Fujitsu Optimizes Evacuation Center Management to Mitigate COVID-19 Risk with AI in Joint Field Trial in City of Kawasaki

Aug 24, 2020 11:47 JST

Source: Fujitsu Ltd

News Facts:
- Evacuation centers in cities face significant crowding in the event of a natural disaster, and ensuring the safety of evacuees amidst the COVID-19 pandemic presents unprecedented challenges
- Fujitsu will conduct a field trial at an evacuation center in Japan to simulate COVID-19 risk against existing emergency response protocols in cooperation with Tohoku University, the University of Tokyo, and the City of Kawasaki in the latest milestone for their joint disaster management project
- In the future, crowd flow simulation tech and AI image recognition solutions used in the trial will also be considered for infection mitigation for other facilities, events, and large gatherings

TOKYO, Aug 24, 2020 - (JCN Newswire) - The International Research Institute of Disaster Science at Tohoku University, Earthquake Research Institute at the University of Tokyo, the City of Kawasaki, and Fujitsu Limited have today announced that they will conduct a joint field trial of AI-based technologies to optimize disaster evacuation center management in Kawasaki City, Japan, amidst the COVID-19 pandemic with a drill on Monday, August 31. This project marks the latest milestone for the four parties, which have been promoting their Joint Project Aiming for Tsunami Disaster Risk Reduction Using ICT in the Kawasaki Coastal Area since November 2017. The field trial will rely on crowd flow simulation technology and AI image analysis to identify measures to mitigate the risk of COVID-19 infection for evacuees by helping decision-makers reduce exposure to the "three Cs": closed spaces, crowded places, and conversations in close proximity.

Fig 1: Evaluation of the simulated risk of infection due to differences in evacuation behavior of evacuees

Fig 2: AI detects number of evacuees and other attributes

Ahead of the trial, the four parties will use crowd flow simulation technology developed in advance to visualize infection risk, which varies depending on the volume and proximity of people entering the evacuation center. The simulation assumes that a certain number of evacuees are infected with COVID-19, offering disaster management decision-makers insights into how to alter their evacuation center management plans accordingly. On the day of the trial, Fujitsu's AI image analysis solution will be deployed on-site to automatically collect information including the number of inbound evacuees and their attributes from video data captured by cameras installed near the evacuation center entry points, delivering real-time information on the level of crowding at the evacuation centers. This data will be transmitted to the City of Kawasaki Disaster Response Headquarters, enabling early and appropriate responses to reduce infection risk from the "three Cs."

Field Trial Outline

Date/Time:
Monday, August 31st 13:00-16:00

Location:
Kawasaki Municipal Tonomachi Elementary School

Participants:
About 60

Trial Overview:
Employees of Kawasaki City will conduct a drill including the opening of the evacuation center as well as the reception of evacuees, in accordance with guidelines for combatting COVID-19 developed by the City of Kawasaki authorities. Evacuation center management plans and protocol will be examined with crowd flow technology that simulates the potential spread of infection, and the effectiveness of AI to count the number of evacuees entering the site will be evaluated. The trial will also offer evacuation center management personnel experience managing the process considering the additional safety challenges posed by the COVID-19 pandemic.

Crowd flow simulation technology to estimate infection risk

This technology, developed by the International Research Institute for Disaster Science at Tohoku University, the Earthquake Research Institute at the University of Tokyo, and Fujitsu Research Institute, can simulate changes in the risk of infection due to differences in crowd flow by incorporating a COVID-19 infection risk assessment function into the existing technology, which reproduces crowd flow under various conditions. With this technology, it is possible to predict how the infection will spread along with the flow of people, assuming that there is a person infected among the evacuees--people that remain within a certain distance of the infected person for a certain period of time are designated as a potential risk of infection. The four parties will use this technology prior to the field trial to simulate various evacuation plans in the event of a COVID-19 outbreak. Specifically, they will evaluate how the risk of infection changes according to differences in the management of evacuation centers, including the number of evacuation center reception points and the number of receptionists, and consider the appropriate implementation plan for the management of evacuation centers according to the situation.

AI image analysis solutions that automatically measure congestion at evacuation entry points

The trial will leverage Fujitsu's AI image analysis solution Fujitsu Technical Computing Solution GREENAGES Citywide Surveillance V3. This solution makes it possible to confirm information including the number and other attributes of inbound evacuees regardless of whether they are wearing masks from images taken by cameras installed near the designated entry points, and to visualize the level of congestion at the evacuation center in real-time. The effectiveness of the system for accurately grasping the situation on the front lines is evaluated by collecting this data in real-time at the City Disaster Response Headquarters.
Future Plans

The four parties will study measures to ensure safer evacuation while mitigating infection risk from COVID-19 based on the insights gained through this field trial. They will also investigate the possibility of detecting congestion and reducing the risk of infection not only during natural disasters but also at facilities and events where large crowds gather, as well as the applicability of the technologies developed and evaluated this time.

Source: Fujitsu Ltd
Sectors: Enterprise IT, Artificial Intel [AI]