TY - GEN
T1 - Flood Area Estimation Using Personal Location Data - Case Study of Japan Floods in 2018
AU - Hiroi, Kei
AU - Yoshida, Takahiro
AU - Yamagata, Yoshiki
AU - Kawaguchi, Nobuo
N1 - Funding Information:
This research was supported by the MIC/SCOPE #172106102.
Publisher Copyright:
© 2019 IEEE.
PY - 2019/3
Y1 - 2019/3
N2 - This research proposes a flood area estimation method in urban areas using personal location data. Many studies have investigated the estimation of flood levels; however, the majority of these previous works are based on the flood monitoring data. The main cause of flood disaster death is drowning due to evacuation delay in the area where observation equipment is not installed, so it is difficult to estimate flood occurrence by the previous method based only on monitoring data. In this research, we propose an estimation method that does not rely on monitoring data and instead estimates flood areas using GPS data collected from smartphones owned by the affected people. This method detects anomaly areas to analyze temporal and spatial changes of an area where a personal movement during a flooding event differs from that during regular times from personal location data. For anomaly detection, we use a dynamic time warping method with fixed window size and inequality metrics to estimate the area where an anomaly event occurred in 2 km grids. We applied this method to Kurashiki city, Okayama Prefecture, where there 52 people died during a flood that occurred in Japan in 2018. Our method found that, at the flooding time, the anomaly occurrence was estimated correctly in the area where inundation actually occurred.
AB - This research proposes a flood area estimation method in urban areas using personal location data. Many studies have investigated the estimation of flood levels; however, the majority of these previous works are based on the flood monitoring data. The main cause of flood disaster death is drowning due to evacuation delay in the area where observation equipment is not installed, so it is difficult to estimate flood occurrence by the previous method based only on monitoring data. In this research, we propose an estimation method that does not rely on monitoring data and instead estimates flood areas using GPS data collected from smartphones owned by the affected people. This method detects anomaly areas to analyze temporal and spatial changes of an area where a personal movement during a flooding event differs from that during regular times from personal location data. For anomaly detection, we use a dynamic time warping method with fixed window size and inequality metrics to estimate the area where an anomaly event occurred in 2 km grids. We applied this method to Kurashiki city, Okayama Prefecture, where there 52 people died during a flood that occurred in Japan in 2018. Our method found that, at the flooding time, the anomaly occurrence was estimated correctly in the area where inundation actually occurred.
KW - Anomaly Detection
KW - Disaster Estimation
KW - GPS Data
UR - http://www.scopus.com/inward/record.url?scp=85067912967&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85067912967&partnerID=8YFLogxK
U2 - 10.1109/PERCOMW.2019.8730882
DO - 10.1109/PERCOMW.2019.8730882
M3 - Conference contribution
AN - SCOPUS:85067912967
T3 - 2019 IEEE International Conference on Pervasive Computing and Communications Workshops, PerCom Workshops 2019
SP - 285
EP - 291
BT - 2019 IEEE International Conference on Pervasive Computing and Communications Workshops, PerCom Workshops 2019
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2019 IEEE International Conference on Pervasive Computing and Communications Workshops, PerCom Workshops 2019
Y2 - 11 March 2019 through 15 March 2019
ER -