarios. The results indicate that VR-based training improved concentration and gave participants a level of control over the environment. Sacks et al. (2013) also adopted similar systems, except for the upgraded version EON Studio V7 for construction safety training. This system satisfies considerable but not all the requirements for a complete IVR environment, as indicated by (Bailenson et al. 2008). It does not completely prevent all users from observing the physical environment; also, at any one time, only the user wearing the tracker glasses was able to observe the environment from their perspective (the others were unable to control their movement as they were limited to the sense of supporting them).

IVR enables position and head tracking and can render a distinct image for each eye, creating visible signals for deep insight. IVR, compared to a mirror, also improves the scope of the visual area of view. These characteristics are essential for defining the IVR benefits and learning experiences (Makransky and Petersen, 2021). IVR differs from augmented or mixed reality technology, according to Loomis et al. (1999), in that the former completely isolates the user's mind within the virtual environment and completely blocks out the outside world, while the latter allows the user to experience the real and the virtual environment concurrently.

4.1.2. Desktop-based VR

In the early stages, Digital Virtual Reality (DVR) emerges as the predominantly used VR technology within the realm of construction engineering education and training (Wang et al., 2018). Bae et al. (2015) assert that DVR is the most cost-effective and least immersive VR system, requiring the least sophisticated components. It can also be called Window on World or Fish tank system. DVR in CHS is commonly applied in training and education. According to Chen et al. (2007), DVR technology allows virtual activities to be carried out on a regular computer monitor. It displays a three-dimensional virtual world on a desktop monitor, without the need for extra tracking equipment. DVR depends on the users’ perception and spatial capacities to observe what occurs around them. A user can conduct most tasks with DVR technology using keyboards and mice. Relying solely on keyboards, monitors, and mice, DVR is perceived as a cost-effective option in comparison to other VR-related technologies (Wang et al., 2018).

Notable examples of distinct DVR advancements include V-REALISM, developed by Li et al. (2003), and the Interactive Construction Management Learning System (ICMLS), designed by Sawhney et al. (2000). To aid with maintenance engineering training, V-REALISM was created. It creates geometrical models by adopting CAD, which is then displayed via the OpenGL programming platform. It improves interaction and navigation in the virtual world by integrating a ranking system for these models. This is considered the primary noteworthy contribution of V-REALISM (Li et al. 2003). Likewise, ICMLS was designed to manage the disconnect between training and physical on-site activities linked to the construction approaches and equipment. ICMLS is a web-based tool for creating virtual models using VR modeling language (VRML). It displays proper operations through web-based computing and discrete-event simulations (DES) (Sawhney et al.2000). Mawlana et al. (2015) indicated that ICMLS could be engraved into construction engineering education and training as it will enhance on-site construction. DVR development is reasonably stable, with inventions concentrating on virtual laboratories and 3D computer prototypes to motivate students and increase their knowledge (Glick et al., 2012; Vergara et al., 2017).

4.1.3. BIM-enabled VR

BIM is connected to developing and operating 3D objects, including applicable properties and details (Gheisari and Irizarry, 2016; Song et al., 2018). The applicable properties and details mainly refer to the critical data needed in realistic building construction throughout its life cycle. This encompasses the planning, designing, construction, operation, and maintenance phases. BIMVR depends on the model, highlighting the links behind different categories of VR and data binding to simulate building operations and activities. One of the essential features of BIM is Visualization (Wang et al., 2014). BIM data can be accessed in an immersive visualization world, and factors such as material type and cost will be analyzed to create an effective construction design. By examining the design information, all the BIM model components, such as mechanical, electrical, and plumbing (MEP), structure, and architecture, can be considered more detailed (Wang et al. 2018). With BIMVR, for example, users may explore the area created, experience the BIM model in a virtual environment without being constrained by gazing at 2D drawings, and take construction designs into a 3D virtual world with all relevant architectural elements.

Users can migrate from the traditional 2D drawing design systems to those in BIMVR virtual world using Autodesk Revit Live. This will help maintain the building management information integrity in the virtual world before the structure is constructed to have a good knowledge of how the design components will materialize (Wang et al. 2018). One of the most significant benefits of BIMVR is the model’s ability to reflect immediate modifications. Xie et al. (2011) indicated that VRML’s conventional VR models might have challenges integrating concurrent information. The compatibility issue may cause such difficulties. To further aid in the decision-making process, several strategies have been developed. For instance, Woodward and Hakkarainen (2011) created a software system that visualizes construction site activities by integrating schedule information with 3D models. Additionally, Park et al. (2016) introduced an interactive building anatomy modeling (IBAM) system, enabling users to engage with building components within a VR environment. Developers can also integrate a fixed question-and-answer game to improve the training experience in its application in CHS.

4.1.4. 3D game-based VR

3DGVR is a game-like computer-based training scenario via interactive, multi-user operating technologies and networks, integrative visuals, etc. It aims to improve interactions among users. The 3DGVR helps students collaborate and interact through tasks relevant to their real-life scenarios (Dickinson et al., 2011; Lin et al., 2011; Li et al., 2012; Nikolic et al., 2015). Besides concentrating purely on the immersive impact, 3DGVR focuses better on the interactions of game objects. For example, we may use a game engine's physics simulation component to precisely describe collision reactions. To reduce the detection procedures’ complexity in 3DGVR, ray-tracing methods and simplified collision boundaries are utilized (Wang et al., 2018). It makes collision detection computation easier and reduces complexity for complicated objects like construction cranes and excavators.

Guo et al. (2012) created an online platform safety training technology that is game-based, enabling users to adopt input devices, like game controllers, mice, keyboards, etc., to perform virtual activities like material delivery and equipment operation. The primary advantage of the system lies in the cost-effectiveness of having replicated trials readily accessible. For instance, with the adoption of the game-based method, diverse schedules and approaches to operating the equipment can be tested. Testing can uncover potential concerns, such as health and safety considerations (Wang et al., 2018). Le et al. (2016) designed a game-based training system to manage construction defects. Revit Architecture was used to create the virtual components, and Linden Scripting Language was adopted to represent the close-to-reality defect scenarios. In this system, the trainees were taught based on their knowledge of construction defects. They were given various scenarios and asked to pinpoint defects and likely tasks that could cause them. The test showed positive outcomes in terms of performance and interactivity (Wang et al., 2018).

4.1.5. Augmented Reality

The AR system, which accounts for 32%, is applied in the instruction and inspection process in VR applications in construction safety. In adopting this system, real-world objects are utilized as AR interface components, and their manipulation is adopted for virtual content interactions (Dong et al. 2013). Collaborative AR, as defined by Kim et al. (2012), involves participants sharing an environment that contains both virtual and real-world items. In that manner, the interactions are among users, not just between the AR system and a user. AR provides a direct or indirect image of a real environment by utilizing sensory technologies, incorporating video and sound alongside AR information. According to Fonseca et al. (2014), people find it easier to interact with AR objects and change their location, attributes, and scale, in a manner that seamlessly aligns with the real world, as opposed to a VR setting. Consequently, multiple studies have asserted that AR systems have the potential to introduce new interaction possibilities and enhance active students’ engagement (Behzadan and Kamat, 2013; Shirazi and Behzadan, 2015; Ayer et al., 2016).

For instance, Chen et al. (2011) adopted ARToolKit to create an AR system to teach students about their capacity to identify spatial objects. Diverse 3D models may be projected onto the real world via the AR system, contributing to the enhancement of students' education (Chi et al., 2013). Additionally, a number of apps have been created to incorporate AR into mobile devices, taking advantage of the evolving nature of mobile gadgets to facilitate learning (Wang et al., 2018). For example, Williams et al. (2015) provided training in context awareness through a mobile AR (MAR) environment. Shirazi and Behzadan (2015) developed the CAM-ART mobile context-aware AR tool for an undergraduate course in construction engineering. In the CAM-ART AR system, the interactions between objects are defined using JavaScript logic, while the content definition was realized using a static extensible markup language. To improve the process of construction, Kim et al. (2012) designed an AR-based system by modifying how equipment is being operated. The benefits of AR in this study are that the system improves visualization from the users’ viewpoint, and they can identify surrounding constraints.

4.3.1. Identification and monitoring of construction hazards

Most traditional safety performance indicator estimation techniques are manual and based on arbitrary assumptions. (Hinze, 2005; Teizer and Vela, 2009). These methods depend on big manual data-collecting actions; thus, data are gathered at a lower frequency (for example, once a month) and whenever incidents are recorded (Navon, 2005; Choudhry et al., 2007). These methods are costly and produce extremely little data sets for adequate and successful project control, and are prone to data entry mistakes (Teizer and Vela, 2009). Identification of hazards is based on conventional origins on drawings, desktop concepts, heuristic knowledge, and accident cases reported. These methods are being adopted to design preventive measures against anticipated safety hazards via project meetings (Bahn, 2013).

To surmount the challenges of these manual activities, VR environment modelling, and visualization emerged to identify hazards that improve users’ interactive and immersive experience (Hadikusumo and Rowlinson, 2002; Perlman et al., 2014). Continuous construction site safety monitoring through an automated system is increasingly regarded as a highly effective method for ensuring ongoing safety compliance (Park et al., 2017). The construction managers who will act upon the data may immediately access it after the automated monitoring system has collected it and formatted it into a structured manner (Rebolj et al., 2008). VR technology that identifies hazards on-site helps construction workers avoid dangerous zones. Most hazard avoidance systems use real-time tracking, which alerts workers of dangerous areas and approaching equipment; meanwhile, hazard identification systems focus on developing personal identification skills (Moore and Gheisari, 2019).

According to Moore and Gheisari (2019), on-site construction hazard monitoring incorporates VR systems with location trackers, providing workers with direct and valid safety information. For instance, Kim et al. (2017) designed an AR system that augments the view of workers utilizing the real-time tracking and Google Glass system. Workers were visually notified while approaching equipment, and equipment operators were made aware of their presence. Cheng and Teizer (2013) investigated how construction workers' situational awareness may be enhanced by combining real-time tracking with virtual reality technology. Construction personnel and equipment locations were tracked using radio-frequency identification (RFID) tracking devices, and the data was shown in a virtual environment. The prototype analysis revealed positive outcomes; the authors believed that with such technology, workers could evade construction site hazards and be trained in how the virtual environment works by adopting “close call” simulations.

Li et al. (2015) adopted VR and real-time tracking for safety and efficiency training. Pre-defined hazard zones and RFIDs helped workers by alerting them with a beep from their hardhats when moving close to heavy machinery or entering a dangerous zone. Furthermore, equipment and workers’ real-time locations, which managers and safety professionals could monitor, were represented in a virtual environment called the “Virtual Construction Simulation System.” The authors argued that AR is essential for the future of construction training and danger detection, claiming that real-time data visualization displays a limited amount of realism. Zhou et al. (2013) investigated the use of 4D visualization in detecting safety threats as they changed throughout the course of an underground metro construction project, based on a lifecycle study of the project. Data and risks were numerically analyzed using the 4D visualization and mathematical equations. The results of the study indicate that the project's detection of safety threats is facilitated by real-time information tracking and visualization. However, it was noted that risks and building sites are ever-evolving, making it challenging to manage this inherent aspect using the suggested method.

4.3.2. Construction safety education training

Safety education and training effectively improve safety control (Sawacha et al., 1999). This generally constitutes on-site and off-site training. While off-site training does not provide employees hands-on learning opportunities, on-site training is inefficient and can cause disruptions to routine construction work, which lowers production. VR can enhance safety education and training by furnishing visualized data and providing hands-on training in virtual off-site (Guo et al., 2017). Even though on-site safety training can be more effective (Pedro et al., 2016) because it involves hands-on experience and direct involvement, it can be time-consuming, expensive, and even dangerous due to actual site circumstances, such as schedule clashing, access problems, weather conditions, overriding safety concerns, and liability issues (Wang and Dunston, 2011). VR technologies provide unique opportunities for efficiently educating and training students or novices with higher cognition levels and lower risks.

The usage of various kinds of VR in workers’ training is rampant. Aviation simulators are a famous example, and VR is efficient for training in road safety (Thomson et al., 2005). VR simulation training for medical operations is also common because it evades the intrinsic risk of training on humans and other living creatures (Sutherland et al., 2006). A number of technologies, such as gaming, VR, AR, and BIM, have been created to enhance construction safety education and training procedures. For instance, Pedro et al. (2015) designed a QR code-scanning technology to provide safety information to university students using smart devices. Although developing the VR elements and classifying the safety details is deemed time-consuming, the results are encouraging. Students’ inspiration and attention to learning are developed in VR-related training.

Training on-site is difficult due to its hazardous nature, and failure on construction sites certainly eliminates the possibility of training. Construction sites training facilities, such as those in the Netherlands, UK, and Australia, can be built specifically to simulate construction environments (de Vries et al., 2004). VR technologies present the chance to enlighten workers about accidents and risky situations as part of construction safety training, as seen in Figure 7. Workers can evaluate a problem, determine the action to take, execute the action, and instantly monitor the outcomes. This leads to psychological information handling, which stays in the memory for a long time. Lucas et al. (2008) study on safety education for construction equipment operators stands out as one of the limited instances where VR has been utilized for education and training in construction safety. An other example is ForgeFX (2012), a general VR industrial safety teaching game developed for the California State Compensation Insurance Fund that includes a component that replicates handling materials to build a timber roof for a house. Ku and Mahabaleshwarkar (2011) suggested the prospect of adopting ‘Second Life’ in construction safety education and training and the countless issues.

1. Download : Download high-res image (698KB)
2. Download : Download full-size image

Figure 7

Perdomo et al. (2005) state that several research have investigated VR as an addition to traditional construction education and training. Sampaio et al. (2010b) investigated the applicability of CAD, VR, and 3D modeling in AEC education. A 3D game that let college students assume the role of safety inspectors and assess potential hazards at work sites was the subject of Lin et al.'s (2011) investigation. Li et al. (2012) also designed a 4D Interactive Safety Assessment, a unique safety-evaluation technique that used virtual scenarios to evaluate construction personnel. The mechanism offered workers virtual hazard conditions linked to their tasks and a spectrum of potential activities of choice. In order to help students comprehend health and safety in trenches, Dickinson et al. (2011) developed a serious game that allows them to explore the 3D world. Michael (2006) defined serious games as games with the main purpose of education instead of entertainment. Based on the evidence from the literature, the adoption of VR in construction safety education and training and the understanding of their usage and efficiency is limited. Some of the issues for researchers have been the substantial expenditure needed to organize the IVR system, including pedagogical and animation elements, the 3D scenery, and the demand to get construction workers to a specified facility (Sacks et al., 2013).

4.4.1. Infrastructure

Challenges associated with the device’s weight can negatively impact its acceptance by users, which will hinder the broader VR adoption (Prabhakaran et al., 2022). According to Yan et al. (2018), as the HMDs’ weight increases, the pressure load and the users’ subjective discomfort increase. HMDs’ weight has been linked with visual discomfort and related injury experiences, as noted by Yuan et al. (2018). Similarly, several studies have highlighted the substantial influence of view angle (Arthur and Brooks, 2000) and display brightness (Vasylevska et al., 2019) on the level of immersion and the user's task performance. Despite these challenges, it is suggested that HMD developers can overcome these challenges without compromising processing capabilities, thanks to advancements in the current chip revolution. The resolution of these challenges is anticipated in the coming years (Metz, 2017).

4.4.2. Algorithm development

In contrast to industries like automobile and manufacturing, where standardized processes are more common, every construction project is unique, with distinct requirements from stakeholders. Consequently, customized algorithms are essential for each construction project to ensure that the virtual world and associated systems align with the specific needs of the stakeholders. Besides, developing an immersive, interactive, illustrative, intuitive, and informative virtual world needs a particular skill set; this is referred to as one of the challenges in the literature (Prabhakaran et al., 2022). As the construction sector is encouraging a more accurate and efficient process, there is a precise need for construction professionals to update their programming skills constantly. Given the existing shortage of skills in the construction sector, as highlighted by Sawhney (1999) and Liu et al. (2015), there is a pressing need for the industry to attract a skilled workforce. Failing to do so may result in the sector relying on third-party developers to supply virtual content, potentially leading to cost overruns.

4.4.3. Virtual content modelling

Creating VR content that is genuinely inviting and captivating is a challenge that demands a significant skill set and funds. This challenge may not be as vital for those companies that have embraced BIM workflow as it is for other companies that are just aspiring to adopt BIM. While BIM models undergo multiple iterations before being imported into a virtual world, the extensive tasks involved in building modeling can be efficiently executed using BIM authoring tools integrated into BIM products. Users may integrate BIM models straight into the Unity game engine with the use of contemporary software tools like Unity Reflect. However, it's worth noting that additional efforts are required for material enhancement, behavior assignments, and post-texturing, which can be resource- and time-consuming activities Prabhakaran et al., 2022).

4.4.4. General health and safety

General health and safety challenges related to VR, like physical (unnatural postural demands, immersion injuries, hygiene), physiological (convergence cardiovascular changes, visual asthenopia signs, and symptoms), and psychological (change in psychomotor performance, addiction, stress), are significant problems while utilizing these technologies (Costello, 1997). Any potential VR user in the VR-CHS must choose the technology based on this specific task and comprehend the possible health and safety implications related to the system. Nevertheless, users should pay attention to these technologies’ health and safety guidelines provided by the manufacturers and ensure they comprehend them. Also, a sequence of undesirable symptoms like headaches and nausea are usually the immersive environment side effects (Weech et al., 2019). According to Prabhakaran et al. (2022), it is anticipated that most of these challenges could be alleviated with the wireless connectivity between computers and HMDs, advancements in display quality like OLED, locomotion methods adopted like teleportation, and lightweight hardware.

4.4.5. Interoperability

The optimization of the VR-CHS process greatly depends on addressing the critical issue of system interoperability in virtual world development game engines such as Unity 3D and Unreal Engine (Rahimian, 2019). Middleware has been employed by several software providers, including Unity (unity3d.com), to bridge this gap recently. Nevertheless, these developments need additional iterations and improvements using middleware applications as they are still in their infancy. Further, with the Unity Reflect introduction (Reflect, 2019), facilitating an immersive experience has become unchallenging with the BIM models’ transfer jointly with their meta-data into the Unity game engine. However, developing interaction, which demands tailored algorithms, remains challenging.