Evacuee behavior analysis for real time disaster detection

Sudden occurrences of serious disasters such as fires, earthquakes, and terrorisms have caused the deaths of many people around the world. One factor is the failure to escape from disaster, as they may not have recognized the danger in time. To cope with this problem, we have been developing ERESS (Emergency Rescue Evacuation Support System). By ERESS, the information of disasters can be broadcast to evacuees immediately by Mobile Ad-hoc NETworks (MANET). ERESS is enable to provide evacuees with information of the location of exits routes and disaster spots. ERESS is updated continually via real time processing. Even when the disasters may spread, people will be able to evacuate from dangerous areas safely and rapidly. This paper develops important portion of ERESS say Pre-ERESS, for automatic disaster detection in real time. Pre-ERESS uses tablet computers and sensors. It will be able to analyze the behavior of humans in real time from experiments. It will also be able to detect disasters from Panic-type disaster experiments. Hence, we will show the validity of the Pre-ERESS from two types of experiments, and show the availability of the system.