FreeSense

Abstract

Human detection aims to monitor how people are moving in an area of interest. There are many potential applications such as asset security monitoring, emergency management, and elderly care, etc. With the development of wireless sensing technique, Wi-Fi-based human detection method carries great potential due to advantages of pervasive accessibility and coverage flexibility. Previous studies have investigated the detection of human movements via signal variations. However, affected by noises, such as multi-path effect and device difference, existing approaches cannot achieve high accuracy and low false alarm rate at the same time. In this paper, we propose FreeSense, a novel Wi-Fi-based approach for human detection. Different from previous studies that characterize the variation of temporal wireless signals or calculate the deviation of Channel State Information (CSIs) from a normal profile, we will detect human movements by identifying whether there is any phase difference between the amplitude waveforms of multiple receiving antennas. In addition, we also model the sensing coverage for movements of different granularities in open space and propose a method to estimate the coverage range. Extensive experiments demonstrate that FreeSense can achieve an average false positive rate (FP) of 0.53% and an average false negative rate (FN) of 1.40%. The coverage range estimation method can achieve an average accuracy of 1.36 m, sufficient to guide the deployment of devices for human detection indoors.