Using artificial intelligence and big data to predict the level of crowding at evacuation centers up to six hours in advance, a team including researchers from the National Research Institute for Earth Science and Disaster Resilience (NIED) has begun developing a system to prevent the spread of the novel coronavirus during disasters.
The team will verify the effectiveness of the system in cooperation with local governments to achieve efficient evacuation in the face of the spread of infections.
Shibaura Institute of Technology and Weathernews Inc., a weather information company, are also part of the team, which has been jointly developing the system since February last year, when the virus began spreading in Japan.
The system uses a form of AI called the disaster support chatbot, otherwise known as SOCDA -- which stands for "social-dynamics observation and victims support dialogue agent platform for disaster management" -- to automatically collect and analyze disaster-related information posted on communications app Line. In the event of heavy rain or an earthquake, the chatbot grasps the current locations of evacuees and their preferred evacuation sites. This data is then analyzed together with big data such as population distribution in the surrounding area, estimated flooding, the movement of people and vehicles, and damage to roads, to predict the degree of congestion at each evacuation center up to six hours in advance.
The results of the analysis will be shown on a map as the congestion rate and immediately shared with the affected municipalities. Based on this information, municipalities can consider setting up additional shelters and distribute relief supplies efficiently.
Evacuees will also receive information on evacuation centers from the chatbot via Line, and will be able to avoid the centers that are expected to be overcrowded. The chatbot will be able to ask health questions such as, "Do you have a fever?" and recommend that evacuees with fever or other symptoms go to evacuation centers with quarantine areas.
The verification of the system will be conducted in collaboration with Nagano Prefecture, which was severely flooded during Typhoon No. 19 in autumn 2019.
After the torrential rains in Kyushu in July last year, local governments reduced the capacity of evacuation centers to avoid the Three Cs, resulting in some residents being denied admission to the centers.
Yuichiro Usuda, director general of the Center for Comprehensive Management of Disaster Information at NIED, said, "We hope to make the system widely used and realize a society where disaster victims can evacuate appropriately even in the midst of the coronavirus pandemic."
Read more from The Japan News at https://japannews.yomiuri.co.jp/
