AI-based fire evacuation guidance system introduced

At the subway station of Daejeon City Hall on Monday morning, a red light installed between the subway platform and the first basement level came to life all of a sudden, signaling a fire. As the lights of the subway station went out, passengers seemed lost in the abrupt chaos. Then, a green laser started beaming from below, guiding the evacuation way for passengers. The monitor that typically informs imminent arrivals of trains guided the passengers to follow down the platform into stairs instead of the ones on fire. Passengers moved quickly along the green laser line, safely running away from the fire.

An AI research team led by Dr. Han Hyeong-seok, a senior researcher at Korea Institute of Machinery and Materials, revealed an AI-based evacuation guidance system and put on a demonstration of it at Daejeon City Hall Station. His team developed a system providing information on the safest evacuation routes in case of a fire based on data accumulated by AI learning. The researchers taught two things to the system. Firstly, the AI system eval‎uated the level of urgency based on a set of information including temperatures, carbon monoxide and smoke densities, which had been gathered by 30 IoT sensors installed inside the subway station. Subsequently, they recorded the level of safety at fire spots based on firemen’s feedback, which the AI system was made to learn and offer the safest exit routes in case of an emergency.

When a fire occurs, based on the information it has learned, the AI system finds the optimal routes of evacuation, and offers guidance for passengers using 130 laser indicators installed on the ceilings. “In case of a fire, 70% of the lights go out,” explained a senior researcher. “Passengers naturally panic, so it is vital to guide their way by using laser, which can penetrate through the smoke and reach the floor for visibility.”

This is part of a cooperation program between Daejeon City and a government-financed research institute, which is designed to tackle issues in their community. Under the scheme, the researcher team is to transfer technologies to local companies, and Daejeon-based companies receive support for commercialization of technology.


shinjsh@donga.com