Human Mobility Estimation Following Massive Disaster Using Filtering Approach

Abstract

Real-time estimation of people distribution immediately after a disaster is directly related to disaster reduction and is also highly beneficial in society. Recently, traffic estimation research has been actively performed using data assimilation techniques for observation data obtained from mobile phones. However, there has been no research on data assimilation technique using real-time gridded aggregated observation data obtained from mobile phones, which are available and can be used to estimate population flow and distribution in a metropolitan area during a large-scale disaster. In this research, population distribution in an urban area during a disaster was estimated using gridded aggregated observation data obtained from mobile phones, using particle filter. The experimental results indicated that the particle filters enabled high-precision real-time estimation in the Kanto district.