STSM- a model to detect and predict large crowd anomalies for optimized path recomendation

Abstract

Cultural diversity of large crowds is one of the major concerns when participants overstep the predefined guidelines. Such behaviors eradicate crowds' safety, resulting massive casualties. Advent of tracking devices and smartphones with multiple sensing abilities can leverage to capture crowds' real-time spatio-temporal (ST) data to serve emergency service plans. In this paper, we present a Spatio-Temporal Service Model (STSM) that can detect and predict anomalies with in a very large crowd. The model correlates crowd's real-time information with past user's ST-Data and their traffic mobility, and alerts all stockholders and recommends optimized path to shelter points and safe exits for any possible disaster. As a case study, tracking devices were deployed to capture spatio-temporal information of about 2970 vehicles used for pilgrims' mobility during Hajj 2016 event. The data analysis is summarized and basis the functionalities of the model for future pilgrimage.