MiamiMapper: Crowd analysis using active and passive indoor localization through Wi-Fi probe monitoring

Abstract

Crowd analysis and monitoring in large congregations is an important problem related to public safety and planning. Researchers have long been using image and video for effective crowd analysis. However, with the advent of more sophisticated technologies, crowd monitoring has been attempted using GPS, RFID as well as Bluetooth based systems. Indoor crowd monitoring in public buildings, such as, airport, museum, theaters, etc. has not been widely studied. In this paper, we propose a non-invasive Wi-Fi based approach for indoor crowd analysis which works by passively monitoring the probe packets generated by Smartphones. Since, smartphones nowadays can be unequivocally associated with a human user, counting unique MAC addresses of smartphones can generate a fairly close estimate of number of users in the indoor space and their movement patterns. Although indoor localization using wireless technology has been extensively studied, using probe packets for crowd monitoring has rarely been investigated. In this paper, we have developed the MiamiMapper system, from COTS hardware and software, for crowd analysis in indoor environment using passive monitoring of probe packets emitted by wireless devices. Our system can localize users and thereby detect a crowd with an average accuracy of 7.28 meters which is 75.33 % more accurate than GPS based systems for indoor environment.