The use of virtual reality as a tool for professional training enables access to knowledge in a more active way. Applying it in high-risk situations training encourages finding the most efective way to share knowledge. With this objective in mind, video game tutorials as a method of learning in virtual environments are becoming increasingly popular. This study proposes a methodology in which virtual reality tutoring techniques are applied to enhance learning in training for radiation emergencies. A virtual reality experience involving a high-risk situation has been developed. In this experience two groups of professionals involved in such emergencies participated, one group were trained with a tutorial, and the other group faced the simulation without it. The results show that participants who received tutoring were able to solve the test correctly in a more efective way.
Radiation emergencies are high-risk and stressful situations in which an accident involving radioactive material has occurred. They difer from nuclear emergencies in that radiation emergencies occur outside nuclear power plants. For example, accidental exposure to stolen radiation sources or accident during transport of radioactive materials. Given the presence of radiation, it is important for professionals in this sector to be prepared to act without errors that could negatively afect health. Therefore, training must be conducted to enable professionals to acquire knowledge and skills necessaries to address these situations. However, these simulations have a high economic and organizational cost. This lowers the frequency of conducting drills, which could negatively impact the eficiency of the professionals’ learning process. Furthermore, using radioactive elements in real simulations is extremely dangerous. Therefore, an alternative to these elements is to indicate the radiation levels of the environment with piece of papers. This kind of partial solution limits the efectiveness of the exercise because radiation detector devices are not used, not even virtually. Drills also cannot replicate dangerous situations with precision. For instance, an earthquake causing a nuclear power plant explosion and its radiation cloud approaching the population. These diferences from the real world prevent trainees from having full immersion, which limits the successful performance of the exercise. Therefore, many studies and programs have started using diferent technologies to ensure efective and safe training.
Virtual reality is becoming increasingly accessible to the public. This technology is primarily used in the entertainment industry but can also be employed for learning and training purposes. One of the greatest challenges of this technology is user adaptation. Virtual reality aims to recreate the real world and all the interactions that can take place within it. To achieve this, individuals need a process of adaptation to use this technology naturally. VR developers ofer tutorials covering these initial concepts to assist in the adaptation process. However, these tutorials are often quite generic and subject to change based on the video game or application specifications. Therefore, the development of a specific tutorial phase is necessary. Through the tutorial, it is expected that the user will be able to use the application successfully. In this way, the user will adapt to the technology, and an improvement in the final exercise results is expected.
Tutorial design is original from the video game industry. Each day, these entertainment systems become more complex. Users often need a learning process to fully enjoy the experience. Consequently, nowadays, almost all commercial video games feature a tutorial to meet these needs. As a result, extensive research has been conducted on tutorial design and implementation. The goal is to create a satisfying experience with a non-intrusive learning process so that users can learn while being entertained.
In this work, a virtual reality tool has been developed for training professionals in radiation emergencies. In this experience there are a main scenario and two preliminary tutorial phases. In these tutorial phases, the user will learn how to interact with their environment and the objectives they must be met for a successful execution of the exercise. An experiment session was conducted to test the efectiveness of the learning provided by this tool. 6 sector professionals were evaluated. These professionals were divided into two groups. One group were trained beforehand with the tutorial, while the other team received no training. The analysis was conducted by observing the total duration, interactions with elements, completed objectives, and movement eficiency. The trained group achieved better results in all the analysed variables. While the sample size is limited by the number of available professionals, the results indicate that the benefits of the tutorial phase are clear. Nonetheless, the developed application and applied tutoring techniques demonstrate a notable improvement in user performance.
Professionals facing risky situations must undergo prior training. Efective learning poses many
challenges when preparing a training simulation. Therefore, since 2010, research in virtual
reality for conducting these processes has seen significant growth, as noted by Radhakrishnan
et al. [
However, humans are still not used to this technology. This often leads to adaptation problems. Usability issues can result in improper training [9] or even worse outcomes compared to traditional methods [10]. Adapting to this technology is a widespread issue. Companies like Meta 1 and Steam 2 are working to improve the usability of their products with initial tutorials. However, these tutorials are generic and subject to modification based on the specific needs of the application being used. This leads commercial video games like Skyrim VR (Bethesda Softworks, 2017), Beat Saber (Beat Games, 2019), or Windlands 2 (Psytec Games Ltd, 2018) to create their own tutorials. This problem is also observed in the field of science. Various research projects involving virtual reality have developed generic tutorials [11] or specific ones to teach users how to interact with the environment [12, 13, 14]. Adaptation is not the only issue that necessitates the application of tutorial techniques. The complexity of video games is also significant [ 15]. Considering the complexity of responding to risky situations, the implementation of such tutorials is considered necessary.
Tutorials can be presented to the user in diferent ways. Frommel et al. [
It is important to understand how information is presented to the user. Kao et al. [
Tutorials can also incorporate cues to guide the user’s next steps. In this regard, notifications
in the head-up display are the most efective, although they can be intrusive, they are always
detected by the user. This is crucial if you want the user to perform specific actions. Another
way to prevent the user from getting lost is by using visual cues [
Virtual reality is a powerful tool for professional staf training. However, usability and the 1Meta First Steps: https://www.meta.com/es-es/experiences/1863547050392688/, accessed on 26/9/2023. 2SteamVR: https://store.steampowered.com/app/250820/SteamVR/?l=english, accessed on 26/9/2023. complexity of concepts require the application of certain techniques to enhance learning. These techniques can be derived from tutorials, the majority of which originate from the world of video games.
As previously mentioned, a virtual reality application has been developed for training professionals in emergency situations. Within this tool, a tutorial has been implemented with the aim of enhancing the learning process. It is expected an improvement in the final performance. The results are based on the analysis of user behavior, which is examined through the analysis of predefined set of objectives and interactions.
There is a wide variety of potential scenarios classified as radiation emergencies. For this study, it has been decided to recreate an emergency where a car accident has occurred. A van was transporting radioactive material for industrial radiography. The accident has occurred within a city. In this scenario, there is an element emitting low levels of radiation, an injured person, and a witness. Additionally, there are trafic cones to stop trafic and radiation detectors to measure the environment radiation. The exercise consists of some objectives that the professional should accomplish. The objectives aim to list the basic actions that an involved personnel must carry out based on the operating protocols. The professional should gather information about the incident, secure the area, address the problem while considering the radiation of the environment, and report what has happened. These tasks were the following:
The specific objectives were gathering information from characters, protecting and talking to the injured person, cutting of trafic, inspecting the van, checking the radiation levels, informing the head ofice and return home.
Most of these objectives have a natural completion order, and the user must follow it. However, depending on the situation, the user may consider another order is more suitable. For instance, it was important to ask witnesses about what happened, but it was not necessary to do so in this precise order. In order to complete all the proposed objectives, the user must be able to interact with elements in the environment. The interactions were aiming or stop aiming, interacting with an object, using correctly or incorrectly an object, picking up or dropping an object, positioning correctly or incorrectly, talking to a character and moving.
The virtual environment takes place outdoors and covers extensive area. The user needs to
move freely in order to complete the objectives. Therefore, it has been decided to give freedom
of movement to the participant [
The virtual reality application consists of three scenarios: the tutorial sandbox scenario for learning interaction, the guided emergency scenario to understand the objectives, and assessment scenario in which the efectiveness of the solution if checked.
Tutorial sandbox scenario is created for users to learn to interact with the elements of the
environment (see Figure 1). To achieve this, they will be presented with a linear scenario. To
prevent user disorientation, directional cues have been placed on the walls of the environment,
indicating the direction to follow [
The user must face the described emergency on guided emergency scenario. This phase teaches
the user the steps they need to take to respond to an emergency successfully. The application
will guide the user with the objectives they must fulfill at each moment. These objectives are
ordered, and the user is obligated to follow the instructions. The user is informed about the
current task through a panel that appears on the head-up display. This panel appears when a
new task needs to be performed. This way of notifying the user is intrusive but efective [
Finally, assessment scenario is a phase where the user must put their knowledge into practice (see Figure 2). In this scenario, the same emergency as in guided emergency phase has happened. The diference between these phases is that assessment takes place in a diferent location; the position and distribution of elements are diferent, and the user is not guided. In this unguided phase, the user is free to fulfil the objectives they deem appropriate and in the order they choose. The data obtained from this unguided phase will be used for the final analysis.
The impact of experimenting the tutorial phase was analysed through a practical experiment. An HTC Vive Pro 2 virtual reality wired headset was used for this purpose. To do this, participants were divided into two groups. The first group, called the untrained group, only completed the assessment scenario. On the other hand, the second group, the trained group, went through all the scenarios. This way, the trained group learned to interact with the elements in the environment, to move, and to understand the objectives they had to achieve. However, participants in the untrained group had to discover all of this on their own.
Participants were informed about the study’s goal and were told that they could end their participation at any time. Then, participants were briefed about the virtual setup and informed about the possibility of a simulated radiation leak. Controller buttons were explained to the participants before starting the experience. After this, participants were introduced to the virtual experience. The trained group began with the tutorial sandbox scenario and then moved on to the guided emergency. There was no time limit in both of these scenarios. Once they completed these two phases of the experience, the user was introduced to the assessment scenario. For this phase, participants had a maximum of 15 minutes. The untrained group only experienced the assessment scenario. The maximum duration was the same.
At the end of the experiment, users filled out a questionnaire about personal data and information regarding their skill and previous experience with virtual reality. Additionally, in this questionnaire, they were asked to express their personal opinion about the experience. This information will be used to improve the tool for future applications and experiments.
A total of 6 adult professionals participated, all of whom were male members of the Spanish Military Emergency Unit. 50% of them had never used virtual reality before, but 100% had previously participated in real radiation emergency simulations. Participants B, D, and F had never used virtual reality. Participants C and E had used it only once. Participant A had used this technology a few times. Participants were divided into two groups, 3 users for each one.
This quantitative analysis aimed to identify the behaviour diferences between trained and untrained group achieve this, the data collected from the assessment phase was compared. All the results indicate a positive learning and adaptation process for the trained group.
The average duration (see Table 1) for the trained group has been 346.33 seconds, while the untrained group took an average of 809 seconds. The standard deviation of the duration for the untrained group has been lower.
Participants in the trained group tend to fulfill more objectives while performing fewer interactions and movements (see Table 4). The average number of objectives completed (see Table 3) for the trained and untrained teams are 8.33 and 5.67, respectively. In both cases, the standard deviation is below 8%. The trained group has been closer to achieving the 9 objectives. The trained group takes an average of 41.58 seconds to complete an objective, while the untrained group takes 142.68 seconds.
The average number of interactions with the elements conducted by the trained group is 209, and for the untrained team, it is 632 (see Table 3). In both cases, the standard deviations are quite high, especially for the trained group. User E has interacted with the elements more than twice as many times as their peers D and F. These groups have achieved a ratio of 25.09 interactions per task for the trained group and 111.46 interactions per task for the untrained group.
The trained group has used teleportation more frequently (see Table 5). On average, teleportation has been performed 233.33 times for the untrained group and 133 times for the trained group. Unlike the untrained users, trained users have attempted to teleport directly to locations where the objectives are located (see Figure 3). Additionally, the teleportation distance in the virtual environment and the distance the user has moved in the real experimental environment have also been calculated. The sum of the teleport distance for the trained group has been lower, both in the virtual environment and the real world (see Table 6). For each objective, the untrained group has made 41.15 teleports and covered a total distance of 111.77 meters. On the other hand, for each objective, the trained group has made 15.97 teleports and covered a total distance of 47.46 meters.
As shown in Table 1, the diference in durations is clear. The untrained group tended to use all the available time. The minimum duration of the untrained group is higher than the maximum duration of the trained group (see Table 2).
An analysis of the fulfilled objectives shows that the trained team is more eficient in achieving almost all the objectives in a shorter time compared to the untrained team. Considering that the experience comprises a total of 9 objectives, it can be said that the trained team has performed well in resolving the exercise. None of the users in the untrained team gathered information by speaking with the witness or informed to the head ofice. In the case of the trained team, participant D did not stop trafic, and participant F also did not speak with the witness. It seems that participants consider that seek external sources of information regarding the incident is not important.
In the case of performed interactions, the average number for the trained team is 66.8% lower. Furthermore, this group has managed to complete almost all objectives with a lower number of interactions. However, it is important to note that in several instances, users unintentionally repeated interactions. This is because there are small objects, and the system is highly sensitive.
Therefore, the system detects that the user has performed actions numerous times, such as pointing at a specific object they want to interact with. Interactions performed on the same object within the same second are considered system errors and have been excluded. However, a more reliable study is needed to identify which interactions can be considered noise. Users B and E are the ones who have performed the most interactions in each group (see Table 4). This has led to an increase in the average for each group and in the standard deviation. In this variable, it can also be observed that the user who interacted the least in the untrained group did so more times than the user who interacted the most in the trained group
The trained group performs a more eficient movement. This group has achieved lower results in number of teleportations and lower movement distance, both in the virtual and real-world scenarios. This shows that this group needed to travel less to successfully resolve the emergency. Furthermore, observing Figure 3, it can be seen that trained participants take a more direct route between points of interest in the environment.
The furthest users are B and E. B is the one who has completed the fewest objectives, while E has completed the most. Analyzing user B, it can be observed that they have had the most number of interactions. However, in terms of movement, they have moved the least. This shows that most of this group participants was focused on achieving nearby objectives, and they did not move to find new objectives. On the other hand, participant E is the one who has had the most number of interactions and has moved the most in order to fulfill all objectives. Unlike participants in the same group, this behaviour has led the participant to successfully complete the exercise.
Virtual reality enables professionals to undergo efective training for health-threatening situations. The presented research has developed a virtual reality application for this purpose. To assess the efectiveness of the tool, an experiment was conducted with professionals involved in these emergency situations. In the experiment, a group of participants was trained with a tutorial stage, and another group did not receive any training. The results demonstrate that the group receiving the tutorial performed better in the assessment stage. The sample size is relatively limited, but the participants in the experiment are professionals involved in radiation emergencies.
This work focuses on the learning process in emergency situations. To complete the study,
several extensions and further research will be considered in future work. First, it is necessary to
observe the persistence of learning over time. To do this, the same professionals will be asked to
participate in future experiments, allowing for an analysis of their memory and learning. Second,
there are numerous possible radiation emergencies. Therefore, studies will be conducted with
other scenarios. This variety of scenarios will allow for the study of user behavior regardless
of the specific situation. Finally, these situations are dangerous and increase the stress levels
of professionals. The physiological state of the participants may be related to their behavior
and the learning process during simulations. Stress should be analysed and taken into account.
One way to detect the stress and anxiety of participants is by using biometric sensors [
This work has been partially supported by the projects ADARVE (SUBV20/2021), funded by the Spanish Council of Nuclear Security; CANTOR (PID2019-108927RB-I00), funded by the Spanish Ministry of Science and Innovation; and EA-DIGIFOLK (101086338), funded by the European Commission. Safety Science 166 (2023) 106226. URL: https://www.sciencedirect.com/science/article/pii/ S0925753523001686. doi:10.1016/j.ssci.2023.106226. [4] S. M. Sakowitz, M. R. Inglehart, V. Ramaswamy, S. Edwards, B. Shoukri, S. Sachs, H. KimBerman, A comparison of two-dimensional prediction tracing and a virtual reality patient methods for diagnosis and treatment planning of orthognathic cases in dental students: a randomized preliminary study, Virtual Reality 24 (2020) 399–409. URL: https://doi.org/10. 1007/s10055-019-00413-w. doi:10.1007/s10055-019-00413-w. [5] H. Engelbrecht, R. W. Lindeman, S. Hoermann, A SWOT Analysis of the Field of Virtual Reality for Firefighter Training, Frontiers in Robotics and AI 6 (2019). URL: https://www. frontiersin.org/articles/10.3389/frobt.2019.00101. [6] A. Villar Rubio, C. León, User performance analysis in guided and non-guided stressful virtual reality scenarios, 2023. doi:10.54941/ahfe1004038. [7] D. Narciso, M. Melo, J. V. Raposo, J. Cunha, M. Bessa, Virtual reality in training: an experimental study with firefighters, Multimedia Tools and Applications 79 (2020) 6227–6245.
URL: https://doi.org/10.1007/s11042-019-08323-4. doi:10.1007/s11042-019-08323-4. [8] R. M. Cliford, S. Jung, S. Hoermann, M. Billinghurst, R. W. Lindeman, Creating a Stressful Decision Making Environment for Aerial Firefighter Training in Virtual Reality, in: 2019 IEEE Conference on Virtual Reality and 3D User Interfaces (VR), 2019, pp. 181–189. doi:10.1109/VR.2019.8797889, iSSN: 2642-5254. [9] D. Checa, I. Miguel-Alonso, A. Bustillo, Immersive virtual-reality computer-assembly serious game to enhance autonomous learning, Virtual Reality (2021). URL: https://doi. org/10.1007/s10055-021-00607-1. doi:10.1007/s10055-021-00607-1. [10] N. Gavish, T. Gutiérrez, S. Webel, J. Rodríguez, M. Peveri, U. Bockholt, F. Tecchia, Evaluating virtual reality and augmented reality training for industrial maintenance and assembly tasks, Interactive Learning Environments 23 (2015) 778–798. URL: https: //doi.org/10.1080/10494820.2013.815221. doi:10.1080/10494820.2013.815221, publisher: Routledge _eprint: https://doi.org/10.1080/10494820.2013.815221. [11] I. Miguel-Alonso, B. Rodriguez-Garcia, D. Checa, A. Bustillo, Countering the Novelty Efect: A Tutorial for Immersive Virtual Reality Learning Environments, Applied Sciences 13 (2023) 593. URL: https://www.mdpi.com/2076-3417/13/1/593. doi:10.3390/app13010593. [12] G. Makransky, T. S. Terkildsen, R. E. Mayer, Adding immersive virtual reality to a science lab simulation causes more presence but less learning, Learning and Instruction 60 (2019) 225– 236. URL: https://www.sciencedirect.com/science/article/pii/S0959475217303274. doi:10. 1016/j.learninstruc.2017.12.007. [13] N. F. Kleven, E. Prasolova-Førland, M. Fominykh, A. Hansen, G. Rasmussen, L. M. Sagberg, F. Lindseth, Training nurses and educating the public using a virtual operating room with Oculus Rift, in: 2014 International Conference on Virtual Systems & Multimedia (VSMM), 2014, pp. 206–213. doi:10.1109/VSMM.2014.7136687. [14] A. Bhargava, J. W. Bertrand, A. K. Gramopadhye, K. C. Madathil, S. V. Babu, Evaluating Multiple Levels of an Interaction Fidelity Continuum on Performance and Learning in Near-Field Training Simulations, IEEE Transactions on Visualization and Computer Graphics 24 (2018) 1418–1427. doi:10.1109/TVCG.2018.2794639, conference Name: IEEE Transactions on Visualization and Computer Graphics. [15] E. Andersen, E. O’Rourke, Y.-E. Liu, R. Snider, J. Lowdermilk, D. Truong, S. Cooper,