Capturing people mobility with mobile sensing technology for disaster evacuation

Abstract

In this paper, we propose a client-server-service-based system that provides crowd detection and mobility capture. Crowd detection is to detect and calculate the density of crowds within a specified area. Mobility capture is to track the direction of the people. If a warning mechanism is added, the system can prevent or dissolve the crowd to avoid accidents in public places by sending alerts when the detected number of people exceeds a certain set percentage. The technology also plays a big role after the disaster, by calculating the path with relatively small population density in the disaster area, the high-density crowd can be led to a safe area avoiding secondary injury by crowding. Compared to common methods, such as cameras, our proposed system has the advantages of low cost and location flexibility. The system can detect any area without pre-deployed, as long as there is a sufficient number of users involved. In this article, we conducted several experiments in real environments to determine if the system can accurately capture crowd information and route tracking.