Indoor Events Monitoring Using Channel State Information Time Series

Abstract

By sensing wirelessly the radio propagation environment and analyzing the channel state information (CSI), one can extend human senses beyond our traditional reach and enrich the insight into the surrounding environment and activities, with or without line-of-sight. On one hand, different indoor activities bring distinctive perturbations to wireless radio propagations. On the other hand, thanks to the nature of multipaths, indoor environmental information is contained and embedded in the wireless CSI. Since the occurrence of an indoor event lasts for a certain period of duration and repeats a similar transition pattern among different realizations, information is embedded not only in each instantaneous CSI sample, but also in how CSI changes along time, e.g., the CSI time series. Inspired by that, this paper proposes an indoor monitoring system that monitors the occurrence of different indoor events in real time with commercial WiFi devices, by exploiting the temporal information embedded in the CSI time series. Through extensive experiments, this paper studies the robustness of the proposed system to variabilities in event instances and human motion interference, and its long-term performance in a one-month test.