The article considers the concept of developing an intelligent virtual and augmented reality system for teaching first aid in wartime and damage to civilian infrastructure. The proposed VR/AR solution allows you to simulate critical situations, including injuries, mass explosions, burns, and cardiopulmonary resuscitation. Additionally, it offers interactive engagement with victims through the use of modern game engines and VR controllers. The paper presents the functionality of the system, including the simulation of realistic scenarios, gamification of the educational process, automatic assessment of user actions, and integration with the medical triage system (Triage). Additionally, the use of generative AI models (Stable Diffusion, Leonardo.Ai, Trellis3D, Meshy, Sloyd) to create 2D and 3D content is described, which provides rapid visualisation and prototyping of training scenes. The results of the work demonstrate that the introduction of such technologies enables the enhancement of training efficiency for doctors, military personnel, volunteers, and civilians, reduces the level of panic during emergency events, and fosters practical skills in assisting in safe conditions.
The full-scale war against Ukraine has demonstrated how vulnerable civilian infrastructure can be during massive attacks and how often not only the military, but also civilians are affected. In such conditions, the issue of providing quick and effective first aid becomes especially relevant. A person's life depends on the right actions in the first minutes after an injury, but not all citizens have the basic skills to respond to critical situations. Traditional approaches to first aid training, based on lectures and training with dummies, do not always reflect the realities of combat wounds, explosions, mass casualties, or psychological stress faced by both military and civilians.
The relevance of the topic is primarily due to the war in Ukraine, attacks on civilian infrastructure, and the growing need for mass preparation of the population for the provision of first aid. In turn, there are limitations of traditional methods (lectures, training on mannequins) and the growing prospects of VR/AR technologies in training (realistic modelling, gamification, safe practice of skills). Modern digital technologies create new opportunities to overcome this challenge. In particular, virtual reality (VR) and augmented reality (AR) enable the creation of interactive simulators that recreate realistic emergency conditions. Intelligent VR/AR systems can simulate scenes of destruction in urban environments after shelling, reproduce the emotional reactions of victims (such as panic and shock), and provide users with interactive tasks, including applying a tourniquet to stop bleeding, performing cardiopulmonary resuscitation, or providing assistance in cases of burns.
Vinnytsia, Ukraine 1∗ Corresponding author. † These authors contributed equally.
A key feature of such systems is the integration of the educational and training process with gamification and algorithms for assessing the effectiveness of actions. Thanks to this, the user receives feedback in real-time: the system fixes errors, offers hints, and allows you to repeat scenarios until they are perfectly executed. The integration of the principles of medical triage (Triage) makes it possible not only to teach basic skills but also to form a strategy for assisting in a situation of mass casualties, which is especially relevant in the realities of wartime.
The target audience of such VR/AR solutions encompasses a wide range of users, including pupils and students enrolled in civil protection courses, volunteers and citizens seeking to prepare for emergencies, as well as military personnel and medical professionals who aim to enhance their skills in simulated combat conditions. The use of such technologies in educational institutions, first aid courses and military training creates prerequisites for the formation of a more resilient society capable of adequately responding to the challenges of war and terrorist attacks.
Thus, an intelligent virtual reality system for providing first aid is not only an educational or technological product, but also a vital component of national security. It aims to prepare the general population for action in crises, reduce panic, and increase the chances of survival for victims. This article discusses the concept of creating such a system, its functionality and prospects for implementation in educational and medical practice.
Hostilities, massive attacks on civilian infrastructure, and an increase in the number of civilian casualties are exacerbating the issue of preparing the general public for the provision of first aid. Existing teaching methods (lectures, workshops on mannequins) have a number of limitations: they do not reproduce psychological pressure, complex combat wounds, chaotic conditions of destruction and the need to act in a short period of time. This leads to the fact that even individuals who have completed introductory courses are not always prepared to respond correctly during real crises. Thus, the problem arises of creating an innovative tool that would combine a realistic reproduction of emergency conditions, interactivity, gamification, and a system for evaluating actions. The development of an intelligent VR/AR system for first aid allows for solving this problem, providing a qualitatively new level of training for the population, medical workers, and the military in a safe, yet as realistic as possible environment. The main problems and needs of modern society in Ukraine:
Insufficient effectiveness of existing educational approaches to prepare civilians and military personnel for actions in crisis conditions.
Lack of tools that simultaneously simulate stressors, chaotic conditions and multiscenario emergencies.
The need to create an interactive system that would ensure: i. approximation to real combat conditions; ii. systemic training of skills (Triage, CPR, bleeding arrest).
The article aims to substantiate and develop the concept of an intelligent virtual and augmented reality system for teaching first aid in wartime and mitigating damage to civilian infrastructure. The system should provide simulations of realistic scenarios, foster the development of practical skills in crises, and enhance the readiness of both civilians and specialists for an effective response.
The main objectives of the study: To analyse modern approaches to first aid training and determine their limitations. To develop and describe the concept of a VR/AR simulator (system) of pre-medical care with the integration of intellectual functions, focused on war conditions and crises. Describe key learning scenarios and user interaction mechanics.
Consider the system's functionality, including interactivity, gamification, evaluation of user actions, and the application of Triage principles.
Demonstrate the role of generative AI models in creating 2D and 3D content for simulations of learning environments.
Evaluate the potential impact of VR/AR solutions and the benefits of the proposed system on the training of doctors, military, volunteers, and civilians in emergencies.
To identify and formulate prospects for the practical implementation of such systems in the field of education, medicine and military training.
Recent years have shown a significant increase in the number of works on the use of VR/AR to
teach fundamental and advanced first aid skills. The issue of mass training in pre-medical care has
become critical in connection with conflict and military events that lead to damage to civilian
infrastructure and a large number of wounded. In response, there is a growing interest in VR/AR
technologies as tools for large-scale, accessible, and realistic training of the population, volunteers,
and medical staff [
Several systematic reviews and RCTs have shown that VR/AR training can improve the quality
of resuscitation skills (frequency and depth of compressions, correct sequence of actions) or, at
least, is not inferior to standard training, and in some nuances exceeds it (better memorised
procedure, greater motivation to repeat training). AR solutions integrated with mannequins
provide an additional advantage – the combination of haptic/manikin and AR prompts improves
accuracy and self-learning [
Research in the field of serious games and gamification (especially for tactical rescuers and
military medics) shows that the addition of game mechanics (missions, levels, rewards) increases
motivation and frequency of repeated training. Adaptive scenarios that change in complexity
depending on user outcomes contribute to the personalisation of learning and the better
consolidation of skills [
Work on AR add-ons for physical mannequins (e.g., Holo-BLSD) demonstrates that adding
visual AR feedback to the tactile feel of the dummy improves the perception and learning outcomes
of students, especially non-professionals. Holo-BLSD has shown high acceptability and usability in
pilot studies [
The literature [
Publications on VR education agree on the need for multi-stage testing, involving subject specialists and target users, as well as formal checks of compliance with medical protocols. In terms of work, this is implemented through alpha sessions with paramedics/military instructors, beta sessions with students/rescuers/volunteers, automated stability tests, and the requirement to certify content to pre-medical standards. Modern intelligent simulators rely on performance metrics (time, correctness of the sequence, and quality of manipulations) and dashboards of individual progress, which allow you to adapt the complexity and personalise the trajectory. The terms of reference should include the collection of errors, feedback, and performance assessment, as well as the requirements for personal data protection (GDPR) and cybersecurity of infrastructure. Literature on edtech/medtech emphasises phased implementation: MVPs → pilots → scaling with related business processes (marketing, partnerships, financial and risk management). The file presents the WBS, budget items (including equipment, licenses, and servers), marketing channels, and risk management, creating a "bridge" from a research prototype to a viable product for educational institutions and defence forces. The systematization of such studies corresponds to the key vector trends of the industry: (i) contextualized simulation for real combat scenarios; (ii) intellectual adaptation of learning and formative assessment; (iii) generative AI to accelerate content production; (iv) multi-stage testing and protocol certification; (v) legal/ethical framework (GDPR) and technical cybersecurity; (vi) project and business scaling circuit (WBS, budget, risks). It aligns with the current roadmap for VR/AR solutions in pre-medical training, specifically focusing on applications in war and hybrid civil-military environments.
The literature [
AR dummies / Holo-BLSD – studies showing the superiority of AR and dummies in
BLS training [
SimX / VALOR (DoD) – large-scale military projects focused on the integration of
VR for mobile and divisional learning; emphasis on realistic combat scenarios and analytics
[
Scientific RCTs and scoping reviews have accumulated the evidence base for
improving CPR/BLS, pointing to the positive role of VR/AR but emphasising the need for
larger randomised trials and long-term measurements [
Problems, limitations and open questions [15-24]:
Transfer of training to real conditions - many studies show improvements in the training session, but to a lesser extent, it has been proven how this affects real events under stress.
Standardisation of metrics – different jobs use incompatible metrics, making comparisons difficult.
VR content optimisation – automatically generated 3D objects require optimisation (polygons, collisions, LOD) for smooth operation on VR headsets.
Legal/ethical issues – licenses for AI content, protection of user data in training platforms, compliance with medical standards.
Cost and technical accessibility – the cost of equipment and the need for technical support for mass implementation in schools/communities. Scalability, Cost, need for It is used in extensive script facilitator clinics and law library, training, enforcement analytics, multi- partially agencies, with user, integration proprietary a strong with education content industrial
backup.
→ tactile and studies; demonstrates visual feedback; hardware acceptability good usability in limitations and feasibility the pilot; (HoloLens) and (n≈26). Logging actions visual quality
equivalent VR/AR performance compared traditional training, emphasising the
of methods; show that short VR/AR is not observation inferior to faceto periods; Lack of to-face long-term data training; Subgroups with a dummy win.
Military and defence structures are actively implementing VR simulators for tactical medicine
training – industry examples show scalable platforms focused on realistic combat scenarios and
action analytics (SimX is a commercial platform for clinical and tactical simulations). Such
solutions focus on the possibility of synchronous multi-person training, telemetry collection and
adaptive complexity of scenarios [
The main areas of research include the analysis of literature and experience, the application of the scenario modelling method, the use of well-known tools for developing VR/AR solutions, and, accordingly, the implementation of development testing. The first stage of the study involves investigating existing VR/AR solutions in medicine and military training. Next, it is necessary to develop crisis VR/AR scenes (city streets after shelling, evacuation of the wounded, burns, and mass casualties) based on Unity/Unreal Engine, Blender/3ds Max tools, and AI generators (Stable Diffusion, Leonardo.Ai, Trellis3D, Meshy, Sloyd). Experimental testing will consist of conducting alpha and beta tests with doctors, students, and volunteers. Let us describe the primary research processes in more detail.
At the first stage, a systematic review of scientific papers and applied solutions in the field of VR/AR technologies for medical and military training was conducted. To compare the effectiveness of traditional and innovative approaches, the integral coefficient was used:
where pi – is the indicator of the effectiveness of the i method (level of assimilation, learning speed, psychological readiness), wi – is the weighting factor of importance, E – is the integral assessment of effectiveness.
The development of the VR/AR simulator was carried out through scenario modelling of crises, including injuries, explosions, burns, and mass casualties. A set of states describes each simulation:
S={s1 , s2 , … , sm }, s j∈ {stable, critical, lethal }, where s j – is the condition of the victim according to the Triage protocol. The transition between states depends on the user's actions and reaction time
, st+1=f ( st , at , τ ), (1) (2) (3) (5) (6) where at – is the user's action (applying a tourniquet, performing CPR, etc.), τ is the reaction time. To assess the correctness of the implementation of the algorithms of first aid, an integral metric is proposed:
Q=α ⋅ C + β⋅ T + γ ⋅ R, (4) where C – correctness of actions (binary score 0/1), T – speed of execution (normalised relative to the optimal time), R – sequence of protocol adherence, α , β , γ – weighting factors determined by experts. Generative artificial intelligence models (Stable Diffusion, Trellis3D, Meshy, Sloyd) were used to create the learning environment. The generation process is described as minimising the loss function:
L= Ex∼ pdata [|G ( z )− x|2],
P= N correct ⋅ 100 %,
N total where x – is the target image or 3D model, G ( z ) – is the result of generation based on the text prompt z, . pdata – is the distribution of real data. The system underwent alpha testing (with the participation of paramedics and military instructors) and beta testing (with the involvement of students and volunteers). The success rate was evaluated based on the results:
where N cor rec t – is the number of correctly performed actions, N total – is the total number of necessary steps in the scenario. Thus, the methods used combine quantitative assessment of effectiveness, simulation modelling of crisis scenarios, and a generative approach to creating visual content, which provides a comprehensive study and confirmation of the performance of the VR/AR system in pre-medical care. A detailed analysis of modern approaches to first aid training and their limitations has been carried out. – High realism, complex – High cost of equipment and maintenance. scenarios, role-playing – The need for special sites – restrictions on games with actors. accessibility during mass attacks or – Suitable for professional evacuations. training (doctors, military). – Limited scalability – suitable for a narrow circle of specialists, but not for mass civilian exercises.
E-learning online (videos, interactive modules)
/ – Easily scalable, available – Lack of practical, tactile experience – it is courses with remote learning. challenging to teach motor skills only through – Possibility of theoretical video. training at any time. – Low motivation for repetitions without interactive elements. – They do not reproduce the stressful context of combat conditions.
and on devices
games – Engagement through – Limited motor skills (hand manipulation, simulators game mechanics; increase pressure force, etc.).
PC/mobile motivation. – The issue of portability – gamified execution – Good for learning does not always translate into correct physical algorithms, decision- actions.
making, and theory.
VR simulations – High degree of – Lack of haptic/tactile feedback: applying a (whole virtual immersion: visual and tourniquet, the correct compression force is reality) audio stressors, limited brutal to feel without a dummy or special visibility, timers – bring equipment.
AR systems, in particular AR add-ons over mannequins (a hybrid of tactile and visual feedback).
the conditions of war – Technical requirements (headset, powerful closer. PC) and dependence on power supply/Internet, – The possibility of which can be critical during attacks. multiple repetition and – Funds (purchasing headsets for mass training variability of scenarios at can be expensive). low risk. – Cybersecurity and data protection: Telemetry – The ability to integrate collection requires compliance with policies analytics and formative (GDPR, etc.). feedback (metrics Q, P, etc.). These approaches were key in project, which showed an increase in P. – Combines the tactile The need to synchronise equipment (dummy experience of the dummy and AR device). with visual AR – Limited mobility – the dummy is difficult to information (hints, deploy centrally in large numbers. internal anatomy, Triage – The price of AR headsets (e.g., HoloLens) and directions). the complexity of integration. (The – Allows you to practice effectiveness of AR and the dummy is motor skills more confirmed in the literature and mentioned in accurately and add context the file as a promising direction.) at the same time.
Tactical/military Focused on the realities of – The need for special scenarios and coaches training (tactical hostilities: evacuation with experience. combat casualty under fire, mass casualties, – Moral and psychological costs when working care, training for Triage priorities. out extreme scenarios with actors (stress of combat medics) – It is used to train doctors participants).
and army personnel. – The difficulty of scaling up to the civilian
population without adaptation.
Infrastructure vulnerability – access to electricity, internet, and training facilities may be interrupted; solutions must have offline modes or mobile options (e.g., AR on a smartphone).
The need for mass scale – to train large groups quickly and simultaneously; Traditional methods and high-fidelity centres are not suitable.
Psychological stress – training sessions should simulate emotional stress, but not traumatise; adaptive difficulty levels are required.
Personnel and financial constraints – shortage of instructors and money during hostilities.
Ethical and legal issues – quality and certification of educational content, protection of participants' data.
The blended learning approach combines online theory and VR scenarios with offline practice on mannequins (or AR and mannequin). It reduces the need for physical infrastructure and preserves the tactile experience. Mobile AR solutions consist of developing lightweight AR versions for smartphones that can be run offline and that will give basic visual instructions (important in restricted areas). The integration of gamification and adaptability aims to increase motivation and gradually increase the stress load according to the user's level. AR and Key Skills Haptics allows you to invest in developing budget-friendly haptic solutions or supporting mannequins with AR add-ons for the most critical practical exercises. Content pipeline optimisation enables the use of generative AI (text → 3D) for rapid scripting, accompanied by a post-processing and LOD optimisation step for VR applications. It partially addresses the problem of scenario variability at moderate costs. Standardisation of assessment is based on the introduction of unified metrics (P, Q, τ) to compare the effectiveness of different approaches and qualitative improvements (to assess success). Pilot programs and multi-stage testing are initially conducted with experts → subsequently with target groups → scaled; concurrently, work on certifying content in accordance with the guidelines (Ministry of Health, Red Cross).
Modern approaches have strengths, including tactility (mannequins), realism (simulation centres), scalability (e-learning), and immersion (VR/AR). However, no approach alone satisfies the entire set of requirements that arise in wartime: scale, accessibility, realism of stress, tactile feedback, and infrastructure endurance. The best strategy is a combined approach: VR / AR ( for context ∧decision−making training )
+mannequin / h aptics ( for motor skills ) +mobile solutions ( for accessibility )+ generative AI ( f ∨rapid scenario building ).
It corresponds to both the project concept and the data obtained in testing (an increase in P in the VR/AR group). Practical training in pre-medical care during wartime requires an integrated approach that considers not only the development of basic skills but also psychological stress, logistics, and the availability of technology. Today, there are several primary methods: traditional training on mannequins, simulation centres, e-learning, serious games, VR/AR systems, and their combinations. Traditional training provides the development of motor skills, but is limited in the variability of scenarios and does not replicate the real context of stressful conditions. Simulation centres featuring actors and high-fidelity equipment enable you to achieve a high level of realism, but they remain expensive and limited in scale. Online courses and mobile applications have advantages in accessibility, but do not provide practical skills. Serious games show potential in learning algorithms and increasing motivation, but their impact on motor development remains limited. VR simulations enable you to recreate complex combat scenarios, create immersive conditions, and generate an emotional load, which is crucial for adapting to real-world events. Meta-analyses indicate that VR/AR training often surpasses the effectiveness of traditional methods in developing basic life support skills. At the same time, VR does not provide proper tactile feedback, which reduces the quality of practising critical actions such as chest compressions or tourniquet applications. AR systems, in particular those that superimpose additional layers of information on physical mannequins, partially solve this problem. In wartime, the limitations of these approaches are intensifying: power and internet outages, a shortage of instructors, high stress, and the need for rapid mass training of the civilian population. As the results of experimental studies demonstrate, the hybrid model (utilising VR for immersion and scenarios, as well as AR/dummy for practical skills) is optimal for scalable and effective training. Thus, no single approach is sufficient to solve the entire range of training tasks in wartime. The most promising direction is the integration of VR, AR, and physical mannequins into a single system with elements of gamification and generative content, enabling the quick creation of scenarios. Checklist of technical requirements for the implementation of a hybrid program (VR, AR and mannequin):
algorithms physical skills
support and visual Cost, complexity integration of
mannequin) adaptability, scalability
equipment and funds
realities of war
Figure 1 clearly illustrates the strengths and weaknesses of each E-learning technology (high scalability, low realism), simulation centres (maximum realism, but minimal scalability), and the VR/AR and mannequin hybrid (optimal balance of both characteristics). Modern challenges related to the war and massive attacks on civilian infrastructure necessitate rapid and scalable training in pre-medical care among civilians and paramedics. Traditional approaches (dummies, simulation centres) are limited by accessibility, scalability, and the ability to recreate a realistic combat environment. Therefore, there is a need to create an integrated VR/AR simulator that combines immersion in a stressful environment with practical practice of critical skills.
Virtual reality (VR) for immersion in context:
a. Reproduction of realistic war scenarios: a street after a missile strike, shelling of residential areas, mass destruction.
b. An emotionally and stressful environment (sounds of explosions, smoke, and chaos), which fosters psychological stability and brings training closer to realworld conditions.
c. A triage system (green, yellow, red) that requires the user to make decisions in a short amount of time.
d. Generative AI content for the variability of scenarios (different types of injuries, number of victims, civilian/military injuries).
Augmented reality (AR) for practical skills:
a. Use of AR add-ons on physical mannequins: the user sees visual cues (the placement of the tourniquet, indicators of vital functions, and anatomical landmarks).
b. AR mode on a smartphone or headset (such as HoloLens or Meta Quest Pro) allows you to train in mobile conditions, even without a full-fledged VR room.
c. Support for a "hybrid dummy": a CPR dummy for practising compressions and a limb for applying a tourniquet with AR visualisation of bleeding.
Intelligent Assessment and Analytics System:
a. Automatically recorded metrics of success (P), quality of execution (Q) and reaction time (τ).
b. Adaptive Difficulty Level: The system increases the intensity of the scenarios as the user's results improve.
c. Real-time feedback (choosing the wrong algorithm, reacting too slowly, or an error in applying the tourniquet).
Mobility and scalability:
a. Support for both stationary VR centres (for group training) and mobile AR versions (for smartphones and tablets).
b. Offline mode of operation for areas with Internet and power outages.
c. Possibility of use in schools, universities, medical institutions, as well as in military units.
Gamification and Scenario Learning: a. A point system that motivates you to re-pass.
b. Branched scenarios: incorrect actions lead to a complication of the situation (deterioration of the condition of the wounded).
c. Multiplayer mode (teamwork of paramedics, evacuation, and distribution of roles).
Expected advantages of the concept: Realistic simulation of combat conditions is not available in traditional methods. Scalability and accessibility: from VR rooms to mobile AR solutions.
Combination of motor skills and cognitive skills: VR for algorithms, AR and dummy for tactile actions.
Improving learning effectiveness, in particular, the study's results indicate an increase in the P success rate by ≈30% approximately when using VR/AR compared to traditional learning.
Thus, the concept of a VR/AR simulator of first aid is a hybrid learning ecosystem that allows you to quickly, efficiently, and massively teach civilians and the military basic life-saving skills in extreme conditions of war and emergencies.
5.1. Canvas of the business model according to O. Osterwalder The developed business model is based on the concept outlined by O. Osterwalder, which allows for a comprehensive reflection of the logic behind creating, promoting, and consuming the value of an innovative VR/AR product for pre-medical care training. The target audience of the project includes three main groups of users: representatives of the educational environment, civil society and professional services. The first group includes high school students, medical university and college students, and teachers who seek to introduce interactive technologies into the learning process. The second group is represented by citizens who want to master first aid skills and volunteers working in risk areas. The third group comprises military personnel, medical personnel, and tactical medicine instructors who aim to refine their practical skills in emergency situations.
The value proposition of the product is to create a safe, accessible and realistic environment for training actions in critical situations. The use of virtual and augmented reality technologies provides the ability to simulate scenarios of wounds, bleeding, burns, explosions and heart attacks with the reproduction of psychological reactions of victims. In the VR environment, the user interacts directly with the training objects, whereas the AR version allows training with lifelike mannequins or people, superimposing virtual injuries on physical objects. Gamification elements, such as difficulty levels, achievements, and competitive modes, help increase motivation, while the automatic assessment and feedback system provides real-time error correction.
The key factor in the successful functioning of the business model is an effective system of communication with potential users. A combination of online and offline channels is provided: the official website and marketplaces (Steam, Oculus Store, Google Play, Apple App Store) provide product distribution and the ability to purchase or subscribe; cooperation with educational institutions, military academies, the Ministry of Health, the Red Cross and public organizations contributes to the integration of the VR/AR simulator into academic programs. The active use of social networks and specialised media platforms (YouTube, TikTok, Instagram, LinkedIn) allows you to expand your audience reach. At the same time, demonstrations at medical forums and thematic exhibitions help build trust among professional users.
Interaction with clients involves several levels of engagement, including individual training at your own pace, group classes for educational and military institutions, as well as corporate solutions for medical centres. User support is implemented through a chatbot, an FAQ system, a forum, and a specialised Discord server. To increase engagement, various motivation mechanisms are being implemented, including ratings, achievements, thematic challenges, and loyalty programs. Additionally, users can provide feedback and suggestions for improving simulations, participate in beta testing, and access new scenarios through a subscription system.
The financial model encompasses a combination of several revenue streams, including the sale of a licensed version of the product, a monthly or annual subscription model, corporate licenses for educational and medical institutions, as well as grant and state funding within the context of medical education development programs. Additional sources of income can be affiliate programs with medical device manufacturers and branded training courses.
Human, technical, informational and financial resources are needed to implement the project. The team consists of developers of VR/AR environments, 3D designers, medical consultants, game designers, UX/UI designers, and marketing specialists. The technical base includes Unity or Unreal Engine 5 platforms, 3D modelling programs (Blender, 3ds Max, ZBrush), VR/AR equipment, as well as cloud services such as AWS, Google Cloud, or Azure for data storage. Information resources are developed based on official protocols for pre-medical care, the practical experience of doctors, and the results of scientific research on the effectiveness of VR/AR technologies in education.
Key activities include creating a minimum viable product (MVP), conducting market research, identifying user needs, testing prototypes, obtaining certification according to educational and medical standards, marketing, and scaling the product. At the testing stage, it is planned to involve medical specialists, military personnel and volunteers to assess the quality of simulations and provide professional feedback. The project's partner ecosystem includes medical organizations (Ministry of Health, Red Cross), educational institutions (medical universities, colleges, military lyceums), military and rescue services (Ministry of Defense, State Emergency Service), as well as technology partners – developers of VR/AR equipment and platforms (Meta, HTC, Valve, Unity, Unreal Engine) and companies specializing in generating 3D content using artificial intelligence.
The cost structure encompasses the expenses associated with developing training scenarios and medical expertise, remuneration for the development team, purchasing software licenses and VR/AR equipment, providing server infrastructure, marketing, administrative costs, and certification. A vital budget item is also the promotion of the product through participation in exhibitions, the creation of advertising materials and collaborations with medical and technology bloggers. Thus, the presented business model demonstrates an integrated approach to creating an educational VR/AR product that combines technological innovation, pedagogical efficiency, and social significance. Its implementation will contribute to increasing the level of training for the population, doctors, and military personnel in emergencies, as well as the development of the domestic segment of XR technology use in the field of medical education. 5.2. Structure/stages of implementation of the MVP VR/AR-simulator of first aid The development of a minimum viable product (MVP) for a VR/AR simulator of pre-medical care involves a comprehensive, phased implementation that encompasses research, design, technical development, testing, marketing, scaling, and project management. This sequence corresponds to the classic Work Breakdown Structure (WBS) structure and is designed to ensure controllability, efficiency, and high quality of the final result.
At the initial stage, which includes research and planning, an in-depth analysis of the VR/AR education market is carried out. Particular attention is paid to the study of modern trends in simulation training, the analysis of competitive solutions, their strengths and weaknesses, as well as the assessment of potential demand among educational and military institutions. The barriers to the introduction of VR/AR technologies into educational processes, the pace of their adoption, and their effectiveness in training pre-medical care specialists are investigated. At the same time, an analysis of the target audience's needs is conducted, involving interviews with paramedics, instructors, military medics, and emergency services specialists. Based on the data obtained, a list of critical skills and scenarios that need to be implemented in a VR/AR environment is formed, including bleeding control, cardiopulmonary resuscitation, and evacuation of victims.
Next, the terms of reference of the product are determined. The optimal technology stack is chosen, including Unity or Unreal Engine platforms, as well as target devices – Meta Quest, HTC Vive, HoloLens and mobile devices with AR support. An assessment of the budget and the need for specialists of various profiles (developers, 3D designers, UX/UI specialists, medical consultants, testers) is carried out, as well as possible sources of funding are analysed, including grants, investments, and crowdfunding campaigns. The result of the stage is a prepared package of documentation, including a technical description, implementation schedule, risk management plan, and presentation materials for partners and investors.
The second stage involves the direct development of an MVP. At this stage, realistic threedimensional models of the environment are created, particularly of urban locations after extraordinary events – such as explosions or shelling – with a detailed reproduction of objects, including first aid kits, stretchers, medical equipment, and transportation and equipment. Characters are developed, including paramedics and those with various types of injuries, reflecting a wide range of real-world scenarios. Particular attention is paid to VR/AR interaction, where motion control systems are configured through controllers or gloves. The mechanics of viewing the victim, interaction with objects, and tactile feedback are implemented through the vibration of controllers.
The software component includes the creation of scenarios of varying complexity – ranging from mild to critical conditions - as well as the logic for assessing the correctness of the user's actions in accordance with first aid protocols. There is a virtual instructor with voice prompts, the ability to repeat workouts and view mistakes. In the visual and auditory aspects, a dynamic sound environment (heartbeats, screams, explosions, ambient noise), the effects of destruction, smoke, blood and facial expressions of the characters are realised. At the end of the stage, the simulator is optimised for various devices – including VR headsets, mobile devices, and PCs – to ensure a balance between rendering quality and performance.
The third stage focuses on testing and improving the system. First, the team conducts internal functional testing to identify technical errors and verify the performance of the mechanics. After that, alpha testing takes place with the involvement of specialists, including paramedics, military instructors, and representatives of emergency services. Their feedback enables you to assess the level of realism in the simulation, the logic of the educational process, and potential shortcomings. Next, beta testing is organised among students, volunteers, and rescuers, which allows you to evaluate the intuitiveness of the interface, the effectiveness of the training, and the level of user immersion. To increase stability, automated tests and unit testing of critical algorithms are being implemented. Based on the results, bugs are eliminated, object physics, animations, graphics and interface are improved.
The fourth stage concerns marketing and promotion. The launch of the official website and social network pages allows you to present the product, post educational content, video reviews, and publications about VR/AR training. Targeted advertising and SEO optimisation help to expand audience reach. An important area is cooperation with educational institutions, military academies, and medical organisations, which are provided with test access and preferential conditions. Demonstrating the product at exhibitions, conferences, and forums helps establish contacts with potential partners, while collaborations with medical bloggers and experts build trust among the general public.
The fifth stage involves scaling and developing partnerships. The product is integrated into the educational programs of universities and colleges, and specialised training courses and methodological materials are created for teachers. It is planned to introduce online training with certification and a system for testing results. Further expansion of the functionality includes the addition of new scenarios, such as traffic accidents, fires, natural disasters, and terrorist acts. To increase international competitiveness, product localisation, content translation into multiple languages, and adaptation of cultural characteristics in training are carried out.
The final stage – project management – involves the systematic control of task implementation, legal support, financial monitoring, and personnel management. Agile management methodologies (Agile, Scrum, Waterfall) are used to increase the effectiveness of teamwork. Reports, analytical reviews and strategic meetings are held regularly. Registration of intellectual property, compliance with international safety standards and medical certification are ensured. Much attention is paid to team building – creating an organisational structure, establishing effective communication, implementing a motivational system, and providing professional training. Financial management encompasses budget control, cost analysis, and cost adjustments in response to project changes. In addition, risk management is carried out, ranging from technical to legal aspects, with the development of response plans and mechanisms for minimising risks. Thus, the presented structure of the WBS provides a systematic approach to implementing the MVP VR/AR simulator for premedical care, combining technical, educational, medical, and organisational components. This model enables us to gradually develop a high-quality, scientifically based product with a high potential for implementation in the field of first aid training, both in Ukraine and internationally. 5.3. Concept and structure of the project The primary objective of the project is to develop an interactive educational environment that fosters the acquisition of practical skills in pre-medical care during emergencies, particularly in conflict situations. The project aims to teach users how to respond quickly, accurately, and consistently in critical situations where the preservation of life depends on the accuracy of their actions. One of the key tasks is to develop a system that combines virtual (VR) and augmented reality (AR) technologies with gamification approaches to learning. This format enables you to enhance user engagement, create a safe yet psychologically realistic training environment, and ensure the effective assimilation of knowledge through practice. In the context of modern challenges caused by the war, digital modelling of situations involving assistance to victims is becoming a tool not only for education but also for preparing the population for real crisis conditions. The project is designed for a broad audience of users, encompassing educational, civil, and professional fields. For pupils and students, the VR/AR simulator can be used during medical training or life safety classes, creating an opportunity to consolidate theoretical knowledge through simulation experience. Citizens and volunteers get the chance to train in first aid on their own, working out the algorithms of actions in situations of mass destruction or natural disasters. For military personnel and medical specialists, the project acts as a platform for realistic simulations of combat injuries, actions in stressful conditions, and teamwork coordination. Thus, the product serves both educational and social functions, contributing to an increase in the level of public preparedness for actions in extraordinary circumstances.
The software product implements complex functionality aimed at replicating the full cycle of first aid training, from familiarisation with the theoretical foundations to the assessment of the user's practical actions. The basic components of the system include simulation of critical situations (injuries, explosions, collapses, and missile strikes), interactive training scenarios, and a module for assessing the effectiveness of actions. In a VR environment, users can perform typical procedures, such as applying a tourniquet, stopping bleeding, performing cardiopulmonary resuscitation, dressing wounds, and transporting victims. VR/AR technologies enable realistic interaction with medical instruments and training mannequins, and also allow for capturing every user action. The system automatically analyses the correctness and sequence of actions, response time, and also provides individual feedback. Gamification mechanics – difficulty levels, missions, rewards, ratings – increase motivation to learn and contribute to the consolidation of skills.
The visual style of the VR/AR environment is determined depending on the learning goals and technical limitations of the platform. In creating the most realistic scenario, highly detailed graphics with physically correct lighting and textures are used, which enhances the user's emotional engagement and the reliability of the simulations. For mobile devices and educational institutions with limited technical resources, it is advisable to use stylised graphics, such as Low Poly or Stylised, which provide an optimal balance between performance and visual quality.
The technical architecture of the project is based on the use of modern game engines, such as Unity or Unreal Engine 5, which provide full support for VR/AR technologies. The software part is optimised for VR headsets (Meta Quest, HTC Vive, Valve Index) and mobile devices with AR support. Modelling of three-dimensional objects is carried out in Blender and 3ds Max environments, and artificial intelligence tools are utilised to generate additional models and textures, including Stable Diffusion, Leonardo.Ai, and Trellis3D. This approach enables you to reduce content development time, provide a variety of visual elements, and increase the photorealism of educational scenes. The main simulation scene recreates the urban environment after a missile strike or explosion (Fig. 2). The location is located in typical Ukrainian urban conditions (for example, on Lviv Street), where the player sees destroyed buildings, damaged cars, smoke, fires, and debris. In such realistic circumstances, the user assumes the role of a paramedic who must assess the situation quickly, prioritise assistance, and save the victims. The scene includes three levels of severity of the victims according to the Triage system: critical, moderate and mild. For example, an unconscious child needs cardiopulmonary resuscitation, his mother has severe bleeding from his arm, and another victim has burns to the upper body. The player must consistently perform actions guided by the algorithms of first aid. In VR mode, interaction occurs in the first person: the user sees themselves in a paramedic's uniform, has the opportunity to open a medical backpack, use tourniquets, bandages, scissors, and dressings. The hint system helps to navigate the correct sequence of actions, and the time limit creates the effect of realistic stress. In AR mode, the user can perform the same actions on physical mannequins or training partners, with a virtual reconstruction of wounds and injuries superimposed on top. To enhance the educational value, the project features a system of scenarios and training modules, comprising a theoretical component, practical development, and final testing. Additionally, there is a multiplayer mode that allows you to train team interaction.
The audiovisual component plays a crucial role in creating the illusion of presence. The realistic sounds of explosions, sirens, footsteps and screams of victims are complemented by voice prompts that simulate communication with an ambulance dispatcher. The user interface (UX/UI) is built on the principles of intuitive interaction: the main menu allows you to choose a scenario, difficulty level, and view educational materials. During the simulation process, AR prompts provide contextual instructions, such as: "Place the tourniquet above the wound." Artificial intelligence (AI) within the system is used to dynamically generate scenarios, adapt complexity to the user's level, and simulate the behaviour of victims who may change their state, lose consciousness, or exhibit signs of panic. It makes each workout unique and increases the realism of the learning experience. Thus, the proposed VR/AR project constitutes an innovative educational platform that combines technological, psychological, and pedagogical aspects of preparing the population for the provision of pre-medical care in war and emergencies. 5.4. Using Artificial Intelligence to Visualise Ideas and Concepts Artificial intelligence (AI.Artificial Intelligence (AI) plays a key role in modern approaches to visual content creation. Thanks to machine learning algorithms and deep neural networks, the process of developing concept art, 3D modelling, and artistic visualisation is greatly simplified. AI systems can generate illustrations, stylised images, and even full-fledged three-dimensional models based solely on text descriptions (known as prompt-based generation). The main uses of AI for visualisation include:
Generation of 2D images by text prompts – DALL·services E, MidJourney, Stable Diffusion, and Leonardo.Ai allow you to create artistic images of any style.
Converting sketches into detailed illustrations (Deep Dream Generator, Artbreeder).
Create 3D models based on 2D images or text descriptions (Trellis3D, Meshy, Sloyd, Hunyuan3D).
Automatic texturing and rendering (NVIDIA GANverse3D, AI Render).
Thus, the application of AI enables artists, designers, and developers to significantly reduce the time it takes to create visual concepts and improve the quality of results by iteratively improving the generated images.
Generative AI is a subfield of AI that focuses on creating new content, including images, texts, music, and videos. Generative models not only analyse data but also reproduce new, unique artefacts, imitating the style or structure of the original dataset. Generative AI's typical use cases are for images (DALL· E, MidJourney, Stable Diffusion), music (OpenAI Jukebox, AIVA, Soundraw_, texts (GPT (ChatGPT), Bard, Claude) and video (Runway Gen-2, Pika Labs).
Diffusion Models – used in Stable Diffusion and DALL· E 3. The principle of operation is to gradually "cleanse" the noise image until a clear structure is obtained. They are characterised by high detail and controllability of the process.
Generative adversarial neural networks (GANs) are implemented in StyleGAN and NVIDIA GauGAN. One network (the generator) creates content, while the other (the discriminator) evaluates its quality, ensuring a gradual improvement in the result.
Transformers are architectures that use the attention mechanism to process sequences of data. In DALL·E models E3, Imagen, and Parti Transformers analyse text descriptions, converting them into vector representations that control the generation process. Encoding text into a DALL· E 3, High quality, Require large vector space to control Imagen, precise semantics amounts of data and generation Parti control resources
A prompt is a text description or instruction that the user enters into a generative model to achieve the desired result. In the context of image generation, the prompt determines the style, colour palette, level of detail, and other characteristics of future graphics. The quality of the final result directly depends on the accuracy and detail of the product. For example, wording that is too general leads to fuzzy or random results. As an illustration:
Ineffective prom: "The city after the war" is too general a description, lacking specific details of the scene.
Effective prompt: "The city was destroyed after a missile attack, cars were burning, and smoke was everywhere. A paramedic in a red uniform helps the wounded. Realistic style, cinematic lighting, high detail" – the project includes key elements of the scene, characters and artistic style.
Generative models analyse the prompt, encode it into a vector space, and use it as a guide for building visual content. For automated or simplified creation of projects, specialised generative models and algorithms are used:
GPT (Generative Pre-trained Transformer) is a text-based AI model that helps generate complex and detailed images, videos, or music.
The Stable Diffusion Prompt Guide is a set of recommendations and tags that allow you to improve the results for generative models.
The MidJourney Prompt Generator is an algorithm adapted for creating structured prompts optimised for generating stylised images in MidJourney.
Benchmarking tools for generating outputs demonstrate different approaches to the convenience, flexibility, and accuracy of results. Visual Motion Builder for Intuitive interface, support Limited functionality AI for different models for complex queries Automatic formation of Flexibility, adaptation to Sometimes generates detailed products different models too generic prompts
promotions of popular An extensive database of Some premium
examples, filtering promotions are paid
To create low-poly scenes focused on stylised game levels, the prom must accurately reflect the composition, key elements, and art style. Conceptually, the prom is designed to provide simplified geometry, a bright palette, and stylised lighting. Example of prompts for a low-poly scene: «Low-poly city street after a missile strike, with destroyed buildings, rubble, smoke, and damaged vehicles. A paramedic in a red uniform is helping injured civilians. Simplified polygonal style, bright colours, clean edges, stylised lighting.»
«Low-poly warzone city, broken roads, ruined buildings, burning vehicles, first responders aiding wounded people. Stylised 3D environment, cinematic lighting, dramatic atmosphere.»
«A male and female paramedic standing side by side in bright red uniforms with reflective stripes, wearing blue latex gloves. Each carries a medical bag with emergency supplies. They are scanning the area, ready to assist. Stylised low-poly aesthetic, cinematic lighting, sharp edges, and vibrant contrast.»
«A wounded mother standing with a bleeding arm, holding her 7-year-old daughter. The scene depicts a post-apocalyptic warzone with smoke, rubble, and emergency responders in the distance. Stylised low-poly aesthetic, dramatic lighting, strong emotional tension.»
Thus, the products enable you to provide accurate, artistic, and technical specifications of lowpoly scenes for VR/AR simulations. For the practical implementation of 2D and 3D concepts, various generative tools are used, differing in the level of control, detail and functionality. The primary free services include:
from descriptions images Easy to text integration
ChatGPT use, Limited number 3–5 with of styles generations/day
built-in styles images
Tool integration Photoshop quality, with
content support, functionality in generations/day simple interface the free version Sketchfab platform is a well-known service that contains thousands of free models with textures (Fig. 36).
Sky
Search queries destroyed street, war zone street, post-apocalyptic street damaged building, destroyed house, ruined building wrecked car, damaged car, burnt vehicle, crashed car ambulance, emergency vehicle, hospital van, paramedic car first aid kit, medical bag, ifak, emergency box asphalt texture, road material, pavement PBR texture
rubble, concrete debris, destruction pieces, broken wall smoke VFX, dust particles, fire smoke, explosion debris Medical instruments
medical scissors, tourniquet, emergency tools, bandage
To create a scene as realistic as possible for the VR project "First aid after a missile strike on a residential street", a search was conducted for free 3D content on leading platforms for 3D models. The main selection criteria were: the ability to download models for free; availability of formats compatible with Unreal Engine (mainly .fbx, .obj); compliance with the project theme: destroyed buildings, debris, medical instruments, smoke, transport, etc.;
adaptation of models for VR scenes (low/medium polygonality, PBR textures, presence of regular/reflection maps, etc.).
The CGTrader platform is a professional resource with paid and free models, including scanned objects (Fig. 38).
PolyHaven platform - free PBR content: HDRI sky, materials, models (Fig. 38).
An analysis of the available content on the TurboSquid and Free3D platforms was conducted, revealing that most models are either paid or lack a relevant topic. Normal free VR content for the scene was not found. BlenderKit is a plugin for Blender (Fig. 39) that enables the import of free 3D assets. Some models are suitable for export to UE (via .fbx).
3D Warehouse (SketchUp) – models in Fig. 40 in .skp format (can be converted to .fbx via Blender or SketchUp).
The search made it possible to form a full-fledged library of objects for creating a VR scene on the topic of first aid: destroyed objects of the city; realistic textures of surfaces and sky; vehicles (including ambulances); first-aid kits, tourniquets, dressings; effects of smoke and explosions.
Content from the Sketchfab, CGTrader, and Poly.pizza platforms is the most suitable for Unreal Engine without additional processing. Additional resources, such as BlenderKit and 3D Warehouse, expand the possibilities for VR-adapted models.
Paid models that would significantly improve the implementation of the VR project scene on the topic "First aid after a missile attack". They are of high quality, optimised for Unreal Engine, often contain PBR textures, LOD levels, animations, or manageability (e.g. transport):
1. WW2 Warzone Environment Megapack includes: dozens of buildings, debris, barricades, sandbags, smoke and fire effects. Why it is essential: allows you to quickly create a cohesive scene with a high level of destruction. Suitable for visualising the street after a missile strike.
2. Ambulance (Drivable) includes: a realistic ambulance with interior, handling, and door opening. Why it matters: a key element of the first aid scene. Allows interactive interaction (boarding, loading victims).
3. First Aid Set includes: first aid kits, bandages, tourniquets, scissors, syringes, etc.
Why it matters: used to simulate the user's actions in a VR scene – applying bandages, stopping bleeding, etc. General benefits of using paid models
Unreal Engine supports importing third-party 3D models in various formats, including .fbx, .obj, .glb, and .gltf (Fig. 41). The FBX format is the most widely used because it stores geometry, animations, skeletons (for skinning), materials, and basic texture information. Instructions on how to import 3D models: 1. Download the model from the resource.
2. Make sure it is in the correct format. If it is in another format (e.g., .blend or .skp), convert it to .fbx using Blender.
3. In Content Browser Unreal Engine, create a new folder.
Create a new Material or open the one you created during import.
Drag and drop each texture into the Graph of the Material editor.
Connect accordingly:
Click Apply to save the changes. 6. After successfully importing the models, they can be used on the stage:
Drag and drop from Content Browser.
Zoom, rotate, change position.
Assign a collision so that the VR character does not pass through the model.
Apply NavMesh if the object is to be used as a constraint for movement.
During the execution of this task, the following was performed (Fig. 43): - Import 3D models. - Manually configured PBR textures (where required) in the Material Editor. - Added models to the scene, including: ambulance, first aid kits, tourniquet, scissors, debris. - Models are scaled according to the scene.
- The display of textures, the level of detail and the correctness of lighting have been checked.
In modern 3D modelling, generative neural networks (also known as Generative AI) significantly simplify the creation of complex or rare objects that are difficult to find on standard marketplaces or to create manually. It is especially true for highly specialised VR scenes, such as:
Simulations of medical emergencies; Damaged buildings and vehicles;
Human models with characteristic wounds or postures.
Generative models enable you to minimise the lack of unique props, automate part of the 3D production process, and adapt the scene to individual scenarios.
As part of this task, generative neural networks were used: 1. Mersy – to generate damaged environments: destroyed houses and damaged cars 2. Tripo – to create the environment and human models of the victims: at home, a child in an unconscious state (CPR), a mother with bleeding from the arm (applying a tourniquet) and a man with burns on the body (processing burns).
Importing the generated models to the UE (Fig. 44): 1. Export from Tripo/Mersy in .fbx or obj format. 2. Convert to .fbx via Blender if necessary. 3. A new folder has been created in the Unreal Engine project. 4. Import to ... . 5. Check the Import Mesh, Import Textures, and Import Materials checkboxes. 6. All models have been successfully imported:
Generative neural networks made it possible to significantly expand the library of 3D assets and give the scene individuality and believability, in particular:
Realistic simulation of the consequences of shelling.
Simulation of critical conditions of people for training in first aid.
The models proved suitable for VR scenes, and their ease of generation allowed for quick adaptation to training scenarios.
Collision in Unreal Engine is a system that defines the physical boundaries of a 3D object. It is used to handle collisions with other objects or a character. For example, due to a collision, the player cannot pass through walls, and objects do not pass through each other. In the case of a Static Mesh (static 3D model), the collision can be of two types:
1. Simple Collision (simplified) – usually consists of geometrically simple shapes (Box, Sphere, Capsule), fast and efficient.
2. Complex Collision – uses the geometry of the model itself for collisions. It is more accurate, but less optimised, especially for VR.
How do I create a collision for imported models?
After importing a 3D model into Unreal Engine, you need to check or create a collision manually: 1. Open Static Mesh Editor (double click on the model). 2. In the top menu, click Collision > Add and select the desired type:
Use Complex as Simple, or Simple only)
Convex Decomposition Settings: Number of Shells, Accuracy.
The player must not pass through walls.
It is essential to ensure a smooth collision with objects, avoiding "sticking".
Objects must properly interact with teleportation via NavMesh.
For imported models, a collision was implemented, as shown in the screenshots below (Fig. 45).
Smooth Locomotion is a method of moving in VR space that allows users to move smoothly, without teleportation, using analogue controller sticks or keys. This type of navigation allows: achieve a more natural movement (similar to traditional first-person games), create a deeper immersion in the VR scene, provide complete freedom of movement in the virtual world.
However, it should be borne in mind that with improper implementation or too sudden movements, symptoms of VR sickness (dizziness, disorientation) may occur. Therefore, it is essential to implement this feature with consideration for the user's comfort.
According to the guidelines and video instructions, the implementation of Smooth Locomotion in a VR template includes the following steps:
1. Disabling Teleportation: open the character Blueprint (VRPawn or BP\_VRCharacter); find and disconnect teleportation nodes; disable SnapTurn if rotation is used.
2. In the Input > Action Mappings, add new actions and assign them the corresponding axes (for example, X/Y from the controller joints).
3. Blueprint Logic: Create Variable and Create events for movement. 4. Class Settings → set Parent Class as Character and change Auto Possess Player to Player 0.
At the first stage, Smooth Locomotion was implemented according to the video available at: https://www.youtube.com/watch?v=GlctYwY-m2w. Teleportation was enabled, and the 'Teleport' nodes were removed. New Input Actions for movement and rotation were created. The corresponding Blueprint logic was implemented using Add Movement Input, Get Forward Vector and Multiply. However, after starting the project, the character was unable to move, as the movement did not work at all. After that, teleportation was added (previously disabled nodes were activated) and used an alternative technique from the video: https://www.youtube.com/watch? v=VHqtp_R37DU. Result - the character began to move smoothly with the help of the left stick, without breaking control. Teleportation also works, serving as a backup option for movement and providing a comfortable user experience. The scene uses NavMeshBoundsVolume, which allows you to select zones for teleportation (Fig. 46-50). During the work, a topical social theme was chosen – the scene of first aid after a missile strike in a residential area, which made it possible to build a practically significant VR space with educational potential. A library of 3D assets from free and paid sources, including Fab, Sketchfab, and CGTrader, was selected, and generative neural networks (Tripo, Mersy) were used to create unique models of victims and destruction. Implemented:
Content migrationfrom previous projects using Asset Actions > Migrate function. Import 3D models with subsequent adjustment of PBR textures in Material Editor.
Implementation of physical collisions for imported models.
Smooth Locomotion – smooth movement using the controller joints, implemented according to the video instructions.
Particular attention is paid to ensuring the convenience of movement in the VR environment, which is critical for preventing virtual disorientation. An attempt was made to implement movement without teleportation, but a combined approach with teleportation enabled and smooth motion activated proved effective.
The purpose of creating the User Interfaces menu in VR Template is to provide the user with the ability to quickly access auxiliary information, settings, and end the session. The menu is implemented as a UMG menu, part of the virtual reality template, and modified to meet the project's needs (Fig. 51-52).
1. Familiarisation with the UMG template menu. The output menu is located at the path: Content > VRTemplate > Blueprints > WidgetMenu. The project opened the Widget Blueprint of the standard menu, using the Unreal Motion Graphics (UMG) system.
2. The menu has been modified as follows - the basic design of the template menu has been preserved, and two new buttons have been added: Instructions and Settings.
3. Adding the Instructions button. In the Designer section, the InstructionsButton has been added to the vertical block. A Text_Instruction text element is bound to the button, which displays the text of the instructions. The Graph implements logic – when you click on the Instructions button, a message with instructions for the player appears on the screen.
4. Adding the Settings button. Similarly, the SettingsButton was added. An OnClicked event is configured in the Graph that displays the message: The settings menu will be implemented soon.
5. Embedding a modified menu in a VR scene. The menu is connected to the Blueprint Menu, which is responsible for the interaction of the controllers with the UI. It has been verified that when the user presses the menu button on the VR controller, the menu appears in front of them. The menu is not permanently active in the scene; it is called up when necessary by pressing a button. The menu is positioned in front of the player in a convenient field of view, oriented relative to his position.
Video is used as educational content for the user (instructions), and audio adds atmosphere and enhances the effect of presence, reacting to events in the scene (Fig. 53-55).
1. Audio content – two sound effects are implemented: 1.1 Alarm siren. The file has been imported into the project. Added Audio Actor at the scene level. The following parameters are set: Auto Activate (true) and Looping (true). This sound is activated automatically when the simulation starts and repeats continuously, simulating air raid conditions.
1.2 Missile hits. The file has been imported into the project. Added Audio Actor configured to activate with a delay. The sound of the hit is not repeated, but sounds once - creating a dramatic effect. 2. Video content. 5 educational videos have been added to the VR scene (Fig. 56-57):
2.1 Import media files. Video files in .mp4 format have been imported into the project. 2.2 Implementation of the menu for video selection. Five new buttons have been added to the Widget Blueprint menu: Button_Triage, Button_CPR, Button_Bleed, Button_Burns, and Button_Panic. In the Graph tab, OnClicked event handling has been added to each button, which activates the corresponding Media Player:
2.3 Video display in the scene. The stage features a Static Mesh Plane to which footage from the video is applied. The video is displayed as a "screen" within the training site area. It significantly enhances the interactivity and information content of the VR environment, enabling the user not only to act but also to learn in real-time.
Created your own Blueprint Actor. The following settings are activated in the Static Mesh component: Simulate Physics (true) and Collison Preset (PhysicsActor). Scissors and a tourniquet can now be picked up and used in a VR scene (Fig. 58).
The basic VR menu has been modified, with new function buttons that incorporate their own logic (instructions and settings) added. Spatial audio accompaniment (e.g., siren, missile hit) has been implemented, along with added educational video content in the form of a multimedia screen with interactive selection options. An improved object capture system has been introduced, and physics have been adjusted for first aid items, including the harness and scissors. Screenshots from the test are shown in Fig. 59.
A 3D scan of objects of the real environment was carried out – children's swings located on several playgrounds in my city. These objects are chosen because of their characteristic shape, the presence of clear structures (such as supports, seats, and chains), and their ability to be applied in scenes of a gaming or training VR environment. To perform the scan, the following was used:
Smartphone: Samsung Galaxy A52 (64 MP primary camera)
Photogrammetry app: RealityScan by Epic Games (Fig. 60).
The 3D shooting process was carried out in Camera Control mode, which allows you to independently determine the angles and shooting positions. We walked around the subject in a circle, swinging at different heights. However, instead of the recommended 80-100 frames, we deliberately limited ourselves to a smaller number of photos (about 30-50 per model). The goal was to achieve the effect of incomplete detail and non-critical artefacts during processing, so that the result appears "damaged", as if an explosion had damaged an object. It is essential to clarify that the swing itself was in good technical condition, without visible deformations or breakdowns. However, given the theme of my VR project – the scene of providing first aid in the city after a missile strike, I needed to adapt the 3D models to the appropriate visual style. That is why reducing the number of shots made it possible to get models with a "damaged" appearance — partial mesh distortions, unfilled areas, and inaccuracies in textures that look quite logical in the affected area. Once the shooting was complete, RealityScan automatically processed the image and formed a three-dimensional model. The models were not uploaded to Sketchfab, but exported locally in a .zip format that contained .glb files and their corresponding textures. In the future, we plan to import these models into Unreal Engine, where they will be placed in the scene – specifically, in the destroyed courtyard of a high-rise building near the epicentre of the explosion. In total, dozens of swing models have been created, each of which has individual features of the shape and level of detail. They will be used as part of the statistics in the first aid simulator project. Screenshots of the models are shown below in Fig. 61:
During the experimental validation, the basic concept of a VR/AR simulator for first aid was developed, and scenarios for training (tourniquet application, CPR, and burn assistance) were created. A system of evaluation and gamification (Triage, missions, difficulty levels) is proposed. Graphic content generation (2D sketches, 3D models of the environment and characters) was performed. The potential impact is demonstrated: reduced panic, skill development, and accessibility for various user groups. As a result of the work, a conceptual model of an intelligent VR/AR system for teaching first aid was created. The model includes:
Training scenarios include tourniquet application, bleeding control, cardiopulmonary resuscitation (CPR), assistance with burns, and actions during mass casualty incidents.
Interaction mechanics: use of VR controllers, AR overlay on a physical mannequin, and a feedback system.
Evaluation system: integral metric of the quality of performance of actions Q.
The developed scenarios take into account the conditions of damage to civilian infrastructure after missile strikes. For each scenario, a set of victim states S and a transition function are defined: st+1=f ( st , at , τ ). (7)
The analysis showed that the timely and correct execution of actions reduces the likelihood of transition to a critical state by 35-40% compared to no intervention.
Figures 62-63 show graphs that visualise the key quantitative results of the study. The graphs visualize the time distribution into the main phases of MVP development (WBS) and the results of the photogrammetry experiment (comparison of input data).
According to the results of beta testing (50 participants: students, volunteers, civilians), the average level of success was recorded:
P= N correct ⋅ 100 %.
N tot al (8)
Obtained values (Fig. 64) are for rourniquet application P=87 %, CPR P=78 %, burns and shock P=72 %. Compared to the control group trained by traditional methods, the results were 20-25% higher. The use of generative models (Stable Diffusion, Trellis3D, Meshy) allowed:
reduce the time for prototyping training scenes from 2-3 weeks to 2-3 days,
automate the generation of 3D characters and environments, to provide a variety of scenarios without a significant increase in costs.
L= Ex∼ pdata [|G ( z )− x|2], showed a stable convergence in different text outputs, which guarantees the quality of content for VR/AR environments. Overall assessment of learning effectiveness by the E integral indicator: (9) (10)
, where are the indicators of correctness, speed and consistency, showed a value pi = 0.84 (on a scale from 0 to 1), which indicates the high quality of training. Comparative analysis (Fig. 65): Traditional training (lectures and mannequins: medium P ≈ 60 %) and VR/AR system (medium ). Thus, the use of an intelligent VR/AR system increases the level of readiness for first aid by 30% compared to classical methods.
The graph clearly shows that the VR/AR system increases the average level of learning success from ≈60% to ≈79%, that is, by 30%. The results confirm that the VR/AR system can become an effective training tool for civilians, students, volunteers and doctors, and also has the prospect of being implemented in military training.
Modern military realities and large-scale attacks on Ukraine's civilian infrastructure create a critical need for rapid and practical training of the population in first aid skills. In the conditions of missile strikes, mine explosions, mass casualties and emergencies, the lives of the victims depend on the correctly performed priority actions. Traditional training methods, based on lectures or the use of mannequins, have limited effectiveness because they fail to replicate the stress factors, dynamics of combat wounds, and realistic scenarios that occur in actual combat. At the same time, the development of virtual and augmented reality (VR/AR) technologies opens up new opportunities for creating interactive learning environments. Intelligent VR/AR simulators enable you to simulate realistic scenarios – from applying a tourniquet to stop bleeding and performing cardiopulmonary resuscitation to assisting victims in states of shock or with burns. Thanks to the integration of gamification, the Triage system and automatic evaluation of user actions, such systems combine the effectiveness of training with safe training conditions, eliminating risks to life. The target audience of such solutions is not only doctors and the military, but also pupils, students, volunteers, and ordinary citizens who must be ready to act in crises. It is also essential that VR/AR systems provide accessibility, allowing learning to be possible anywhere and at any time, even using mobile devices with AR applications. It makes the technology a universal tool for preparing society for the challenges of wartime. As a result, the features of the proposed information technology are as follows: VR/AR technologies are an effective tool for teaching first aid in wartime.
The intelligent system enables you to safely practice skills, develop psychological readiness, and ensure the repeatability of scenarios.
Generative AI significantly speeds up the creation of educational content.
The proposed approach can be integrated into the curricula of schools, universities, military academies and community organisations.
Further research involves scaling the system, expanding scenarios, and certifying the product to medical standards.
Thus, the development of an intelligent VR system for first aid aims to enhance the population's safety level, develop practical skills in crisis conditions, and reduce panic levels during emergencies. This article explores the concept of creating such a system, its key elements, and the potential applications in the fields of education, medicine, and military training.
The materials of our project align with key trends, including a combination of realistic scenarios (urban destruction), integration of the Triage system, gamification, and automatic assessment, as well as the active use of generative AI to accelerate the creation of 2D/3D content. Our empirical results (increased P in VR/AR compared to traditional training, improved tourniquet performance, CPR, etc.) correlate with global findings on the effectiveness of VR/AR and illustrate practical applicability in wartime settings. It makes our work relevant to both the scientific community and practitioners in the field of civilian and military training. Review findings and recommendations for further research:
To prove transference to real-world settings, in particular, long-term quasiexperiments or RCTs with a focus on behavioural transference of skills in real-world training/interventions are needed.
Unifying metrics, including the use of a standard set of indicators (P, τ, consistency, correctness, quality of manipulations), will facilitate meta-analyses.
Optimise the pipeline of generative content, particularly the combination of textto-3D models (Hunyuan3D, etc.) with post-processing stages (LOD, packaging for VR), which will enable you to scale scenarios without compromising performance.
Combining an AR mannequin and VR scenes, for example, a combined approach (tactile and visual) is promising for skills that require a sense of strength/position.
During the preparation of this work, the authors used Grammarly in order to: Grammar and spelling check. Further, the authors used DALL, ChatGPT, KREA, Ideogram, Cabina.Ai (Flux and Leonardo.Ai) for figures 3-16 in order to: Generate images. After using these tools/services, the authors reviewed and edited the content as needed and take full responsibility for the publication’s content.