Device-free human monitoring using channel state information

Abstract

An innovative approach to analyze the CSI information available in Wi-Fi based systems for the understanding of device-free human actions in indoor areas is discussed. The method is based on the representation of the CSI data as "strings" and on the analysis of their distance in the complex domain in order to identify different human behaviour and actions. The scattering of the human body is exploited to solve the inverse problem of human behavior reconstruction from the electromagnetic field measurement. A preliminary experiment using a couple of IEEE 802.11 wireless devices is presented to recognize the posture of a device-free human in a non-controlled indoor scenario.