Indoor fires cause thousands of deaths and millions of dollars in property damage each year in Taiwan. Fire is an inevitable disaster as a result of how to assist people escape from fires is a critical challenge. Under the continued sup-port by the National Science and Technology Council (NSTC) in Taiwan, Pro-fessor Pei-Hsuan Tsai and her team collaborate to develop an artificial intelli-gence-based fire auxiliary system to provide online and real-time guidance for victims of fires, to real-time predict fire spreading, and to prevent fires by au-tomatic environmental inspection. The biggest advantage of this system is that it can be run in parallel with the existing fire protection system without re-building the entire fire system. By introducing artificial intelligence technology to optimize the current fire protection system, it is expected to promote the development of smart buildings in the future and improve people's living safe-ty!
The fire protection system should have functions including guiding escape, collecting fire site information and environmental safety assessment. However, traditional fire protection systems generally have the following three important problems: First, they adopt pre-planned escape paths and cannot dynamically change the escape paths according to the real-time conditions of the fire scene. Second, most of the real-time fire data comes from sensors, and the fire scene is collected through the network. However, in the fire scene, the sensor net-work equipment is easily affected by high temperature combustion, gas dis-turbance and other factors, resulting in interruption phenomenon, such as sen-sor device or network damage that hinders fire scene data collection. Third, most of the current environmental fire protection inspections rely on profes-sional firefighters to go to the scene to conduct environmental assessment and inspection according to fire regulations. This requires a lot of manpower. Be-sides, the environment may change dynamically from time to time, such as changing the position of large furniture, stacking a large number of flammable substances and cartons, that will affect the fire safety of the environment. It cannot be detected and reminded to improve from time to time by depending on human inspection. In view of this, artificial intelligence-based fire-fighting assistance systems focus on the development of the following three core tech-nologies:
Escape from fire by online indoor navigation
In terms of assisting fire escape, Professor Tsai's team designed indoor path planning algorithms considering where people and fires are to select a safer and closer exit and establish an escape route that effectively avoid dangerous areas nearby fires. Compared with other existing methods, the online indoor navigation effectively reduces casualties.
Keep away fire by real-time fire spreading prediction
Correct fire decision-making relies on real-time fire information that mostly sourcing from sensor network. However, sensors easily crashed or network probably disconnected in fires. Professor Tsai’s team developed a real-time fire spreading prediction to supplement the insufficiency of real-time fire infor-mation when sensing data are unavailable.
Prevent fire by automatic indoor inspection
An ounce of prevention is worth a pound of cure. At present, fire prevention mostly relies on regular inspection by firefighters that are high labor costs. In view of this, Professor Tsai's team developed an automatic indoor inspection - "Automated Building Fire Simulation Based on Digital Twin Architecture", which can quickly virtualize the indoor environment and automatically gener-ate models through the assistance of video artificial intelligence. Combined with fire simulation to discover the hazard factors that induce fire or affect es-cape.
Now should be a good time for Taiwan's industry to get involved in smart buildings and smart cities. Living quality and safety are being improved all over the world. However, most of the buildings and constructions in large cit-ies with a long history are at a relatively backward stage of fire information equipment. The fire-fighting system tools designed and developed by Professor Tsai's team and the introduction of artificial intelligence technology can com-plement existing apartment buildings, factory warehouses, and fire safety defi-ciencies such as fire early warning and escape response, improve the survival rate of victims and reduce property losses. Integrating cyber physical system and artificial intelligence computing into the original fire protection system can lead the machinery, automation and embedded software industries to enter the smart building market with deep development prospects and future prospects in Asia and the world.
Reference on International Publication:
1. An-Fong Lee and Pei-Hsuan Tsai*, "A Resource-saving Shelter Selection Ap-proach for Large-scale Area Emergency", submitted to IEEE Systems Journal and under major revision.
2. G.R. Shih and P.H. Tsai*, " Safest-path planning approach for indoor fire evacuation ", submitted to International Journal of Disaster Risk Reduction and under major revision.
3. Rong-Guei Tsai, Yi-Yuan Tsai, Pei-Hsuan Tsai*, "Automation tool for home fire safety check". IEEE Sensors Letters, 2021, 5(12), 1-4. doi: 10.1109/LSENS.2021.3124800
Media Contact:
Yen-Hui Liang
Program Manager
Department of Engineering and Technologies
National Science and Technology Council
+886-2-27377525
yhliang@nstc.gov.tw
Yen-Hui Liang
Program Manager
Department of Engineering and Technologies
National Science and Technology Council
+886-2-27377525
yhliang@nstc.gov.tw
