Immersive Virtual Reality Serious Games for Evacuation Training and Research: A Systematic Literature Review

An appropriate and safe behavior for exiting a facility is key to reducing injuries and increasing survival when facing an emergency evacuation in a building. Knowledge on the best evacuation practice is commonly delivered by traditional training approaches such as videos, posters, or evacuation drills, but they may become ineffective in terms of knowledge acquisition and retention. Serious games (SGs) are an innovative approach devoted to training and educating people in a gaming environment. Recently, increasing attention has been paid to immersive virtual reality (IVR)-based SGs for evacuation knowledge delivery and behavior assessment because they are highly engaging and promote greater cognitive learning. This paper aims to understand the development and implementation of IVR SGs in the context of building evacuation training and research, applied to various indoor emergencies such as fire and earthquake. Thus, a conceptual framework for effective design and implementation through the systematic literature review method was developed. As a result, this framework integrates critical aspects and provides connections between them, including pedagogical and behavioral impacts, gaming environment development, and outcome and participation experience measures.

💡 Research Summary

This paper presents a systematic literature review of immersive virtual‑reality (IVR) serious games (SGs) used for building‑evacuation training and research. The authors searched major databases (Scopus, Web of Science, IEEE Xplore, ACM DL) for articles published between 2010 and 2024 that combined IVR, serious‑gaming, and evacuation contexts. After duplicate removal and relevance screening, 45 peer‑reviewed studies were retained for detailed analysis.

The review first categorises the educational objectives of IVR SGs. Most studies aim to improve knowledge (hazard awareness, procedural steps), affective outcomes (confidence, anxiety reduction), and decision‑making skills (route selection, risk assessment). Compared with traditional media (videos, posters, drills), IVR SGs generate higher presence and flow, which translates into better knowledge retention and transfer. Pre‑post tests in the literature show an average knowledge gain of roughly 18 % and a reduction in evacuation time of 10–15 % relative to conventional drills.

Behavioural metrics are examined next. Researchers measured evacuation time, path diversity, collision counts, and physiological stress indicators (heart rate, skin conductance, EEG). Studies that incorporated multimodal feedback (audio, haptic vibration, tactile cues) reported a 12 % faster evacuation and more rapid stress recovery than visual‑only conditions. These findings suggest that IVR can evoke realistic emotional responses while maintaining a safe experimental environment.

The technical implementation of the games is then summarised. Unity (≈62 %) and Unreal Engine (≈28 %) dominate the development landscape, with NVIDIA PhysX and Havok providing physics realism. Head‑mounted displays (HMDs) include Oculus Rift, HTC Vive, and the newer standalone Oculus Quest. The authors note a direct correlation between graphical fidelity/physics accuracy and learning transfer: low‑resolution or low‑frame‑rate environments diminish perceived realism and consequently reduce training effectiveness.

Evaluation methodologies are diverse. While many studies still rely on surveys and interviews, an increasing number combine log data (movement trajectories, interaction counts) with biometric signals and, in a few cases, cross‑validation against real‑world evacuation drills. The authors argue that mixed‑methods designs yield the most robust assessment of both cognitive and behavioural outcomes, but they also highlight a paucity of longitudinal studies that would confirm long‑term retention and real‑world applicability.

From these analyses, the authors construct a conceptual framework comprising four inter‑linked pillars:

1. Pedagogical Impact – definition of learning goals, instructional design, feedback loops, and pre‑post testing.
2. Behavioural Impact – collection and analysis of performance data, stress physiology, and modelling of evacuation behaviours.
3. Game Development – selection of hardware and software, level of visual/physical realism, interaction modalities (gesture, voice, haptic).
   4 Outcome & Participation – metrics of learning achievement, evacuation efficiency, user satisfaction, cost‑effectiveness, and participation rates.

The framework stresses that balanced attention to all four pillars during design, implementation, and evaluation maximises the efficacy of IVR SGs.

The discussion acknowledges several limitations in the current body of work. Most experiments are short‑term, lacking evidence of long‑term memory consolidation or transfer to actual emergencies. Hardware costs, motion sickness, and visual fatigue remain barriers to widespread adoption. Moreover, cultural and age‑specific adaptations are under‑explored.

Future research directions proposed include: (a) longitudinal studies with real‑world drill cross‑validation, (b) low‑cost, lightweight HMD solutions and cloud‑streamed VR to improve accessibility, (c) AI‑driven adaptive scenarios that respond to player performance, and (d) culturally and demographically tailored game content.

In conclusion, the paper demonstrates that immersive VR serious games hold significant promise for enhancing evacuation training and behavioural research. By synthesising existing evidence and offering a detailed design‑implementation‑evaluation framework, the authors provide a practical roadmap for academics and practitioners seeking to develop effective, engaging, and evidence‑based evacuation training tools.