For my Masters final project I developed a VR-based adaptive simulation for firefighter training, utilising Learning Analytics and Reinforcement Learning systems to analyse performance, generate feedback and adjust the experience accordingly. There are many crucial procedures for engaging a fireground, which are exercised in life-threatening scenarios. The inherent dangers associated with live-fire training suggest that an effective VR alternative would be highly valuable.

My research project aimed to answer this question: “Can VR, LA and ML be combined to provide effective situational and communication training for firefighters?”

The academic research process involved a complementary study with user-testing. After being on-boarded, test subjects were evaluated inside the VFT, measuring their performance as they received iterative automated feedback. Subjects demonstrated a notable improvement in technique and understanding of firefighting procedure.

The VFT was designed with both trainee and instructor in mind. Using particle effects and realistic environments, I was able to emulate key aspects of firefighting, providing an immersive platform for conducting training of procedure, communication and technique. I built a dedicated interface for instructors which allowed them to monitor and control the simulated fireground, allowing them to take the role of overseer in the training experience.

A secondary system was developed alongside the full-realism simulator, in which instructors could quickly create their own courses and place fires for the trainee to engage. This was particularly useful for testing a trainees ‘primary search’ technique, a standardised approach to scanning a site for danger.