Abstract: Pervasive applications such as digital memories or patient monitors collect a vast amount of data. One key challenge in these systems is how to extract ``interesting" information. Because users cannot anticipate their future interests in the data when the data is stored, it is hard to provide appropriate indexes. In this paper, we present an automatic approach to identify ``interesting" data using users' location information. As location tracking devices such as GPS are readily available, digital memories or other pervasive systems may record location information along with the data. Given the location information, our \tm Tag system identifies ``interesting" days. We evaluated Tag using a real wireless trace and demonstrated its capabilities. Using Tag, we were able to identify days when mobility patterns change and differentiate days when a user follows a regular pattern from those when she shows a irregular pattern. We also discovered general mobility characteristics such as most users have one or more repeating mobility patterns and repeating mobility patterns do not depend on certain days of the week, except for no activity for weekends.
Keywords: wireless network, mobile computing, mobility, location, security, anomaly detection
No online copy available.
Copyright © 2007 by the authors.