Tracking People in Indoor Environments

Abstract

Tracking the movement of people in indoor environments is useful for a variety of applications including elderly care, study of shopper behavior in shopping centers, security etc. There have been several previous efforts at solving this problem but with limited success. Our approach uses inexpensive pressure sensors, placed in a specific manner, that allows us to identify multiple people. Given this information, our algorithm can track multiple people across the floor even in the presence of large sensor error. The algorithm we develop is evaluated for a variety of different movement patterns that include turning and path crossing. The error in correct path detection is shown to be very small even in the most complex movement scenario. We note that our algorithm does not use any a priori information such as weight, rfid tags, knowledge of number of people, etc.