Firefighter Training Simulation

Project Description:

This project, developed for the China government, is a firefighter training simulator utilizing advanced AI to create highly realistic fire scenarios. The AI maps the environment, predicting fire spread based on material properties and environmental conditions. A standout feature is the scenario mapping editor, allowing users to assign specific materials to nodes, influencing the fire's behavior and enabling diverse training simulations.
Originally, the project was set to expand into VR for a more immersive experience. However, due to the COVID-19 pandemic, the project has been paused and remains on hold to this day.

My Role in the Project - Senior Unity Developer:

- Developed the core system architecture.
- Created a custom, performant fire renderer.
- Implemented the fire spread system, which dynamically reacts to different environmental materials.
- Built a realistic smoke system with behavior based on the environment.
- Developed an editor tool to assign material properties across the entire environment.
- Handled server communication and exception management.

Challenges and Solutions:

The fire simulation needed to be both highly realistic and optimized, as it could spread across the environment and had to run on WebGL, a platform that demands efficient performance. I designed a custom particle system that achieved the same level of performance as Unity's default system.

Additionally, working with server connections on WebGL posed challenges, particularly with handling internet connections. After extensive research, I successfully integrated server communication using JavaScript integration to ensure reliable connectivity.