Details published in the Journal of Building Engineering reveal a new approach to fire evacuation, moving beyond simply identifying the closest exit. Researchers at the National Institute of Standards and Technology (NIST) have developed “Safe Step,” an AI model that forecasts fire evolution to dynamically update safe evacuation routes within a building. Unlike previous algorithms focused on shortest paths, Safe Step considers how hazards accumulate over time, redirecting occupants away from exits that may become unsafe as a fire spreads. “Fires can grow and spread,” said Hongqiang “Rory” Fang, a research associate at NIST and first author of the journal paper. “Our model forecasts how the fire is evolving and can help update emergency exit displays to direct people toward the safest exit.” Some smart buildings are currently testing the technology with dynamic emergency exit displays utilizing real-time temperature and air quality data.
Safe Step: AI for Predicting Fire Spread & Evacuation
The ability to accurately predict a building fire’s spread and guide occupants to safety represents a significant advancement in fire safety technology. Researchers at NIST have developed an AI model to achieve this. The model utilizes reinforcement learning, a type of artificial intelligence, to make decisions through trial and error, learning evacuation routes based on building layouts and data from NIST’s own fire simulation tools. Crucially, Safe Step doesn’t require real-time fire simulations; it relies on live sensor data measuring temperature and air quality to continuously refine its recommendations. Researchers quantify safety using the fractional effective dose (FED) of toxic gases, aiming to minimize occupant exposure over time. Testing against traditional algorithms in single-level buildings demonstrated Safe Step’s consistent ability to identify safer evacuation routes, even when initially closer exits become compromised.
For example, the model can anticipate a scenario where a nearby exit becomes unsafe as a fire intensifies, redirecting occupants to a more distant, but ultimately safer, alternative. NIST is now working to expand the model’s capabilities to handle multi-story buildings and incorporate multiple agents to simulate individual occupant behavior and potential congestion points, potentially coordinating evacuation routes with firefighter access.
Reinforcement Learning Guides Safe Evacuation Route Selection
A new artificial intelligence model is factoring in the evolving danger of a fire to dynamically guide building occupants to safety, going beyond simply identifying the nearest escape. This contrasts with traditional algorithms that prioritize shortest paths based on initial conditions, potentially leading people into harm’s way as conditions change. “We asked ourselves, ‘Can we build a better algorithm that predicts how the fire evolves, and in a way that helps save more lives?’” said NIST mechanical engineer Wai Cheong Tam. Consider a scenario where a fire begins across a hallway; a conventional algorithm might direct someone to cross immediately for the closest exit. However, Safe Step can foresee a situation where that exit becomes untenable as the fire intensifies.
“Fires can grow and spread,” said Hongqiang “Rory” Fang, a research associate at NIST and first author of the journal paper. “Our model forecasts how the fire is evolving and can help update emergency exit displays to direct people toward the safest exit.”
Fractional Effective Dose (FED) Minimizes Hazard Exposure
NIST researchers are refining fire evacuation strategies, focusing instead on minimizing cumulative hazard exposure using a metric called the fractional effective dose (FED). This approach, detailed in the Journal of Building Engineering, moves beyond traditional algorithms that prioritize shortest paths, instead accounting for the evolving toxicity of a fire’s byproducts during an escape. The team’s AI model, dubbed Safe Step, utilizes FED to quantify the severity of fire hazards experienced by evacuees over time; a lower FED indicates reduced exposure and, consequently, greater safety. Safe Step doesn’t merely calculate a safe route at a single moment; it continuously assesses risk based on real-time data from building sensors monitoring temperature and air quality. This dynamic assessment is critical, as a previously clear path can quickly become hazardous as a fire grows.
The model might instead recommend a longer route to an exit further down the hallway, recognizing that the initial, closer option could become untenable by the time an evacuee reaches it. This predictive capability is achieved by assigning numerical values to hazards, allowing the algorithm to select the path with the lowest overall FED. Researchers are now expanding Safe Step’s capabilities to handle multi-level buildings and model the interactions of multiple evacuees, aiming for a system that can coordinate safe egress for all occupants while also facilitating firefighter access.
“This research is still in the early R&D stage, but it represents an important step toward intelligent firefighting where effective use of advanced technologies can protect property and save lives,” said Fang.
Dynamic Exit Displays & Multi-Agent System Development
The integration of artificial intelligence into building safety systems is moving beyond theoretical models and into practical testing; dynamic emergency exit displays are now being evaluated in select smart buildings, offering occupants real-time guidance during emergencies. These displays are not simply illuminated signs, but responsive interfaces utilizing data from sensors monitoring temperature and air quality to assess exit safety and dynamically redirect people away from compromised routes. This proactive approach represents a significant shift from traditional evacuation strategies that prioritize the shortest path, regardless of evolving hazards. This will allow the model to account for congestion and coordinate access for firefighters, potentially directing evacuees to less crowded exits while ensuring emergency responders can effectively enter the building.
